Citation
A New Technological Era for American Agriculture

Material Information

Title:
A New Technological Era for American Agriculture
Creator:
United States. Congress. Office of Technology Assessment.
Publisher:
U.S. Congress. Office of Technology Assessment
Publication Date:
Language:
English
Physical Description:
446 p. : ill. ; 28 cm.

Subjects

Subjects / Keywords:
Agricultural innovations -- United States ( LCSH )
Agriculture -- Technology transfer ( LCSH )
Agriculture -- Environmental aspects -- United States ( LCSH )
Agricultural biotechnology -- United States ( LCSH )
Food -- Quality ( LCSH )
Genre:
federal government publication ( marcgt )

Notes

General Note:
The report concludes that these technologies have the potential to provide new solutions to many agricultural problems. The challenge, however, will be whether government, industry, and the public can strike the proper balance of direction, oversight, and use to allow these technologies to flourish. Congress will be faced with many issues and choices as American agriculture moves into this new era.

Record Information

Source Institution:
University of North Texas
Holding Location:
University of North Texas
Rights Management:
This item is a work of the U.S. federal government and not subject to copyright pursuant to 17 U.S.C. §105.
Classification:
Y 3.T 22/2:2 Ag 8/6 ( sudocs )

Aggregation Information

IUF:
University of Florida
OTA:
Office of Technology Assessment

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

A New Technological Era for American Agriculture August 1992 OTA-F-474 NTIS order #PB92-216118 GPO stock #052-003-01290-1

PAGE 2

. Recommended Citation: U.S. Congress, Office of Technology Assessment, A New Technological Era for American Agriculture, OTA-F474 (Washington, DC: U.S. Government Printing Office, August 1992). For sale by the U.S. Government Printing Office Superientendent of Documents, Mail Stop: SSOP, Washington, DC 20402-9328 ISBN 0-16 -037978-4

PAGE 3

Foreword American agriculture is entering a new technological era that holds great promise. Biotechnology and advanced computer systems have the potential to increase productivity, enhance the environment, improve food safety and quality, and bolster U.S. agricultural competitiveness. Many of these new technologies will be available in the 1990s. But their introduction will be under circumstances unlike any met by past technologies. Uncertainties over these new technologies raise questions of potential impacts on food safety and the environment, and possible economic and social costs. Nevertheless, there will be a push for some of these technologies-biotechnology in particular to be used commercially, adopted by industry, and accepted by the public. Congress requested the Office of Technology Assessment to examine emerging technologies that may be available to American agriculture in the 1990s, their potential for industry, and consequent policy issues. This report analyzes the technologies and related policy issues Congress may need to resolve. The analysis includes an assessment of adjustments industry must make to capitalize on the new technologies, the scientific and institutional issues relevant to food safety and environmental risk and benefit, and the implications for intellectual property rights and science policy. The report concludes that these technologies have the potential to provide new solutions to many agricultural problems. The challenge, however, will be whether government, industry, and the public can strike the proper balance of direction, oversight, and use to allow these technologies to flourish. Congress will be faced with many issues and choices as American agriculture moves into this new era. This OTA report for Congress is the fourth and final report in a series begun in 1990. The study was requested by the Senate Committee on Agriculture, Nutrition, and Forestry; the House Committee on Government Operations; and the House Committee on Agriculture. The first report issued was Agricultural Research and Technology Transfer Policies for the 1990s, the second report was U.S. Dairy Industry at a Crossroad: Biotechnology and Policy Choices, and the third report was Agricultural Commodities as Industrial Raw Materials. Findings from these reports were relevant to the issues debated for the 1990 Farm Bill. OTA greatly appreciates contributions of the advisory groups, authors of background papers, reviewers, and other contributors to this study who were instrumental in defining key issues and a range of perspectives on them. Their participation does not necessarily represent endorsement of this report, for which OTA bears sole responsibility. Directo r iii

PAGE 4

A New Technological Era for American Agriculture OTA Project Staff Roger C. Herdman, Assistant Director, OTA Health and Life Sciences Division Walter E. Parham, Program Manager Food and Renewable Resources Program Michael J. Phillips, Project Director Marie Walsh, Senior Analyst Laura R. Meagher, Senior Analyst Kevin W. OConnor, Senior Analyst Lawrence R. Jones, Analyst Laura Dye, Research Assistant Susan Wunder, Editor Administrative Staff N. Ellis Lewis, Office Administrator Nellie Hammond, Administrative Secretary Carolyn M. Swarm, P.C. Specialist

PAGE 5

Assessment Advisory Groups Advisory Group on Animal Technology and Related Issues Dale Bauman Professor Department of Animal Science Cornell University David Berkowitz Director Technology Transfer and Coordination Staff Food Safety Inspection Service U.S. Department of Agriculture Roger Blobaum Director of Americans for Safe Food Center for Science in the Public Interest Thomas T. Chen Associate Director Center of Marine Biotechnology University of Maryland Gary Cromwell Professor Department of Animal Science University of Kentucky Stanley Curtis Professor Department of Animal Scienc e University of Illinois Dennis Dykstra Vice-President of Quality Assurance Jack Frost Golden Plump Poultry Terry Etherton Professor Department of Dairy and Animal Science Pennsylvania State University Don Gill Regents Professor and Extension Animal Nutritionist Oklahoma State University William Hansel Distinguished Professor Department of Veterinary Science Louisiana State University Michael Hansen Research Associate consumer Policy Institute Consumers Union Norman Harvey Farmer Florence, VT Burke Healey southern Cross Ranch Davis, OK Walter Hobgood Director of Animal Nutrition and Health Monsanto Co. Maynard Hogberg Head Department of Animal Scienc e Michigan State University Gerald Isaacs Chairman Department of Agricultural Engineering University of Florida Daniel D. Jones Deputy Director Office of Agricultural Biotechnology U.S. Department of Agriculture James Kleibenstein Professor Department of Economics Iowa State University John Kopchick Professor Edison Animal Biotechnology Center Ohio University Edward Korwek Partner Hogan and Hartson Darold McCalla President Granada Biosciences, Inc. Tom McGuckin Associate Professor Department of Economics New Mexico State University Dave Meisinger Product Manager for New Products Pitrnan-Moore, Inc. Jan E. Novakofski Associate Professor Department of Animal Scienc e University of Illinois B.I. Osburn Associate Dean for Research School of Veterinary Medicine University of California Dennis Powers Director Hopkins Marine Station Stanford University Rodney Preston Professor Department of Animal Science Texas Tech University Vernon Pursel Research Leader Agricultural Research Service U.S. Department of Agriculture Allan Rahn Professor Department of Animal Science Michigan State University Caird Rexroad Research Leader Agricultural Research Service U.S. Department of Agriculture Tim Rose Fanner Lyons, KS Andrew Rowan Assistant Dean School of Veterinary Medicine Tufts University Colin Scanes Professor and Chairman Department of Animal Science Cook College, Rutgers University Thomas Sporleder Professsor Department of Agricultural Economics and Rural Sociology Ohio State University Charles Strong Extension Biotechnology Coordinator University of Georgia Michael T omaszewski Professor/Extension Dairy Specialist Department of Animal Science Texas A&M University Edward Veenhuizen Manager of Animal Science Projects Lilly Research Laboratiories Eli Lilly and Co. v

PAGE 6

Advisory Group on Plant Technology and Related Issues Michael Adang Associate Professor Department of Entomology University of Georgia Garren O. Benson Professor Department of Agronomy Iowa State University Norman Brown President FBS Systems Inc. J.A. Browning Professor Department of Plant Pathology Texas A&M University John Burke Research Leader Agricultural Research Service U.S. Department of Agriculture Raghavan Charudattan Professor Department of Plant Pathology University of Florida Sharon K. Clark President Minnesota Corn Growers Assoc. Ken Cook Vice-President for Policy Center for Resource Economics James Davis Vice-President of Development and General Counsel Crop Genetics International Willard Downs Professor Oklahoma State University Lynn Forster Professor Department of Agricultural Economics and Rural Sociology Ohio State Univeseity Nicholas Frey Product Development Manager Specialty Plant Products Division Pioneer Hi-Bred International James Fuxa Professor Department of Entomology Lousiana State University Robert Hall Associate Professor Department of Plant Science South Dakota State University Chuck Hassebrook Director Center for Rural Affairs Dale Hicks Professor Department of Agronomy and Plant Genetics University of Minnesota Carol Hoffman Research Scientist Institute of Ecology University of Georgia Donald Holt Director Agricultural Experiment Station University of Illinois Marjorie Hoy Professor Department of Entomology University of California Jeffrey Ihnen Attorney-at-Law Venable, Baetjer, Howard, and Civiletti George Kennedy Professor Department of Entomology North Carolina State University John C. Kirschman President FSC Associates Ganesh Kishore Research Manager Monsanto Co. Jerry R. Lambert Professor Department of Agricultural Engineering Extension Service/USDA Clemson University Sue Loesch-Fries Assistant Professor Department of Plant Pathology Purdue University Gaines E. Miles Professor Department of Agricultural Engineering Purdue University Kevin Miller Farmer Teutopolis, IL Kent Mix Rancher Larnesa, TX Ian Munro Director Canadian Centre for Toxicology Michael W. Parka chairman Food Research Institute University of Wisconsin Philip Regal Professor Department of Ecology and Behavioral Biology University of Minnesota Jane Rissler Director of Biotechnology National Wildlife Federation Douglas Schmale Farmer Lodgepole, NE Allan Schmid Professor Department of Agricultural Economics Michigan State University Ronald Smith Extension Specialist Auburn University Steve Sonka Professor Department of Agricultural Economics University of Illinois Kent K. Stewart Professor Department of Biochemistry and Nutrition Virginia Polytechnic Institute and State University Nicholas D. Stone Associate Professor Department of Entomology Virginia Polytechnic Institute and State University Christen Upper Professor/Research Chemist Agricultural Research Service U.S. Department of Agriculture Department of Plant Pathology University of Wisconsin Ann Vidaver Professor Department of Plant Pathology University of Nebraska William Wilson Professor Department of Agricultural Economics North Dakota State University Paul Zomer Director of Bioherbicide Research Mycogen Corp.

PAGE 7

Advisory Group on Agricultural Research and Related Issues Norman Berg (retired) Formerly with Soil Conservation Service U.S. Department of Agriculture Donald Bills Deputy Area Director for Product Quality and Development Institute Agricultural Research Service U.S. Department of Agriculture Patrick Borich Director Cooperative Extension Service University of Minnesota Lucas Calpouzos Professor School of Agriculture California State University Mary Carter Associate Administrator Agricultural Research Service U.S. Department of Agriculture Neville Clark Former Director Texas Agricultural Experiment Station Texas A&M University Arnold Denton Senior Vice-President Campbell soup co. Chester T. Dickerson Director of Agricultural Affairs Monsanto Co. Catherine Donnelly Associate Director Agriculture Experiment Station University of Vermont Jeanne Edwards Past Member National Agricultural Research and Extension Users Advisory Board W.P. Flatt Dean College of Agriculture University of Georgia Ray Frisbe IPM Coordinator Department of Entomology Texas A&M University Paul Genho National Cattlemens Association Deseret Ranches of Florida Robert Heil Director Agricultural Experiment Station Colorado State University Jim Hildreth (retired) Formerly with Farm Foundation Vemer Hurt Director Agricultural and Forestry Experiment Station Mississippi State University Terry Kinney (retired) Formerly with Agricultural Research Service U.S. Department of Agriculture Ronald Knutson Professor Department of Agricultural Economics Agricultural and Food Policy Center Texas A&M University William Marshall President Microbial Genetics Division Pioneer Hi-Bred International, Inc. Roger Mitchell Director Agricultural Experiment Station University of Missouri Lucinda Noble Director Cooperative Extension Cornell University Susan Offutt Executive Director Board on Agriculture National Research Council Dave Phillips President National Association of County Agents Jack Pincus Director of Marketing Michigan Biotechnology Institute Dan Ragsdale Former Research Director National Corn Growers Association Roy Rauschkolb Resident Director Maricopa Agricultural Center University of Arizona Alden Reine Dean Cooperative Agricultural Research Center Prarie View A&M University Grace Ellen Rice Legislative Director American Farm Bureau Federation Bob Robinson Director of Agricultural and Trade Analysis Division Economic Research Service U.S. Department of Agriculture Jerome Seibert Economist Department of Agricultural Economics University of California Keith Smith Director of Research American Soybean Association William Tallent Assistant Administrator Agricultural Research Service U.S. Department of Agriculture Jim Tavares Associate Executive Director Board on Agriculture National Research Council Luther Tweeten Anderson Chair for Agricultural Trade Department of Agricultural Economics and Rural Sociology Ohio State University Walter Walla Director Extension Service University of Kentucky Tim Wallace Extension Economist Department of Agricultural Economic-s University of California Mike Wehler Vice President National Pork Producers Council John Woeste Director Cooperative Extension Service University of Florida Fred Woods Agricultural Programs Extension Service U.S. Department of Agriculture NOTE: OTA appreciates and is grateful for the valuable assistance and thoughtful critique by the advisory group members. The groups do not, however, necessarily approve, disapprove, or endorse this report. OTA assumes full responsibility for the report and the accuracy of its contents. vii

PAGE 8

Chapter 1 Overview and Summary Photo credit: Grant Hellman, inc.

PAGE 9

Contents Page ADVANCING TECHNOLOGIES FOR AGRICULTURE . . . . . . . . . Biotechnology . . . . . . . . . . . . . . . . . . . Advanced Computer Technologies . . . . . . . . . . . . . . IMPACTS OF THE NEW TECHNOLOGIES . . . . . . . . . . . . Production Measures . . . . . . . . . . . . . . . . . . Agribusiness, Farm Labor, and Rural Communities . . . . . . . . . . Management . . . . . . . . . . . . . . . . . . . . Food Quality . . . . . . . . . . . . . . . . . . . . Intellectual Property Rights . . . . . . . . . . . . . . . . MAJOR FINDINGS AND OPTIONS . . . . . . . . . . . . . . Environmental Safety . . . . . . . . . . . . . . . . . Food Safety . . . . . . . . . . . . . . . . . . . . Public Sector Research . . . . . . . . . . . . . . . . . SUMMARY . . . . . . . . . . . . . . . . . . . . 3 3 6 8 9 9 10 11 11 12 12 19 27 31 Figure Figure 1-1. Jurisdiction and Coordination of Environmental Policy for Biotechnology-Derived Agricultural Products . . . . . . . . . . . . . . . . Tables Table Page . 14 Page 1-1. Estimates of Crop Yield and Animal Production Efficiency . . . . . . . 9 1-2. Projected Annual Rates of Growth (1990) . . . . . . . . . . 9 1-3. Federal Agencies Primarily Responsible for Food Safety . . . . . . . . 20 1-4. Total Research Funding for State Agricultural Experiment Stations, Selected Years . 28

PAGE 10

Chapter 1 Overview and Summary Technological innovation has played a significant role in transforming American agriculture in the past and again promises major impacts on the U.S. food production and processing industries. The transition from horsepower to mechanical power ( 1920 1950) boosted the productive capacity of agriculture even as farm labor requirements decreased dramatically. From 1950 to 1980 agricultural productivity rose further as chemical fertilizers, feed additives, and pesticides increased yields and helped farmers control pests and disease. Biotechnology and advanced computer systems now are ushering American agriculture into a new technological era. These technologies have the potential to increase U.S. agricultural productivity and competitiveness, enhance the environment, and improve food safety and quality. Many of the new technologies will be commercially viable in the 1990s. However, they will not automatically be put to use. Todays public increasingly questions whether technological change is always good or needed and is voicing new concerns about the safety of the food supply, the environment, and the changing structure of agriculture. These issues as well as declining public confidence in institutions create an atmosphere in which agricultural biotechnology may not readily be approved for commercial use or adopted by industry. Lack of public acceptance could prevent some technologies from being used even if they are approved by regulatory agencies. To avoid this fate, agricultural biotechnology must meet rigorous scientific standards of safety and efficacy. And, institutions regulating these products must satisfy unprecedented demands for accountability. This report focuses on the new technologies for agriculture and the related issues that policy makers most likely will face during this decade. Part I identifies advances being made in agricultural biotechnology for crops, animals, and food processing, and in computer technologies to improve agricultural management. Part 11 analyzes ways in which these technologies might improve agricultural productivity and discusses certain adjustments that industry will need to make to capitalize on this potential. Part 111 considers scientific and institutional issues relevant to environmental benefit and risk assessment of biotechnology. Part IV focuses on food safety and quality issues, presenting institutional, scientific, and public perspectives on these issues. Finally, Part V analyzes some of the implications of the technologies for intellectual property rights and science policy. ADVANCING TECHNOLOGIES FOR AGRICULTURE Biotechnology Biotechnology, broadly defined, includes any technique that uses living organisms or processes to make or modify products, to improve plants or animals. or to develop microorganisms for specific uses. It rests on two powerful molecular genetic tools: recombinant deoxyribonucleic acid (rDNA); and cell fusion technologies. Using these techniques, scientists can isolate. clone, and study the structure of an individual gene and explore the genes function. Such knowledge and skills allow scientists to exercise new control over biological systems, leading to significant improvements in agricultural plants and animals. Plant Technologies Each year in the United States, weeds, insects, and disease (as well as weather and soil conditions) significantly decrease potential crop yields and cost farmers billions of dollars in lost revenues. New approaches to control pests include the use of biological agents to manage pests and the application of biotechnology to produce plants with new genetic characteristics. Biological control of pests is the use of living natural enemies to reduce pest populations to levels lower than would otherwise occur. The classical (searching native lands for control agents to pests of foreign origin) and augmentation (periodic release of control agents to increase populations) approaches are the most commonly used biological control tactics. To date, biological control has been most successfully used in orchards and vegetables; efficacy in field crops has been limited. Insect and weed control using biological control agents has been most successful; use of biological agents to control disease is lagging. Traditional selection and breeding approaches, as well as new biotechnology approaches are being used to improve the control and range of biological control agents. Several biocontrol agents currently are available or could be in the next 10 years, but the field is not sufficiently advanced to replace most pesticides in that time. New tissue culturing and genetic engineering tools combined with traditional agricultural research methods are allowing scientists to alter plants to have greater dis-3

PAGE 11

4 l A New Technological Era for American Agriculture ease, insect, and weed resistance; to withstand environmental stresses such as cold, drought, and frost; to develop value-added products from agricultural commodities; and to improve understanding of plant resistance and of the interactions of plants, pests, and biological control agents in the agro-ecosystem. Genetic Engineering of Plants for Insect Control Traditional breeding programs have and will continue to produce insect-resistant or insect-tolerant varieties of crops. However, the tools of biotechnology can be used to selectively engineer plants for this trait. Candidate genes must code for proteins that are stable in the plant cell and insect midgut; have high activity against target insects; and are safe for non-target invertebrates and animals. Genes coding for trypsin inhibitors and for bacterial Bacillus thuringiensis (Bt) toxin are two possible candidates. The gene coding for the Bt toxin has been cloned and inserted into plants; transgenic plants producing Bt toxins are expected to be commercially available by the mid to late 1990s. Genetic Engineering of Plants for Weed Control Improved understanding of the mechanisms of action of herbicides is leading to the improved ability to design herbicides effective against some plants (target weeds) but inactive against others (nontarget weeds or crops). The lack of naturally occurring resistance genes in crops Photo credit: Richard Nelson, Samual Roberts Noble Foundation Transgenic tomato plant expressing the coat protein gene of tobacco mosaic virus (left) and control plant (right). limits the ability to use traditional breeding methods to develop herbicide tolerant crops; however genetic engineering techniques can overcome these constraints. The first herbicide tolerant crops are expected to be available commercially by the mid 1990s. Genetic Engineering of Plants for Disease Control Biotechnology is being used to elucidate the mechanisms by which pathogenic organisms cause disease and to engineer plants with enhanced disease resistance. Genes coding for virus coat proteins (i. e., the proteins that make up the shell that surrounds viruses) can be genetically engineered into plants to elicit resistance to infection by the source virus, and in some cases to related viruses having similar coat proteins. Several plant viral coat proteins have been transferred to plants to confer resistance. Genetically engineered dicotyledonous plants resistant to certain viruses are expected to be available commercially by the mid 1990s. But virus resistant monocotyledonous plants will probably not be available until the late 1990s or early 21st century. Plants resistant to bacteria and fungi are not expected to be developed until the end of the decade and not available commercially until after the year 2000. Animal Technologies Biotechnology has the potential to improve feed efficiency, reduce losses from disease, and increase reproductive success in all sectors of the livestock industry. Advances in growth promotants, reproductive technologies, and animal health will play a major role in enhancing the efficiency of animal agriculture and the quality of its products. Growth Promotants-Currently used growth promotants such as anabolic steroids and antimicrobial compounds will continue to be used in the livestock sector. However, rDNA techniques are being used to produce new products such as a new class of protein hormones called somatotropins. Porcine SomatotropinPigs administered porcine somatotropin (pST) for a period of 30 to 77 days show increased average daily weight gains of approximately 10 to 20 percent, improved feed efficiency of 15 to 35 percent, decreased adipose (fat) tissue mass and lipid formation rates of as much as 50 to 80 percent, and concurrently increased protein deposition of as much as 50 percent without adversely affecting the quality of the meat. Prolonged release formulations and daily injection produced similar growth rates and feed efficiencies. PST is currently being reviewed by Food and Drug Administration (FDA) for commercial use.

PAGE 12

Chapter lOverview and Summary .5 CONTROL T Photo credit: Terry Etherton, Pennsylvania State University Comparison of pork loins that show the effect of pigs treated with porcine somatotropin (pST). The loin-eye area of the loin treated with pST is 8 square inches; the control is 4.5 square Bovine SomarotropinBovine somatotropin (bST) is currently undergoing FDA review for use in lactating dairy cows to increase milk production. While individual gains rely on the management ability of the producer, on average, gains of about 12 percent are reasonable. Bovine somatotropin does not alter the composition of milk. The fat, glucose, protein, mineral, and vitamin composition of milk fall within the range of values normally observed in milk from cows not supplemented with bST. Bovine somatotropin decreases pregnancy rates (proportion of cows becoming pregnant), increases days open (days from parturition to conception), but does not alter conception rates (services per conception). These observed effects are similar to those occurring in highproducing cows that do not receive bST. Implications of using bST in dairy production are discussed more thoroughly in the OTA publication U.S. Dairy Industry at a Crossroad: Biotechnology and Policy Choices. Reproduction TechnologiesThe field of animal reproduction is undergoing a scientific revolution. For example, in the cattle industry it has become possible to induce genetically y superior females to shed large numbers of eggs (superovulation); and to fertilize these eggs in vitro with the sperm of genetically superior males. Each resulting embryo can then be sexed and split to produce multiple copies of the original embryo. Each of these new embryos can then be frozen for later use, or transferred to a recipient cow whose reproductive cycle has been synchronized to accept the developing embryo. The recipient cow carries the embryo to term and gives birth to a live calf. It may be possible in the near future to sex the sperm rather than the embryo, and to create more inches. copies of each embryo than currently is possible. New techniques being developed will make it easier to insert new genes into the embryos to produce transgenic animals. Embryos produced by new reproductive methods are being marketed, although as yet no transgenic animals are available. Transgenic AnimalsThe combination of new reproductive technologies with recombinant DNA technologies (the identification. isolation, and transfer of selected genes), provides opportunities to produce transgenic animals efficiently and cost effectively, and to improve livestock quality more rapidly than could be done with traditional breeding. Some transgenic livestock may contain genes that improve growth characteristics or resistance to disease. These new developments also have human medical implications. It may be feasible to produce important human pharmaceuticals in livestock. Transgenic animals can also serve as a powerful research tool to understand genetic and physiological functions, and to provide a model system to study human disease. For example, pigs display striking physiological similarities to humans and because of this, transgenic pigs are currently being developed to serve as a model system to understand and treat gastrointestinal cancers. Commercial availability of transgenic animals is not expected before the year 2000. Animal Health Technologiesimprovements in animal health will lead to considerable cost savings to the animal industry. Biotechnology rapidly is acquiring a prominent place in veterinary medical research. New vaccines include those created by deleting or inactivating

PAGE 13

6 l A New Technological Era for American Agriculture the genes in a pathogen that cause disease. The first genedeletion viral vaccine to be approved and released for commercial use was the pseudorabies virus vaccine for swine. Many currently used diagnostics tests are costly, time consuming, and labor intensive, and some still require the use of animal assay systems. Monoclinal antibodies and nucleic acid hybridization probes can be used to produce simpler, easily automated, and highly sensitive and specific diagnostic procedures. At least 15 different rapid diagnostic tests based on monoclinal antibodies are on the market or soon will be. Food Processing Technologies Historically, the food processing industry has had to accept and adapt to heterogeneous raw materials. Biotechnology can be used to tailor food crops to meet food processing and consumer needs. For example, new plant tissue culture techniques can be used to produce food flavor and coloring ingredients. These methods potentially could replace production and extraction of these ingredients from plants. Genetic engineering can also be used to alter food characteristics. Genes coding for enzymes involved in starch and lipid biosynthesis are being isolated and cloned, enhancing the prospects of engineering plants with specific compositions of starch and oil. And, genetic engineering is being used to eliminate toxins, allergic compounds, or off-flavor components in plants, and to delay ripening of tomatoes. New biotechnology products are being developed for food manufacturing and monitoring of animal products for food safety. For example, a genetically engineered version of the enzyme rennet, which is normally extracted from the forestomach of calves, has recently been approved by FDA for use in cheese manufacturing systems. Bacteria and yeast strains engineered to convert waste products such as blood, bone, and milk whey into useful products could decrease the costs associated with their disposal. For example, engineered yeast strains are capable of fermenting the lactose in whey to value-added products, such as vitamin C, biofuels, or pharmaceuticals. Food safety monitoring will be enhanced by the development of nucleic acid probes and monoclinal antibodies; raw materials, ingredients, and finished products can be analyzed for the presence of pathogenic organisms and chemical and biological contaminants. Detection kits are also commercially available for monitoring several pesticides, antibiotics, and bacterial contaminants. Advanced Computer Technologies Since the industrial revolution, agricultural systems have intensified, and agricultural productivity has increased significantly along with farm size. Labor-saving devices on farms have increased output per worker several fold, and advances in understanding and application of biological principles have boosted agricultural yields significantly. With increased production, however, farm management becomes correspondingly more challenging and complex. In general, methods for making management decisions have failed to meet this challenge. As a Photo credit: Calgene, Inc Tomatoes with genes that delay ripening (left) and control (right) 3 weeks after harvest.

PAGE 14

Chapter 1Overview and Summary l 7 result, many decisions are uninformed and many agricultural systems poorly managed. The application of advanced computer technologies to agricultural management can help remedy this situation. Improved access to information will allow farmers to monitor progress more effectively and to determine suboptimal performance. For example, close monitoring of animal performance will allow early detection of diseases and can help reduce stress in animals. Overall, advanced computer technologies can provide managers with the ability to systematically determine the best decision rather than arrive at decisions in an ad hoc fashion. Optimal decision making requires a holistic view of a farm enterprise, factors that affect it, and probable consequences of management decisions. Thus, a farmer deciding whether to plant a specific crop on a specific field should weigh the profitability of the crop as well as overall farm needs (e. g., nutrition requirements if it is an animal enterprise). The decision will impact land sustainability and the need to use certain pest-control strategies. By-and-large, computers have had little impact on production agriculture to date. Predictions that every farmer would own a computer by 1990 have not come true. Few farmers have computers and those that do use them primarily for book keeping and general calculations (e. g., ration balancing). The largest impact of computers in American agriculture has been in support industries. Using computer networks and tracking systems, equipment dealers can provide faster service, and feed dealers are better able to manage feed inventories. Most of these advances have come from directly adopting general business software with little or no input from the agricultural academic community. The primary agricultural application of advanced computer technology by the mid 1990s will be ad hoc expert systems (i.e., computer programs that use knowledge to solve well-defined problems). Problem diagnosis expert systems currently are under development, and farmers will have a cadre of these systems at their disposal to diagnose diseases and to evaluate production performance. These systems generally will not be integrated with one another and each will consider only one aspect of a problem. Integrated systems that solve production problems while considering economic consequences will not become available until the later part of the decade. The primary use of expert systems within the next 5 years may be by agribusiness which will be able to leFarmer and consultant examine data from COMAX (COtton Management eXpert) computer program. verage the cost of adopting these technologies across a number of farms. Using expert systems to increase service to farmers may change the role of some professionals. For example, expert systems can help veterinarians take an epidemiological approach to solving problems. It will also allow some diversification in services provided. For example, animal nutritionists may be more likely to become involved in consulting for the crop program when aided by an expert system. Computer-based sensors will be used on a limited basis to collect real-time data for expert systems. The primary use of sensors will be for monitoring weather and field conditions for crop management. Expert systems will help farmers interpret these data and suggest appropriate management strategies such as irrigation, fertilization, or pesticide treatment. Another technology likely to see application by the mid1990s is full-text retrieval systems. It will be possible for farmers and Extension personnel to have a CDROM with all of the latest publications at their fingertips. Using a full-text retrieval system, they will be able to retrieve pertinent information that will help them improve their decisions. For example, when a farm experiences a corn mycotoxin problem, the owner-operator can access an information base to find relevant literature. Robots for highly specialized, labor-intensive tasks will begin to be applied to agriculture in the late 1990s. This would include robot transplanting of seedlings, pork carcass sectioning, and harvesting of fruits and vegeta-

PAGE 15

8 l A New Technological Era for American Agriculture Photo credit: Gerald Isaacs, University of Florida An experimental fruit picking robot uses a machine sensor and a computer to locate individual fruit for detachment. Approximately 3 seconds per fruit are required. bles. Robots for milking cows, however, may reach commercial application by the mid-1990s. IMPACTS OF THE NEW TECHNOLOGIES The new era of biotechnology and advanced computer technologies will be faster paced than previous technological eras. A more rapid pace of technological change will be fostered by major changes in public policy regarding technology. One of the most important changes was the granting of property rights for new plant varieties, new life forms, and computer software. Patent rights were extended to new plant varieties by the enactment of the Plant Variety Protection Act of 1970. This was followed in 1980 by the U.S. Supreme Court ruling in Diamond vs. Chakrabarty that investors in new microorganisms, whose inventions otherwise met the legal requirements for obtaining a patent, could not be denied a patent solely because the innovation was alive. This decision opened the door to patent a broad range of potential new products of the biotechnology era. Capping this series of policy changes was the amendment to the Copyright Act in 1980 that made explicit provisions for computer programs as (literary) works of authorship. In previous technological eras most technologies were capital intensive and substituted for labor and land. Many emerging biotechnologies will substitute for conventional purchased inputs. For example, biopesticides will replace some chemical pesticides in plant insect control, biotechnology-improved animal disease vaccines likewise will replace some existing vaccines. On the other hand, some biotechnologies will compliment existing technologies. An example is the genetic transformation of plants to incorporate desired traits. In this case, conventional plant breeding will still be required for incorporation of biotechnology-induced traits into commercial lines, for continued plant improvement selection, and for seed multiplication. In addition, for the foreseeable future, chemical fertilizers will remain important in crop production. As with past technological eras, successful adoption of specific biotechnology innovations will result in additional profits for some, at least the early adopters. As in the past, increased profits will result mainly from reductions in real production costs per unit of output. This, in turn, can increase productivity and the competitive position of U.S. agriculture. As with past technological innovation, biotechnology is expected to be supply-increasing in the aggregate. The implications, however, can be quite different for different farms. Late adopters of the new technology, for example, will be faced with lower product prices. This is because early adopters have already reduced their production costs, enjoyed increased profits in their period of initial adoption, and are ready to respond to the next wave of technological innovation. Increased supplies are generally associated with lower prices. Consequently, nonadopters often have higher costs while facing lower prices for their products. Successful use of technologies of this new era most likely will require changes in the production process and may require a higher quality of management. This may mean increased human as well as monetary capital. Less educated farmers with limited capital resources may find it difficult to implement the new technology successfully. Thus, the new technologies may widen the gap between capital-limited and capital-rich farm operators. Many advancing technologies are approaching commercialization. In crop agriculture, biotechnology research has advanced at a much faster rate than anticipated just a few years ago, and transgenic crops are currently undergoing field trials. in animal agriculture, vaccines and diagnostics are on the market or will be soon. Growth promotants are going through the regulatory process. Reproduction technologies are advancing at a rapid pace and cloned embryos are currently being marketed. Transgenics are still in the future but considerable strides are being made in the use of livestock to produce high value pharmaceuticals. These technologies and others will impact agriculture in a number of ways.

PAGE 16

Chapter 1Overview and Summay) l 9 Table l-lEstimates of Crop Yield and Animal Production Efficiency by 2000 Less new Most likely More new Actual technology technology technology 1990 2000 2000 2000 Crops Cornbu/acre . . . . . . Cotton-lb/acre . . . . . . Soybeansbu/acre . . . . . Wheatbu/acre . . . . . ., ., Beef Lbs meat/lb feed . . . . . . Calves/100 cows . . . . ... Dairy Lbs milk/lb feed . . . . . . Lbs milk/cow/year . . . . . Poultry Lbs meat/lb feed . . . . . . Eggs/layer/year . . . . . . Swine Lbs meat/lb feed . . . . . . Pigs/sow/year . . . . . . 116.2 600.0 32.4 34.8 113.8 NA 32.6 37.7 128.5 708.0 33.7 42.6 141.6 NA 36.4 53.8 0.143 90.0 0.146 93.750 0.154 96,221 0.169 102.455 1.010 14,200.0 1.030 17,247.200 1.050 19,191.600 1.057 20,498.800 0.370 250,0 0.373 250.500 0.389 258.0 0.428 273.125 0.154 13.900 0.174 14.420 0.181 15.750 0.196 17.791 NOTE: OTA expresses its appreciatio nto Yao-chi Lu in deriving the estimates for this table, NA = Not available. SOURCE: Office of Technology Assessment, 1992, and Phil Coiling, Agriculture Research Service, U.S. Department of Agriculture, for their assistance Table l-2Projected Annual Rates (1990-2000) of Growth agriculture is in milk production. Since 1960, the annual rate of growth has been about 2.0 to 2.5 percent. OTAs 1985 projection (24,200 pounds of milk per cow by year 2000) was higher than its current one (19,200 pounds of milk per cow by year 2000). A major reason for this change is the slowness to market of bovine somatotropin. In 1985, bST was predicted to be commercially available in 1987. BST had yet to be approved by the Food and Drug Administration as of early 1992. Less new Most likely More new technology technology technology Corn Cotton Soybeans Wheat Beef Lbs meat/feed Calves/cow. Dairy Lbs milk/feed Milk/cow/year. Poultry Lbs meat/feed Eggs/lay/year. Swine Lbs meat/feed Pigs/sow/year. 0.210/0 NA 0.06 0.80 1.000% 1.66 0.39 2.02 1.97% NA 1.16 4.36 0.21 0.41 0.74 0.67 1.67 1.30 . . . Efficiencies in crop production will about match historical trends or climb slightly, and for the most part will exceed OTAs 1985 projections. This, in part, reflects the movement of many of the new technologies from the laboratory to the field at a much quicker pace than thought possible in the mid1980s. Even though rates of growth may accelerate during the 1990s, the absolute quantity of yields will, for the most part, be lower than projected in the mid1980s. This is due, in part, to the fact that many of the early biotechnology inputs will be substitutes for chemical inputs and, hence, the absolute gain in efficiency will in many cases be negligible. Yields are expected to improve in the latter part of the decade as more is learned about the genetic make up of plants. 0.20 1.94 0.39 3.01 0.46 3.67 . . . 0.08 0.02 0.51 0.32 1.46 0.89 . . . 1.22 0.37 1.62 1.25 2.41 2.47 . . . NOTE: OTA expresses its appreciation to Yao-chi Lu and Phil Coiling, Agriculture Research Service, U.S. Department of Agriculture, for their assistance in deriving the estimates for this table. NA = Not available. SOURCE: Office of Technology Assessment, 1992. Production Measures Agribusiness, Farm Labor, and Rural Communities The advance of agricultural biotechnology will play an important role in increasing agricultural productivity at about the historical rate of the last two decades. (See tables 1I and 1-2. ) The most dramatic increase in animal Historically, the commodity-oriented agribusiness sector has been driven by economic forces to produce at

PAGE 17

. 10 l A New Technological Era for American Agriculture maximum efficiency and to maintain low costs. This has resulted in a system that is effective at converting undifferentiated commodities into low-cost food. Today this sector is undergoing change inspired, in part, by the evolution of more demanding and differentiated food consumers. in response, retailer strategies have emerged that focus on improving service to the consumer. information technology has facilitated the shift of marketing efforts toward the discovery of consumer preferences. To respond to a more consumer-oriented environment, input suppliers may need to explore how information technology can facilitate the coordination of activities needed to assure particular attributes. Information technologies in the future may facilitate new business strategies by providing improved information flows and by facilitating coordination of production and marketing activities. To date, input suppliers have experienced more consequences of the new technologies than any other part of the agricultural industry. In anticipation of biotechnology-enhanced seed, chemical and seed input industries have transformed structurally. Multinational chemical and pharmaceutical companies have acquired almost all the major seed companies. Concentration of input industries increases the potential for monopoly power, hence the potential for exploiting farmers in their purchase of improved inputs. The trend toward vertical integration in agriculture and toward proprietary production processes could result in a captive market for some biotechnology products. For example, a genetically engineered seed might be produced by a large, vertically integrated chemical-seed company with specified inputs such as fertilizer, pesticides, and herbicides produced only by that company. Where product quality is influenced strongly by biotechnologies (i. e., pork by pST); and where highly specialized new markets are formed (i. e., for pharmaceuticals), increased incentives for production-marketing links via contracting and other forms of vertical integration can be expected. The advancing biotechnology and information technologies generally will shift labor from farming as has been true of past technologies. Newly emerging technologies will displace less farm labor than mechanization, but the farm labor force will have to be substantially more skilled than in the past. For example, a key requirement of the new information technology will be computer literacy. Programs to support skill upgrading of the farm labor force will be needed to capture fully the potential benefits of the new technologies. Photo credit: Grant Heilman, Inc. Production of lean meat with porcine somatotropin (pST) will give meat packers a strong incentive to vertically integrate or contract with farmers. Economic pressures will be strong for most swine producers to adopt pST or exit the industry. The emergence of biotechnology and computer technologies will most likely spur on the decline of many small farms and agriculturally dependent rural communities. Moreover, increased demand by many farmers for one-stop shopping centers for farm suppliesincluding those involving biotechnologies and information technologiesmay reduce the viability of business enterprises in smaller communities. These enterprises will need to diversify into nonfarm-related economic activities if they are to remain economically viable. Management The new technologies will demand greater attention to management issues than have technologies in the past. For crop agriculture, in particular, a systems approach to the use of genetically engineered plants and biocontrol technologies will be needed. Concern about pest resistance to technologies that control pests is reaching a high level. Many chemical technologies are ineffective today because of pest adaptation caused by poor management strategies. As products from biotechnology are used to control pests, management strategies for delaying or pos-

PAGE 18

Chapter lOverview and Summary l ll sibly avoiding pest adaptation need to be identified. Evidence exists already that insects are quite capable of adapting to Bt, one of todays most popular genetically engineered protein toxins. At present, there is some information to establish general guidelines about the judicious use of engineered crops with insect and pathogen resistance and herbicide tolerance. However, to establish more detailed guidelines will entail generating a body of empirical knowledge relevant to these products. And, an effective educational program designed to bring these results to the agricultural industry and the public is needed. For animal agriculture research results clearly show the extent of response achieved from technology depends heavily on the management capability of the producer. Use of somatotropins, for example, may require altering the animals diets. Administration of somatotropin to lactating cows may require extending the reproductive cycle. As important as these management issues are, a more pressing issue is that of animal welfarewith or without biotechnology as a complicating factor. Much of the success in increased productivity in agriculture has been the result of lowered costs through the use of confinement systemswhich some have coined factory farming. The question from an animal welfare perspective is whether we have gone too far. The impact of biotechnology on animal well-being is perhaps the most challenging issue genetic engineering raises. The technology is most likely impact neutral in that one could use biotechnology to enhance animal wellbeing as well as compromise it. Clearly, biotechnologys impact depends on what is done and its effect. If it is used judiciously to benefit humans and animals, with foreseeable risks controlled, and the welfare of animals kept in mind, it is morally defensible great benefit. Food Quality Information about food quality can be and can provide provided through labeling, brand names, price, and grades. Food grades are used to classify products according to certain quality characteristics and are established by the U.S. Department of Agriculture (USDA). In particular, they sort a group of foods with heterogeneous characteristics into lots of more uniform characteristics. Biotechnology will challenge the relevance of grades since this new technology is capable of producing products of uniform high quality. For example, as discussed above, pST reduces backfat thickness and increases protein deposition in hogs, resulting in a final product that is more desirable to a health conscious society. Current USDA grading criteria based, in large part, on backfat thickness and degree of marbling will not be relevant since there will be little. if any, difference from animal to animal in these characteristics in products produced with the new technology, For a grading system to be useful, new grading criteria will be needed. What these new criteria should be and how they will be measured are open to question. An argument can be made for providing quality information via labels to consumers and dispensing with USDA grades for most, if not all, agricultural products. Intellectual Property Rights Intellectual property protection is one of the most important incentives for the commercial development of biotechnologyand computer-related processes and products. Patents and other forms of intellectual property (plant breeders rights, trademarks) provide this protection. Patents may be issued in the United States for microorganisms, plants, and nonhuman animals. U.S. patent law is the most inventor-friendly statute in the world: if Congress takes no action regarding patentable subject matter, broad protection for inventions created by biotechnology will continue. The Patent and Trademark Office (PTO) issued its first patent on an animal in 1988. No further patents have been issued since, and the backlog of applications at PTO now numbers at least 160. Since the status of patent applications is, by law, confidential, no way exists to determine when or if the patent office will issue subsequent animal patents; or whether such patents will have agricultural applications. Congress, through its oversight responsibilities, may require PTO to explain the present status of any such patent applications. Rapid technological advances in computer software is challenging the intellectual property laws in the United States and internationally. Copyright law offers straightforward remedies for the literal copying of program code, although enforcement remains a problem. Functional aspects of computer programs pose difficult questions for application of copyright. The protection of software-related inventions by patent is a fairly recent development and is controversial. PTO faces considerable challenges in examining applications for computer-related inventions. An incomplete data base of prior art for computer-related inventions makes it difficult for examiners to judge whether an application describes a novel invention. Improving the database of prior art is one important means of improving the quality of the examination but will be difficult because so much of what

PAGE 19

12 l A New Technological Era for American Agriculture constitutes prior art has been in the form of products, not literature or issued patents. MAJOR FINDINGS AND OPTIONS For any new technology, it is important to weigh the potential benefits against the risks and possible costs of its widespread adoption. Biotechnology-related risk assessment focuses on the planned introduction of genetically modified organisms into the environment (environmental safety) and on the consumption of products derived from biotechnology (food safety). In many ways this is a difficult time for a new technology to emerge. Negative experiences with nuclear and chemical industries have made the American public wary of new technologies, and confidence in institutions has eroded. For these reasons, and because the consequences of environmental introductions of genetically modified organisms cannot be predicted with certainty, biotechnology has been subjected to extensive, apprehensive scrutiny and regulatory oversight. Many institutions will choose to go the extra mile to ensure public confidence as some policy issues are resolved. In making policy decisions it remains important, nonetheless, to distinguish clearly between the technical basis for assessment and regulation of technology-related risks, and what might or might not be done as an extra step to maintain public confidence. Balancing safety and institutional credibility against economic competitiveness will be a skill much in demand throughout the decade. Environmental Safety Findings Adequacy of a Knowledge Base for Risk Assessment Analysis-After several years of experience with planned introductions, a consensus is growing among scientists that the risks of planned introductions of genetically modified organisms into the environment can, for the most part, be assessed with available analytical capabilities. Although risk assessment is itself a relatively young field, the capacity to identify and weigh risks and benefits in a structured and analytical way has matured rapidly in recent years. Based on experience with other technologically oriented issues such as pollution and its control and food safety, risk assessment as a field has generated principles and methodologies that can be adapted for planned introductions of recombinant-DNA modified organisms in the environment. The fields of community ecology, population biology, population genetics, evolutionary theory, and agricultural sciences as well as others have contributed to our current understanding of the ecology of planned introductions. Decades of research in life history dynamics, competition, characteristics of colonizing species or disturbed habitats, disease resistance, and gene flow have provided a basis for risk assessment of planned introductions. Thus, while it is impossible to assess the exact consequences of any specific planned introduction, the fact remains that ecological understanding combined with risk assessment methodologies make it possible to analyze the potential risk of each introduction before it is allowed to take place. Adequacy of a Knowledge Base for Science-Based, Risk-Based Regulations-Reports of the National Research Council, the Ecological Society of America, and the Scope document of the Office of Science and Technology Policy (OSTP) and the Council on Competitiveness all advocate science-based and risk-based regulations of biotechnology applications. The implementation of such regulations draws on the ability of regulators to conduct adequate risk assessments, which in turn rests on the knowledge base and technical capabilities discussed above. Regulatory oversight rests with Federal agencies, with varying degrees of involvement by state regulatory personnel. USDAs Animal and Plant Health Inspection Service (APHIS) has taken the lead in designing a process for the evaluation of possible risks and benefits when a specific planned introduction of a genetically engineered plant is proposed. Technical information to be provided by an applicant is clearly defined, so that a thorough, science-based risk assessment can be performed. Technical personnel in fields such as genetics and ecology have joined the staff of APHISs Biotechnology, Biologics, and Environment Program (BBEP), to ensure vigorous assessments. State regulatory personnel are drawn into the process so that they can provide additional technical information specific to local habitats and add an additional perspective. The Environmental Protection Agencys (EPA) Office of Pesticide Programs (OPP) has extended its review processes under the Federal Insecticide Fungicide and Rodenticide Act (FIFRA) to planned introductions of microbial pesticides; it also cooperates with USDA-APHIS in reviewing proposals for introduction of pest-resistant plants. EPAs Office of Toxic Substances (OTS) has recently published draft regulations to cover planned introductions of genetically modified microorganisms; significant controversy exists as to whether these regulations are indeed scienceand risk-based, or whether they sim-

PAGE 20

Chapter lOverview and Summary .13 ply single out biotechnology for attention because it is biotechnology. The final status of these regulations, as well as their implementation processes, is not yet known. State agencies have yet to be pulled into EPA regulatory processes to the extent accomplished by USDA. Extent That Regulations Are Product-BasedReports of the National Research Council and the Ecological Society of America stated that the techniques of biotechnology are not themselves inherently risky or unmanageable. In line with these findings, the early Coordinated Framework, the document that established responsibilities of Federal agencies that regulate biotechnology derived products, and the principles put forth by OSTP and the Council on Competitiveness recommend that biotechnology not be regulated as a process. Rather, a central tenet for biotechnology regulation is that the various products of biotechnology should be regulated, just as are products of other technologies. The product/process distinction has generated a great deal of controversy in the past. However, as the experience base with biotechnology has grown, the premise of judging each product on its own basis rather than automatically implementing special regulations, has gained wide acceptance. The extent to which this premise has been implemented, however, varies among agencies. Though its focus is on plant pests, USDA-APHIS has been able to include along with other organisms under its purview any vector, vector agent, donor organism, recipient organism, or any other organism or product produced through genetic engineering if it can be defined as a pest. This product-selective approach makes it possible for regulated articles to become exempted from special review as evidence indicates their safety. Under FIFRA, EPA-OPP also has applied an existing mandate to products of biotechnology, specifically plants engineered to produce compounds aiding them in resisting pests. By pulling these pesticidal plants under the rubric of its oversight for pesticides, EPA-OPP seems in one sense to be focusing on the product rather than the process by which it was generated. However, a question exists as to whether or not *pesticides is the appropriate category into which to place these particular products, especially since naturally occurring plants produce some anti-insect compounds (see next section). To assume authority over plants genetically modified to be resistant to pests, EPA-OPP seems to have chosen to look only at plants that have gone through a biotechnology process, leaving naturally-occurring pest-resistant plants alone. Under the Toxic Substances Control Act (TSCA), EPAOTS has promulgated draft regulations for oversight of microorganisms that do not fall under other authority. However, under these draft regulations, essentially all microorganisms other than those modified through biotechnology techniques are automatically exempted from review, whereas those modified through biotechnology techniques are labeled new and therefore subject to regulation. When the only products subjected to special review are biotechnology products, a question arises as to whether or not the regulations are contradicting the scope principles by focusing on process. The draft regulations under TSCA have been charged by some with automatically and unfairly assigning a special riskiness to organisms modified through biotechnology, while exempting organisms that are known to be potentially dangerous but not produced through a biotechnology process. This discrepancy, and perhaps its final resolution, underscores a central tenet of regulationthat regulation should be based on scientifically determined risk. Appropriate Review Authority for Plants Genetically Modified for Pest ResistanceUnder the Coordinated Framework (figure l-l), which established the responsibilities of Federal agencies with regard to biotechnology, EPA-OPP took on authority for plants into which genes coding for compounds toxic to insects had been introduced. The premise was that these were special pesticidal plants that presented risks to the environment, food, and human health similar to traditional chemical pesticides applied externally in large volumes to plants. This premise is questioned for several reasons. Compounds toxic to insects that are part of plant tissue do not cause pesticide run-off and other such environmental problems (so long as they are alive); they are distinctly localized. Furthermore, most of the compounds are not complex, like many synthetic compounds, and may well be more readily biodegradable. Another key argument with the premise of singling out plants genetically modified for enhanced resistance to pests is that all plants have natural pest resistance characteristics. Selection pressures over evolutionary time have favored the spread of genes in natural populations that code for characteristics unattractive or harmful to insects. Making a distinction between genetically modified plants and natural plants that are pest resistant, calling the former pesticidal plants and the latter simply plants is in fact arbitrary, not science-based. If the pesticidal plant premise is disallowed, an argument then exists that EPA-OPP is not automatically the best home for regulatory review of such plants.

PAGE 21

Figure I-l Jurisdiction and Coordination of Environmental Policy for Biotechnology-Derived Agricultural Products. a I President I Biotechology sciene National Enviromental coordinating Committee ( B scc), Policy Act (NEPA) now Biotechnology Research I Subcommittee (BRS) I coordinated framework Scope document united states Department of Agriculture (USDA) Animal and Food office of Plant Heatlh safety and Agricultural Inspection Inspection Biotechnology/ Service Agricultural Assessment (APHIS) (FSIS) Biotechnology Program Research (NBIAP) Advisory Committee (OAB/ABRAC) I I I USDA statutes EPA statues FDA statues Plant Pest Act FederaI Insecticide, Fungicide, Food, Drug, and Cosmetic Act (FDCA) Plant Quarantine Act and Rodenticide Act (FIFRA) Noxious Weed Act Toxic Substances Control Act (TSCA) Virus-Serum-Toxin Act Federal Meat Inspection Ad Poultry Products Inspection Act a OSTP, the Council on Competitiveness, and OMB do not have direct oversight over the Federal agencies; the connections shown here are those of influence through law, key policy documents, or review. SOURCE: Office of Technology Assessment, 1992.

PAGE 22

Chapter lOverview* and Summary l 15 Finally, EPA-OPP has in the past dealt with chemicals and, to a small but growing extent, microorganisms. These are the areas of staff expertise, for the most part, not plant ecology. The latter is the strength of USDAAPHIS. In fact, USDA-APHIS currently takes the lead in assessing applications for field trials of plants genetically modified for enhanced pest resistance. In consultation with EPA-OPP personnel, USDA plant scientists employ their plant expertise and their established review system toward this end. Although companies and universities have moved ahead and conducted tests, the unclarified status of OPPs approach to large-scale commercialization worries these parties as well as State governments. Treating all crop plants as pesticides would take an immense toll in State government time and personnel; yet States cannot plan because they have not as yet received guidance from EPA as to what is coming. Appropriateness of TSCA for Biological CommercializationCan or should a law written for chemicals, specifically TSCA, be used to cover living organisms? Essentially, this is happening as the traditional role of gap filler played by TSCA is applied to planned introductions of microorganisms used for purposes other than as pesticides. Approval for the introduction of microorganisms rests on determination that they will not harm human health or the environment. Microorganisms themselves are not toxic; neither are they likely to be applied in the volumes typical of chemical applications. Instead of persisting as do many synthetic chemical compounds, living organisms are eminently biodegradable. However, because they can potentially reproduce themselves and spread in the environment, their use brings up concerns different from those aroused by chemicals. TSCA could be stretched to cover microorganisms. However, biologically trained staff will have to be given the authority to develop the procedures and requirements of the office. Managers will have to acknowledge the differences between microorganisms and chemicals, and back up their biologically trained staff accordingly, when different treatments are devised. Paradigmatic shifts in management policy need to occur if EPA is appropriately to adapt to living organisms those laws, premises, and procedures originally designed for chemicals. EPAs ability to evidence such flexibility is questioned. Managing Risks of Large-Scale IntroductionsAs agricultural biotechnology moves toward commercialization and large-scale planned introductions, the combination of several approaches can maximize benefits and minimize risk. Technically sound implementation of science-based regulations are critical to risk management, as are technically competent regulatory personnel. In addition, specific scientific and agronomic methods are needed to manage risks of particular planned introductions. Examples are methods to reduce the chances for horizontal gene transfer or to diminish the survival potential of any non target recipient of an introduced gene. Scientists are exploring ways in which the gene of interest, or supplementary genes transferred along with it, can be designed to constrain the potential for transfer (a kind of internal, genetic containment system). Agronomic methods can also be used to manage identified risks. For example, physical or spatial barriers could be put in place between a field of genetically modified crop plants and the adjacent field or surrounding natural vegetation. While this sort of barrier would probably not be necessary in most cases, in particular cases where gene flow was of concern (perhaps for canola), this could be useful. Other mechanisms could be used as well, such as surrounding a field of genetically modified plants with barriers of a trapping species that attracts any pollinators that might otherwise carry genes from one of the modified crop plants to other plants. The actual need for such separations whether spatial, or temporalcan be determined by assessing the risk of gene flow or of establishment of genetically modified organisms. Photo credit: Grant Heilman, Inc. A traditional approach to isolation of plants is to spatially separate desired plants from other plants. Similar guidelines for spatial separation have been applied to transgenic plants as well.

PAGE 23

. 16 l A New Technological Era for American Agriculture Risks of Genetically Modified Plants or Microorganisms Becoming PestsAny novel organism potentially represents some level of risk to the environment, whether that organism is naturally occurring or genetically modified. However, the likelihood of a genetically modified plant or microorganism actually becoming a pest is relatively low. The long history of agriculture shows that current crops are not likely to become established as weeds. Long established mechanisms for containment in agricultural systems have been highly successful in the United States. Furthermore, recombinant-DNA modified organisms, unlike wild, naturally occurring organisms, are designed to exist only in a specific environmental regimethe nurturing surroundings of a cultivated field. Microorganisms modified for agricultural purposes are constrained somewhat like plants, although they are not so dependent on cultivation for continued survival. However, the extensive agricultural experience with microorganisms has not resulted in a pest problem. To become a pest, an agricultural plant or microorganism has to exist independently of cultivationoutside the planted field. Several steps are necessary to its success; each one, from dispersal to the production of viable, competitive offspring, is not likely to occur. Potential for Gene Transfer or Cross-Hybridization Between Genetically Modified Plants and Wild PlantsCross-hybridization, the crossing of two plants of different species to produce fertile offspring, is a rare phenomenon. While gene transfer between individuals of the same species is straightforward, gene transfer between different species is not; their genetic compositions are usually sufficiently different that they do not line up and match well for the key molecular and cellular events of reproduction. Even if a transferred gene were involved in such a cross, it would be cast onto an alien genetic backgroundits expression could be problematic. Most crop species in the United States do not have indigenous weedy relatives with which they could crosshybridize. Canola is the only major crop for which related weedy species exist in the United States. The possibility of cross-hybridization is greater in other countries, where crop species and related weedy species do coexist. Developing countries, in particular are the centers of origin for many crop species. As it exports agricultural biotechnology capabilities, the United States has at least a moral responsibility to provide advice to developing countries as to the management of risk from cross-hybridization. Options 1. The tools of biotechnology offer great potential to American agriculture; regulatory treatment of agricultural products derived with such tools will play a dominant role in any related gains or losses in economic competitiveness. Science-and risk-based regulation of products can help ensure safety while not impeding the economy. Congress could directFederal regulatory agencies to make science-based, risk-based regulation of biotechnology products (not process) a unifying policy across agencies. This would be a clear message to the executive branch that Congress expects a unified approach across Federal agencies based on the product not on the process of biotechnology. Communication through interagency groups would help to ensure a common approach based on scientifically determined product risk. This approach can help protect health and environment and, at the same time, should generate a comprehensive, workable regulatory apparatus for incorporating the tools of biotechnology into American agriculture. However, EPA will need to address their shortage of technical staff needed to conduct technical risk-based reviews. No scientific evidence exists to justify Congress directing agencies to review and regulate biotechnology as a process, rather than the products produced by it. Nevertheless, EPA-OTS has been accused of regulating the process of biotechnology, not the products, in its proposed rules. If agencies were to ignore the use of risk assessment of products and automatically penalize any efforts made using biotechnology, several impacts would likely occur. Industries and universities probably would agency -shop, orienting their efforts toward the agency with the clearest analytical assessment of science-based risksthat agency will be the least arbitrary and the most predictable, an approach certainly favored by industry. The agency regulating biotechnology as a process sends out an obvious negative message to industry and perhaps an equally important, if more subtle, message to the public. Regulations based on the assumption that biotechnology is inherently unpredictable and highly risky can lead to reverse public reactions and political pressures that may be detrimental to the economic competitiveness of American agriculture. 2. Enhanced pest resistance in crops is one of the most promising applications of new biotechnology tools. Obstacles to its development could send a negative message to agribusiness, slowing its incorporation of

PAGE 24

Chapter 1Overview and Summary* .17 biotechnology as a mechanism fostering increased economic competitiveness. Congress could keep the oversight authority for plants genetically modified, for enhanced pest resistance under EPA Office of Pesticide Programs (OPP), but direct EPA to strengthen OPP. If oversight of pesticidal plants introduced at a largescale is to be handled effectively by OPP. several changes would need to occur. Technical staff with plant expertise would need to augment current staff; definitions would have to be clarified, given that some naturally occurring plants contain more "pesticidal compounds than will the products of biotechnology: communication with Statelevel implementors would need to be improved immediately; and a clear approach would have to be articulated so that the public, industry and academia would know where the agency stands and how it will implement its policy. Congress could direct USDA-APHIS to regulate large scale introducation of plants genetically modified for enhanced pest resistance. Since USDA-APHIS-BBEP has taken the lead for field tests of plants genetically modified for enhanced pest resistance, APHIS could handle large-scale introductions. This has the advantages of centralizing plant oversight and making effective use of an already well functioning technical staff and organizational unit. The chief disadvantage would be a disruption in the original Coordinated Framework, which ascribed authority to EPAOPP. Congress could direct EPA to work with USDA to develop a similar model of operation and to report on progress to Congress within a specified period of time (e.g. 6 months). Despite disadvantages of forcing two very different offices to work closely together, this has the advantage of allowing USDA to handle any risk concerns related to planned introductions, while allowing EPA to continue to handle food-safety concerns related to pesticidal toxins in the food supply. USDA has established a strong track record for taking the lead in field tests of pestresistant plants. 3. TSCA is a statute explicitly designed to regulate activity for commercial purposes. Academic research, therefore, has been exempt from TSCA oversight. The proposed draft rules for microorganisms, however, greatly expand the regulatory net. One rationale for including academic research is that sometimes universities engage in technology transfer or patent filing, or receive research funds from companies. Obviously, the effects of microorganisms being placed in the environment by a university scientist are no different from the effects of those same microorganisms being placed in the environment by an industry scientist. Concern exists, however, that the draft rules could have a negative impact on academic research. l Congress could allow the proposed ru1e to stand, placing the same requirements on academic research as on industrial research. Subjecting universities to the requirements placed upon companies seems contrary to the Congressional intent behind TSCA. It could have significant negative impacts on university research. Faced with the added bureaucracy and high costs entailed by this rule, the majority of university researchers might deliberately avoid planned introductions of genetically modified organisms. This would leave industry in charge of an area of research that could continue to benefit from broad, objective, openly published study. Such a situation would inhibit the production of new knowledge for use in future risk assessments. However, it is an arbitrary decision to exclude universities automatically from oversightthe release of organisms that pose a risk should be regulated regardless of who conducts the release. Congress could direct EPA to develope an oversight mechanism for planned introductions as an alternative to the proposed TSCA rule. Universities could make use of their already existing system of oversight committees and institutional biosafety officers to regulate biotechnology field trials in house. Just as the Institutional Biosafety Committees (IBCS) review laboratory research involving recombinant DNA, they could review proposals for planned introductions. It would entail education of laboratoryoriented personnel as to the ecological considerations of field release, as well as possible expansion of committee membership to include appropriate disciplines. Serving on an IBC is a time-consuming effort for university personnel. Many feel that there are already too many university committees on which they must serve. Use of IBCS to provide oversight is a possible tradeoff for the university between being able to conduct this research or not. Congress could direct EPA-OTS to develop special procedures to minimize or eliminate any unwarranted regulatory burden on universities, to ensure that public research continues in this area, and to report to Congress on the method selected and its results.

PAGE 25

18 l A New Technological Era for American Agriculture This option would still hold public scientists accountable but would be aimed at lessening the regulatory burden if the appropriate procedure is used. Several possible procedures exist. One possibility would be that the agency funding the research would have the responsibility for monitoring and reviewing the work. As part of the funding contract, the principal investigator would agree to follow EPA guidelines on management and to contact EPA if the need arose. This makes it possible for the funding agency to monitor the project and enforce regulations through the distribution of funds. Another approach is to streamline the application for public researchers. For example, an abstract of a grant proposal could be required to contain specific information that would be sufficient to trigger important questions that arise about the project from EPA. Another possibility would be for EPA to set aside a budget for reimbursement of costs incurred in filing an application. However, even if a cost-savings mechanism is developed, a bureaucracyminimizing mechanism will also be necessary if Congress desires to encourage public researchers and their home institutions to conduct the objective research that will contribute further to our knowledge base. Congress could amend TSCA to exclude universities or to provide alternative means to regulate academic research. An argument can be made for including academic researchers. Obviously, genetically modified organisms released into the environment by a public researcher have the same effect as the same organism placed into the environment by an industry scientist. On the other hand, concern exists about the legal precedent that could be set by extending TSCAs scope to noncommercial research and that it could have a negative impact on research. An application fee for a single field trial costs between $180,000 and $600,000. Even the lower cost is more than most universities or research grants are able to cover. Even though companies have personnel and a budget to cope with regulatory processes, universities for the most part do not have regulatory policy offices or the budget for filing applications. However, if universities and industry worked together, industry would benefit by not having universities file applications. Congress could make its intent for universities clear by stating it in legislative language through TSCA. 4. As large-scale planned introductions become imminent, companies are looking to the regulatory agencies for guidance as to how to proceed. Clear guidance is critical to commercial development of agricultural biotechnology. Congress could direct EPA-OPP and OTS to clarify their regulatory approaches to large-scale introductions and report back to Congress on their approaches within u specified period of time. Interagency work groups, as well as the leadership of EPA, can orient efforts toward assisting EPA staff in clarifying the regulatory guidelines. A flexible approach seems appropriate. Clarifying regulatory guidelines would be particularly helpful to agribusiness working with pesticidal plants or microorganisms other than microbial pesticides. USDA-APHIS-BBEP could provide model mechanisms for clear communication of requirements, use of input from outside the agency, addition of technologically-trained personnel, and creation of an effective structure as well as clarification of direction. Congress could direct EPA to continue on its present course. This is basically a status quo option. It would mean a continuation of the lack of clarity of regulatory policy for potential applicants at the large-scale stage. This lack of predictability could have a negative impact on industry. The absence of applications to EPA-OTS for environmental releases under TSCA over the last year illustrates industries response to lack of predictability in the regulatory arena. It also undermines public confidence in the ability of regulatory agencies to regulate biotechnology. Congress could conduct over-sight hearings of EPA and USDA regarding regulatory policy for large-scale release. Oversight hearings could assist the agencies to develop policy to meet congressional intent for regulating these products even though the regulatory agencies have stated that current laws are sufficient for regulation of products derived from biotechnology. This could help clarify differences in laws written primarily for chemicals instead of genetically modified organisms. 5. Institutions handling new technology must win public confidence and be responsive to public concerns. A balance between maintaining the public interest and ensuring industry competitiveness must be achieved. Congress could direct EPA and USDA to emphasize: 1) increased input of public participation into their Systems; 2) an open process; 3) scientifically sound procedures communicated clearly to other scientists; and 4) follow-up on appropriate cases.

PAGE 26

Chapter lOverview and Summary .19 Most systems can be made sounder when external input is factored into decisions. External advisory committees, hearings, and informal workshops are examples of mechanisms by which Federal agencies can obtain such input. EPA-OPP for example, cosponsored workshops on transgenic plants to gain scientific advice as they deliberated their approach to pesticidal plants and has used its scientific advisory board in deliberations over TSCA draft rules. USDA-APHIS has held a variety of conferences and workshops on planned introductions, stressing public input and State officials input. In fact, USDA-APHIS has made State input an integral part of its review process; EPA could wisely adopt this approach in OPP and OTS. By developing scientifically sound procedures for determining data needs and communicating them clearly, an agency can build an accessible database and contribute to and benefit from the input of the scientific community. USDAs Agricultural Research Service is complementing the work of APHIS by building a database on field tests. Parties concerned about a new technology want to know that potentially problematic cases are being subjected to close follow up. While USDA and EPA can and do impose monitoring requirements on field tests, both agencies could benefit from implementing more extensive follow upon specific cases that might prove troublesome (perhaps by monitoring indicators identified for a Possible worst-case scenario). This is, of course, time consuming. However, if implemented, it should be used in a rigorous manner, so that undue burdens are not placed on straightforward cases, yet so the public feels secure in the knowledge that problematic cases will be tracked after introduction. Congress could require regulatory velop explicit plans for building public report those plans to Congress. agencies to deconfidence and This option would give agencies maximum flexibility. It would allow for the evolution of regulation based on the experience of the agency. Moreover, this approach would allow for a solution to be developed within the agency as opposed to it being imposed on the agency from outside. Reporting the plan to Congress would allow the public to express its opinion and to exert pressure on the agency to change those parts of the plan found to be unacceptable. On the other hand, this process is time consuming for the agencies and Congress. With the large demands on Congress, some members probably would be concerned that it was not the best use of their time. If regulatory agencies fail to maintain public confidence, new Law(s) or congressional oversight could be established to satisfy the public demand for account ability. This option is relatively drastic and could have several disadvantages. Managing a system from the outside invites logistical and other difficulties. Moreover, the tendency with this approach would be to freeze procedures at a particular moment. This could hamstring the natural and positive evolution of regulation, such as the gradual extraction of generic principles from case-by-case reviews. More generally, this approach would be more in the nature of imposed management rather than a solution developed within the agencies, and as such, its own credibility may be weakened. However, it is an option that could ensure accountability to the public if regulatory agencies are incapable of doing so themselves. Food Safety Biotechnology is not so different from previous agricultural technologies as to raise novel scientific issues concerning the safety of foods. What is substantially different, however, is the climate in which this new class of technologies is being introduced. Society in general is more skeptical of the need for new technologies. Scientific illiteracy combined with a lack of knowledge about agriculture and biology leads some people to misunderstand how and why these technologies will be used. Society is also skeptical of how new technologies are developed and regulated. Scandals involving institutions that develop and regulate these technologies have shaken the publics confidence in the ability of these institutions to carry out their activities responsibly. Public confidence will sink further if the public feels that food safety standards are too lax, are fraught with scientific uncertainty, or are not adequately enforced. In addition, uncertainty exists within industries as to how new food technologies will be regulated (table 13). FDA policy has been a long time in the making for biotechnology-derived products. EPA has yet to establish guidelines on data requirements to establish residue tolerances for pesticidal plants, and USDAs Food Safety and Inspection Service (FSIS) has not established guidelines concerning transgenic animals. Genetically engineered products, plants in particular, are approaching commercialization at a faster rate than was anticipated even 5 years ago. These agencies no longer have the luxury of long time frames in which to articulate policy. An end to the uncertainty over how these products will be regulated is needed. Additionally, general need exists

PAGE 27

20 l A New Technological Era for American Agriculture Table 1-3Federal Agencies Primarily Responsible for Food Safety Agency Principal statutory authority Responsibilities Food and Drug Administration Federal Food, Drug, and Cosmetic Act Safety/quality/effectiveness of animal feeds and drugs, and all foods except meat and poultry USDA-Food Safety and Inspection Service Federal Meat Inspection Act and the Safety/wholesomeness/accurate labelFederal Poultry Products ing of meat and poultry products Inspection Act USDA-Agricultural Marketing Service Egg Products Inspection Act Safety/quality of egg products and shell eggs Environmental Protection Agency Federal Insecticide, Fungicide, Safety of pesticide products Rodenticide Act Federal Food, Drug, and Cosmetic Act Pesticide residue tolerance in food/ feeds National Marine Fisheries Service and Agricultural Marketing Act Voluntary seafood inspection Food and Drug Administration SOURCE: Office of Technology Assessment, 1992. to regain public confidence in the regulatory agencies responsible for determining the safety of new biotechnology products. Findings Establishment of Federal Regulations and Guidelines Concerning Biotechnology FoodProductsin the first half of the 1980s, it was anticipated that animal biotechnologies would be developed more quickly than plant biotechnologies because more was known about animal physiology than plant physiology. However, several scientific breakthroughs have speeded progress toward transgenic plants and some are now in various stages of field testing. As transgenic plants approach commercialization, scientific guidelines for assessing their safety will be needed. Further delay in establishing Federal regulations and guidelines could cause a competitive disadvantage to industry, as well as continue to undermine public confidence in the ability of regulatory agencies to establish a clear policy concerning biotechnology. FDA is now wrestling with the question of whether to classify all, none, or some transgenic plants as food additives and to require a food additive petition for these foods. In May 1992, FDA published a preliminary proposal regarding the regulation of new varieties of genetically modified crops. This policy states that FDA is concerned with the characteristics of the food product and not with the method used to produce the product. Thus, new genetically modified crop varieties will not automatically be required to obtain a food additive regulation. New varieties that do not contain new toxicants, elevated levels of inherent toxicants, altered nutrient composition or bioavailability, or enhanced allergenic potential may be regarded as not significantly different from conventionally produced new varieties that are generally regarded as safe. These varieties could be marketed without premarket oversight by FDA. The adulteration clauses of the Federal Food, Drug, and Cosmetic Act could be used to remove these varieties from the market if FDA disagrees with a firms safety evaluation. Varieties that contain substances (either gene expression products or unintended products) that differ significantly in structure, function, and composition from substances currently contained in foods may be required to obtain a food additive regulation. The lack of a priori oversight of some new varieties, however, may still leave considerable uncertainties in the minds of the public, at least for the first generation of products developed. Public confidence in the process may still require at least a minimum review of the product prior to commercial release. Such review may consist of notifying FDA of the development of a transgenic crop and provision of a minimum level of data so that FDA can make a determination as to whether a food additive petition will be needed. Such a notification process could be open to the public so that any significant concerns can be identified. Additionally, public interest groups have expressed opposition to the policy and have threatened legal action to prevent its implementation. The policy is currently open to public comment, and could be subject to revision. Congress may yet be required to intervene in the development of food biotechnology regulations if differences cannot be resolved in a timely fashion. If such action is needed, several options are available to Congress. Public Confidence in the Decision making Process One method of enhancing public confidence in the regulatory process is to make that process open and accessible and to increase public participation in the process. Opponents of increased public input in regulatory decisionmaking processes argue that citizens lack the training needed to understand complicated scientific and technical

PAGE 28

Chapter lOverview and Summary .21 issues, and as such their participation only delays the agencys decisionmaking without offering any offsetting benefits. Critics also fear that public representatives may act in emotional and irrational ways and make unreasonable demands. Those who support increased public input argue that such input is invaluable in establishing the legitimacy of regulatory decisions. Indications also exist that public participation can increase the comprehensiveness of agency decisions by encouraging the agencies to focus on a wider range of issues and values than they normally would. Lastly, it is hard to justify no public participation in regulatory processes in a democratic society. The public will not make the regulatory decisions that is the responsibility of the State and Federal agencies whose statutory authority requires them to ensure a safe and wholesome food supply. However, public confidence that these agencies are fulfilling their responsibilities will be enhanced if there are mechanisms available for public questions and concerns to be heard and addressed prior to decisionmaking by the regulatory agency. At present, public input into the regulatory process consists of notification and comment procedures and participation on advisory committees. Recent revelations that companies have withheld negative research results from regulating agencies have also undermined public confidence and raised serious questions about the process used in making safety assessments. Currently, manufacturers of technology submitted to the regulating agency for approval also perform the safety assessment following guidelines established by the agency. This situation creates potential conflicts of interest. Most companies are honest, but given the current climate of public skepticism, the appearance of impropriety may be sufficient to prevent consumer acceptance of a new technology. Given the lack of public understanding about biotechnology, doubts about the validity of the safety data used to make regulatory decisions for this new class of products could be substantial. There may be merit in considering a safety assessment process that includes independent testing of products. Tradeoffs Between Industry Competitiveness and Societys Right to be Informed About Health and Safety issuesPublic interest groups argue that industry claims too much scientific data as confidential business information (CBI) when submitting a new technology for agency approval, thereby limiting the amount of health and safety data available to the public. On the other hand, industry feels that there is too little protection of proprietary data by Federal regulatory agencies. Achieving the proper balance between protecting proprietary rights and disclosing health and safety data to the public is a delicate undertaking. Disclosure practices are regulated by the Trade Secrets Act and the Freedom of Information Act. The Trade Secrets Act of 1982 subjects government employees to criminal penalties for the disclosure of proprietary data unless authorized by law. The Freedom of Information Act (FOIA) of 1982 permits agencies to protect trade secrets and commercial and financial information that is confidential. Both laws seek to protect information that would be of commercial value to a firms competitor. However, a congressional order mandates that EPA and FDA release some types of scientific data in certain circumstances. The FDA has restrictive CBI policies. Although Congress has mandated that health and safety testing data for new drugs can be released after another manufacturer becomes eligible to sell the drug unless extraordinary circumstances are shown, little data are actually released. This is in part because FDA defines extraordinary circumstances to include any claim that the data are CBI, such as a claim that it could be used by competitors in foreign countries. While FDA usually does not release safety data, it did in the case of bovine somatotropin (bST). For the first time in FDA history, FDA published an article in a peer reviewed scientific journal detailing how FDA reached its conclusion that bST was safe for human consumption. Specific safety data were presented. Additionally, the National Institutes of Health (NIH) and FDA hosted a scientific meeting with public participation to discuss food safety concerns of bST. FDA has also published an article explaining why FDA granted GRAS status to the genetically engineered enzyme chymosin. Thus, FDA has shown that it is possible to release such information when it is in the public interest. FIFRA protects CBI, but allows release of health and safety testing data for registered pesticides. Also, data concerning production, distribution, sale, or inventories of a pesticide maybe released in connection with a public proceeding if disclosure is in the public interest. Thus, FIFRA permits the release of health and safety data after the decision is made but not during the process. After notification of a food additive or pesticide registration petition has been published, under FOIA, requests for safety data can be made. However, sometimes it is not possible for agencies to determine whether or not information is CBI in the time allotted to them to make a regulatory decision. Attempts to mitigate these

PAGE 29

22 l A New Technological Era for American Agriculture problems include requesting that companies restrict their CBI claims and that they justify their claims of confidentiality at the time they submit a petition. Decisions to disclose CBI focus on whether or not such disclosure will be harmful to the company. No attempt is made to weigh this harm against the publics right to be informed about health and safety issues that might affect them. Other countries, most notably Canada, have taken the approach that disclosure of health data is authorized if it is in the public interest as it relates to public health, public safety, or protection of the environment and if it clearly outweighs in importance the financial loss to the competitive position of a company or person. Enforcement of Regulations-Research indicates that a significant factor in public lack of confidence in regulatory agencies is concern that regulations are not adequately enforced. For example, although Federal law bars sale of produce with pesticide residues above Federal tolerances, recent studies show that consumers are willing to pay for labels assuring them that these tolerances are in fact not exceeded. If the public is to regain trust in regulatory agencies, enforcement of regulations will need to be improved. This will be difficult as biotechnology becomes a new focus of public concern and a new arena of regulatory responsibility. The regulatory agencies do not have the resources to increase enforcement activities significantly. A recent General Accounting Office study found that the regulatory agencies involved in food safety had fewer staff, less funding and a larger workload in 1989 than in 1980. Available resources already are being stretched, and must be spread even thinner to develop new multiresidue assay procedures and sampling methodologies for tracking genetically modified organisms. A new approach to food safety assessment must be developed as well. Traditional approaches to safety assessments of food additives are inappropriate for the assessment of whole foods because large enough quantities of the food cannot be fed to test animals without invalidating the results of the test. New assay and testing methods applicable to genetically modified foods will thus be needed, and this will require additional agency resources. LabelingMany consumers have expressed a desire that food products developed with biotechnology be so labeled. However, while consumers express a desire to have accurate and verifiable labels, many of them are not willing to pay much for those labels. For example, approximately one-third of consumers do not seem willPhoto Credit: U S Department of Agriculture, Agricultural Research Service Chemist evaluates a screening assay for residues. New analytical methodology will need to be developed for biotechnology-derived foods. ing to pay anything for labels; another 5 to 10 percent of consumers seem willing to pay as much as 50 percent higher food prices for labels. Most consumers seem willing to pay 5 to 10 percent more for labels. Clearly a labeling proposal that is expensive will not be popular with most consumers. FDA has stated in its preliminary policy that generic labeling of biotechnology food products will not be required but selected products may require labeling. Such products include those for which nutritional composition has been altered or potential allergens introduced. International CoordinationThe United States annually imports billions of dollars worth of food products, many from countries that also use biotechnology in their food industries. If U.S. food safety regulations concerning biotechnology substantially differ from other countries regulations, difficulties could arise. U.S. producers will likely beat a competitive disadvantage if U.S. policy

PAGE 30

Chapter 1Overview and Summary l 23 is substantially stricter than that of other countries. Enforcement will be difficultno generic methods exist to detect genetic modification. Reliance on the word of other countries that their products contain no biotechnology-derived constituents may or may not be acceptable. If U.S. regulations are substantially less stringent than those of other countries, then the U.S. agricultural export market could suffer. Agricultural commodities are a major export of the United States. Thus, international coordination will be paramount. Preliminary FDA policy is consistent with international organizations working papers and reports on food safety assessment procedures for genetically modified organisms. Options 1. FDA and EPA no longer can delay the development of final regulations and guidelines because transgenic plants are approaching commercialization. FDA has the choice of requiring a food additive petition for all, some, or no transgenic plants. Congress could monitor the development of regulations and conduct oversight hearings of FDA and EPA to determine why final regulations and guidelines do not exist and to have them report back to Congress with recommendations in these areas within a specfied period of time. This would be a strong signal to the executive branch that Congress is concerned about the delay in providing guidance to the private sector for these new technologies. An oversight hearing would provide the agencies with an opportunity to explain their rationale and concerns in establishing regulations for these new products and allow Congress the opportunity to provide guidance and direction to the agencies. Congress and the Executive Branch through EPA, FDA, and USDA have a number of options for regulating transgenic organisms. The following part of Section 1 illustrates options available. Congress or FDA could establish categorical exclusions to the requirement of a food additive regulation for certain transgenic organisms and require a case-by-case approach for the remaining products. Essentially, this is the policy chosen by FDA. Transgenic organisms that involve gene products that are widely present in the current food supply, and do not introduce new toxicants, elevate levels of existing toxicants, alter the composition or bioavailability of nutrients, or transfer allergenic components, and that use safe marker and promoter sequences can be excluded from the need for a food additive regulation. These products do not introduce new food compounds into the food supply and they have no unintended effects. Therefore, FDA states that they can be classified as GRAS because they are equivalent to traditional new varieties that historically have been given GRAS status. Only products that contain components that are significantly different in structure, function, and composition may be required to obtain a food additive regulation on a case-by-case basis. This option is a risk based option that requires extensive safety testing for products that are not normally found in the food supply, and less testing for products that contain substances already widely consumed. It places responsibility for the initial food safety assessment with industry. Lack of FDA oversight, especially for the first generation of biotechnology-derived food products, may raise public concerns. A number of public interest groups have indicated their opposition to this policy. Option: Congress or FDA could establish a policy similar to the preliminary policy articulated by FDA, and include a formal notification procedure. Such a policy would require the establishment of a system for notifying FDA when a new transgenic crop is marketed. As currently outlined, FDA policy allows firms to determine if a new variety contains components that are already widely consumed. Thus, firms can make a determination about the GRAS status of new biotechnology products without consulting FDA. In the beginning, it is quite likely that most firms will consult FDA prior to marketing a new biotechnology-derived variety, but they are not required to do so. This situation is likely to create considerable apprehension among the public. Thus, a formal system of notification may be desirable. The notification process could include safety data the company used to determine that the product was GRAS. Such data includes the identity of the host and donor organisms, information on the genetic construct, and information on the physiology of the gene product. Additional information required could include compositional data. A comparison of nutrient and toxic component levels in transgenic and counterpart traditional crops could be included, as well as data on allergens. This type of information will be available in the development of transgenic organisms and is required for a company to make its determination of the regulatory status of the product. Thus, requiring this information to be on record with FDA should not present undue burdens on industry. However, requiring FDA to review and act on this information for all transgenic crops will place a strain on the agencys re-

PAGE 31

24 l A New Technological Era for American Agriculture sources. Most likely FDA will need additional resources to implement this policy. The notification process could be open to the public so that they can raise concerns and issues regarding transgenic organisms. It may also be useful for FDA to use an advisory committee to comment on the data presented. If an advisory committee is used, representatives from the public could be included along with technical representatives. Such a policy might be effective for the safety assessment of the first biotechnology food products developed. It would allow FDA to provide at least minimal oversight over all biotechnology food products, assure the public that scientific information is available, and thus, might alleviate some public concern. In the short run, such a policy may appear to result in unnecessary regulation of these products. However, it may be the price industry must pay to have their products accepted by the public, at least in the initial stages of commercializing biotechnology food products. Congress or FDA could require a food additive petition for all transgenic crops. This policy would force all transgenic food products to undergo a premarket safety approval process. It would only be based on a risk assumed to be inherent in the process of genetic engineering, an assumption not supported by scientific data. This policy would likely delay commercialization of transgenic crops already being developed and possibly could inhibit the development of additional transgenic crops. On the other hand, this policy would not be inconsistent with a broad interpretation of the food additive definition. And it probably would soothe some consumer fears and uncertainties about these products. Congress or FDA could establish some categorical exclusions of transgenic food products from the requirement of a food additive petition, and could require all other biotechnology products to meet the requirements of a food additive petition. Once again categorical exclusions might include transgenic crops that do not contain components that are significantly different from those currently present in the food supply and for which unsafe, unintended components have not been introduced. This policy would be more risk based than requiring all transgenic organisms to meet the rigors of a food-additive petition, because transgenic organisms that are essentially the same as products that have historically been viewed as safe would not be required to undergo premarket approval. This policy would ease some of the burden on industry. There may still be public apprehension with respect to those products that have been excluded. Congress or FDA could establish a policy in which the gene expression product is classified as a food additive if the same traditionally processed product would have been classified as such. It could exclude from the food additive definition gene products that would not have been classified as a food additive if produced by traditional means. Gene products that might be excluded as food additives are those that would code for agronomic functions such as drought resistance. This policy is based more on the intended use of the gene product rather than any safety risk that the gene product may pose, but would be consistent with how FDA has historically interpreted the food additive amendment. It would, however, be difficult to justify on scientific grounds. Congress or FDA could establish a policy that the requirement for a food additive petition for transgenic organisms be determined on u case-by-case basis for each transgenic organism. Such a policy would allow FDA to provide oversight of all biotechnology products. This would provide the public with an assurance that all transgenic organisms would be reviewed by FDA. However, continuation of this type of policy indefinitely could overwhelm FDA, since the number of products that could be developed is large. At some point, FDA will likely need to categorize some products as GRAS, just as it does with chemical additives. Congress or EPA could establish guidelines for the safety evaluation required to establish pesticide tolerances for whole plants. Currently, EPA does have guidelines for transgenic pesticidal microorganisms, but has yet to establish such guidelines for whole plants. Transgenic plants producing pesticidal compounds, such as Bt producing plants, are completing small-scale field trials. Guidance from EPA for dealing with such plants no longer can be delayed. Establishment of safety guidelines will require a new assessment paradigm (discussed later). Additionally. because States, FDA, and USDA enforce pesticide tolerances, EPA needs to work closely with appropriate agencies in establishing tolerances. EPAs work with States needs improvement in this area. Only recently has EPA even begun to compile a list of contact persons in State agencies. This ignoring of States could easily lead to State laws that are incompatible with Federal regulations, or

PAGE 32

to gaps in State authority or expertise to carry out Federal regulations. Congressional hearings and oversight may be necessary if EPA does not improve this situation. Congress or USDA -FSIS could establish guidelines concerning transgenic animals. USDA-FSIS plans to release guidelines in the near future concerning the slaughter of experimental animals in which gene transfer attempts failed. Guidelines concerning the slaughter of transgenic livestock are still in early draft form. Of particular interest will be guidelines concerning the slaughter and potential food use of transgenic animals that produce pharmaceuticals. FSIS and FDA have established a joint committee to deal with issues that jointly affect the two agencies. Careful monitoring of how successful this committee is may be required. 2. Public confidence in the regulatory process needs to be enhanced. Making the regulatory process open and accessible to the public and above reproach is a key factor in providing trust and confidence in the decisionmaking process. Congress could direct agencies (FDA, USDA) to establish mechanisms to allow for increased public participation and to report their results to Congress within 1 year. This option sends a clear message to the agencies that Congress is concerned about the publics view of regulatory agencies and that the public should be more involved in the decisionmaking process. It gives maximum flexibility to the agencies to determine the method of incorporating the publics input. A number of mechanisms are available. For example, Federal agencies could establish criteria by which local agencies can be notified any time significant risk or unique questions arise that are pertinent to them. Agencies may wish to adopt a procedure similar to that used by FIFRA, i.e., notification of petitions received, and if public interest warrants, an informal hearing. Increasing public participation will require increased resources and risk politicizing decisions, but could also enhance public confidence in the regulatory process. It might cost less in the long run. Congress could direct the agencies to crease the use of advisory committees for decisions involving biotechnology and to change the composition of their membership to increase the number of nontechnical public representatives. For FDA, advisory committees could help establish GRAS and the minimum information needed for food additive applications of genetically engineered whole foods. These committees could be used as a first screening mechanism to see if a food additive petition is actually needed. Public meetings help assure the scientific validity of the process. EPA might also use advisory committees to establish tolerances for genetically engineered plants with pesticidal properties. This might be helpful since in-house expertise to handle this responsibility seems to be lacking. Advisory committees might also prove useful to USDA in establishing a policy on transgenic animals. The credibility of any advisory committee will be enhanced if it includes public representatives. FDA may need to consider granting current nonvoting members of its advisory committees the right of full voting membership. And they may need to expand the list of technical fields beyond MDs from which experts are drawn. Use of advisory committees presents some logistical problems and requires additional resources, but provides expertise that currently may be missing. Additionally, the possibility that non-technical representatives will pursue political agendas and unnecessarily delay committee decisions exists. However, used properly, such representatives can focus the attention of the committee on issues that might otherwise be overlooked and provide legitimacy to committee decisions. l Congress could direct the agencies (EPA, FDA, USDA ) to change the notification procedures for advisory committee meetings. The standard method of notification for advisory committee meetings involves publication in the Federal Register. Few members of the public know what the Federal Register is, much less read it regularly. Also, notices published are written by and for those individuals knowledgeable in the field and, thus, the general public might not be clear as to what the issue is. Additionally, most meetings are held in Washington, DC. Agencies could have committees convene in different cities and publish announcements, other then the Federal Register, that are more likely to be noticed by a wider public. Such activities are likely to be more expensive than current ones, however, but make the decision-making process more accessible to the public. Congress may wish to appoint a task force to study the role of independent safety testing of biotechnology products. Independent testing is unlikely to be popular with industry, however, a growing perception exists that companies are withholding negative data and that the safety 297-937 0 92 2 OL 3

PAGE 33

26. A New Technological Era for American Agriculture review conducted by regulatory agencies is made without accurate and complete data. Enhanced authority to subpoena data by regulatory agencies, most notably FDA, could be useful. Additionally, it may be worthwhile to consider establishing independent testing of products. FDA, for example, rather than companies could choose outside investigators to perform selected safety assessments, and these contractors could report results directly to FDA rather than the companies. A study to consider the broad range of implications of such a change would be warranted before implementation. 3. Public interest groups argue that industry claims too much scientific data as confidential business information (CBI), and that this restricts the amount of health and safety data available to the public. Industry argues that there is too little protection of proprietary data and that this situation adversely affects their competitive position. Achieving the proper balance between protecting proprietary rights and disclosing health and safety data to the public is a delicate endeavor. l congress could encourage FDA to publish more scientific review articles and hold public meetings in cases that generate public interest. Clearly it is possible for FDA to release considerable health and safety information to the public as it has done for bST. The public controversy surrounding this product apparently outweighed any competitive disadvantage presented to the firms producing bST. Such a policy might prove useful in responding to public concerns about other biotechnology products and potentially could enhance the accountability and credibility of FDA decisions. Congress could conduct oversight to provide increased guidance to regulatory agencies attempting to encourage firms to reduce CM voluntarily. Congress could monitor whether health and safety data are being made available as products approach commercialization or if firms withdraw their voluntary cooperation and claim more data as CBI. If firms increase CBI claims, Congress could direct Federal agencies to require firms to justify CBI claims when a petition is submitted rather than waiting until a FOIA request is made. Currently, firms realize that it takes regulators longer to determine the validity of CBI claims than the time allotted to make regulatory decisions. This could encourage some firms to make CBI claims of data that in fact are not confidential. Congress could also direct agencies to facilitate reconsideration of a decision if CBI data are released after a regulatory decision is made and causes public concern. Currently, firms can avoid public disclosure of data during the regulatory process simply by claiming confidentiality and know that the regulatory decision will not be reconsidered. If the decision is allowed to be reconsidered, firms may reduce their CBI claims. Industry will oppose increased disclosure of safety data because it will erode their competitive position, On the other hand, with the current climate of public skepticism of new technologies and regulatory agencies, increased industry accountability and public disclosure of safety data may be required of business. Congress could liberalize the CBI policy. Congress could direct FDA to release data it is currently authorized to release but generally does not. Congress could consider adopting a regulatory policy similar to that used in Canada which would weigh any harm to the company against the publics right to be informed about safety concerns. Current policy considers only the harm to firms. As a last resort, Congress could force the disclosure of health and safety data. Once again the potential harm to the competitive position of companies must be weighed against the publics right to be aware of potential safety risks and to regain public confidence in the regulatory process. Industry probably will object to an easing of CBI policy. Public support, on the other hand, may be equally strong for disclosure. 4. Genetically modified foods will require a new paradigm for food safety evaluations. Changes in data needs, assay procedures, and sampling methodologies will be required. Congress could fund the development of new analytical methodologies and assay procedures through the National Institutes of Health (NIH). New analytical methods for whole food assessments must be developed if FDA is to determine the safety of genetically modified crops, and to monitor foods once they are marketed commercially. NIH, in coordination with FDA, could provide funding to develop food analytical technologies, These new technologies and assessment procedures would be useful in determining the safety of genetically engineered foods and could also enhance research programs such as the designer foods project (a component of cancer research) and nutritional programs.

PAGE 34

Chapter lOverview and Summary l 27 Congress could provide funds to NIH for the development of databases detailing the normal range of nutritional and toxic components of food. Major nutrients and toxic substances in food have been identified, but additional information is needed to assess these food components, such as the quantities at which they normally are present in foods and their chronic impacts on humans. Assessment of such information will be needed to determine whether genetically modified foods present greater safety risks than do foods currently consumed. Congress could direct FDA and EPA to request that assay procedures developed by firms to detect additives be readily adaptable for use under field conditions. Currently, when firms submit a food additive petition or a pesticide registration, they are required to provide an assay method to detect residues or additives in food. Generally, the method provided applies to a single residue and requires sophisticated instrumentation for identification and quantification. Agencies might require multiresidue assay methods that are more readily usable under field conditions than they are today. The residues would have to have some similar characteristics for a multiresidue technology to work. Development of such assay methods may create technical difficulties and are likely to create added costs for industry. However, they would improve monitoring and enforcement activities of regulatory agencies, an issue of particular importance to the public. 5. Surveys clearly show that consumers desire additional information about the foods they consume. Labeling is a method to provide this information, especially for those concerned about foods produced from biotechnology. Congress could mandate that all food products containing constituents derived from biotechnology be so labeled. This would satisfy the desire of the public to be able to identify foods derived using biotechnology. But it probably would be expensive to provide labels and difficult to verify label information. No generic means exists today to identify whether a food constituent, such as a kernel of corn that will be ground into meal, has been genetically engineered or not, and it is unlikely that such a method can be developed. Consequently, genetically modified products would have to be kept segregated throughout the market to be able to assure the public as to whether their food contains such products or not. This is not now the case for many bulk commodities, such as grains, and entirely new marketing structures would need to be developed. Increased vertical integration of agricultural industries would likely occur. And, significant government resources would be needed to enforce mandatory labeling and the added expense would be passed along to consumers. Thus, guaranteeing that a product does not contain any products derived from biotechnology could become expensive. Based on current research, it is not clear that consumers would be willing to pay that added expense. Congress, through research and extension agencies, could encourage niche markets to be established to satisfy the concerns of those willing to pay higher prices for labeled food signifying that it does not contain genetically engineered food. An alternative to passing the high cost of verification along to all consumers is to establish a higher priced niche market for biotechnology-free foods that would satisfy needs of some consumers. Such a market would be similar to the current organically produced food market. Organic produce is higher priced than traditionally grown produce but provides an alternative product to consumers who are willing and able to pay higher food prices. Recent legislation has been enacted to help resolve some problems involved with organic produce such as a lack of a standard definition, grower certification and oversight procedures. Such a policy might also work for biotechnology-free food products, and would have the advantage of passing the extra costs along only to consumers willing to bear them. Public Sector Research It is becoming increasingly difficult for the land-grant system to carry out its historic mission. In addition to the increasingly specialized nature of the research conducted, pressures from outside the system are building. Changing political support, resource base, and institutional frameworks combined with the development of revolutionary new technologies will put pressure on the land-grant system to change dramatically. Historically, political support for the agricultural research system has come from the farm and rural population. For this reason, agricultural research has focused heavily on increasing the productivity of agriculture. However, this traditional base of support has been steadily eroding, and urban groups have put pressure on the system to shift research priorities to such areas as water quality, human nutrition, food safety, and sustainable agriculture.

PAGE 35

. 28 l A New Technological Era for American Agriculture The development of biotechnology and advanced computer technologies has the potential to revolutionize the way in which agricultural research is conducted, and to provide powerful tools to help address social problems. The scientists who conduct research using these technologies will need a thorough grounding in the basic disciplines that underlie them. Today only a small proportion of academic agricultural scientists have this background. Moreover, for advanced computer technology research to reach its potential, it will need to be identified as a research priority and universities must be encouraged to develop a promotion and tenure system that recognizes more than a publication record for research accomplishments. In addition, multidisciplinary teams involving basic computer sciences, systems design, and traditional agricultural sciences need to be encouraged. To this end, development of nationally recognized centers of excellence, similar to those developed for biotechnology, need to be considered. In general, agricultural research is underfunded. Estimates of the social rate of return to public-sector agricultural research investments range from 35 to 145 percent, indicating a significant underinvestment in this type of activity by the public sector. There has also been a slight, but potentially significant shift in the source of funding for agricultural research at land-grant universities (table 1-4). The States, which provide the majority of the funding for research at these universities, have been constrained in spending by the recession of the early 1990s. Few States have increased funds for research and many have cut funding in this area. USDA funding, the second largest single contributor to agricultural research, has remained basically stagnant, barely keeping up with inflation. Funding from the private sector for university research, on the other hand, has been increasing in the form of industry-supported research, and from the sale of products by universities. Currently, these sources of income represent about 13 percent of the total funding for agricultural research, but have increased by 60 percent since 1982. The product sales category is a potentially lucrative source of funding for universities. Legal and institutional changes have made it easier for universities to capitalize on their research, since now they can retain title to any federally funded technology the university develops. Incentives to privatize the benefits of university innovation could shift the university further toward private funds, especially if public funds do not keep pace with increased needs. Changing clientele, funding bases, technologies, and institutional structures will create new demands on the land-grant system. Decisions need to be made on how land-grant universities can best serve society in this new era. Findings The Uniqueness of Land-Grant UniversitiesLandgrant universities differ from other universities in their legislated mission to address research on the problems of society. Some argue that the land-grant system has, in part, already abandoned its mission, as agricultural researchers increasingly work for disciplinary laurels rather than societys benefit. Others argue that the system deTable 1-4Total Research Funding for State Agricultural Experiment Stations, Selected Years a (in millions of dollars) USDA Other Product Year USDA b competitive Federal d State e Industry sales Other f Total 1982 . . . . 161.3 5.5 77.8 522.2 57.0 58.5 70.0 952.3 1984 . . . 174.9 6.1 81.7 591.4 64.1 61.3 79.8 1,059.3 1986 . . . . 174.4 11.9 110.8 704.3 78.1 62.9 89.8 1,232.1 1987 . . . . 175.6 16.8 114.9 732.5 87.4 68.4 104.2 1,299.8 1988 . . . . 187.0 19.3 115.0 770.0 91.2 77.8 114.1 1,374.2 1989 . . . . 194.0 21.9 130.4 827.6 101.2 82.4 132.1 1,489.6 1990 . . . . 203.6 20.0 143.9 877.9 113.8 91.6 145.7 1,596.5 a Funding is for the State Agricultural Experiment Stations only and does not include the 1890 universities, the Schools of Veterinary Medicine, or the Forestry Schools. Funding is in current dollars. b USDA includes Hatch, Mclntyre-Stennis, Special Grants, Evans-Allen, Animal Health, and miscellaneous other funds administered by the Cooperative State Research Service. c USDA competitive is the USDA competitive grants program. d other Federal includes funding from Federal agencies excluding USDA and includes funding from NIH, NSF, AID, DOD, DOE, NASA, TVA, HHS, PHS, etc. State is state appropriations. Other includes funding from nonprofit organizations, and contracts and cooperative agreements administered by USDA. SOURCE: Inventory of Agricultural Research, Cooperative State Research Service, U.S. Department of Agriculture, Washington, DC, various years.

PAGE 36

Chapter lOverview and Summary l 29 fines societys problems too narrowly, placing too much emphasis on increasing agricultural productivity and too little on nutrition, environmental, and rural problems among others. Some also argue that too much attention has been given to production agriculture and not enough to postharvest technologies, value-added products, consumer preferences, and agribusiness problems. No easy answers exist as to what types of research should be conducted with public funds. What is clear, however, is that as the traditional clientele (i.e., farmers) continues to shrink. greater demands will be placed on the system to address the needs of other groups. To be able to do so may require some difficult choices concerning research mix, with some traditional research programs being eliminated and some new programs initiated. Research Funding Based on Mission FunctionsIn recent years the land-grant system almost exclusively has embarked on a program to increase public funds through competitive grants. Relatively little attention has been given to securing other types of funding such as Hatch formula funds. This strategy is questionable for the landgrant system in the long run. Research conducted in conjunction with this study suggests that the most appropriate funding policy is a healthy mixture of formula funds and competitive grants. The results indicate that different funding mechanisms may be more appropriate for the different functions or goals of land-grant universities. For example, if the goal is to increase cuttingedge basic research, increased funding for competitive grants might be the best approach. If the primary goal is to enhance research applicable to problem solving or to train future researchers, the more stable and locally controlled Hatch formula funds may be the more appropriate mechanism. The appropriate allocation of these two types of grants depend on the relative priorities given to the three missions of land-grant universities. Potential Privatization of Research at Land-Grant UniversitiesTwo new sources of research funds are private sector investment and product sales. Constrained and basically stagnant research budgets provide many incentives for universities to increase funding via these mechanisms, but the development has raised many concerns. For example, incentives to privatize university innovations for the benefit of the university rather than society could conflict with the mandated mission of the university. Using public resources to reap private gains raises many ethical questions. Allowing individual researchers to share in the profits of their publicly funded work and encouraging universities to produce consumer products opens the door to potential abuses. Certainly, potential exists for conflicts of interest. There may be financial conflicts if individual researchers are allowed to capture the returns of their innovations. To some extent, this situation already exists in that researchers use public funds to generate new knowledge that can be sold to the private sector in the form of consulting fees. But there is a distinction between providing expertise to potentially multiple clients and having a vested interest in the development of one or several products by companies. Universities also may face conflicts of interest. The credibility of the university may suffer if it is viewed as being too cozy with industry. If public universities are viewed as being more concerned with their own private good than with the public welfare, then the public may not maintain its support for the university. One underlying principal of scientific research is the free exchange of research results. Concern arises that with increased potential to earn income from research, the results of research will become more proprietary. Moreover, research results may not be freely or readily exchanged if a researcher, university, or industrial sponsor attempts to patent the results or seek additional private-sector funding. Given the level of underinvestment in agricultural research and the stagnation of public-sector funding for this activity, the extra revenue earned from product sales could provide great benefits for the university and for society. Whether those benefits will be attained will depend on how the revenue generated from commercialized activities is used. The extra revenue could be used to fund socially underfunded research or to enhance the teaching capacity of the university. The new arrangements may enable universities to contribute to economic development in ways not previously possible. Whether or not the funds are used for such purposes will depend on how well university administrators are able to maintain a sense of priority for the overall research and teaching program, and whether they have the administrative skills to keep scarce resources allocated to the proper ends. Policy Options 1. The new partnership between the public and private sectors potentially can revitalize agricultural research, but could also bias the overall research endeavor and damage the credibility of universities. Research and close monitoring will be needed to understand the changes occurring within the landgrant system and to ensure that they are not undermining the system as a whole.

PAGE 37

30. A New Technological Era for American Agriculture Congress could require the U.S. Department of Agriculture to monitor the increased private-sector, funding of agricultural research and to prepare an annual report for Congress containing the data. Currently, little is known about the extent of privatesector funding at land-grant universities and the nature of the relationship between the universities and the private sector. Congress could conduct oversight hearings periodically on this issue. Furthermore, Congress could direct USDA to collect data from the land-grant universities on the extent of public-private collaboration, to prepare an annual report for Congress containing the data, and to provide guidelines on the appropriateness of various public-private sector research collaborations. Congress could direct USDA to require land-grant universities to establish an explicit policy with regard to research sponsored by the private sector and report that policy to Congress. The USDA would require each university using private-sector research funds for agriculture to establish a specific policy as to how those funds are used based on a broad policy established by the land-grant system. Establishing an advisory board that includes members of the public in setting priorities for research funded from the private sector might be an effective mechanism. This would help to increase public confidence that the university is using funds to solve problems that confront society. 2. High rates of return to public-sector investments have been reported by numerous studies, including past OTA reports. This indicates that public sectorresearch funding is below optimum rates. Congress could increase public-sector support of agricultural research. Increasing public-sector support of agricultural research might help to lessen the pressure on land-grant universities to obtain funds from the private sector. Given the high rate of return on public-sector funding of agricultural research, funding increases probably would prove beneficial. Congress could maintain or decrease public-sector funding for agricultural research. Federal funding for agricultural research has been relatively flat for the last 30 years. As a consequence, States have picked up the increased costs of conducting agricultural research. It is difficult for States today to take on an ever increasing share of public supported research. If the Federal Government continues to reduce its contribution to research funding, land-grant universities must look for alternative sources of funding. Private-sector funding from specific industries or individual firms or product sales from technologies developed by the university are the most likely sources of additional research funds. The impact of this shift in support is not known but needs further analysis. 3. Recent research indicates that public-sector funding mechanisms should be goal oriented. Congress could appropriate funds for agricultural research through funding mechanisms bused on welldefined agricultural research goals. The land-grant system provides teaching, extension, and research functions. Preliminary research suggests that Hatch formula funds are more suited to teaching and extension activities and competitive grants more suited to basic research. By appropriating funds according to goals to be achieved, Congress could improve the effective use of public funds. Congress could maintain the current emphasis of increased funds for competitive grants and level or decreased funding of formula and intramural funds. Implicitly, this would indicate that Congress places greater emphasis on basic research than on adaptive research, extension, and teaching activities. Evidence does not exist that the lack of basic research is the primary constraint to the ability of land-grant universities to fulfill their historic mission of addressing research aimed at solving societal problems. Congress could extend competitive grants to extension and teaching curricula development. A strong case can be made for formula funding of agricultural research. However, if the only acceptable political form of increased funds is competitive grants, then expanding these grants to include adaptive research, extension and teaching could be considered. Balanced funding of basic research, adaptive research, teaching, and extension would significantly strengthen the landgrant universities and help them meet their multiple missions more effectively. Congress could award certain competitive grants to basic research that clearly shows ties to adaptive research. This would be a clear signal that Congress considers the original mission of land-grant universities to be appropriate today. Currently, most grants for basic research are not tied directly to adaptive research. Thus, it is

PAGE 38

difficult to differentiate between funding provided by the National Science Foundation (the major funding agency for basic research) and the U.S. Department Agriculture. 4. The public is increasingly losing confidence in land-grant universities credibility, and credibility needs to be restored. Development of a more missionoriented system with increased public input could help to restore confidence in the system. The OTA report Agricultural Research and Technology Transfer Policies for the 1990s addresses this issue in some detail and provides specific options that suggest changes in the system to make it more mission oriented. Those options are incorporated here by reference. Some of the options were incorporated into the 1990 Food, Agriculture, Conservation, and Trade Act of 1990( 1990 Farm Bill). SUMMARY Newly emerging biotechnologies and information technologies hold great promise for American agriculture and can provide solutions to many problems. In the decade of the 90s, however, public concerns about the environment, food safety, industry structure, and institutions will focus on these emerging technologies. Whether these technologies will be accepted and flourish, or stagnate, will depend in large measure on how U.S. public institutions resolve the complex problems of regulatory oversight and on whether scientists and policy makers can allay public concerns about biotechnology in particular.

PAGE 39

Part I The Advancing Technologies

PAGE 40

Chapter 2 Emerging Plant Technologies

PAGE 41

Contents Page TOOLS AND TECHNIQUES OF BIOTECHNOLOGY . . . . . . . . . 38 Biotechnology Techniques Used To Create Transgenic Plants . . . . . . . 38 Other Biotechnology Techniques . . . . . . . . . . . . . . . 42 Application of Biotechnology Techniques To Create Transgenic Plants . . . . . 44 Biotechnology in the Food Processing Industry . . . . . . . . . . . 49 TOOLS AND TECHNIQUES OF BIOLOGICAL CONTROL . . . . . . . . 50 Approaches Used in Biological Control . . . . . . . . . . . . . 50 Use of Biological Control Agents To Control Pests in the United States . . . . . 54 SUMMARY . . . . . . . . . . . . . . . . . . . . 60 CHAPTER PREFERENCES . . . . . . . . . . . . . . . . 61 Figures Figure Page 2-1. Identification and Isolation of Desired Gene . . . . . . . . . . . 40 2-2. Gene Transfers With Bioblaster . . . . . . . . . . . . . . 41 2-3. Plant Tissue Culture Technology . . . . . . . . . . . . . . 42 2-4. Antisense Technology . . . . . . . . . . . . . . . . 43 2-5. Preparation of Monoclinal Antibodies . . . . . . . . . . . . 44 Tables Table Page 2-1. Transgenic Crops Produced . . . . . . . . . . . . . . . 39 2-2. Current Targets for Crop Modification for Herbicide Tolerance . . . . . . 47 2-3. Virus Coat Proteins Engineered Into Plants . . . . . . . . . . . 48 2-4. Disease Resistance Genes Introduced Into Plants . . . . . . . . . . 49 2-5. Use of Tissue Culturing To Improve Food Characteristics . . . . . . . 50 2-6. Use of Parasite or Predator Insects To Control Insect Pests in the United States . . 54 2-7. Pathogens Used To Control Insects in the United States . . . . . . . . 55 2-8. Control of Weeds by Microbial Agents in the United States . . . . . . . 57 2-9. Microbial Herbicides Commercially Available or in Development in the United States 58 2-10. Use of Insects To Control Weeds in the United States . . . . . . . . 58 2-11. Biological Control Agents Commercially Available To Control Plant Disease in the United States . . . . . . . . . . . . . . . . . . . 59

PAGE 42

Chapter 2 Emerging Plant Technologies Each year in the United States weeds, insects, diseases, and poor weather conditions significantly lower crop yields. On average, major crop production in the United States achieves only about 22 percent of the yield theoretically possible under ideal conditions, based on genetic potential. Approximately 69 percent of this loss is due either to unfavorable climate and production using inappropriate farm management practices or poor soils. However, weeds, insects, and disease result in an annual average loss in total yield of 2.6, 2.6, and 4.1 percent respectively (6, 7, 8, 39). Seventy-one percent of crop insurance payments paid in the United States (from 1939 to 1978) were for crop losses caused by drought, excessive water, and cold (6, 8). The financial value of these losses is staggering. Diseases in fruits, vegetables, grains, and oilseeds result in annual average losses in value of 17, 13, 11 and 13 percent respectively. For some highly perishable fruits, such as raspberries, blackberries, and cherries, losses from disease are estimated to be 38, 34, and 24 percent respectively of their total value. Annual losses in the United States due to viral diseases alone are estimated to be $1.5 to $2.0 billion dollars (5). A recent study estimated that crop diseases resulted in lost revenues equal to approximately 15 percent of the total crop in North Carolina. This value, if extrapolated to the United States as a whole, would result in losses of approximately $12.6 billion per year (8, 28). Loss in value due to weeds has been estimated at 10 to 20 percent of the total crop value; nearly $16 billion per year. Approximately $5 billion is spent annually to control weeds on farms and in rangelands, forests, and waterways ( 10, 26). Traditional approaches to managing these problems have included the use of traditional breeding techniques to develop new crop varieties resistant to pests and better adapted to geoclimatic conditions. cultural practices, and the application of chemicals. Pest management is complicated by the fact that plant pests continuously adapt to new management techniques. The need to develop new approaches to control plant pests is paramount. New pest management methods being developed focus on biological approaches, including the use of biotechnology to alter the plant genome and the use of biological control agents. Approaches that focus on improving the plants ability to withstand adversity in general involve genetically modifying the plant to have new characteristics. Scientists genetically modify organisms by altering or adding to an organisms genetic information with the intent to improve the physical characteristics of the organism. The genetic material of living organisms is composed of deoxyribonucleic acid (DNA). 1 The universal nature of genetic material enables scientists to transfer genetic material between species that are normally not sexually compatible, and can be used to modify microorganisms (e. g., bacteria, viruses, and fungi), animals, insects, and plants. The genetic modification of plants can be accomplished using three different types of techniques: classical, cellular, and molecular (29). The classical methods of genetic modification include those associated with traditional plant breeding. Such methods include: fertilization of sexually compatible plants coupled with the preferential selection of those plants containing the desired characteristics, the use of chemicals or radiation to mutate the genetic material such that the mutated organism possesses preferred characteristics, and traditional cell culturing of plant sex cells such as anthers (the plant organelle that contains pollen) ovules, and embryos. Cellular techniques involve regenerating a whole plant using culturing techniques, but unlike classical methods, the cellular techniques use tissue cells other than sex cells. Techniques include: cell fusion, in which two sexually incompatible plants are hybridized, and somaclonal variation, 2 which involves selecting plants that have been regenerated from undifferentiated plant cellssuch plants often differ significantly from the parent plants. 1 The exception (o this statement are the viruses whose genetic material is composed of ribonuclcic acid (RNA), rather than DNA. 2 Plants arising from the culturing of undifferentiated cells often differ strikingly from each other and from the parent plant from which the culture was derived. In some unknown way, the process of culturing cells releases a pool of genetic diversity. Possible explanations of this phenomena include chromosome breakage and reunion, DNA rewmmgement, and point mutations. The amount o( ~ariation that occurs is affected by some factors that can be controlled, such as the length of time the cells are cultured, the genotype of the tissue, the medium, and the culture conditions ( 15. 30). -37-

PAGE 43

38 l A New Technological Era for American Agriculture The molecular techniques include those most commonly associated with biotechnology. Selected genes are isolated and transferred to a host organism using vectors (a piece of DNA that helps to incorporate a new gene into a host organism) or direct transfer techniques such as microinjection, electroporation, or particle guns. Molecular techniques allow for the transfer of selected genes between sexually incompatible species of the same type of organism, or between different types of organisms such as between plants and bacteria. This chapter will focus on advances made in the use of biological methods to enhance crop production. Emphasis will be given to the use of molecular techniques and the use of biological control agents to enhance both pest resistance and the ability to improve crop production in less-than-ideal conditions. 3 TOOLS AND TECHNIQUES OF BIOTECHNOLOGY Biotechnology can be broadly defined as the use of living organisms to alter other organisms. In a practical sense, biotechnology is a set of tools that allow researchers to manipulate genetic material. These tools allow researchers to develop products that could not have been previously produced, and to explore new research questions that significantly expand our scientific knowledge. This section will describe some of the most important tools of biotechnology. Biotechnology Techniques Used To Create Transgenic Plants Transgenic crops are those crops whose hereditary DNA has been augmented by the addition of DNA from a source other than parental germplasm, using recombinant DNA techniques. The primary goals of transgenic crop research is to produce crops with improved ability to resist pests (i. e., disease, weeds, and insects); improved ability to grow under less-than-ideal soil and climate conditions; and to improve the quality characteristics of crops (e. g., by changing the oil composition of oilseed crops). Many advances have been made that improve scientists ability to create transgenic plants, and several major crops grown in the United States have been successfully transformed (table 2-1). Production of transgenic crops with improved characteristics, however, is constrained by insufficient knowledge of the appropriate genes for transfer; the knowledge base in plant biochemistry and physiology has not kept up with the development of molecular biology and transformation technologies. To create a transgenic plant, scientists must: 1. 2. 3. 4. 5. isolate and purify the gene to be transferred, find appropriate mechanisms (i.e., vectors or nonvector mechanisms) to transfer the gene into plant cells, attach appropriate regulatory sequences to ensure proper expression of the new gene in the plant, insert proper genetic markers to identify those cells that have been transformed, and regenerate the transgenic cell or tissue into a complete plant. Advances and methods used to accomplish each step will be described below. Gene Identification, Isolation, and Purification Isolating a single gene is complicated by the fact that a DNA sample obtained from a plant usually contains many genes. Researchers must be able to separate the one gene of interest from all of the other genes. Once isolated, the gene of interest is multiplied (cloned) to produce enough genetic material for subsequent uses. The process used to isolate and multiple the gene of interest is generally referred to as shotgun cloning because the process allows for the replication (cloning) of the entire genome (the sum of all genetic information contained in the chromosomes) of the organism. A sample of DNA is first cut into small pieces, some of which may contain the desired gene. Special enzymes (restriction endonucleases) are used to cut the DNA at specific sites such that each piece has the same types of ends (figure 2-1 ). Pieces of DNA that have been cut with the same enzyme can be glued together regardless of the source of the DNA. This feature allows, for example, pieces of DNA from plants to be pasted together with DNA pieces from bacteria. It also allows scientists to paste DNA fragments into molecular vectors, pieces of DNA capable of inserting foreign genetic material into a cell. Scientists use vectors to help isolate and purify specific genes. Commonly used vectors include bacterial plasmids (circular pieces of DNA that can be easily in Because of the large quimtity of research on these technologies. this chapter will cite mainly OTA commissioned background papers and other review articles.

PAGE 44

Chapter 2Emerging Plant Technologies 39 Table 2-1Transgenic Crops Produced Grains and oilseeds a Fruits and vegetables Other Cotton Tomato Alfalfa Rice Sugar beet White clover Sunflower Potato Poplar Soybean Peas Lotus Rapeseed Lettuce Arabidopsis Corn Cucumber Petunia Cabbage Tobacco Asparagus Walnut Carrot Pear Celery a Wheat and barley have not yet been successfully transformed, but it is anticipated that these crops WiII also be amenable to genetic engineering by the mid-1990s. SOURCE: Office of Technology Assessment, 1992. serted into bacterial cells where they can replicate) and bacteriophages (viruses that infect bacteria). 4 To isolate a gene from an organism, the DNA sample of the organism is cut into many pieces, and all of these pieces are inserted into vectors (e.g., bacterial, plasmid, or bacteriophage). The vectors are then inserted into bacterial cells. As the bacteria reproduce, the vectors containing the pieces of the organisms DNA are also reproduced. This process results in the production of multiple copies of the organisms DNA, which is contained in the vectors. Now scientists have enough copies of genetic material to begin isolating the vectors that contain only the genes of interest. Isolation of the appropriate vectors is accomplished using a probe, a sequence of genetic material that recognizes the desired gene. The probe is used to identify the vectors containing the desired gene. These selected vectors can then be reintroduced into bacteria, where they are replicated many times to produce millions of copies of the desired genes. The desired gene can then be removed from the vector in quantities sufficient to perform subsequent genetic modifications (41 ). The above procedure can be easily applied to organisms that possess small genomes, such as bacteria, but is more difficult to apply to more complex organisms such as plants, whose genome size is huge. Additionally, difficulties occur as a result of the lack of knowledge concerning the functions of many plant genes, which precludes the development of probes. Because of these difficulties, additional methods are being developed to improve the isolation of plant genes. Restriction Fragment Length Polymorphism (RFLP) mapping is used to identify and clone plant genes and to further our understanding of the function of plant genes. RFLP maps take advantage of the fact that corresponding sites in the DNA of individual plants may differ as a result of mutations (referred to as polymorphisms). These polymorphisms can be identified and correlated with known markers (i. e., genes whose function have been identified), which helps to identify the general location of an unidentified gene (2 1 ). This procedure identifies the approximate location of a specific gene within the plant genome, which limits the amount of plant DNA that must be searched to isolate that specific gene. Once the general location of a specific gene is located, isolating the specific location of the gene depends on other methods still under developments RFLP maps are being made for corn, potato, tomato, rice, bean, pine, soybean, wheat, barley, sorghum, alfalfa, and Arabadopsis (27). Mechanisms To Transfer Purified Genes Into Plant Cells Once a gene has been isolated and purified, it can be transferred to create a transgenic plant. For many dicotyledonous plants (i. e., plants having two seed leaves (cotyledons) and net-veined leaves, such as soybeans), the Ti plasmid of certain strains of the soil bacterium Agrobacterium tumefaciens is commonly used as a vector to insert foreign genes into the plant. Unfortunately, Ti plasmids cannot be used to transform monocotyledonous plants (i.e., plants having a single cotyledon and parallelveined leaves), which includes most of the major cereal crops (e. g., corn, rice, wheat) (27). Vectorless methods have been developed to transform cereal crops. For example, chemicals (e. g., polyethylene glycol or calcium phosphate) and physical methods (e.g., electrical stimulation) are used to make plant cells leaky so that genetic material can flow in. These approaches have been used successfully to transfer foreign genes into rice and corn (27). 4 Plasmids are commonly used to construct cDNA libraries (see ch. 3 ) and bacteriophages are used to construct genomic libraries. 5 Methods being developed include chromosome walking in which successively smaller overlapping portions of the RFLP fragment are isolated until one walks to the desired gene. This method is constrained by the fact that RFLP fragments may still be too large to clone by the conventional methods described above (27). Another method is called gene tagging, which uses a transpmon (a piece of DNA capable of moving around in the genome) to activate the gene of interest. The gene can bc located by locating the transposon. Use of this method is inhibited by the size of the plant genome, the lack of transposons for many crops, and the fact that the transposon is often naturally present in multiple copies in crops (27).

PAGE 45

40 l A New Technological Era for American Agriculture Figure 2-1 Identification and Isolation of Desired Gene DNA containing gene to be isolated Recombinant DNA Molecules Multiplication of Bacteria Containing the Desired Gene To Yield Many Identical Copies of Fragments SOURCE: Office of Technology Assessment, 1989

PAGE 46

Chapter 2Emerging Plant Technologies 41 Figure 2-2Gene Transfers With Bioblaster driving the projectile carrying new genes and tungsten powder against the I I The bolistic method is an alternative vectorless method of gene transfer. This method uses a particle gun to shoot high-velocity microprojectiles coated with DNA into a plant (figure 2-2). It has been used to transfer genes to tobacco, soybean, and corn (27) and can be used to transfer genes to the plant cell nucleus (where the chromosomes are located) and potentially to other cell organelles that contain genetic material, such as the chloroplast (e. g., genes involved in photosynthesis) and the mitochondria (e. g., cytoplasmic male sterility genes used in the development of some hybrid crop varieties). Currently, there is little control over where the foreign gene is inserted into the host plant. New methods are being developed to target the insertion site, but the frequency of success is low. Use of Selectable Markers To Identify Transformed Plants Cells that have foreign genes inserted need to be differentiated from those that have not been transformed. Scientists use markers to identify the transformed cells. The most commonly used marker is the kanamycin resistance gene. Cells containing this gene are resistant to the antibiotic kanamycin and will grow on a culture medium containing high levels of that antibiotic. Untransformed cells not containing the kanamycin resistance gene will not grow on this medium. Genes coding for herbicide tolerance can also be used as a selectable marker to differentiate transformed plants from those that have not been transformed. Use of Promotors To Control the Expression of the Foreign Gene Once a foreign gene has been incorporated into the genetic material of a plant, it must still function properly. Scientists use promotors (regulatory genes) to control when and where in the organism the gene is turned on. To date, most transgenic plants contain constitutive promotors, which means that the foreign gene is expressed equally in all tissues and at all development stages. Scientists are trying to isolate promotors that turn the inserted genes on only in specific tissues at certain development stages of the plant, and at a specific time. For example, it is desirable to direct the expression of insect tolerance genes only to the tissues eaten by the insect, such as leaves. The most commonly used plant promotor to date is derived from the cauliflower mosaic virus and is mostly constitutive. However, promotors that respond to light, heat, wounds, and oxygen deficiency, and that show tissue specificity for seeds, pollen, root nodules, and tubers are being identified (27). Understanding the molecular basis of promoter-mediated regulation of gene expression as well as isolation of promoters with varying specificities of expression is critical for the development of new generations of plant-based biotechnology products. Use of Tissue Culture To Regenerate Transformed Plants Once a plant cell or tissue has been genetically transformed, it must be regenerated into a complete plant. Advances in plant tissue culturing techniques have now made it possible to regenerate many of the most important crops (figure 2-3). Early genetic modification research used protoplasm culturing to regenerate the transformed plant cells. Protoplasts are formed by enzymatically removing the outer wall of plant cells. These protoplasts are genetically transformed using the tools of biotechnology, then coaxed into forming a cell wall and eventually growing into a complete plant. However, such regeneration is difficult to achieve with many plant cells, which has lead to the development of callus culturing and cell-suspension methods. Callus tissue cultures originate from tiny pieces of tissue snipped from seedling shoots or other appropriate plant parts. The tissue is placed in a petri dish containing plant hormones and other plant nutrients. The cells grow

PAGE 47

42 A New Technological Era for American Agriculture Figure 2-3Plant Tissue Culture Technology immature fruit 1 I i SOURCE: S.K. Harlander, University of Minnesota and divide, forming a mound of undifferentiated cells called a callus. When transferred to a regeneration medium, the cells in the callus differentiate into roots and shoots, which then grow into plants. Thousands of plants can be regenerated from one piece of tissue, but the process is labor intensive and expensive. Methods for the growth of cell suspensions allow for the regeneration of plants from single cells rather than clumps of tissue. Tissues can be agitated in a flask containing a liquid medium, causing the cells to separate. In the appropriate medium, these cells will form somatic embryos that differentiate into entire plants. Embryo suspensions have been used to regenerate wheat, sorghum, and corn (27). Callus culturing and cell-suspension methods allow for the use of a variety of plant tissues (e. g., leaves, stems, shoot tips, or cotyledons) from many plant species to be used to regenerate new plants. And, Agrobacterium particle gun technologies or other direct methods can be used to transform these tissues. Thus, most major crops can now be genetically engineered and regenerated to complete plants. Other Biotechnology Techniques Biotechnology is most closely identified with the use of recombinant DNA technologies to produce transgenic crops as described above. However, other technologies, some of which also involve the use of recombinant DNA, will also play a significant role in the development of new plant technologies. Some of these technologies are described below. Antisense Technology Antisense technology is a powerful research tool that enables scientists to study the physiology and development of organisms. It is also useful in the production of transgenic crops that have new characteristics (37). For example, this technology is being used to prevent softening in tomatoes (see Biotechnology in Food Processing). The power of the technique lies in its ability to eliminate or reduce the expression of a gene in an organism. An analogy that might help to explain how this technology works is to view the expression of a gene as being similar to reading a sentence. For the sentence to make sense, it must be read in a certain direction; sentences that are read backwards, for instance, dont make sense. Gene expression is similar, A gene must be read in a certain direction to produce a gene product that makes sense to the organism (i. e., it is a functional compound). The antisense technology consists of incorporating into an organism a synthetic gene that reads backwards (i. e., a product is made that doesnt make sense to the organism). The expression product of this backward-reading gene is a mirror image of the expression product of the same gene when it is read forward. When the expression products of the forward and backward genes meet, 6 they stick together, thus inactivating the product of the forward-reading gene (figure 2-4). Thus, the antisense technology can be used to inactivate selected genes in the plant. Use of the technique, however, is constrained by the need to know the precise nucleic acid sequence of at 6TechnicaHy, when a gene is expressed, it is first copied and modified to a second compound called messenger ribonucleic acid (mRNA). The mRNA then serves as the template for the subsequent production of proteins. It is the mRNA, rather than the protein, that meets and causes the inactivation.

PAGE 48

Photo credit: U.S. Department of Agriculture, Agricultural Research Service. Molecular biologist at UC/USDA Plant Gene Expression Center successfully transferred new genes into cells of corn using a gene gun. least a portion of the gene that codes for the expression product to be inhibited. Polymerase Chain Reaction The Polymerase Chain Reaction (PCR) technology enables scientists to rapidly generate large amounts of genetic material from a trace amount, which would otherwise be too small to analyze. PCR is an enzymatic process carried out in repeated cycles, each of which doubles the amount of DNA present. Small flanking sequences of DNA are identified on each end of the DNA sequence that is amplified. These flanking sequences are then used to create complementary strands of DNA that serve as primers. These primers are then annealed to the flanking sequences, and when appropriate enzymes and nucleic acids are added under the proper conditions, a new DNA strand is formed beginning at the primer and extending across the sequence of DNA to be replicated, such that a copy of this sequence is made. This methodology is rapid, sensitive, and relatively easy to carry out; about 25 cycles can be carried out in an hour. PCR reduces the difficulty of isolating and manipulating specific DNA 7 The spleen cells are fused in the presence of an agent. such a polyethylene giycol, to myclorna cellstumors of B Iynlpht)c}te origin. X Alternatively the hybrid cells can be grown as tumors in the peritoneal cavltics of mice where very high levels of antibody accumulate in the ascites fluid surrounding the tumor.

PAGE 49

44 l A New Technological Era for American Agriculture Figure 2-5Preparation of Monoclinal Antibodies Myeloma cells are mixed and fused with B lymphocytes SOURCE: Office of Technology Assessment, 1988. antibody production methods). It is this purity that makes monoclinal antibodies so useful. Application of Biotechnology Techniques To Create Transgenic Plants The tools of biotechnology are allowing researchers to explore new means to control plant diseases, insect pests, and weeds. Tissue culturing and genetic engineering, combined with traditional agricultural research methods, The products of this fusion are grown in a selective medium. Only those fusion products which are both immortal and contain genes from the antibody-producing cells survive. These are called hybridomas. Hybridomas are cloned and the resulting cells are screened for antibody production. Those few cells that produce the antibodies being sought are grown in large quantities for production of monoclonal antibodies. are allowing scientists to alter plants or biological control agents to achieve enhanced efficacy and host range in controlling plant pests. Biotechnology is also being used to improve a plants ability to withstand environmental stresses, such as cold, drought, and frost, improve the shelf-life of fruits and vegetables and is being used to develop value-added products from agricultural commodities (e. g., increased carbohydrates. modified oils, and proteins that contain essential amino acids). In addition to developing new products, the tools of biotech-

PAGE 50

Chapter 2Emerging Plant Technologies l 45 Photo credit: U.S. Department of Agriculture, Agricultural Research Service. Plant molecular biologist examines successful results of the cloning of a gene necessary for plants to synthesize ethylene, the ripening hormone. More recently, scientists have blocked this gene, producing genetically engineered tomatoes that ripen on demand. nology are expanding the knowledge base of plant resistance and the interactions of plants, pests, and biological control agents with the rest of the ecosystem. Genetic Engineering of Plants for Insect Control Traditional breeding programs have successfully produced varieties of alfalfa, cotton, corn, rice, sorghum, soybean, and wheat that have been resistant to, or tolerant of, key pests and will continue to play an important role in developing insect resistant plants for some time in the future. However, the tools of biotechnology have created the possibility of selectively engineering plants for insect resistance. Biotechnology will permit the transfer of resistance genes into plant species for which the resistance gene is not inherent. Biotechnology is also being used to improve the understanding of mechanisms by which plants are resistant to insects. Few genes known to produce insecticidal proteins have been identified. Candidate genes must code for proteins that are stable in the plant cell, are not rapidly digested when consumed by insects, have high activity against feeding target insects, and are safe for nontarget invertebrates and animals. Insecticidal proteins produced by the spore-forming bacteria Bacillus thuringiensis (Bt) are among the few known to meet these criteria, The Bt bacteria produces crystals that contain compounds toxic to insects. Insects feeding on plants contaminated with Bt bacteria ingest the crystals, which are dissolved in the insect midgut, releasing the protein toxPhoto credit: Monsanto Co. Tomato plants that show one stripped by caterpillars and one not. The plant not stripped contains the Bacillus thuringiensis toxin gene. ins. Different strains of the Bt bacteria produce insecticidal toxins specific to Lepidoptera (butterflies and moths) only, to Diptera (flies and mosquitoes) only, to Coleoptera (beetles) only, and to both Lepidoptera and Diptera. Genetic engineering is being used to improve the delivery of the Bt toxin to insect pests by incorporating the insecticidal gene into other vectors (see Biological Control of Anthropoids: Pathogens) or by transferring the insecticidal gene directly to plants. Genes coding for the Bt insecticidal protein have been cloned and inserted into tobacco, tomato, and cotton plants among others ( 1). Transgenic plants producing Bt insecticide are expected to be commercially available by the mid to late 1990s. Genes for some insect trypsin inhibitors have also been cloned. Trypsin inhibitors are compounds that, when present in large amounts, may reduce the ability of an insect to digest plant material. Some plants, such as the seeds of cowpeas and beans, contain large quantities of trypsin inhibitors (i.e., 1 to 2 percent of the total protein), and the levels in plant leaves may be increased in response to mechanical damage or insect feeding. Trypsin inhibitor genes derived from tomatoes have successfully controlled the growth of insect larvae when transferred to tobacco plants. Transgenic plants genetically engineered to produce trypsin inhibitors may be available by the end of the decade ( 1). Genes that code for lectins and for arcelin are also potential candidates to confer insect resistance to transgenic crops. Lectins are sugar-binding proteins found in the seeds of peas and common beans. They are effective against bean weevils and cabbage weevils. Arcelin is

PAGE 51

46 A New Technological Era for American Agriculture produced in the seeds of wild beans and is toxic to bean bruchid pests (1). Genes coding for insecticidal proteins other than Bt toxins and trypsin inhibitors must be identified. RFLP maps are being used in tomatoes, for example, to discover the location of insect resistance genes in plants. The development of tissue-specific promotor sequences and promotors that respond to selected environmental stimuli are needed to improve the efficacy of insect control. Genetic Engineering of Plants for Weed Control The presence of weeds in crops decreases productivity and crop quality. To control weeds, farmers commonly apply herbicides. Most herbicides act by inhibiting key enzymes in photosynthesis or other essential plant biosynthetic pathways. Plant species respond differently to herbicides depending on the sensitivity of plant enzymes to the herbicide or the ability of the plant to metabolically inactivate the herbicide. These abilities explain why herbicides are often effective against either grassy or broadleaf plants, but not both (26). Herbicide manufacturers would like to develop broadspectrum herbicides active against all economically important weeds, but their efforts have be constrained because broad-spectrum herbicides not only kill weeds, but they injure crops as well. Two approaches have been taken to minimize crop damage when using broad-spectrum herbicides. One approach is to use herbicide antidotes, compounds that enhance the metabolic inactivation of herbicides in plants (19, 20). Few such antidotes have been discovered, however, and it is unlikely that this approach will yield significant success in the near future. The alternative approach is to develop crop varieties that are resistant to the herbicide used. Traditional methods have been used successfully to develop herbicide-tolerant crops. Tissue culture and plant regeneration techniques have produced tobacco and soybean varieties tolerant to sulfonylurea herbicides and corn varieties tolerant of imidazolinone. Attempts to develop herbicide-tolerant crops using tissue-culture techniques are most successful when the herbicide affects only one compound in a plant biosynthetic pathway (i. e., it has a single target site) and a mutation in that compound confers herbicide tolerance without affecting the growth of the plant, or when the mutation of a single plant gene increases the ability of the plant to inactive the herbicide or to absorb less of the herbicide. Use of these methods is constrained by the lack of naturally occurring herbicide tolerance genes in crops (26). Genetic engineering techniques overcome the lack of naturally occurring herbicide resistance genes in plants by allowing for the transfer of these genes between crop species. Thus, crops tolerant to a specific herbicide (but not all herbicides) can be developed. Three different approaches have been taken to engineer crops successfully for herbicide tolerance, the first of which are expected to be commercially available by the mid 1990s (table 22). One approach relies on making the crop produce excess quantities of the enzyme normally affected by the herbicide. By producing an excess quantity of the enzyme, a sufficient quantity is still available to catalyze important plant biosynthetic pathways even though some of the enzyme has been inactivated by the herbicide. Excess production can be achieved by inserting several copies of the gene coding for the enzyme into the plant, or by using promotor sequences that cause excessive expression of the genes coding for the enzyme, This method has been used successfully to produce crops tolerant to glyphosate and phosphinothricin (26). The most commonly used approach to produce crops tolerant to herbicides is to alter the gene coding for the enzyme affected by the herbicide in such a way that the resulting altered enzyme is still effective in the plant, but is not inactivated by the herbicide. This altered gene is then inserted into the plant where it produces an altered enzyme that confers herbicide tolerance. This approach has been used to produce crops tolerant to glyphosate, sulfonylureas, phosphinorthricin, atrazine, and imidazolinone. The third approach is to transfer to plants those genes that code for enzymes that inactivate herbicides. This approach has been taken to confer plant tolerance to bromoxynil, 2,4-D, and phophinothricin. An alternative approach to weed control is to develop crops that produce their own herbicides. These plantproduced herbicides, called allelochemicals, can be either volatile organic compounds released into the air or soil where they can be absorbed by the weed or nonvolatile organic compounds released as root exudates or leachates of other organs, such as seeds. Most volatile allelochemicals are terpenoids whose secretion increases with rising temperatures, while most nonvolatile allelochemicals are aromatic chemicals (26). Significant research is still needed before crops can be engineered to produce allelochemicals. Alternatively, it may be possible to identify and use plants known to naturally produce allelochemicals as cover crops or in low tillage

PAGE 52

Chapter 2Emerging Plant Technologies l 47 Table 2-2Current Targets for Crop Modification for Herbicide Tolerance Research Commercial Weed/crop Herbicide institution introduction targets Atrazine Bromoxynil Betanal 2,4-D Dicamba Glyphosate Imazapyr Metribuzin Basta Sulfonyl ureas Ciba Geigy, Inc Calgene, Rhone-Poulenc Schering Max Planck Sandoz Monanto, Calgene American Cyanamid Molecular Genetics Mobay Hoechst DuPont Not expected to be commercialized Mid 1990s Late 1990s Not a commercial target Late 1990s Mid 1990s Early to mid 1990s Late 1990s Mid 1990s Mid 1990s NA Broadleaf/dicots Broadleaf/sugar beet NA Broadleaf/NA Broad spectrum soybean, rape, cotton, corn Broad spectrum/corn Broad spectrum/ soybean Broad spectrum/ rape, beet, potato, soybean, corn Broad spectrum/ soybean, rape NA = Not applicable. SOURCE: Off Ice of Technology Assessment, 1992 situations to control weeds. For example, it has been shown that certain cucumber strains produce compounds toxic to the weeds proso millet and barnyard grass under field conditions. The possibility of using alleochemicalproducing plants is also being explored in fruit production (33). Understanding the nature of allelochemicals in addition to the advances that have been made in elucidating the mechanisms of herbicide action is expected to enhance the design of future herbicides. Genetic Engineering of Plants for Disease Control Bacteria, fungi, parasitic seed plants, nematodes, insects, and viruses, among other organisms, can destructively alter the structure or physiological processes of plants, resulting in disease. However, plants possess the ability to resist the invasion of pathogenic organisms. All of the plants of a species can be resistant to a pathogen, or certain varieties of a plant species can be resistant to a subspecies of the pathogen (i. e., cultivar specificity). The interaction of bacterial and fungal pathogens with plants is helping to elucidate the mechanisms by which plants resist pathogenic organisms (27). The ability of plants to resist pathogenic organisms involves the complex interaction of genes in both the plant and the pathogen. The interaction of compounds produced by plant resistance genes and genes in the pathogen (i. e., avirulence genes) triggers a hypersensitive response. Plant cells initially infected by the pathogen die, preventing the spread of the pathogen to the rest of the plant. Thus, the pathogenic effects remain localized at the site of initial infection, and disease is prevented from spreading throughout the plant. The mechanisms by which pathogens infect plants are also being elucidated. Pathogenic microorganisms contain pathogenicity genes that produce compounds toxic to the plant and/or allow the pathogen to attach to the plant, penetrate the cuticle and degrade the walls of plant cells, and degrade chemicals produced by the plant in its own defense. These pathogenicity genes can be activated by signals from the plant itself. For example, the presence of cell wall degradation products in plants can trigger the production of enzymes in some pathogenic fungi that degrade the cell wall. In a similar manner, compounds produced by pathogens trigger a response by the plant to the pathogen. Plant defense genes are stimulated to produce compounds that may be toxic to pathogens, reinforce the cell wall, and/or inhibit enzymes produced by the pathogen (27). Efforts are underway to clone and characterize pathogen and plant genes involved with resistance. To date, no plant resistance genes have been cloned, however, avirulence genes from bacteria and viruses but not fungi, have been. Additionally, few plant defense genes have been identified and cloned. Only the gene coding for chitinase, a compound that is toxic to fungi, has been shown to confer disease resistance when transferred to tobacco. Also a compound derived from moths, when

PAGE 53

48 l A New Technological Era for American Agriculture Table 2-3Virus Coat Proteins Engineered Into Plants Photo credit: Richard Nelson, Samual Roberts Noble Foundation. Transgenic tomato plant expressing the coat protein gene of tobacco mosaic virus (left) and control plant (right). transferred to tobacco, decreased the severity of an infection by the bacteria Pseudomonas solanacearum. Given the state of the art, it is highly unlikely that plants resistant to bacteria and fungi will be developed before the year 2000 (27). Greater success has been achieved in developing plants resistant to viruses. Plants have long been known to display cross protection, a phenomena that occurs when plants infected with a mild strain of a virus do not develop severe symptoms when challenged with a stronger strain of the same virus. Cross protection is comparable to immunity in animals, although plants do not have immune systems and the mechanism of protection differs. Although cross protection has been achieved in plants by inoculating individual plants with a mild virus strain, this process is very labor intensive and carries a small risk that the virus strain used will become more virulent and act in a synergistic fashion with other viruses (27). Genetic engineering has been used to avoid these problems. Genes coding for virus coat proteins (i.e., the proteins that make up the shell that surrounds viruses), other Alfalfa mosaic virus Cucumber mosaic virus Potato viruses S, Y, and X Potato leaf roll virus Tobacco mosaic virus Tomato mosaic virus Tobacco rattle virus Tobacco streak virus Soybean mosaic virus Papaya ringspot virus Tomato spotted wilt virus SOURCE: Office of Technology Assessment, 1992. virus proteins, and virus RNA sequences can be introduced into plants to elicit a resistance response (3, 4). Plants engineered with coat protein genes from a specific virus have resisted subsequent infection by the same virus, and in some cases to related viruses having similar coat proteins. Currently, many viral coat protein genes from different plant viruses have been transferred to plants to confer resistance (table 2-3) (4). The mechanism by which protection occurs is not fully understood. Most evidence suggests that the accumulation of viral coat proteins in plant cells interferes with the release of viral RNA needed to initiate infection (4). In addition to viral coat proteins, other viral genes have been transferred to plants. Those having potential for virus control include: genes for virus replication, antisense RNA, satellite RNA, and ribozymes. The antisense technology has also been used to inhibit viruses in plants. Other approaches include transferring satellite RNA sequences (small RNA sequences that depend on helper viruses to replicate and package new virus particles) to plants where they have protected the plant from developing symptoms in response to an infection by the helper virus. Genes coding for RNA sequences that act like enzymes (i.e., ribozymes) have also been transferred to plants where they have cleaved invading viruses (27). Genetically engineered dicotyledonous plants resistant to certain viruses are expected to be commercially available by the mid 1990s. Monocotyledonous plants resistant to viruses will probably not be available until the late 1990s or early the next century. Currently, only a few genes with potential for controlling fungi and bacteria have been identified, cloned, and introduced into plants (see table 2-4). Genetic Engineering of Plants for Thermal and Water Stress Tolerance Progress in improving the tolerance of plants to water and thermal stress will depend, in part, on better ways

PAGE 54

Chapter 2Emerging Plant Technologies 49 Table 2-4Disease Resistance Genes Introduced Into Plants Disease pathogen Gene/plant Fungal. . . . . Chitinase/tobacco Bacteria. . . . . Antibacterial protein from moth/ tobacco, potato Enzyme to detoxify bacterial toxin Viral. . . . . Viral coat protein Other virus genes Satellite RNA RNA enzyme (ribozyme) Antisense RNA SOURCE: Office of Technology Assessment, 1992 of defining and quantifying these stresses as well as nonstress states. Defining these stresses is further complicated by the fact that water stress and temperature stress are not easily separated, particularly at high temperatures. New tools, such as remote and contact sensing, 9 are being developed to detect plant stress (9). The lack of detailed knowledge of the physiology of water and temperature stress tolerance also constrains progress in this field. The root system of the plant exerts major control over water uptake. Little research has been conducted to measure root response to water and thermal stress. Most measurement techniques used to date are disruptive if not destructive to root systems. New techniques are needed to determine factors that affect the distribution of roots in the soil and the ability of the roots to absorb water and transport that water through the vascular tissues of the plant (9). Plant-cell culturing, combined with selection for enhanced ability to adjust the salt and water concentration of plant cells (osmotic pressure), has been shown to be effective in improving drought tolerance. However, while improved sensitivity to osmotic pressure has increased the survival of the plant, it does so at the expense of plant growth and yields (34). Some plants contain genes that code for proteins conferring tolerance to extremes of temperature or drought; these genes are possible candidates for isolation and transfer to other plants through genetic engineering techniques. For example, tobacco cells that are exposed to gradually higher levels of salt synthesize several novel proteins. One such protein is osmotin, whose synthesis is regulated by several mechanisms, including exposure to low water environments or changes in endogenous levels of the hormone abscisic acid (ABA). ABA is known to lower the rate of transpiration from leaves and prevent water loss. The role of osmotin in cellular osmoregulation is now under investigation (9). Some plants, when challenged by elevated temperatures, produce heat shock proteins. Genes coding for several of these proteins have been sequenced and their promoter regions identified. However, the metabolic functions of most of these proteins are not understood, and this constrains their use in biotechnology to improve plant tolerance to elevated temperatures (9). In general, the fundamental research needed to understand the mechanisms of tolerance to thermal and water stress simply has not kept pace with the development of biotechnology tools, and thus, scientists do not currently know what genes to transfer into plants to improve tolerance for these stresses. Thus, genetically engineered plants tolerant to elevated thermal or water stress are unlikely to be developed within this decade. However, antifreeze proteins have been transferred to plants and production of plants with improved cold tolerance may become available within 10 to 15 years. Plants transgenic for antifreeze proteins have the potential to improve cold hardiness by lowering the temperature at which leaves freeze ( 12, 17). Antifreeze proteins from fish are also being used to improve the post-harvest freezing and thawing qualities of fruits and vegetables by inhibiting ice recrystallization in tissues (22). Biotechnology in the Food Processing Industry Historically, the food processing industry has had to accept and adapt to heterogeneous raw materials. Biotechnology can be used to better tailor food crops to meet food processing and consumer needs. Tissue-culture techniques are being used to select or construct crop varieties with improved functional, processing, or nutritional characteristics (table 2-5). Plant tissue-culture techniques can be used to produce food flavor and coloring ingredients. These methods could potentially replace production and extraction of these ingredients from plants ( 15, 18). For example, a private company recently has succeeded in using tissue culture techniques to produce vanilla ( 14). Contact sensing requires contact with plant tissues and may require destruction of at least part of the plant. It involves the direct ctetcrmination of the state of a physical, biological, or chemical quantity. Remote sensing quantitates parameters meusured by using a sensor to detect electromagnetic waves emitted or reflected by plants.

PAGE 55

50 A New Technological Era for American Agriculture Photo credit: U.S. Department of Agriculture, Agricultural Research Service. Framed by drought-dried cornstalks, drought-resistant lima beans stand tall and lush in test plot. Scientists hope that genetic engineering researchers can isolate the genes that give the lima bean such a high degree of drought tolerance. Table 2-5Use of Tissue Culture To Improve Food Characteristics Crop Characteristic Tomato . . . . Increased solids Increased shelf life Carrots . . . . Increased sweetness, crunchiness Celery . . . . Decreased stringiness Corn. . . . . Improved amino acid composition Rapeseed. . . . Decreased saturated fatty acids SOURCE: Office of Technology Assessment, 1992. Genetic engineering is also a means of altering food characteristics. Genes coding for enzymes involved in starch and lipid biosynthesis are being isolated and cloned, enhancing the prospects of engineering plants with specific composition of starch and oil. Genes coding for floral pigment pathways are also being isolated. Plants potentially can be engineered to produce pharmaceuticals such as blood clotting factors and growth hormones. For example, oilseed rapeseed has been genetically engiPhoto credit: DNA Plant Technologies, Inc. Vegi Snax is an example of successful application of plant tissue culture for selection of crop varieties with improved functional, processing, and nutritional characteristics. neered to produce enkephalins (40). In addition, ense technology is being used to eliminate toxins compounds, or off-flavor components in plants, delay ripening of tomatoes ( 15). antis allergenic and t o Biotechnology is also being used to improve microorganisms used as vegetable starter cultures and in brewing and baking (i.e., organisms used in making sauerkraut, pickles, olives, soysauce, wine, beer, and bread) such that these organisms tolerate different temperature and pH ranges. Similar work is being conducted with microorganisms used to produce food ingredients such as acetic acid, citric acid, niasin, vitamin B 12, xantham gum, and monosodium glutamate. In addition, genetically engineered enzymes are being developed to treat food processing wastes ( 18). Finally, biotechnology is being used to develop methods to assay levels of pathogens, toxins, and chemical contaminants in raw ingredients and final products. DNA probes and poly and monoclinal antibody kits are beginning to replace traditional bioassay methods. For example, many of the assay procedures used to detect pesticide residues in food are monoclinal antibody kits ( 18). THE TOOLS AND TECHNIQUES OF BIOLOGICAL CONTROL Approaches Used in Biological Control Biological control of pests relies on using living natural enemies (e. g., parasites, predators, and pathogens) to reduce pest populations to levels lower than would otherwise

PAGE 56

Photo credit: Calgene, Inc. Antisense tomatoes (left) and control (right) 3 weeks after harvest. occur ( 13). Parasitic organisms are those whose development takes place in or on a single host organism; predator organisms are those that consume other organisms as a food source; and pathogenic organisms are those that cause disease in other organisms. Many organisms, including insects and other arthropods (e. g., spiders and mites), bacteria (and related organisms such as rickettsiae and mycopiasmas), viruses, fungi, protozoa, and nematodes are being used as biological control agents to manage weeds, insects, and other arthropod pests, as well as disease organisms in economically important plant species. Biological control methods have been used in the United States on a limited basis for at least 100 years. Approaches used can be classified into three common types-the classical approach, augmentation, and conservation (25). Biological control agents used to control nonindigenous pests, particularly those introduced from other countries, is called the classical approach. When a non-native pest is introduced into a new environment, often there are no natural enemies to control that pest. The classical approach searches the area of origin of the pest and identifies natural enemies. These natural enemies are then introduced into the new environment to control the pest (25). Attempts are made to establish the introduced natural enemies as part of the ecosystem so that pest suppression will be permanent. The augmentation approach focuses on increasing the existing population of indigenous pest enemies. Small numbers of natural enemies can be released periodically, as needed, to increase the indigenous population to levels sufficient to control pest numbers at levels below those that cause serious economic problems. The newly released natural enemies are expected to become part of the ecosystem, and to help suppress more than one generation of pests (25). This approach is similar to administering a booster shot to augment indigenous-pest enemy populations. Alternatively, large numbers of natural enemies can be released at one time with the intent of quickly suppressing the pest population by creating an epidemic-like situation. The control agent (i. e., natural enemy) is not expected to become a permanent part of the ecosystem and the natural enemy is not expected to control more than one generation of the pest. The natural organisms used with this approach are usually microorganisms, such as bacteria and fungi. They are manufactured, formulated, standardized, packaged, registered as pesticides, and applied to pests using methods and tools similar to those used for chemical pesticides. Because of these similarities to chemical pesticides, this strategy is often referred to as the microbial pesticide or inundative approach to augmentation. This approach generally requires regular application because the control agents do not survive between crop seasons. or survive in insufficient number to be effective the next season, or are prevented by other factors from causing significant disease in the pest population ( 10, 16). Conservation practices can be used to protect and maintain natural enemy populations by manipulating the environment. such as altering cropping patterns and farm management practices to enhance the indigenous population, maintaining refuges and providing feeding and nesting sites for natural enemies, and by applying pes-

PAGE 57

52 l A New Technological Era for American Agriculture ticides only when pest populations exceed specified levels ( 16, 25, 35). In general, the classical method of biological control has been the approach most frequently and successfully used to control weeds, insects, and other arthropods in the United States. This is perhaps not surprising given the large number of pests that are of foreign origin. For example, an estimated 39 percent of the 600 most important arthropod pests in the United States are of foreign origin and more than 630 additional foreign arthropods are on the list of lesser pests (36). Based on past history, it is predicted that exotic arthropod species will continue to be added at a rate of about 11 species per year and that approximately 7 of those species will become significant pests. Clearly the classical approach will continue to be a major biological control methodology. Biological control approaches have had limited success against pests in grain and row crops. 10 Biological control has been most successful against naturalized permanent pests in areas of low disturbance (such as rangeland, pastures, forests, and some aquatic habitats) where the targeted pest is the dominant species, and where the end goal is a stable plant community. The poor record of success in grain and row crops is often attributed to the fact that grain crops only persist for short periods of time, during which the natural enemy must discover the crop and become established, must find and attack its host pest, and must increase its population to numbers sufficient to reduce the pest population significantly. The abrupt end of the crop season precludes the establishment of stable interactions between pests and natural enemies in grain crops ( 13, 16, 25). It is perhaps for these reasons that the microbial pesticide approach using fast-acting pathogens has received more research attention than any other biological control approach to pest suppression in grain and row crops. The bacterium Bacillus thuringiensis, which produces compounds that are quickly toxic to some insects, can be used effectively in this manner. The microbial pesticide approach is also being taken to develop fungi that control weeds ( 10, 16). The conservation approach has received the least research attention. Little incentive exists for the private sector to develop these technologies because the product that is developed is management information. Successful development of this approach will most likely fall to public sector researchers. Methods to control communities of organisms in a systemic fashion rather than a single control agent are needed (11 ). Research Needs Extensive research in many disciplines will be required if biological control is to become more widely used. A better fundamental understanding of pest-natural enemy interactions, ecology, and population biology is needed, as well as attention to more applied problems of mass rearing, formulation, and delivery required to make these control agents commercially viable. Successful development will require a multidisciplinary approach and will draw from expertise in many fields, including: systematic (taxonomy), ecology, behavioral science, physiology, genetics, chemistry, and epizootiology (the study of population disease at the population level), among others ( 10, 16, 25, 38). Taxonomic, biochemical. and genetic comparisons of pests from the same or similar species taken from geographic areas of suspected evolutionary origin also are needed. These studies can help identify pests and their natural enemies, improve understanding of the relationship between pest and enemy, and determine the geographic distribution of each. Use of classical biological control methods will be enhanced if techniques can be developed to detect and eliminate parasites and pathogens from the imported natural enemy cultures ( 10, 16, 25, 38). An improved understanding of the natural enemy-pest dynamics and factors that enhance the effectiveness of control is needed. Elucidation of the structure and roles of insect hormones and compounds that attract or repel pests is needed. Additional research is needed to understand the natural enemy population (i. e., infectivity, virulence, specificity of host; biological fitness including survival, persistence, and dispersal; the role of population density, etc.), the pest population (i.e., susceptibility, development of resistance, mechanisms of immunity, population density impacts, and distribution), the effects of the abiotic and biotic environment (i. e., weather, soils, host plants, biotic transport agents, sunlight, cropping patterns, etc.), and the environmental impacts of releasing predators, parasites, and pathogens to control pests (10, 16, 25, 38). A major constraint to using the augmentation approach to biological control is the inability to cost-effectively 10 Recent ~ork With bacu]ovi~~e~ tc) ~c)n[rol insects has been promising and this biological control agent may prove to be an cxceptk)n tO this statement ( 16).

PAGE 58

Chapter 2Emerging Plant Technologies l 53 raise large numbers of parasites, predators, and pathogens. The life cycles of many natural enemies are complex and raising these organisms in an artificial setting is difficult. New mass rearing techniques need to be developed for many biological control agents. For natural enemies that are parasitic insects, laboratory rearing requires maintaining not only the host insect, but the food source of the host insect as well, which may include plants that are themselves difficult to grow. Thus, mass rearing of a parasitic insect requires maintaining both an appropriate plant population and host insect population, a costly arrangement that points to the need to develop artificial diets ( 10, 25). Viruses can also be difficult to mass produce. Viruses are obligate cellular parasites and must be produced within living cells. For viruses that are pathogenic to insects, this can be accomplished either by infecting whole insects or by infecting cultures of continuous cell lines derived from the host insect. Recent advances in insect cell culture is improving the prospects of virus pesticide production. Significantly, most of these advances are being made in the biomedical field rather than the agricultural field, because biomedical industries are using certain classes of viruses (such as baculoviruses) as vectors to express foreign genes for high-level production of biological and pharmaceutical products ( 16). Mass production techniques for fungal spores are also needed. The application of automated systems and robotics to mass production could potentially significantly reduce the cost. Other problems encountered while mass rearing natural enemies include the loss of genetic variability and the loss of effectiveness of species that have been raised for several generations in the laboratory ( 16, 25). The performance of biopesticides in the field has often been highly variable due to environmental factors, interactions with other organisms, and poor delivery to target organism among other problems. Formulation of biopesticides (mixing of the cultured microbial preparation with inert agents to achieve proper dilution, deposition, moisture holding capacity, protection from ultraviolet rays, shelf life, slow release, etc. ) must be improved to increase efficacy in the field. Long-range needs include identifying new control agents, increasing the toxicity of agents against susceptible pests, and expanding the range of hosts of the control agent (10, 16). Delivery systems also need to be improved. Techniques must be designed to promote maximum efficacy and ease of application. New sprayer technologies, application of biopesticides by irrigation methods, and timed release formulations are needed. Finally, a general need exists to assess the efficacy and impacts of control agents after release. Studies using biological control agents have rarely adequately documented efficacy, reliability, and economic feasibility. Population establishment and buildup, degree and timing of feeding damage, plant population density and productivity, plant stress, and nontarget side effects need to be assessed. Any changes in the fitness of the naturalized bioagent need to be ascertained to ensure efficacy and environmental safety. While these questions are pertinent to all biological control agents, they will be critical to regulatory approval of genetically engineered control agents ( 10, 16, 25, 38). Use of Biotechnology in Biocontrol Research Traditional technologies, such as chemicalor ultraviolet-generated mutations followed by selection for desired phenotypic traits, and sexual mating will continue to play a role in producing and identifying natural enemies via improved control capability or host range. Additionally, traditional culture techniques can be used to induce increased secretion of certain toxins and enzymes involved in pathogenesis. However, new biotechnology tools, such as protoplasm fusion and gene transfer, will also be used to improve virulence, sporulation, fitness for survival, infectivity under suboptimal conditions, and production of pesticidal metabolizes; and to expand host range and the tolerance of control agents to certain chemical pesticides ( 10, 16, 25, 38). Biotechnology to improve biological control agents, such as insects and other arthropods, nematodes, protozoans, and fungi, is technologically more complex than biotechnology involving viral and bacterial control agents. Use of genetic engineering in predator and parasitic insects is constrained by the lack of universal vectors or other techniques to transfer foreign genes into the insect, and the lack of useful insect genes that have been cloned. Recombinant DNA techniques are being used to turn slow acting viruses into quick acting viruses, and to increase virus virulence. Genetic engineering is being used to improve the delivery of Bacillus thuringiensis toxin to the pest. Methods include incorporating the toxin gene into bacteria that inhabit seed coatings, roots, or surface films where target insects feed. Genetic engineering in fungi is being used to improve germination, penetration of the insect cuticle, and increase toxicity. Little biotechnology research has been conducted using protozoans and nematodes (10, 16, 25, 38). In addition to enhancing the field efficacy of biological control agents, biotech-

PAGE 59

54 A New Technological Era for American Agriculture nology provides powerful research tools to further our basic understanding of the physiology and biology of these control agents and their environment. Institutions Involved in Biological Control Research Biological control research has been conducted primarily by public sector institutions, such as the U.S. Department of Agriculture (i.e., the Agricultural Research Service, the Office of International Cooperation and Development, the International Research Division, and the Forest Service), the Land Grant University System, and other public and private universities. Other Federal agencies that have supported biological control research include the U.S. Army Corps of Engineers (primarily for aquatic weeds), the Department of Interior (mainly the Park Service), the Department of Energy (through the national laboratory system), and the Tennessee Valley Authority. Selected State Natural Resources or Agricultural departments (notably those of California and Florida) also have supported biological control development. The U.S. Environmental Protection Agency is involved in registering biological control agents as pesticides. The U.S. Department of Agriculture Animal and Plant Health Inspection Service regulates the importation of natural enemies and the environmental release of biological control agents. The State Department also is involved in obtaining permission to search foreign countries for natural enemies of pests imported to the United States, and with negotiating release conditions of natural enemies with Canada and Mexico (10, 16, 25, 38). Private industry interest has been focused primarily on organisms that can be used in microbial pesticide applications, such as Bacillus thuringiensis to control insects, and a few selected fungi (i. e., CASST, COLLEGO, and DeVine) to control weeds. A limited level of privateindustry support exists for the use of predators and parasites to control arthropods. A few small, private firms mass rear parasites and predators for release, but conduct little or no research (10, 16, 25, 38). Use of Biological Control Agents To Control Pests in the United States Biological Control of Arthropods: Parasites and Predators Arthropod (e.g., insects, spiders, mites) damage is a major contributor to crop losses and decreased quality of agricultural products. A wide array of biological control agents can be used to control arthropods, bacteria, viruses, fungi, protozoa, and nematodes. In the United Table 2-6Use of Parasite or Predator Insects To Control Insect Pests in the United States Pest insect Host plant Classical method Rhodesgrass scale. . . . . . Citrus blackfly . . . . . . Walnut aphid. . . . . . . Cottony cushion scale. . . . . Olive scale. . . . . . . . Spotted alfalfa aphid . . . . . Alfalfa weevil. . . . . . . California red scale . . . . . California purple scale. . . . . California yellow scale. . . . . Browntail moth . . . . . . Satin moth , . . . . . . Oriental moth. . . . . . . Elm leaf beetle . . . . . . European pine sawfly . . . . . European spruce sawfly . . . . Larch casebearer . . . . . . Larch sawfly . . . . . . . Augmentation method Mexican bean beetle . . . . . Mealybugs. . . . . . . . California red scale . . . . . Spider mites . . . . . . . Two spotted spider mite . . . . Conservation method European red mite . . . . . Grasses Citrus Walnuts Citrus Olives Alfalfa Alfalfa Citrus Citrus Citrus Forests Forests Forests Forests Forests Forests Forests Forests Soybeans California citrus Citrus Almonds Strawberries Apples SOURCE: Office of Technology Assessment, 1992. States, these agents have been used to control several arthropod species (table 2-6). The classical method of control is the approach used most often, and the greatest success has occurred in more stable habitats such as forests and orchards, rather than row crops. Traditional selection methodologies have been used to identify parasites or predators with improved control capability or host range. For example, such techniques were used to identify strains of a parasitic mite resistant to selected pesticides, which were subsequently released into California almond orchards to control spider mites. Increased pesticide resistance allows this parasitic mite to be used in conjunction with Integrated Pest Management programs that use pesticides to control navel orangeworms above a threshold level. The ability to use this predatory mite in conjunction with other insect control programs increased the acceptance of this parasite for spider mite control (25). Use of genetic engineering in predator and parasitic arthropods is constrained by the lack of universal vectors or other techniques to transfer foreign genes into the arthropod. Current research is focusing on the use of transposons to transfer genes, but transposons may be

PAGE 60

Chapter 2Emerging Plant Technologies .55 Photo credit: U.S. Department of Agriculture, Agricultural Research Service. The parasitic wasp Microplitis croceipes lays her eggs in the tobacco budworm. By putting this natural predator to work, scientists hope to control members of the genus Heliothis, which cause major damage to cotton, corn, soybeans, and other crops. specific to certain species of insects, and thus cannot be used as a universal mechanism to transfer genes to all insect species. Another major constraint is the lack of useful arthropod genes that have been cloned (25). Further development of predator and parasitic arthropods to control pest arthropods is being constrained by several factors. Selection standards for classical control approaches are needed. The economic importance of the target pest is frequently the only factor considered when selecting possible subjects for biological control. Characteristics of the natural enemy itself, such as its suitability of mass rearing at reasonable cost, additional host requirements, impact on beneficial or endangered species, or dispersal characteristics may not be considered (25). Use of augmentation techniques to control pest arthropods with other parasitic and predator arthropods is limited by the lack of artificial diets and subsequent high cost of mass rearing, incomplete information on release methods, lack of rapid and effective monitoring methods, and lack of ability to stockpile or store natural enemies or maintain gene banks. Quality control standards for private firms that mass rear predatory or parasitic arthropods are lacking. Mixed colonies or even colonies of the wrong species have sometimes been provided; in some cases, firms have produced parasitic arthropods unable to fly. Arthropods can be sold without guidelines as to number to release, optimal timing of release, or how to monitor efficacy of release. Professional quality standards and appropriate management information are Table 2-7Pathogens Used To Control Insects in the United States Pest insect Host plant Viruses European pine sawfly . . . . Douglas fir tussock moth . . . Soybean looper. . . . . . Velvetbean caterpillar moth . . . Gypsy moth. . . . . . . Bacteria Japanese beetle. . . . . . Mosquito larvae . . . . . Greater wax moth. . . . . . Fungi Browntail moth . . . . . Plant bug . . . . . . . Aphids. . . . . . . . Spotted alfalfa aphid . . . . Mosquito larvae . . . . . San Jose scale. . . . . . Whiteflies . . . . . . . Protozoa Grasshoppers . . . . . . European corn borer . . . . Nematodes Butterflies, beetles . . . . . Face fly. . . . . . . . Mosquito larvae ., . . . . . Trees Trees Soybeans Soybeans Trees Turf grass NA Beehives Trees Apples Potatoes Alfalfa NA Trees Trees Rangeland Corn Cranberry, Citrus Cattle NA NA = Not applicable, SOURCE: Office of Technology Assessment, 1992. needed (25). Conservation methods to maintain predator or parasitic arthropods are constrained by gaps in the knowledge of the role of natural enemies in crop systems and how best to modify management practices to maintain natural populations. Biological Control of Arthropods: Pathogens In addition to parasitic and predatory arthropods, pathogens can be used to control pest arthropods. Pathogens that have been used to at least partially control arthropods (almost exclusively insects) in the United States include bacteria, particularly different strains in the Bacillus genus; viruses, particularly members of the baculovirus group; fungi; protozoans; and nematodes (table 2-7). Bacillus thuringienses (Bt), discussed earlier, is the pathogenic bacteria most frequently used to control insects. The tools of biotechnology can be used to improve the delivery of the Bt toxin to insect pests. The gene that codes for the toxin can be incorporated into bacteria other than Bacillus thuringiensis; these bacteria may inhabit seed coatings, roots, or surface films where target insects feed. Genes coding for Bt toxins have incorporated in strains of Pseudomonas, a soil bacteria that colonize corn roots, and into Clavibacter xyli, a plant-associated (endophytic) bacterium that grows in the vascular tissues of

PAGE 61

56 l A New Technological Era for American Agriculture Photo credit: U.S. Department of Agriculture, Agricultural Research Service. Entomologist compares an insect ravaged cotton leaf from a control variety with one that has been genetically engineered with a protective gene from Bacillus thuringiensus. plants. The Monsanto and Mycogen Corp. are incorporating Bt toxin genes into Pseudomonas, while Crop Genetic International is working with Clavibacter (1, 16). Genetic engineering techniques are also being used to modify Bt toxin genes to be toxic to a broader range of pests and to be more potent. Traditional selection and screening procedures applied to natural isolates are being used as well, to identify strains of Bacillus bacteria that are either more efficacious or that have different host specificity. These methods will potentially extend Bt use to include control of cotton bollworm, European corn borer, and corn rootworms. Genetically engineered and new, naturally selected strains of Bt are expected to be commercially available by 1995 (1, 16). Viruses are also being used to control insects. Many types of viruses infect insects, but only a few cause pathogenic epizootic diseases that are sufficiently fastacting and widespread to be considered useful for pest control. The first virus to be registered by EPA and produced commercially as a pesticide was a type of baculovirus that forms large polyhedral occlusions within the nucleus of infected cells. It was marketed in the mid 1970s by the Sandoz Corp. under the name Elcar, and was used to control cotton bollworm. Its market was displaced by the new pyrethroid pesticides. It has not been remarketed, although increasing resistance to pyrethroids may lead to renewed commercial interest. Three other baculoviruses have been used by the U.S. Forest Service to control the Douglas fir tussock moth, the gypsy moth, and the European pine sawfly ( 16). Baculoviruses are used to control lepidopterans (butterflies and moths) because they cause widespread lethal epizootic diseases, lead to morbidity within a week of infection, are compatible with other agrichemicals, can be applied by conventional spraying techniques, and are stable on the shelf for extended periods of time (years). Further, the baculoviruses replicate only in arthropods. Each is specific to a host or group of closely related hosts, and must enter and replicate within a specific type of host cell. This specificity is attractive from an environmental control perspective ( 1). Two other viruses of potential usefulness for biological control of insects are the Autograph californica virus and the codling moth granulosis virus. A. californica has a relatively wide range of hosts and could be used to control alfalfa looper, cabbage looper, fally armyworm, beet armyworm, and wax moth. The codling moth granulosis virus could be used to control insects that affect pome fruits and walnuts ( 16). Genetic engineering is being used to make viral pesticides faster acting. Neurotoxin genes that paralyze the pest insect and quickly halt insect feeding are being introduced into baculovirus. Alternatively, insecticidal hormones can be incorporated into the baculovirus to disturb insect development or behavior. The genes that code for an enzyme that regulates juvenile hormone levels in insects; a protein that regulates the release of a major molting hormone; and a protein hormone that elicits several behavioral characteristics during molting all recently have been isolated ( 1). The lack of suitable cloned neurotoxins and insect hormone genes is delaying further progress in improving viral control agents. Promotors that can be recognized by selected host cells of pest insects (i.e., cells of the midgut, for example) are being used to extend baculovirus ranges. The recent discovery that baculoviruses normally contain a gene regulating insect molting hormone activity is leading to the development of baculovirus strains in which this gene has been deleted. These gene-deleted strains have been shown to reduce insect feeding during infection, and to hasten the onset of insect morbidity ( 1). Baculoviruses genetically modified to delete the insect molting hormone regulatory gene are expected to be available before 1995. Baculoviruses engineered to carry

PAGE 62

Chapter 2Emerging Plant Technologies l 57 Table 2-8Control of Weeds by Insect and Microbial Agents in the United States Weed Habitat/crop affected Alligator weed Lantana Musk thistle Northern jointvetch Persimmon Prickly pear cactus Puncture vine Skeletonweed St. Johnswort Stranglervine Water hyacinth Aquatic Rangeland, forest, crops Rangeland Rice and soybeans Rangeland Rangeland Pasture, annual crops Rangeland Range and arable lands Citrus Aquatic SOURCE: Office of Technology Assessment, 1992. insecticidal genes such as insect hormones and neurotoxins could be available in the late 1990s. The only fungus registered and commercially produced for insect control in the United States was Hirsutella thompsonii. This fungus was used to control citrus rust mites, but was not commercially successful primarily because it did not survive storage or transportation. Further, environmental factors, including insufficient moisture, adversely affected its efficacy. Genetic engineering of fungi is now being used to improve germination, improve penetration of the insect cuticle, and increase toxicity ( 16). A major limitation to using protozoans is that they kill insects very slowly, if at all. Generally they affect arthropods by causing chronic disease with sublethal effects, reducing the ability of the arthropod to survive the winter. Nosema locustae, used to control grasshoppers on rangeland, is the only protozoan to be registered and commercially available in the United States ( 16). Research involving nematodes has been increasing. Steinernema carpocapsae has been used in the United States to control some lepidoptera species. It is not effective if applied to vegetation surfaces or other situations where it can dry out, but it can be effective in the soil or in burrows in plant tissues. Dedalenus siricidicola has been used to control woodwasps, even though its action is to sterilize its host rather than kill it. Very little genetic engineering is being used with nematodes. Biological Control of Weeds: Microorganisms and Arthropods Historically, biological control of weeds most commonly has been mediated by microorganisms (mainly bacteria and fungi, see table 2-8) and insects. Worldwide, 89 species of weeds have been controlled using 192 species of introduced organisms (the classical approach); an additional 25 weed species have been controlled using 33 species of native organisms (the bioherbicide approach) ( 10). Pathogenic microorganisms kill or severely debilitate their host plants by causing disease. Pathogenic and nonpathogenic microbes also produce metabolizes that are toxic to plants, and these phytotoxins can also be used as herbicides. For example, the fungus Gliocladium virens, when prepared and applied properly, can release enough of the toxin viridiol in the soil to control pigweed without harming cotton seedlings. The private sector has shown interest in developing microbial herbicides. Two microbial herbicides (COLLEGO and DeVine) are commercially available and four others are undergoing trials for registration as herbicides (table 2-9). Other microbial herbicide candidates are undergoing experimental development. About 107 fungi and 1 bacterium are being evaluated worldwide as bioherbicides ( 10). Additionally, a parasitic nematode, Orrina phyllobia, has been shown to be a practical means to control silverleaf nightshade. Development of a microbial herbicide can take several years. For example, it is estimated that the development of COLLEGO R took 11 years of effort from the time of discovery to commercial availability at a cost of about $1 to $1.5 million. In comparison, a typical chemical herbicide takes 7 to 10 years to develop and costs approximately $80 million. Early research on microbial herbicides is subsidized by public funds, but the expense of large-scale fermentation, toxicology testing, formulation, and registration are borne by industry. In some cases, these costs could prove to be quite high ( 10). Further development of microbial herbicides will require improved mass production, formulation, and delivery systems. Some native pathogens, such as the rusts and certain smut fungi, cannot be artificially grown. Methods to obtain sufficient quantities of these pathogens from infected plants must be developed. Weed pathogens are being genetically manipulated to improve virulence, sporulation, fitness for survival and infection under suboptimal conditions, and production of herbicidal metabolizes; to expand host-range; and to increase tolerance to certain chemical pesticides. For example, it has been discovered that altering a single enzyme (pisatin demethylase) can cause a fungal pathogen, but not a nonpathogenic fungi, to become virulent on new host plants. Genetic engineering techniques are also being used to increase virulence by transferring genes encoding herbicidal phytotoxins to pathogenic microorganisms (10). 297-937 0 92 3 QL 3

PAGE 63

.58 A New Technological Era for American Agriculture Table 2-9Microbial Herbicides Commercially Available or in Development in the United States Herbicide Pest Crop/habitat effected COLLEGO R Northern jointvetch Rice DeVine R Stranglervine Citrus CASST TMa Sicklepod Soybean and peanut BioMal TMa Round-leaf mallow Annual crops Cercospora rodmanii a Waterhyacinth Aquatic Mycoleptodiscus terrestris a Eurasian watermilfoil Aquatic a Undergoing trials SOURCE: Office of Technology Assessment, 1992. Table 2-10Use of Insects To Control Weeds in the United States Weed Crop/habitat affected Classical approach St. Johnswort . . . . Lantana. . . . . . Alligatorweed. . . . Prickly pear cactus. . . Puncturevine . . . . Tansy ragwort . . . Hydrilla . . . . . Purple loosestrife . . . Leafy spurge . . . . Diffuse, spotted and Russian knapweeds. Yellow starthistle. . . . Salt cedar. . . . . Field bindweed . . . Waterlettuce. . . . . Broom snakeweed . . Baccharis neglecta. . . Augmentation approach Waterlettuce. . . . . Purple nutsedqe . . . Range and arable lands Rangelands, forests, and plantation crops Aquatic Rangeland Pastures and annual crops Rangeland Aquatic Range and arable lands Rangeland Rangeland Rangeland Rangeland and forests Various crops Aquatic Rangeland Range and arable lands Tried and discontinued Tried and discontinued SOURCE: Office of Technology Assessment, 1992. Traditional techniques are also used to alter pathogen characteristics. These include chemicalor ultravioletgenerated mutations followed by selection for desired phenotypic traits, breeding, and nonsexual transfer of hereditary properties. Cultural techniques also are being improved to increase secretion of certain toxins and enzymes involved in pathogenesis. In addition to microbial pathogens, insects and other arthropods also can be used as biological control agents for weed control (table 2-10). The relationship between insects and weeds is complex. Some weeds (e. g., St. Johnswort) can be controlled with just one insect. Others may require more than one insect for control. For example, control of tansy ragwort, a poisonous weed found in the Pacific Northwest, is mediated by a moth that defoliates it and a second insect that feeds on its root as a larva and on the resprouting growth as an adult. This relationship between each co-evolved arthropod and its weed host makes each study unique and raises the question of whether scientific expertise will ever be adequate to fully assess the potential for weed control by arthropods (10). Arthropod adults and immature larva and nymphs feed and complete at least a part of their life cycles on certain weeds. In this process, they damage the plants, weakening and reducing their productivity and competitiveness. In general, the feeding activity of immature arthropods is more damaging than that of adult arthropods. The extent of the damage caused by arthropod feeding depends on the particular weed tissues destroyed, the timing of the damage as it relates to the plants growth cycle, and the extent of other plant stresses present. For example, sucking insects and grasshoppers defoliate plants late in the plants life cycle and do not cause as much damage as insects that defoliate plants early in their life cycles. Arthropods that attack the seeds of weeds that cannot reproduce vegetatively are likely to have the greatest impact on weed control. In addition to feeding damage, some arthropods weaken plants by introducing toxins causing cell proliferation and gall formation ( 10). Of the more than 250 naturalized plant species considered to be major weeds, only a few dozen have been considered for classical biocontrol by arthropods. Nonetheless, this approach has been the most common and successfully used method of biological weed control. It is estimated that the control of St. Johnswort by insects has yielded benefits worth approximately $2 million per year. It takes 1 to 4 years to find and clear each insect or other arthropod biocontrol candidate and development costs are estimated at $1 to 2 million. However, the estimated return on research is about $30 for every $1 invested ( 10). Few attempts to control weeds with ar-

PAGE 64

Chapter 2Emerging Plant Technologies l 59 Photo credit: U.S. Department of Agriculture, Agricltural Research Service. Tiny (1/8th inch long) flea beetle, Apthona flava, on leafy spurge is one of several biological control agents tested to combat a costly weed that infests 2 million acres of rangeland in the Great Plains. thropods using the augmentation approach have been tried, and generally they have been discontinued. Traditional selection methods are used to select coldtolerant strains of weed-damaging insects and strains whose larva have higher survival rates in hot weather, and whose prediapause behavior has been altered. Genetic engineering is not currently used to improve arthropods as biological control agents ( 10). Biological Control of Disease Biological control of plant diseases is achieved by decreasing pathogen populations or by preventing the occurrence of infections. Approaches taken include manipulating resident microbial communities to decrease disease (conservation approach) or applying to the plant organisms antagonistic to pathogens (augmentation). Only three plant disease biocontrol agents are commercially available (table 2-11 ) (38). Table 2-1 lBiological Control Agents Commercially Available To Control Plant Disease in the United States Agent Disease controlled Bacteria Agrobacteriurn radiobacter . . Crown gall in dicots (strain K84) Pseudomonas fluorescent. . . Damping off and root rot in cotton Fungi Peniophora gigantea. . . . Root and butt rot in conifers SOURCE: Office of Technology Assessment, 1992, Use of Agrobacterium radiobacter to control crown gall in dicots costs an estimated 1 to 5 cents per plant treated, and less if the seeds are treated. Peniophora giganted applied to freshly cut conifer stumps preempts colonization by the pathogen responsible for root and butt rot, diseases resulting in annual losses of nearly $1 billion. Pseudomonas j7uore.seen.s, sold under the name of Dagger G by Ecogen, controls diseases in cotton. In 1989, it was used on approximately 75,000 acres of cotton in the Mississippi Delta region (38). Diseases that potentially could be controlled in the next decade include take-all disease in wheat, and dampingoff and root rot in crops other than cotton. Yeasts to suppress Penicilliurm and other postharvest pathogens in citrus and other fruit; the bacterium Bacillus subtilis to control brown rot in peaches; and compost amended potting media to control Rhizoctonia and Pythium in nursery stocks are other potential control agents (38). The use of microbial disease control agents has been plagued by inconsistent efficacy in the field. In some cases, agents that have worked in one field have failed to be effective in immediately adjacent fields. The biocontrol agent and pathogen interact in the midst of a vast array of other microorganisms that sometimes decrease the efficacy of the control agent (23, 24). A better understanding of the community dynamics, population and community ecology, population genetics of plant-associated microorganisms and of the mechanisms that regulate the community structure and dynamics of plantassociated microorganisms is needed. Much of the research in the area of biocontrol of plant diseases has focused on improving the understanding of the mechanisms by which biocontrol agents prevent disease. One mechanism of action called interference competition or antibiosis refers to the inhibition of one organism by a metabolic product of another. The use of the bac-

PAGE 65

60 l A New Technological Era for American Agriculture terium Agrobacterium radiobacter strain K84 to control crown gall tumors caused by Agrobacteriurm tumefaciens is an example of this type of mechanism. A. radiobacter produces an antibiotic to A. tumefaciens. Control of takeall disease in wheat by Pseudomonas fluorescens strain 2-79 is another example of antibiosis (38). Peniophora gigantea controls root rot in pine caused by the fungus Heterobasidion annosum, on the other hand, by competing with the fungus for nutrients and space, a process referred to as exploitation competition. A third mechanism, hyperparasitism, occurs when fungal pathogens destructively parasitize another organism. Fungi of the Trichoderma and Gliocladium family, for example, parasitize soil-born plant pathogens such as Rhizoctonia solani and Pythium species. A fourth mechanism of disease prevention by biological agents is hypovirulence. For example, some strains of the chestnut blight fungal pathogen Cryphonectria parasitic (those with reduced virulence) can impart protection to chestnut trees from more virulent strains of this pathogen. Traditional screening techniques are being used to develop fungicide-resistant strains of the fungus Trichoderma, which allows this disease control agent to be used with fungicides so that fewer chemicals need be applied. Strains of the bacteria Pseudomonas syringae pv. tomato, which controls bacterial speck in tomatoes, have been made resistant to copper. The copper resistance allows P. syringae to be used in the presence of copper bactericide. Combinations of P. syringae and copper bactericide gives greater control over bacterial disease than occurs with the biocontrol agent or bactericide treatment alone (31, 32). Pathogenic organisms can become resistant to biological control agents. For example, A. radiobacter controls the plant pathogen A. tumefaciens by producing a compound called agrocin. The gene producing agrocin is carried on a plasmid, which can be naturally transferred to A. tumefaciens. Thus, A. fumefaciens is becoming resistant to A. radiobacter. Genetic engineering is being used to construct mutant strains of A. radiobacter that no longer have the ability to transfer the agrocin plasmid, thus decreasing the potential of A. tumefaciens to develop resistance to this natural pesticide. Protoplasm fusion techniques are also being used to construct strains of Trichoderma harzianum that are more effective than parental strains in controlling Pythium ultimum (38). Biological Control of Frost Damage The temperature at which frost injury occurs in a number of crops is determined by the population density of ice-nucleation-active bacteria on plant leaves. By decreasing the numbers of these bacteria, some protection against frost damage can be achieved. The application of non-ice-nucleating bacteria prior to colonization of ice-nucleating bacteria can effectively prevent the establishment of the ice-nucleating bacteria by limiting the resources (i.e., space and/or nutrients) available to the ice-nucleating bacteria. Ice-minus deletion mutants of the bacteria Pseudomonas syringae have been constructed to control frost. The first planned introductions of genetically engineered bacteria into the environment in the United States involved the field-testing of these ice-minus bacteria. SUMMARY Pest control is a major concern of crop producers in the United States. Each year, pest damage results in billions of dollars of lost revenue to farmers. Poor weather conditions add to those losses. To control pest damage, farmers have traditionally used chemical approaches. Biotechnology is now providing opportunities to use biological approaches such as transgenic plants resistant to pests and better adapted to geoclimatic conditions, and the use of biological control agents. The ability to create transgenic plants with useful agronomic characteristics is constrained by the lack of knowledge concerning plant physiology. Our understanding of plant metabolism has not kept up with the development of biotechnology methods. However, plants resistant to certain insects are approaching commercialization. Most of these plants have a Bacillus thuringiensis toxin gene insert, but some research also is being conducted using insect trypsin inhibitors that disrupt the digestion of feeding insects. Several transgenic Bt plants are undergoing field trials, and it is expected that several companies will begin petitioning EPA for approval for commercial release soon. Plants tolerant of herbicides are being developed to aid the management of weeds. Development of broadspectrum herbicides has been constrained because they not only kill most weeds, but also cause significant damage to crops. Crops tolerant to specific herbicides allow the use of these herbicides in conditions where they previously could not be used, and may allow for the replacement of some environmentally damaging herbicides. Some of these crops are nearing commercialization stages. Transgenic plants are being developed that are resistant to disease. Scientific understanding of the complex interactions between fungi or bacteria and host plants is limited, so much of the early successes have been in

PAGE 66

Chapter 2Emerging Plant Technologies l 61 developing plants resistant to certain viral diseases. Several virus-resistant plants are under development. Development of transgenic plants tolerant to geoclimatic conditions is in the early stages. Research is being conducted to understand the mechanisms of heat and drought tolerance, and to enhance the ability of plants to withstand cold temperatures. However, the successful commercialization of these plants is unlikely to occur before the end of the decade. In addition to engineering crops themselves, there is increased interest in developing biological control agents to manage pests. The use of biological control in the United States, to date, is relatively limited and most successes have involved controlling pests in forests, orchards, grasslands and aquatic environments. Use of biological control in grain and row crops is very limited. However, there is more emphasis placed on developing such products to control weeds, insects, and disease in the major food crops, and improved strains of Bacillus thuringiensis to control insects and a few fungal strains to control weeds are approaching commercialization. More research still is needed to successfully develop other products. The food processing industry will also be affected by biotechnology. Plants are modified for new quality and processing characteristics. For instance, tomatoes with delayed softening characteristics are nearing commercialization. Research is also underway to alter the starch, oil, and protein content of selected crops to more closely reflect consumer preferences and to enhance their processing characteristics for specific end uses. Diagnostic kits are in various stages of development to detect the presence of microorganisms, chemicals, and other contaminants in food products. The development of transgenic plants and biological control organisms offer new approaches to controlling pests and to improving food processing characteristics. However, many issues have been raised concering the development of these products. Some groups are worried about the effects these products will have on small farms, and on food safety and the environment. Additionally, many of these products will require extensive farm management capabilities for effective use. These issues will be discussed in subsequent chapters. 1. CHAPTER 2 REFERENCES Adang, M.J. and Miller, L. K., Genetic Technology for Resistance to Insect Pests, commissioned background paper prepared for the Office of Technology Assessment, 1991. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. Amheim, N. and Levenson, C. H., Polymerase Chain Reaction, Chemical and Engineering News, Oct. 1, 1990, pp. 36. Beachy, R.N. et al., Expression of Sequences of Tobacco Mosaic Virus in Plants and Their Role in Disease Resistance, Tailoring Genes for Crop Improvement, G. Bruening et al. (eds. ) (New York, NY: Plenum Publishing Company, 1987), pp. 169 180. Beachy, R. N., Loesch-Fries, S. and Turner, N., Coat Protein-Mediated Resistance Against Virus Infection, Ann. Rev. Phytopath. 28:451-474, 1990. Bialy, H. and Klausner, A., A New Route to Virus Resistance in Plants, Bio/Technology 4, 96, 1986. Boyer, J. S., Plant Productivity and Environmerit, Science 218:443-448, 1982. Boyer, J. S., Water and Plant Productivity, Water and Water Policy in World Food Supplies, W.R. Jordan (cd. ) (College Station, TX: Texas A&M University Press, 1987), pp. 233. Browning, J. A., Plant Disease Management, commissioned background paper prepared for the Office of Technology Assessment, 1991. Burke, J.J. and Upchurch, D. R., Agricultural Technologies Associated With Plant Temperature and Water Stress, commissioned background paper prepared for the Office of Technology Assessment, 1991. Charudattan, R. and Andres, L. A., Biocontrol for Weeds, commissioned background paper prepared for the Office of Technology Assessment, 1991. Cook, R.J. and Baker, K. F., Nature and Practice of Biological Control of Plant Pathogens, American Phytopathological Society, St. Paul, MN, 1983. Cutler, A.J. et al., Winter Flounder Antifreeze Protein Improves the Cold Hardiness of Plant Tissues, Journal of Plant Physiology>, vol. 135, 1989, pp. 351-352. DeBach, P. (cd.), Biological Control of Insect Pests and Weeds (New York, NY: Reinhold Publishing Co.), 1964. Discover, June 1991, p. 18. Fraley, R. T., Genetic Engineering in Crop Agriculture, commissioned background paper prepared for the Office of Technology Assessment, 1991. Fuxa, J. R., Pathogens for Insect Control, commissioned background paper prepared for the Office of Technology Assessment, 1991. Georges, F., Saleem, M., and Cutler, A. J., Design and Cloning of a Synthetic Gene of the Flounder Antifreeze Protein and its Expression in Plant Cells, Gene, vol. 91, 1990, pp. 159-165. Harlander, S., Biotechnology in Food Processing in the 1990s, commissioned background paper prepared for the Office of Technology Assessment, 1991.

PAGE 67

62 l A New Technological Era for American Agriculture 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. Hatzios, K. K., Biotechnology Applications in Weed Management: Now and in the Future, Advances in Agronomy 36:265, 1987. Hatzios, K. K., Recent Developments in the Physiology and Biochemistry of Herbicide Safeners, Brighton Crop Prot. Conf. Weeds 3:1207-1216, 1989. Helentjaris, R. and Burr, B. (eds. ), Development and Application of Molecular Markers to Problems in Plant Genetics (New York, NY: Cold Spring Harbor Laboratory, 1989). Hightower, R. et al., Expression of Antifreeze Proteins in Transgenic Plants, Plant Molecular Biology, vol. 17, 1991, pp. 1013 Hirano, S.S. and Upper, C. D., Temporal, Spatial, and Genetic Variability of Leaf-Associated Bacterial Populations, Microbiology of the Phyllosphere, N.J. Fokkema and J. van den Heuvel (eds.) (Cambridge, MA: Cambridge University Press, 1986), pp. 235251. Hirano, S.S. and Upper, C. D., Population Biology of Pseudomonas syringae, Ann. Rev. Phytopathol. 28: 1990. Hey, M. A., Use of Parasites and Predators To Control Insect and Mite Pests in Agriculture, commissioned background paper prepared for the Office of Technology Assessment, 1991. Kishore, G. M., Plant Genetic Modification for Weed Control commissioned background paper prepared for the Office of Technology Assessment, 1991. Loesch-Fries, S., Genetic Modification for Disease Resistance, commissioned background paper prepared for the Office of Technology Assessment, 1991. Main, C.E. and Gurtz, S.K. (eds.), 1988 Estimates of Crop Losses in North Carolina Due to Plant Diseases and Nematodes, Department of Plant Pathology Special Publication No. 8, North Carolina State University, Raleigh, NC, 1989, 209 pp. National Academy of Sciences, Field Testing Genetically Modified Organisms: Framework for Decisions (Washington, DC: National Academy Press, 1989), 170 pp. National Research Council, National Academy of Sciences, Board on Agriculture, Genetic Engineer31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. ing of Plants: Agricultural Opportunities and Policy Concerns (Washington, DC: National Academy Press, 1984). Papavizas, G. C., Trichoderma and Gliocladium: Biology, Ecology, and Potential for Biocontrol., Ann. Rev. Phytopathol. 23:23-54, 1985. Papavizas, G. C., Genetic Manipulation to Improve the Effectiveness of Biocontrol Fungi for Plant Disease Control, Innovative Approaches to Plant Disease Control, I. Chet (ed.) (New York, NY: Wiley, 1987), pp. 193-212. Putnam, A.R. and Duke, W. B., Biological Suppression of Weeds: Evidence for Alleopathy in Accessions of Cucumber, Science 185:370, 1974. Quisenberry, J. E., Cartwright, G.B. and McMichael, B. L., Genetic Relationship Between Turgor Maintenance and Growth in Cotton Germplasm, Crop Sci. 24:479, 1984. Rabb, R. L., Stinner, R. E., and van den Bosch, R, Conservation and Augmentation of Natural Enemies, C.B. Huffaker & P.S. Messenger (eds.) Theory and Practice of Biological Control (New York, NY: Academic Press, 1976). Sailer, R. I., History of Insect Introductions, C. L. Wilson and C.L. Graham (eds.), Exotic Plant Pests and North American Agriculture (New York, NY: Academic Press, 1983), pp. 15-57. Sheehy, R. E., Kramer, M., and Hiatt, W. R., Reduction of Polygalacturonase in Tomato Fruit by Antisense RNA, Proc. Natl. Acad. Sci., 1988. Upper, C.D. and Hirano, S. S., Microbial Biocontrol of Plant Diseases, commissioned background paper prepared for the Office of Technology Assessment, 1991. U.S. Department of Agriculture, Losses in Agriculture, USDA Handbook No. 291, 1965, 120 pp. Vandekerckhove, J. et al., Enkephalins Produced in Transgenic Plants Using Modified 2s Seed Storage Proteins, Biotechnology 7:929-936, 1989. Wessells, Norman K. and Hopson, Janet L., Biology (New York, NY: Random House, Inc., 1988).

PAGE 68

Chapter 3 Emerging Animal Technologies Photo credit: Grant Heilman, Inc.

PAGE 69

Contents Page COMPOUNDS THAT PROMOTE GROWTH, ENHANCE FEED EFFICIENCY, AND REDUCE .CARCASS FAT . . . . . . . . . . . . . . . Somatotropin . . . . . . . . . . . . . . . . . . . . Beta-Agonists . . . . . . . . . . . . . . . . . . . Antimicrobial Agents . . . . . . . . . . . . . . . . . Anabolic Steroids . . . . . . . . . . . . . . . . . . REPRODUCTION TECHNOLOGIES . . . . . . . . . . . . . . Estrous Cycle Regulation . . . . . . . . . . . . . . . . Embryo Cloning . . . . . . . . . . . . . . . . . . . Embryo and Sperm Sexing . . . . . . . . . . . . . . . . TRANSGENIC ANIMALS . . . . . . . . . . . . . . . . . Process of Creating Transgenic Animals . . . . . . . . . . . . . Transgenic Poultry . . . . . . . . . . . . . . . . . . Transgenic Swine . . . . . . . . . . . . . . . . . . Transgenic Ruminants . . . . . . . . . . . . . . . . . Transgenic Fish . . . . . . . . . . . . . . . . . . . Research Needs . . . . . . . . . . . . . . . . . . . ANIMAL HEALTH TECHNOLOGIES . . . . . . . . . . . . . . Vaccines . . . . . . . . . . . . . . . . . . . . . Immunomodulators . . . . . . . . . . . . . . . . . . Diagnostics . . . . . . . . . . . . . . . . . . . . FOOD PROCESSING APPLICATIONS . . . . . . . . . . . . . SUMMARY . . . . . . . . . . . . . . . . . . . . CHAPTER PREFERENCES . . . . . . . . . . . . . . . . Figures 65 66 71 72 74 75 75 77 77 79 79 85 85 86 87 87 88 88 90 90 92 93 94 Figure Page 3-1. Bovine Somatotropin (bST) Production Process . . . . . . . . . . 69 3-2. Nuclear Transplantation . . . . . . . . . . . . . . . . 78 3-3. Reproductive Technologies Used To Produce Transgenic Animals . . . . . . 80 3-4. Construction of a cDNA Library . . . . . . . . . . . . . . 81 3-5. Gene Transfer Using Embryo Stem Cell Culture . . . . . . . . . . 84 3-6. Basic Steps in a DNA-Probe Hybridization Assay . . . . . . . . . . 93 Tables Table Page 3-1. Effect of bST on Specific Tissues and Physiological Processes in Lactating Cows . . 67 3-2. Antimicrobial Agents Approved as Growth Promotants for Swine, Poultry, and Cattle in the United States . . . . . . . . . . . . . . . . . . 72 3-3. Anabolic Steroids Commercially Available in the United States . . . . . . 74 3-4. Diagnostic Monoclinal Antibody Kits . . . . . . . . . . . . . 91 3-5. Pathogens for Which Diagnostic Kits Using Nucleic Acid Probes Are Available . . 91

PAGE 70

Chapter 3 Emerging Animal Technologies The U.S. livestock industry is immense, and the costs of running it are correspondingly large. Feed and health care costs for the Nations nearly 100 million head of cattle (beef and dairy), 55 million pigs, 10 million sheep, and 600 million chickens and turkeys amount to billions of dollars annually. Disease and reproductive losses also significantly erode industry profits. Like any industry, livestock producers strive to reduce costs and losses, and to maximize profits. Feed constitutes almost 70 percent of the cost of producing pigs for pork. Improvements in feed efficiency (i.e., a lower quantity of feed consumed per unit of weight gained) and faster weight gain could potentially lower production costs in this and other sectors of the livestock industry. Animal diseases cost the livestock industry billions of dollars each year. For example, anaplasmosis in cows costs an estimated $300 million a year in losses and disease control. The bacterium Staphylococcus aureus, which causes 55 percent of mastitis, costs U.S. dairy producers some $250 million annually. New vaccines and diagnostic kits can help decrease disease in livestock. Other economic losses in the livestock industry result from low conception rates and embryo mortality. Such losses can be minimized by a greater understanding of reproduction as well as by emerging technologies for improving reproductive success. Biotechnology has the potential to improve feed efficiency, reduce losses from disease, and increase reproductive success in all sectors of the livestock industry, in part by furthering our understanding of animal physiology, and in part through the development and commercialization of new techniques and products. The term biotechnology refers to a wide array of techniques that use living organisms (or parts of organisms) to make or modify products, to improve plants or animals, or to develop microorganisms for specific uses (45). Under this broad definition, biotechnology includes many long-practiced technologies, such as animal breeding and cheese, wine, and beer making. Generally, however, the term biotechnology is used in reference to such new technologies as recombinant DNA techniques (also called genetic engineering), cell culture, and monoclinal antibody (hybridoma) methods. The application of these new methods to the livestock industry has already generated a number of products for improving production, animal health, and food processing, and will continue to do SO. Biotechnology is specifically used to produce products that will promote growth and increase feed efficiency and carcass leanness in growing animals, and significantly increase milk production in lactating animals. New reproductive technologies are providing means to rapidly upgrade herd quality. Transgenic animals are being produced to grow faster, have greater disease resistance, and to produce high-value pharmaceutical products. New vaccines and diagnostic kits are being developed to improve livestock health. Biotechnology is also being used to process meat and dairy products and to detect food contaminants that might be present in those products. This chapter presents some new livestock biotechnologies currently under development. 1 COMPOUNDS THAT PROMOTE GROWTH, ENHANCE FEED EFFICIENCY, AND REDUCE CARCASS FAT Compounds currently used in the livestock sector to promote growth and increase feed efficiency, such as anabolic steroids and antimicrobial compounds, will continue to be used. However, new products are also being developed, including protein hormones called somatotropins and catecholamine compounds called beta-adrenergic agents. These compounds increase growth rates in young animals, improve the efficiency with which food is converted to muscle, and significantly reduce carcass fat so that meat products are leaner. Somatotropins also increase milk production in lactating dairy cows. Currently, recombinantly-derived bovine and porcine somatotropins are undergoing Food and Drug Administration (FDA) review for use in lactating dairy cows and pigs, respectively, and one beta-adrenergic agent is undergoing testing for approval in pigs. 1 Because of the large quantity of research on these technologies, this chapter will mainly cite OTA commissioned background papers and other review articles. -65

PAGE 71

66 l A New Technological Era for American Agriculture Somatotropin A hormone is a chemical that is produced by one organ or cell and transported to another to cause a biological effect (i. e., it is a chemical messenger between different cells and organs of the body). Hormones can be steroids, proteins, peptides, or modified amino or fatty acids. About 70 percent of the hormones in blood are protein hormones. Somatotropin is a protein hormone produced by the pituitary gland, a small gland located at the base of the brain. All vertebrates (i.e., animals that have backbones) produce somatotropin. In addition, evidence exists that some nonvertebrate animals, such as shellfish (i.e., oysters, clams, etc.), also produce somatotropin. All major livestock species produce somatotropins unique to each species. Naturally produced bovine somatotropin (bST) contains 190 or 191 amino acids, and each polypeptide can contain either the amino acid valine or leucine at position 126, which gives rise to 4 variants of bST. Pigs produce porcine somatotropin (pST) consisting of 191 amino acids. The amino acid sequence of pST, however, differs from bST at 18 positions. In contrast, bST and ovine (sheep) somatotropin (oST) differ by only one amino acid position (3, 16, 40). Differences in the amino acid sequence of proteins lead to species specificity. The amino acid sequence determines the unique three-dimensional shape characteristic of a specific protein. Only proteins of the appropriate shape bind to a receptor, and thus elicit a biological response. Proteins from one species that differ by many amino acids from the equivalent protein in another species generally do not elicit a biological response in the other species. Conversely, bST and oST that differ only by one amino acid are active in either sheep or cattle. However, human somatotropin differs from pST by 59 amino acids and from bST by 68 amino acids (a 35 percent difference). Bovine, porcine, and ovine somatotropin are not biologically active in humans (20, 23, 49). Mechanism of Action Somatotropins affect growth rate, feed efficiency, milk yield, and the proportion of fat and protein in the carcass. These effects occur in response to the coordination of numerous metabolic pathways by somatotropin. These metabolic effects are both direct and indirect. The direct effects include nutrient partitioning among tissues, most specifically liver and adipose (fat) tissue (table 3-1 ). indirect effects include those mediated by insulin-like growth factor-1 (IGF-I), whose secretion is stimulated by somatotropin. Somatotropin affects glucose metabolism. Glucose is a carbohydrate used as an energy source by many tissues, or as a raw material for the synthesis of other molecules (as in the production of milk lactose). Administration of somatotropin increases blood glucose levels by stimulating glucose production by the liver, and may possibly reduce glucose use for energy by other body tissues. 2 Thus, additional glucose is available for uses such as increased growth or milk production while normal body functions are still maintained. The changes in glucose use by body tissues and glucose production by the liver appear to be caused by somatotropin altering the response of these tissues to acute signals, such as to insulin and other hormones that affect glucose metabolism (3. 16). Somatotropin also adjusts lipid (fat) metabolism. In growing pigs, for example, somatotropin redirects nutrients (primarily glucose) away from fat synthesis to providing energy for lean tissue accretion. The adjustments in tissue lipid metabolism depends on the nutritional status of the animal. If an animals energy (food) intake is greater than its requirements, somatotropin allows for the reallocation of nutrients to support increased lean tissue accretion (growth) or milk production (lactation) instead of storing excess nutrients as body fat. If the animals nutrient intake is equal to or less than its requirements, somatotropin directs adipose tissue to mobilize deposits of body fat so that these energy reserves can be used to support the increased lean tissue accretion (growth) or milk production (lactation). The former situation is more likely to be the case for young growing animals and the latter situation would be typical of lactating cows in early lactation. Like glucose metabolism, adjustments in lipid metabolism result from changes in the way adipose tissue responds to acute signals, such as to insulin and other hormones (3, 16, 40). In addition to the direct metabolic effects that somatotropin coordinates, it stimulates the release of other compounds with metabolic effects, most notably insulinlike growth factor I (IGF-I). IGF-I probably mediates the effects of somatotropin on animals such that the cellular rate of milk synthesis is increased and the rate at which mammary cells die is decreased, thus causing higher daily milk yields for a longer period of time during the 2 Evidence in Iactatlng daiv cows suggests that glucose use by tissues other than the mammary gland is decreased when somatatroPin is administered. It is still not clear whether glucose use by skeletal muscle is decreased in growing pigs (3, 16).

PAGE 72

Chapter 3Emerging Animal Technologies .67 Table 3-lEffect of bST on Specific Tissues and Physiological Processes in Lactating Cows a Process affected during first few days and Tissue weeks of supplement Liver o o Muscle Pancreas 0 ( ( I secretory activity and maintenance of mammary glands blood flow and nutrient uptake synthesis of milk with normal composition production of glucose response to acute signals (e.g., insulin) that allow for greater glucose production mobilization of fat stores to meet needs for increased milk production if nutrient intake is inadequate use of nutrients for fat storage so that they can be used for increased milk production if nutrient intake is adequate response to acute signals (e.g., insulin and other hormones that affect lipid metabolism) that allows for synthesis and breakdown of body fat reserves to be coordinated with changes in use and availability of nutrients uptake of glucose insulin and glucagon secretion reponse to changing glucose levels production of 1,25 vitamin D 3 absorption of Ca, P and other minerals required for milk ability of 1,25 vitamin D 3 to stimulate calcium binding protein calcium binding protein use of glucose by some organs so more can be used for milk synthesis use of fat stores for energy if nutrient supply is Inadequate use of nutrients to make body fat if nutrient supply is adequate insulin and glucagon clearance rates energy expenditure for maintenance energy expenditure consistent with Increase In milk yield (i.e., heat per unit of milk not changed) cardiac output consistent with increases in milk yield productive effidency (milk per unit of energy intake) occur in initial period of bST supplement when metabolic adjustments occur to match the increased use of nutrients for milk. With longer term treatment voluntary intake Inceases to match nutrient requirements. demonstrated in nonlactating animals and consistent with observed performance in lactating cows. SOURCE: D.E. Bauman, Bovine Somatotropin: Review of an Emerging Animal Technology, commissioned background paper for the Office of Technology Assessment, Washington, DC, 1991. lactation cycle (3). In growing animals, IGF-I stimulates cell proliferation in a variety of tissues (bone, muscle, connective, and adipose tissue) and increases protein synthesis in muscle ( 16, 40). Poultry Somatotropin Research using somatotropin to enhance growth and carcass composition in poultry (i. e., chickens, turkeys, and ducks raised for meat and egg production) is limited. Earliest research involved chickens that had their pituitary glands removed. Administration of chicken somatotropin (cST) was shown partially to restore growth. Chicken somatotropin also has been shown to increase circulating levels of IGF-I (40). Administration of cST to broiler chickens 3 ( i .e., chickens marketed at 6 to 7 weeks) has not been shown to influence growth, feed efficiency, or carcass composition. In young (post-hatched) chicks, the binding of somatotropin to its receptors in the liver is very low, whereas in adult chickens high somatotropin binding has been observed. There appear to be low somatotropin receptor numbers and/or receptor affinity for somatotropin during the early stages of chicken growth, potentially up to the time when broiler chickens are marketed. This might provide an explanation as to why cST has little or no effect in young broiler chickens. The basis for this low binding is not known, but some evidence exists that somatotropin itself regulates the number of somatotropin receptors (40). While most studies have reported no enhanced growth in young chickens given cST, one study using daily injections of intermediate doses of native cST did elicit improved growth in 4-week-old broiler chickens. This raises the possibility that diet, frequency of cST administration, molecular form of cST, or dose may be necessary conditions to achieve a growth response in broiler chickens. Thus, it cannot be ruled out that optimal conditions have not been employed in most studies. Based on the evidence to date, however, cST administration appears not to be an effective means of promoting growth or productive efficiency in growing broiler chickens (40). Administration of cST to roaster chickens (i.e., chickens more than 8 weeks old) has been shown to stimulate growth and feed efficiency while reducing carcass fat. The effects of cST on breast meat weight varied depending on the method of cST administration. For example, the weight of the breast meat was reduced when cST was administered in a pulsatile (rhythmic dripping) fashion, but increased when administration was by continuous infusion or daily injection. The extent of growth and of fat tissue accumulation also varied with method of administration and age of the chicken. These results suggest that cST can be used to improve roaster-age Chicken somatotmpin derived from chicken pituitary glands and from recombinant DNA procedures were tried

PAGE 73

. 68 l A New Technological Era for American Agriculture chickens, but that the mode of administration and dose, and potentially diet, need to be optimized to achieve consistent results (40). Turkeys that have had their pituitary glands removed have been treated with bST and cST; neither influenced growth. Administration of chicken or turkey somatotropin to intact turkeys has not been adequately explored. Some evidence exists that bST or pST injections into the egg increase the growth and feed efficiency of male chickens after hatching, and reduce abdominal fat in both male and female chickens. In summary, it has not been definitively demonstrated that somatotropin can be used to improve growth, feed efficiency, or carcass composition of poultry. More research is needed to determine if this is in fact the case, to optimize conditions needed to achieve growth, and to improve the mode of administration. There is a general lack of research on poultry biology and much basic research is needed to understand growth mechanisms in poultry. There is also a need to characterize fully the structure and control of the receptor(s) for chicken somatotropin, to identify the specific amino acid sequence of somatotropin that binds to the receptors, to understand the signal system used for somatotropin to elicit its biological response, and to identify hormones that may counteract the effects of somatotropin in poultry. Given the state of the art, it is unlikely that cST will be available for poultry production before the later part of the 1990s (40). Porcine Somatotropin Pigs administered porcine somatotropin (pST) for a period of 30 to 77 days have been shown to increase average daily weight gains by approximately 10 to 20 percent; improve feed efficiency by 15 to 35 percent; decrease adipose (fat) tissue mass and lipid formation rates by as much as 50 to 80 percent; and concurrently increase protein deposition by as much as 50 percent, without adversely affecting qualities such as taste and texture of meat. Prolonged release formulations and daily injections produced similar growth rates and feed efficiencies. In addition, similar growth rate increases were observed in both barrows (castrated male pigs) and growing gilts (immature female pigs) ( 16). Daily administration of pST to gilts weighing between 110 and 220 pounds did not affect the age at which puberty occurred, the proportion of gilts reaching puberty prior to 240 days, or the pregnancy rate. One study did indicate that with pST administration, ovarian function was impaired in prepubertal gilts, and that the onset of puberty was delayed. Withdrawal of pST restored normal reproductive function ( 16). The minimally effective dose of pST needed to increase growth performance is approximately 20 micrograms of pST per kilogram of body weight per day. In the commercial setting, pigs will likely be treated with pST for about 60 days during the growing-finishing period ( 16). Photo credit: Terry Etherton, Pennsylvania State University. Comparison of pork loins that show the effect of pigs treated with porcine somatotropin (PST). The loin-eye area of the loin treated with PST is 8 square inches; the control is 4.5 square inches.

PAGE 74

Chapter 3Emerging Animal Technologies l 69 For effective use of pST, prolonged release formulations lasting at least 30 days need to be developed. Optimal nutrient requirements need to be determined. Initial data indicate that the diet should contain about 1.2 percent lysine (6). Current corn-soybean meal formulations containing about 16 percent crude protein may need to be supplemented with additional lysine, and perhaps other amino acids. Total feed intake will likely increase by 10 to 15 percent with pST administration. The nutritional requirements of pST-treated pigs is currently being studied by the National Research Council. One study found that porcine somatotropin increased milk production between days 12 and 29 of lactation and the nursing piglets have a greater weight gain which matched the increased milk yield ( 16). However, this increase in milk yield and piglet weight gain has not been consistently observed (2, 8, 9, 10, 11, 42, 43, 44). Also, in some cases, adverse health effects were noted in pST treated sows ( 10, 42). Porcine somatotropin is currently being reviewed by FDA for commercial use. (For additional information on pST and its effects on carcass grades, see ch. 14. ) Bovine Somatotropin Bovine somatotropin (bST) is currently undergoing FDA review for use in lactating dairy cows to increase milk production (figure 3-1 ). While individual milk yields depend on the management ability of the producer, on average, gains of about 12 percent can be expected with bST administration. However, response varies with the stage of lactation. Administration of bST early in lactation (i.e., immediately following parturition and prior Figure 3-lBovine Somatotropin (bST) Production Process SOURCE: Elanco, a division of Eli Lily to peak milk yield) evokes a small or negligible response (3). Administration after peak milk yields evokes a high response due to an immediate increase in milk yield, and a reduction in the normal decline in yields that occurs as lactation progresses. Maximum milk response is achieved with a daily bST dose of about 30 to 40 mg/day. BST does not alter the gross composition of the milk. The fat, glucose, protein, mineral, and vitamin composition of the milk all fall within the range of values normally observed in milk from cows not given bST (3). The relative ratio of nutrient requirements of cows administered bST do not change, but the cow will eat more feed to accommodate the increased milk production. The magnitude of the increase in feed intake depends on how much milk production increases and on the energy density of the diet. BST decreases pregnancy rates (proportion of cows becoming pregnant) and increases days open (days from parturition to conception). Conception rates (services per conception) are not altered. The effects observed are similar to those occurring in high milk producing cows that do not receive bST (3). The implications of using, bST in dairy production are discussed more thoroughly in OTAs 1991 publication U.S. Dairy industry at a Crossroad: Biotechnology and Public Choices (47). A small number of studies using somatotropin to increase growth in growing cattle has been conducted, but research in this area is increasing. Results to date are highly variable due to the fact that studies differ significantly with respect to source and type of somatotropin used; dose and potency of somatotropin; route and frequency of administration; number, sex, type, and age of animals; duration of treatment; level and type of nutrition; and methodology used to determine characteristics measured. Thus, comparisons are tenuous, but on average, administration of somatotropin to growing cattle increases average daily weight gain by 12 percent, improves feed conversion efficiency by 9 percent, increases carcass lean content by 5 percent, and decreases carcass fat content by 15 percent ( 15). Additional long-term studies are needed. Optimal dose, nutritional needs, duration, and withdrawal period before slaughter need to be determined. Ovine Somatotropin A small number of studies has examined ovine somatotropin (oST) or bST for use as a growth promotant in sheep. Because oST and bST are similar on amino acid sequence they both are effective. Like the studies with growing cattle, investigations with sheep vary sig-

PAGE 75

70. A New Technological Era for American Agriculture nificantly in design and methodology. These studies suggest that on average, administration of somatotropin to sheep will increase the average daily weight gain by 18 percent, improve the feed conversion efficiency by 14 percent, increase the carcass lean content by 10 percent, and decrease carcass fat content by 15 percent. Ruminants present some special challenges with regard to supply of amino acids to support high rates of protein accretion. Recent studies with growing cattle and lambs demonstrate that nutritional constraints imposed by rumen fermentation may limit amino acid supply and ultimately the biological response to somatotropin (4, 2 1). Long-term studies are needed, and optimal conditions of somatotropin administration and nutrient requirements must be determined ( 15). Fish Somatotropin Recombinant trout somatotropin injected into yearling rainbow trout increased growth rates by 100 percent as compared to control fish. Body length increased, and the chemical composition of the muscle tissues was indistinguishable from that of the controls (34). However, injection into individual fish is inefficient and different modes of administration are needed. Other studies have tried dipping and incubating test fish in an appropriately balanced salt solution containing fish somatotropin. Results have been encouraging; within 5 weeks, body weight had increased by 1.6 times over that of controls (34). Evidence exists that invertebrates also produce somatotropins. Somatotropin from abalone has been isolated and shown to enhance growth in juvenile abalone. Recombinant trout somatotropin has been shown to increase the size of oysters (34). Somatotropin also can be used to increase growth in finfish and shellfish. Research is needed to determine the most effective and practical means of administration. Large-scale production and purification of recombinant fish somatotropin is paramount. Optimum dose, nutrient requirements, and other related conditions must be established for each target species. Most studies to date have been short-term studies. Long-term studies to understand the effects of somatotropin on fish must be conducted. Given the work that is still needed, it is unlikely that somatotropin will be used commercially in the fish industry before the second half of the 1990s. Somatotropin Related Technologies Recognition of the role that somatotropin plays in growth and milk production has led researchers to search for means to increase endogenous levels of somatotropin in livestock as an alternative to administration of exogenous somatotropin. The production and secretion of somatotropin by the pituitary gland is controlled by another protein hormone called growth hormone releasing factor (GRF). Early studies in pigs involved daily injections of 30 micrograms of GRF. Neither growth rate nor feed efficiency was significantly improved. There was a significant improvement in carcass composition (less fat), although the improvement was not as great as with exogenous administration of porcine somatotropin. Using synthetic analogs of GRF that are resistant to degradation by protease enzymes elicits a greater reaction; daily weight gain and feed efficiency increased, and carcass composition changed in a manner similar to that which occurs with exogenous administration of porcine somatotropin (16). There is some evidence that GRF does elicit some effects that are different than those of somatotropin. For example, a small improvement in the digestibility of dietary dry matter has been observed in GRF-treated cattle and this has not been routinely observed with bST-treatment (3, 16). GRF itself can be produced in bacteria, but some of the synthetic analogs cannot, and alternative methods will be required to produce sufficient quantities for commercial use. It is not expected that GRF will be commercially available before the later half of the 1990s. An alternative way to increase endogenous somatotropin levels is to block compounds that prevent the secretion of somatotropin. Release of somatotropin from the pituitary gland is blocked by a compound called somatostatin. Deactivating somatostatin will increase the levels of somatotropin in the animal. Somatostatin is deactivated by stimulating the animal to produce antibodies to this compound. The process involves coupling somatostatin with another compound that stimulates the immune system in animals. Administration of this coupled compound to an animal causes the animal to produce antibodies that bind to somatostatin and deactivate it, thereby preventing it from inhibiting the release of somatotropin from the pituitary. When used in pigs, this process doubled the concentration of porcine somatotropin and increased growth rates slightly, but it is likely that higher somatotropin levels will be needed to increase growth in pigs significantly. In cattle, use of this method increased growth rates by 10 to 17 percent and improved feed efficiency by 13 percent (16). A third possible way of increasing the effectiveness of somatotropin is to couple somatotropin with a monoclinal antibody specific for somatotropin. In dwarf mice

PAGE 76

Chapter 3Emerging Animal Technologies l 71 that have deficient pituitary glands, a somatotropin-monoclonal antibody complex increased weight gains 400 to 600 percent more than administration of somatotropin alone ( 16). In lactating sheep, a somatotropin-monoclonal antibody complex increased milk production more than somatotropin alone ( 16). The mode of action is not known with certainty. It is speculated that the complex is selectively recognized by different target tissues and receptors in preference to somatotropin alone. It is possible that the monoclinal antibody inhibits the receptor from internalizing the somatotropin, which allows the somatotropin to be active for a longer period of time. The use of monoclinal antibodies from species other than the animal being treated, however, may cause an immune response by the animal. Beta-Agonists Beta-agonists (also called beta-adrenergic agonists) are compounds similar to adrenaline. They are generally of two types, the beta1 agonists that stimulate cardiovascular functions and the beta-2 agonists that regulate smooth muscle function. Beta-agonists are currently used in humans to control bronchial asthma and to relax premature uterine contractions. Beta-agonists can also act as repartitioning agents. They redirect nutrients away from the formation of adipose tissue (fat deposits) and towards muscle growth (48). Almost all cells have beta-adrenergic receptors. interaction of beta-agonists with the cell membrane receptors initiates intercellular responses that affect fat and protein metabolism and accretion. Beta-agonists are not currently approved for use as livestock growth promotants in the United States. At least three companies have tested beta-agonists to promote growth and enhance carcass leanness in meat-producing animals. Beta-agonists tested include clenbuterol and cimaterol in lambs, beef, swine, and broilers (American Cyanamid); salbutamol in swine (Glaxo Animal Health, United Kingdom) and ractopamine hydrochloride in finishing swine, beef and turkeys (Eli Lily and Co.). Results of early studies with clenbuterol, cimaterol, and salbutamol were variable and available evidence suggests that none of these compounds are under development as growth promotants for livestock application (48). Eli Lilly and Company is developing ractopamine hydrochloride to enhance carcass leanness and promote growth in meat-producing animals. In finishing swine (i.e., pigs weighing 100 to 250 pounds), ractopamine is administered as a feed additive, at doses of 5 to 20 parts per million (ppm), usually for a period of 42 to 49 days. Ractopamine is registered under the trade name Paylean, and is currently undergoing FDA review (48). Trials involving finishing pigs were conducted in the United States, Canada, and several other countries worldwide. Ractopamine increases the rate of daily weight gain (maximum of 8.9 percent), decreases feed consumption (average of 3.9 percent), and improves feed conversion (up to 12.3 percent over untreated controls). s Additionally, two measures of carcass leannessloin eye leanness and the 10th rib fat thicknessimproved by a 14.9 percent increase and 13.6 percent decrease, respectively. Total lean content of the carcass increased from 50.9 percent to 56.9 percent as determined by total carcass dissection. Swine with superior genetics for leanness show a greater response to ractopamine than those with low lean-gain potential. Visual and taste panel evaluations of meat palatability characteristics from the ractopaminetreated pigs appear to be unchanged (48). While total feed consumption decreases slightly, use of ractopamine requires crude protein levels greater than current National Research Council recommendations for finishing swine. Rations containing 16 to 20 percent crude protein or lysine equivalent appear to optimize the growth performance response to ractopamine. However, carcass leanness effects are seen at lower crude protein levels. Addition of fat to the diet, a common practice in swine, did not affect carcass leanness, daily weight gain, or feed conversion responses to ractopamine (48). Some reports have indicated that beta-agonists cause hoof lesions in swine. No such effects were observed in another study with ractopamine given in amounts up to 25 times the highest intended level of use (550 ppm). Similarly, at three times the intended use level (60 ppm) during the finishing phase, there were no observed effects on the subsequent percent of gilts in heat, the percent 4Clenbuterol is cumently marketed in Europe, Mexico, Canada, South America, and Asia as a veterinary prescription drug to treat bronchial and smooth muscle disorders in animals (primarily race horses and sheep). It has not been approved for use in the United States. Salbutamol is marketed as an anti-asthmatic in humans ( 17, 48). Twelve trials involving 1278 barrows and gilts were fed rations of 16 percent crude protein and administered ractopamine as a feed additive in quantities up to 20 parts per million.

PAGE 77

72 l A New Technological Era for American Agriculture farrowing rate, the number of live or dead newborn pigs, the 21-day pig weaning weight, or gilt weights at the end of the nursing period (48). Antimicrobial Agents Biotechnology is being used to produce new compounds that can enhance livestock production, but traditional means will continue to be used for the same purpose. One such traditional method is the addition of antimicrobial agents to livestock feed. Antimicrobial agents are compounds that, when administered in low concentrations, suppress or inhibit the growth of microorganisms. Antimicrobial agents include antibiotics (naturally occurring substances produced by yeasts, molds, and other microorganisms) and chemotherapeutic (substances that are chemically synthesized). Copper also has antibacterial properties when present in relatively high concentrations. Antimicrobial have been widely used as feed additives for swine, poultry, beef cattle, and dairy calves since the early 1950s and numerous trials have been conducted during that time to document the efficacy of antibiotic use. Approximately half of the 4.65 million kilograms of antibiotics and chemotherapeutic sold in the United States in 1988 were for nonmedicinal use (12). In the early 1980s, it was estimated that approximately 75 percent of pig feeds, 80 percent of poultry feeds, 60 percent of feedlot cattle feeds, and 75 percent of dairy calf feeds contained antimicrobial agents (12). An estimated 90 percent of all feedlot cattle are administered antibiotics ( 12). Today, approximately 88 percent of the antibiotics used in livestock are given at subtherapeutic levels to promote growth, improve feed utilization, reduce mortality, reduce liver abscesses, and improve reproductive efficiency. Currently, 14 antibiotics and 6 chemotherapeutic have been cleared by the FDA for use as livestock feed additives (table 3-2). The exact mechanism by which antimicrobial stimulate growth is not known with certainty. Three mechanisms have been proposed: a metabolic effect, a nutritional effect, and a disease control effect. Various antimicrobial have been shown to affect water and nitrogen excretion, to inhibit oxidation reactions that require magnesium ions, and to increase protein synthesis in muscle cells. However, none of these metabolic effects is significant enough to account for the observed increases in growth (12). The nutritional effect is based on the premise that certain intestinal microbes synthesize vitamins and amino acids essential to animals, while others compete with the Table 3-2Antimicrobial Agents Approved as Growth Promotants for Swine, Poultry, and Cattle in the United States a Antibiotics Chemotherapeutics Bacitracin zinc (S,P,C) Arsanilic acid (S,P) Bacitracin methylene Carbadox (S) disalicylate (S,P) b Sodium arsanilate (S,P) Roxarsone (S,P) Sulfamethazine (S,C) Bambermycins (S,P) Sulfathiazole (S) Chlortetracycline (S,P,C) Lincomycin (S,P) Erythomycin (P) Lasalocid (C) c Monensin (C) c Oxytetracyline (S,P,C) Penicillin (S,P) Streptomycin (S,P) Tiamulin (S) Tylosin (S) b Virginiamycin (S,P) a The letters in parenthesis refer to the species for which the drug is approved; S = swine, P = poultry, and C =cattle. b Bacitracin methylene disalicylate and tylosin are also approved in cattle to reduce liver abscesses. c Lasalocid and Monensin are approved for use in poultry to control coccidiosis. SOURCE: Office of Technology Assessment, 1992. host animal for these nutrients. Shifts in the intestinal population of bacteria associated with the use of antibiotics could result in greater availability of nutrients for the host animal. Some antibiotics have been shown to stimulate yeast growth and bacteria that produce vitamins while reducing population levels of lactobacilli, bacteria that require amino acids in the same proportions as pigs and chicks. Increased intestinal wall thickness and total gut mass, thought to be caused by bacterial invasion or toxins, are reduced by antibiotics. This decreased mass possibly leads to greater nutrient absorption and increases diversion of energy and nutrients away from heat production by the gut to body growth. Evidence exists to support the hypothesis that the dietary protein requirements of animals administered antibiotics are lower than those of control animals. The most striking evidence in support of the nutritional effect is seen with the ionophore class of antibiotics, which causes an increase in propionic acid and a decrease in acetic acid in the rumen. Biosynthetic pathways using propionic acid are energetically more efficient than those using acetic acid, which could account for the marked reduction in feed requirements per unit of gain for animals administered the ionophores. The most widely accepted theory as to how antimicrobials promote growth is the disease-control effect.

PAGE 78

Chapter 3Emerging Animal Technologies l 73 Antibiotics control subclinical disease, thereby allowing animals to more closely approach their genetic growth potential. The fact that antibiotics stimulate growth more in young animals than older animals provides some support for this theory because young animals have lower immunological competency and are more susceptible to disease. Also, the degree of the growth response is strongly influenced by the cleanliness of the living environment and the disease load of the animals involved. Most of the research concerning antimicrobial is conducted at the pharmaceutical firms that develop these products. Research at universities evaluates the efficacy of already approved antimicrobial agents under different housing, management, and feeding programs. Some clinical studies of compounds in development are also conducted at universities. Current research is focusing on the development of new antimicrobial, new techniques for screening and evaluating the safety of antimicrobial, detection of residues in meat, and the possible spread of antimicrobial resistance. Genetic engineering techniques can be used to alter the production of antibiotics by bacteria and to develop nucleic acid probes for use in safety evaluation. Other research is focusing on ways to improve the efficiency of nutrient utilization and microbial fermentation in the gastrointestinal (GI) tract. Techniques that modify membrane function in bacteria can increase the transport of ions and substrates into bacterial cells, which could enhance digestion in ruminants. Alternatively, the use of live antagonistic microorganisms in feed can be used to maintain the optimal microflora. More efficient methods of delivering antimicrobial, including intraruminal delivery devices, boluses, and rotation of two or more agents, are being developed. The compatibility and synergism of antimicrobial combinations and the effect of the diet are also being explored (12). Antimicrobial Use in Poultry Antimicrobial use in chickens up to 4 weeks old increases growth rate and feed efficiency by approximately 7 and 4 percent, respectively. Older chickens also show improvement, although not as high. Young turkeys have shown improved growth rates and feed efficiency of approximately 13 and 7 percent, respectively. When antimicrobial are used in laying hens, egg production improved by up to 4 percent, the feed required per dozen eggs was reduced up to 5 percent, and matchability improved about 3 percent. Similar results were obtained in turkeys. Antimicrobial use also appears to reduce mortality (12). Antimicrobial Use in Swine In pigs, antimicrobial have been shown to increase growth rates, reduce feed requirements per unit of weight gain, and reduce mortality and morbidity. Smaller (younger) pigs respond more to antibiotics than heavier pigs. Antibiotics have been found to improve growth rate of pigs weighing between 7 and 25 kg by 16 percent and to reduce the amount of feed required per unit of gain by 7 percent. In slightly heavier pigs (from 7 to 49 kg), the improvements in weight gain and feed efficiency were 11 and 5 percent, respectively. Over the entire growingfinishing period, antibiotics improved weight gain by 4 percent and feed efficiency by 2 percent. Improvements in growth rates, feed efficiency, and mortality rates from antibiotic use are greater under farm conditions than in highly controlled test conditions at universities and research stations. In addition, the effectiveness of antibiotics has not diminished over 40 years of use ( 12). Copper gives growth rate and feed-efficiency utilization rates similar to those of antimicrobial, and in young pigs a combination of copper and antimicrobial appears to have an additive effect. Antimicrobial are not usually continuously administered to breeding animals, but during certain critical stages of the reproductive cycle, such as at the time of breeding, administration of antimicrobial can improve conception rates (by about 7 percent) and increase litter size (by about a half a pig). Use of antimicrobial at farrowing reduces the incidence of uterine infections. Data also indicate a slight improvement in the survival and weight gain of nursing pigs that have been given antimicrobial in prefarrowing and lactation diets. Evidence also exists that the withdrawal of antibiotics from animals that have been administered antibiotics for a long time is associated with a reduction in reproductive performance ( 12). In the last 5 years, two new antibiotics were cleared for use in swine. Three more antibiotics are currently under development (12). Antimicrobial Use in cattle In beef, growth rates have increased up to 5 percent, and feed efficiency gain has increased up to 7 percent with antimicrobial use. Antimicrobial are also commonly used to reduce, by nearly half, the incidence of liver abscesses. Animals with abscessed livers gained

PAGE 79

74. A New Technological Era for American Agriculture weight more slowly than those without abscessed liversabout 1/3 pound per day less. Antimicrobial can be used to improve weight gain in dairy calves. but no general beneficial response has been noted in lactating cows (12). Anabolic Steroids Steroids are a class of lipid compounds composed of four interconnected rings of carbon atoms linked with various functional groups. Some steroids act as vitamins while others act as hormones. The anabolic steroids used to promote growth are estrogens and progesterone (female sex hormones) and androgens (male sex hormones). Steroids have been demonstrated to promote growth, increase feed efficiency, increase lean meat production, and reduce carcass fat. These hormones have been demonstrated to have growth-promoting properties in beef, sheep, swine, poultry, and fish. Such effects are greatest in ruminants. Anabolic steroids were first approved for use in livestock in 1954. Currently they are approved for use as growth promotants in the United States only for beef and sheep. It is estimated that 10 percent of heifers and 60 percent of steers are treated with anabolic steroids as calves; 70 percent of stocker cattle; and 90 percent of feedlot cattle are administered anabolic steroids (35). Anabolic steroids reduce the cost of producing beef by an estimated $17 per head, and a complete ban on anabolic steroids in the United States would result in an estimated net-return loss of $2.4 to $4.1 billion in beef and sheep products (35). Anabolic steroids are used in the United States either singly or in combination, with the most common method of administration being a prolonged release implant inserted at the base of the ear (see table 3-3). A combination estradiol-trenbolone acetate implant is currently under FDA review. The mechanisms by which steroids act in livestock are still not known with certainty, despite the fact that these compounds have been used for nearly 40 years. It has generally been postulated that estrogens stimulate the production and release of somatotropin from the pituitary gland, and that the increased sornatotropin, in concert with insulin, increases the uptake of amino acids and the synthesis of muscle protein (35). New studies indicate, however, that estrogens and somatotropins are additive, and act independently, and therefore it is unlikely that the action of estrogens occurs via elevated levels of endogenous somatotropin. This evidence has led to the proposal of alternative hypotheses. One such proposal postulates that because there are estrogen receptors in bovine skeletal muscle, estrogens could directly bind to these receptors and stimulate protein synthesis (35). Alternatively, estrogens may stimulate the somatotropin receptor sites in the liver; greater binding and receptor capacity has been observed following estradiol administration. However, estrogens do not elicit an anabolic response in rats despite the fact that they stimulate somatotropin release and there are estrogen receptors present in rat skeletal muscles. This evidence suggests that the mode of action of estrogens may in fact be different than any of those hypothesized (35). Table 3-3Anabolic Steroids Commercially Available in the United States Commercial Method of Anabolic steroid name use Estrogens Beta-estradiol Compudose Implant Zeranol a Ralgro Implant Androgens Trenbolone acetate Implant Progesterone Melengesterol acetate Feed additive Combination Beta-estradiol/testosterone Synovex-H Implant Heifer-oid Implant Beta-estradiol/progesterone Synovex-S Implant Synovex-C Implant Steer-oid Implant a Zeranol is technically not an estrogen (its produced by a fungus) but has estrogenic properties. SOURCE: Office of Technology Assessment, 1992

PAGE 80

Chapter 3Emerging Animal Technologies .75 Most androgens have not consistently shown anabolic activity in ruminants, although trenbolone acetate (TBA) used alone, and especially when combined with estrogens. gives good response. TBA significantly elevates plasma estradiol levels, which may explain at least part of its activity. Androgens are thought to work by blocking muscle receptors for another class of hormones, the corticoid hormones. This decreases muscle protein degradation and turnover, rather than increasing protein synthesis (35). The pharmaceutical industry conducts most anabolic steroid research. Universities conduct some research concerning the mechanism of action of steroids and work in conjunction with the pharmaceutical industry to conduct clinical trials. Current research is focusing on using combinations of steroids and on methods to improve timedrelease implants so that they release lower levels immediately following implantation and continue to release for a longer period thereafter. Researchers are also exploring the possibility of administering androgens to pregnant ewes and cows in the hope of increasing growth potential in the offspring (a process known as imprinting). Imprinting has been shown to improve growth. feed efficiency, and carcass leanness in female offspring, but leads to no observed changes in castrated male offspring (35). A clearer understanding of the mechanism of action of anabolic steroids is needed. Research is also needed to determine the optimum dose of steroids required to maximize anabolic response. Current dosage rates are 14 to 36 mg for estrogens, 200 mg for progesterone, 200 mg for testosterone, and 140 to 200 mg for trenbolone acetate, administered by implants lasting for 90 to 120 days. These doses are probably lower than those that would yield maximum growth; however, to change dosage would require FDA approval (35). Determining optimal dosage for maximum anabolic effects might also help determine the mode of action of these steroids and whether steroids are additive in effect with other hormones. Further research is needed to determine the nutrient requirements for maximum response and to determine the effects of steroids on meat marbling. Anabolic steroids do not appear to affect the texture, flavor, juiciness. or cooking loss of meat, but some controversy remains concerning the effect of steroids on carcass quality, marbling, and carcass grade, particularly with respect to TBA/ estradiol combination (35). REPRODUCTION TECHNOLOGIES The field of animal reproduction is undergoing a scientific revolution. For example, it is currently possible to induce genetically superior cows to shed large numbers of eggs (superovulation). It is also possible to fertilize these eggs in vitro with the sperm of genetically superior bulls. Each resulting embryo can then be sexed and split to produce multiple copies of the original embryo, frozen for later use, or transferred to recipient surrogate cows whose reproductive cycle has been synchronized to accept the developing embryo. In the near future, it may be possible to sex the sperm rather than the embryo and to create greater numbers of copies of each embryo than is currently possible. Embryos produced by new reproductive methods are currently being marketed. Techniques now being developed will make it easier to insert new genes into the embryos to produce transgenic 6 animals. Although as yet no transgenic farm animals are commercially available, these new technologies are being used to improve the quality of livestock herds more rapidly than could be achieved with traditional breeding. Currently, however, many of these technologies are still relatively inefficient. Estrous Cycle Regulation Research has shed new light on the basic mechanisms controlling egg growth and maturation, and corpus luteum 7 function. This new knowledge is aiding the development of precise methods to regulate the estrous cycle, induce superovulation, and reduce the heavy losses due to early embryo deaths that occur in all domestic animals. Perhaps the most important development in ovarian physiology in recent years is the discovery of the ovarian hormone inhibin, which decreases the ovulation rate. g Some breeds of animals with exceptionally high ovulaAnimals whose hereditary DNA has been augmented by the addition of DNA from a source other than parental germplasm, using recombinant DNA techniques (46). Transgenic animals can be created that possess traits of economic importance including improved disease resistance, growth, lactation, or reproduction. 7 The COWUS lu[eum is a temporary endocrine organ that is produced at the site of ovulation during each estrous cycle. lt produces hofmones needed to maintain pregnancy. x ]nhibln decreases ~)vu]atlon rates by suppressing the secretion of follicle stimulating hormone (FSH), a ho~one produced by the PituitaV gland.

PAGE 81

76 l A New Technological Era for American Agriculture Photo credit: U.S. Department of Agriculture, Agricultural Research Service. Animal physiologist prepares an embryo for microscopic examination before implanting it into an animal. tion rates, such as the Booroola strain of Merino sheep in Australia, are known to have low levels of circulating inhibin. Cattle immunized against inhibin have lower circulating levels in their blood and show increased ovulation rates, The genes controlling inhibin production have been cloned, and the potential exists for producing transgenic animals in which these genes are repressed or deleted (18). Progress has also been made in understanding the control mechanisms that regulate corpus luteum function and its production of progesterone, a hormone that regulates the length of the estrous cycle and helps maintain pregnancy. Regulation of the estrous cycle is needed to ready surrogate mothers to receive embryos, and also to initiate superovulation. Estrous cycle regulation is reasonably well understood and developed in cattle and sheep. Conception rates in treated cows are similar to those obtained with animals bred at naturally occurring estrus. The estrous cycle of pigs appears to be more complex than that of ruminants and the process of controlling the cycle is not as efficient. Currently, superovulation treatments for cattle use highly purified hormones produced by recombinant DNA technology. About 10 viable eggs are produced, on average, per treatment (compared to the 1 egg a cow normally produces per ovulation) ( 18). As new knowledge of the factors controlling egg development and corpus luteum function is applied, the number of viable embryos produced by each superovulation treatment is expected to increase. Once eggs are collected, they are matured and fertilized in vitro. In vitro fertilization occurs only when a capacitated sperm (i. e., a sperm specially prepared to penetrate the egg cell membrane) encounters an egg that is in an optimal maturation state. Great progress has been made in understanding the factors involved in egg maturation and sperm capacitating in livestock. As a result, in vitro fertilization rates as high as 70 to 80 percent are produced in cattle, swine, sheep, and goats, and offspring are successfully produced. Conception rates with superovulated and artificially inseminated eggs in cattle are the same as those obtained by artificial insemination of control animals bred at naturally occurring estrus. Embryos produced with these techniques are currently being marketed. It is estimated that about 100,000 calves are born annually in the United States as the result of embryo transfer techniques. Many more embryos are being exported (41). Early detection of pregnancy can enhance a livestock producers ability to identify and rebreed animals that have not become pregnant. Traditionally, pregnancy has been detected by rectal palpation. This procedure can be conducted at 40 days post breeding, but at this early date the possibility exists of damage to the fetus. In practice, rectal palpation is usually carried out at 60 days or later in cattle. An alternative method is to measure progesterone concentration in milk. Concentration can be measured at 20 days after breeding. However, the process is expensive and results in about 15-percent false positives. A new method under development involves using a radioimmunoassay procedure to detect protein B, a gly coprotein produced by cells of the ruminant placenta (18). High embryo mortality is a major cause of reproductive loss in all livestock. Embryos of all species must signal their mothers in some way to prevent regression of the corpus luteum, so that the progesterone secretion needed to maintain pregnancy can continue. Early pregnancy recognition signaling systems are complex and apparently differ from species to species. In ruminants, compounds similar to alpha interferon may be early signals of pregnancy. Administration of interferon early in pregnancy is being tested as a possible means of reducing 9spm ~-pacitatlon invo]ve5 the Uptake of calcium ions which changes the PH of the sperm

PAGE 82

embryo loss. In mice and humans. platelet activating factor is known to be an early pregnancy reccognition signal. Preliminary data exist to suggest that it may play a role in early pregnancy in sheep and cattle (18). Embryo Cloning Multiple copies of a mammalian embryo were first produced by physically splitting an early embryo into halves. giving rise to identical twins (18). If the embryo) is dividemore than twice, however, few offsprings survive. Thus, no more than four indentical animals can be produced by splitting, and generally only two empryos are produced by this method. This procedure is already used in the cattle empryo transfer industry nearly doubling the number of offspring produced. A more efficient and promising method of producing multiple copies of an embryo is a technique called nuclear transplantation. Basically, the procudure involves the transfer of a nucleus from a donor embryo into an immature egg whose own nucleus has been removed. The recipient egg cell is activated by exposuure to an electric pulse, allowed to develop into a multicelled embryo, and then used as a donor in subsequent nuclei.ir transplantations to generate multiple clones. This procedure (outlined in figure 3-2) has been used successfully with cattle, sheep, and swine. This technique has already produced hundreds of embryos that have been successfully carried to term in cattle. and recloning has resulted in as many as eight calves from one embryo (29). The value of this technique is enhanced by the ability to transfer nuclei successfully from frozen embryos into eggs whose nuclei havc been removed. Conception rates obtained after transfer of embryos produced by neculear transplantation are varible, but rates us high as 50 percent have been obtained. However, embryo losses after transfer are higher than normal, resulting in actual pregnancy rates ranging from 15 to 33 percent (18). Combining the techniques of in vitro fertilization. embryo cloning, and artificial estrous cycle regulation can result in major changes in livestock breeding and in the rates of genetic improvement. Embryo and Sperm Sexing The availability of a technique to preselect the sex of the progeny holds great economic potential for the livestock industry. In the diary industry, females are the major income producers. while in the beef industry, males area economically more valuable. Until recently, no methods existed that provided the degree of separation needed for commercial use. However, recent advances in the seperation of the X and Y sperm, and sexing of the embryo have been made. It has long been a goal of mammalian phsiologists to develop a method to effectively separte X and Y chromosome-bearing sperm to control the sex of the offspring. Most sperm seperation techniques are based on potential differenccs in the size and density of the two sperm types. 10 These methods, however have met with little success (41). Development of cell-sorting techniques based on the differences in sperm size and Fluorescence of sperm DNA (flow cytometric measurements) has provied the first effective mehod to sort the sperm C ellls. Johnsson et al. (22) recenently reported successful serperation of intact vi. able X and Y chromosome-bearing sperm using this method. Although the differece in DNA contents of the X and Y chromosome-bearing sperm in rabbits amounts to only about 3 percent. 94 percent of the rabbits (does) inseminated with X-bearing sorted sperm produced females and 81 percent of the does inseminated with Ybearing sorted sperm produced males. This method has been used to separate X and Y bearing intact sperm of cattle. swine, and sheep with greater than 80-percent accuracy (2 ). Commercial use of this process is limited. at present. by the number of sperm that can be sorted per hour and by increased embryo mortality observed in the embryos produced after insemination with the sorted sperm. Neither of these factors is thought to represent an insurmountable difficulty. The most accurate method of sexing embryos is to create a picture of the number, size, and shape of the chromosomes contained in the embryonic cells, a process called karyotyping. However. this method requires removal of about half of the cells of early stage embryos, which decreases embryo viability and limits the number of embryos that can be transferred. Another method uses antibodies 11 to defect proteins (antigens) unique to male embryos. This method is not damaging to the embryos and encouraging results have been obtained in one laboratory; however, the technique yields variable results and has not been widely adopted ( 18). 1 Mcthocl\ used arc differential wxiimentation techniques including differential velc)clty sedimentation, free-t]ow clcc[roptlorcsis, amt con~ection counter-stream ing galvanization. 11 The antibodies are attached ( labeled) to a tlourescent cxmlpound to allow for detection.

PAGE 83

78 l A New Technological Era for American Agriculture Figure 3-2Nuclear Transplantation An embryo is nonsurgically removed from a donor cow or is produced by in vitro fertilization. Individual embryo cells are removed. Each embryo cell is injected into a specially prepared egg cell that has had its nucleus removed. An electric pulse is administered to cause fusion. Each egg cell is grown to a multicell embryo at which point the cloning procedure can be repeated or the embryo can be transplanted to a cow that eventually gives birth. SOURCE: Office of Technology Assessment adapted from R S Prather and N L First. Cloning Embryos by Nuclear Transfer, Genetic Engomeeromg of Animals, W Hansel and B J Weir (eds.), Journal of Reproduction and Fertility Ltd Cambridge, UK, 1990, pp 125-134

PAGE 84

Chapter 3Emerging Animal Technologies .79 Photo credit: U.S. Department of Agriculture, Agricultural Research Service, Animal physiologist checks swine sperm cells on video monitor to evaluate their motility, a procedure that precedes laser X-Y sperm separation. More recently, the sex of bovine and porcine embryos has been determined by untempting to match fragments of DNA that are contained only on Y (male) chromosomes with the same DNA fragments in the embryo. Due to its chemical structure, a fragment of DNA will combine with a second DNA fragment that has a corresponding nucleic acid sequence. Therefore, a fragment of DNA that is specific to males can be used as a probe to identify male DNA fragments in the embryo. Combined with technologies that produce multiple copies of the DNA fragments, this method determines the sex of the embryo using only a few cells. It is rapid (about 6 hrs) and extremely accurate (up to 95 percent). but may be overtaken by the rapidly developing capability to separate X and Y chromosome-bearing sperm ( 18). TRANSGENIC ANIMALS The new reproductive technologies of superovulation, in vitro egg maturation and fertilization, nuclear transplantation. and embryo sexing can, and are being used to upgrade livestock herds. When these technologies are combined with recombinant DNA technologies (the identification, isolation, and transfer of selected genes), it becomes possible to produce animals containing foreign DNA in their germ lines (transgenic animals). (See figure 3-3. ) The tools of biotechnology provide the opportunity to develop transgenic livestock that contain genes coding for improved growth charticteristics, lactational performance, and resistance to disease and stress. Transgenic animals have human medical implications as well. It may be feasible to produce important pharmaceuticals in livestock. Only certain human drugs can be chemically synthesized or produced by bacteria, because some compounds undergo modifications after the protein has been produced (referred to as post-translational modifications). Animals are capable of performing these modifications. but bacteria are not. Transgenic animals can also serve as powerful research tools to understand genetic and physiological functions, and provide a model system with which to study human disease. The production of transgenic animals is inextricably linked to the new reproductive technologics discussed in the previous section. lndeed, it is impossible to produce animals containing foreign DNA in their germlines without first manipulating the embryo and transferring it to a recipient animal. Process of Creating Transgenic Animals The process of making a transgenic organism is similar for plants and animals, and many of the tools and methodologies used are the same. As in plants, to create transgenic animals. the gene being transferred must first be identified and purified. Appropriate mechanisms (vector or nonvector) must then be found to transfer the gene into the animal cell, and appropriate regulatory sequences must be included to ensure proper expression of the gene. Unlike plant cells that are regenerated into whole plants by tissue culturing techniques, animal embryos (with the exception of fish) must be transferred to surrogate mothers for development and birth. Gene Identification and Purification The methods used to isolate and purify animal genes for transfer are the same as those used in plants, and have been described in detail in chapter 2. The method described in chapter 2 is the creation and screening of genomic libraries, libraries of DNA fragments that contain all of the genetic material of the chromosomes. An alternative approach is to create what is called a complementary DNA (cDNA) library. This method can also be used in plants, and it is frequently used in animals. Genes are composed of DNA, and they code for proteins. But, before the protein is constructed, several intermediate steps occur. The DNA of the gene is first transcribed and processed into another compound called messenger ribonucleic acid ( mRNA). It is the mRNA that serves as the actual template for the production of

PAGE 85

Host cell preparation using Gene transfer. superovulation combined with Gene insertion is random. natural mating or artificial I by splitting I Early stage embryo (less than 64 cells) been synchronized to receive the embryo I J Cloned DNA obtained by gene identification and isolation SOURCE: Off Ice of Technology Assessment. adapted from J P Simons and R.B. Land, Transgenic Livestock, J Reprod Fert. Suppl. 34:237. 1987

PAGE 86

Chapter 3Emerging Animal Technologies l 81 proteins. Messenger RNA is not identical to the genomic DNA. This is because there are sequences of DNA contained within the gene that do not code for protein. After the DNA of the gene is transcribed to mRNA, these noncoding regions are snipped out and thrown away. Thus, the mRNA contains the coding regions, but not the noncoding regions of the genomic DNA. Special enzymes exist that can use the mRNA as a template to create DNA that has a complementary sequence to the mRNA. This new DNA is called complementary DNA (cDNA). It is identical to the sequence of the genomic DNA with the exception that, like the mRNA from which it was derived, it contains the protein coding regions, but not the noncoding regions of the genomic DNA (see figure 3-4). Thus. a library of cDNA sequences can be constructed from mRNA rather than the chromosomal DNA used to construct genomic libraries. The mRNA that serves as the protein template for the desired gene can be obtained from tissues that express high levels of the protein. For example. if one wanted to find the gene that produces insulin, a reasonable approach would be to extract the mRNA from the pancreas, which produces very high levels of insulin. This high level of insulin production means that there is a significant amount of mRNA for insulin. Also, because the pancreas is specialized for insulin production, mRNA for other proteins, say for example, somatotropin, may not be present in large quantities. Thus, the use of cDNA libraries decreases the amount of genetic material that must be searched to identify the gene of interest. The process of looking for a particular gene is tantamount to looking for a needle in a haystack. Use of a cDNA library, as opposed to a genomic library, provides a smaller haystack that must be searched. It might seem at first glance that the best method to use would be to construct cDNA libraries rather than genomic libraries. However, limits exist to the use of cDNA libraries. To construct both cDNA and genomic libraries, it is important to know the structure, sequence, and function of the protein for which one is trying to isolate the gene that codes for it. The lack of knowledge concerning the sequence and function of important proteins is the major constraint to the isolation and purification of the genes coding for those proteins. Additionally, construction of a cDNA library is easiest when tissues exist in the organism that specialize in the high-level production of the protein coded for by the gene that is being isolated. This method does not offer significant advantages when the protein is produced in low quantities by nearly every cell in the organism. Figure 3-4Construction of a cDNA Library Chromosomal DNA consists of reguiatory sequences noncoding sequences Each gene may contain several coding and noncoding sequences. When genes are expressed, the DNA is copied to a strand of RNA. Enzymes snip out and discard the noncoding sequences to form messenger RNA (mRNA) that contains ony the coding sequences. noncoding in the cell. sequences To construct a cDNA library mRNA mRNA is isolated from the cell. I I reverse transcriptase SOURCE: Office of Technology Assessment, 1992 Also, evidence exists that genes that do not contain the noncoding regions do not function as well as genes that contain the noncoding sequences (5, 7, 33). While the functions of the noncoding sequences are not known

PAGE 87

82. A New Technological Era for American Agriculture with certainty, they may play some role in the regulation and expression of the gene itself. Therefore, incorporating cDNA genes that do not contain the noncoding regions into transgenic animals results in the genes not being expressed as well as a genomic gene. Unfortunately, many of the animal genes that have been isolated and purified are cDNA genes rather than genomic genes. Thus, the tradeoff is that it may be easier to isolate and purify cDNA genes than genomic genes, but they dont work as well when used to create a transgenic organism (5, 7, 33). Gene Transfer Once an animal gene has been purified, it must be transferred to the host animal cell. Genes can be transferred using direct transfer methods (e.g., microinjection, electroporation, chemical) or vectors ( i.e., viruses). The first transgenic animals created were mice in 1980 (37). Since then, transgenic cattle, sheep, swine, poultry, and fish have been produced. The most common method used to produce transgenic animals is microinjection. This method involves directly injecting cloned DNA into a fertilized egg. 12 The cytoplasm of cow and pig embryos is opaque, and the embryos must first be centrifuged to locate the nucleus; otherwise the procedure for cows and pigs is similar to that used in mice, rabbits, and sheep (36). Fish embryos are surrounded by a tough membrane called a chorion, and this membrane first must be removed before DNA can be injected. Even with the removal of the chorion, the nuclei are not visible and so the DNA is injected into the cytoplasm. Injection into the cytoplasm rather than the nucleus requires greater amounts of DNA (34). Other direct transfer methods attempted include the use of short electrical pulses (electroporation), or chemicals to make cell membranes permeable to the passage of large molecules such as DNA. These approaches have been used with sperm as well as eggs. The possibility of using sperm as a method to incorporate new genes into a species is an exciting prospect. One research group has reported using this method successfully to create transgenic mice that passed the new gene on to their offspring (27). Other researchers, however, have not yet been able to duplicate this result. The use of electroporation methods in fish have resulted in up to 40 percent of the embryos becoming transgenic and this approach may be far more useful in fish than microinjection. Another approach being attempted in fish is the use of liposomes, vesicles contained in the phospholipid layer of cell membranes, as a means to encapsulate foreign DNA for entry into the cell. This method has not yet yielded any successes (34). Poultry reproduction is significantly different from that of other livestock species. By the time the fertilized egg is layed, the developing embryo may already contain as many as 60,000 cells. This precludes using the microinjection technique because the number of cells that might incorporate the injected DNA could be small. Additionally, only some of the cells that incorporate the foreign DNA will express it. Attempts have been made to inject DNA directly into unfertilized eggs still in the ovary, but this method did not yield any transgenic offspring (24). As a result of the deficiencies of direct gene transfer methods in poultry, a vector system has been developed. The most commonly used vector is a retrovirus. The gene that is to be transferred can be incorporated into the retrovirus. The host animal cell can then be infected with the retrovirus incorporating the new gene. Retroviruses are attractive vectors because only a single copy of the virus is integrated into a chromosomal site. Retroviruses also tend to be either species specific or to infect only a few closely related species. Two types of retroviral vectors have been developed. Replication-competent retroviruses are those that are capable of self-replicating. These viruses have been successfully used in chickens. One-day-old embryos were infected with the retrovirus and transgenic chickens were hatched. Furthermore, the virus successfully infected germ line (sex) cells, and the new gene was passed on to the transgenic animals offspring (24). Replication-defective viruses lack the genes necessary for self-replication. These viruses cannot reproduce without the presence of a helper vector. The retrovirus is engineered in such a way that it contains all of the normal viral genes except those needed to package its own genetic material. The helper vector (also engineered) possesses the genes needed for packing retroviral genetic material, but does not include the other viral genes (i. e., genes that enable it to infect cells and cause virulence). Introduction of the retrovirus and the helper vector into host cells provides all of the elements needed to enable the retrovirus carrying the desired gene to infect and incorporate that gene into the host chromosomes. This method is considered safer than using replication-comIZ Specifically, the DNA is in.jccted into the male pronuchm of the fertilized egg. The pronuclcii are the cg.g and sperm nucleii prwmt after the sperm penetrates the cgg membrane.

PAGE 88

petent ectroviruscs because the rcplictition-detective ectrovirus can only be infective and spread to other cells if the helper vector is present. However, there is a small possibility that the helper vector nd replicaton-defective retrovirus might recombine to form a replication-competent retrovirus. Additionally the DNA sequences carried by replication-defective retoviruses are not incorproated in the germ lnes of hickens. hence they are not passed to the offspring. Improved replication-defective retrovirus vectors are needed (2-I). A number of transgenic cattle, pigs, sheep, chickens. and fish have been produced using direct transfer methods (almost exclusively microinjection) and viral vector methods. However, these techniques have several limi tations. Microinjection techniques are expensive to use and the efficiency of transgenic animal production is very low. For a transgenic animal to be created. embryos must survive the physical manipulation and inflection of DNA. must incorporate the DNA into their chromosomes, and must express the gene product. The percentage of microinjected embryos that actually results in transgenic animals is low, ranging, for example. from (). I to 4.45 percent in sheep and from 0.3 to 1.73 percent in swine (36. 38). The low rate of efficiency limits the study of transgenic livestock because of the high number of donor and recipient females that must be maintained totexperimentation. Efficiency rates are much higher in fish, ranging from 35 to 80 percent, because fish undergo external fertilization and do not require in vitro culturing of the embryos and transfer to surrogate mothers. Microinjection techniques are not only inefficient methods of creating transgenic animals. but they also do not provide any control over where the new gene is incorporated into the gernome (26). The site of gene incorporation is random. which also occurs with retroviruses. Because the site of incorporation influences gene expression, random insertion causes reduced control over the ability of researchers to control expression levels. Because of these deficiencies. alternatives to viral vectors and microinjection are being sought. A promising new method for gcnerating transgenic animals has recently been developed in mice and may be applicable to other mammals. This new technique uses stern cells derived from an embryo. Stem cells are normally undifferentiated, that is. they do not become specialized tissue cells such as muscle, brain, liver cells, ect. Howe\er. stem cells retain their ability to become specitilized cells under the proper stimuli ( i.e., they are pluripotent ). 13 Stem cells can be used as vectors to introduce selected genes into a host embryo. This method hits several significant advantages over microinjection methods. the most profound of which is that it is possible to insert DNA at specific, predetermined sites within the genomc of the stern cells (18). Targeted insertion is possible bccause stern cells have an intrinsic ability to recombine similar (homologous) DNA sequences. which results in the replacement of an endogenous gene with the desired gene. Stem cells can also be tested in vitro to ensure that integration of the new gene has occurred before these cells are transferred to a developing embryo. To isolate stem cells (see figure 3-5), an early stage embryo is cultured on a monolayer of specially prepared cells. The proliferating embryo cells are recultured until individual sterns cells can be isolated. These individual stem cells can then be cultured indefinitely. At this stage. DNA sequences containing desired genes can be inserted into the stem cells. 14 A geneticallly transformed stem cell is then microinjected into an immature embryo to produce a chimera, an organism that contains cells from more than one source. It the stem cells are incorporated into the germ lines of these chimeric animals, then these animals can be interbred to obtain offspring homozygous for the desired trait (18). Use of the stem cell method will make it possible to produce a broad range of transgenic animals that could not be produced economically using direct microinjection or viral vectors. Targeted gene insertion also has the significant advantage of allowing host animal genes to be inactivated or removed and replaced with modified forms of the genes, such as ones that are expresses at a higher level, have new patterns of tissue-specific expression, or have a modified biological activity. A host organisms endogenous genes can be inactivated by targeting an insertion into an essential region of the gene. This fact is of particular interest to the livestock industry, because inactivation of genes that have inhibitory physiological effects is likely to result in improvement in a number of productive traits. For example, bovine somatostatin is a hormone that inhibits bovine somatotropin production; inactivation of this gene would ~ plurlpotcncy help ~ake s[cm ccl ]~ ~[[rac[iv~ vectors of DNA transfer. While in tiswe culture. DNA ctin cusily be inserted into stem CCIIS. When stem cells w-e injec[cd Into an early stage embryo. the conditions for tissue specialization arc present, and stcm CCIIS undergo the normal tissue development that occurs as the embryo develops during pregnancy. Thus. using stcm cells provides an efficient means to transfer DNA. IJ Meth{)d~ u~ed include ~ iral infection and use of an electric pu]w to make cl?]] membranes Ietik) (clcctr~)pt)ruti(m ).

PAGE 89

84. A New Technological Era for American Agriculture Figure 3-5Gene Transfer Using Embryo Stem Cell Culture A mouse embryo (blastocyst) is donated. The embryo is cultured onto a thin layer of specially prepared (feeder) cells. The embryo attaches to the cell layer, and the inner cells of the embryo begin to proliferate. Groups of these differentiating embryo cells are separated, recultured and single colonies of embryonic stem cells are identified and transferred. x SOURCE: M.R. Capecchi, The New Mouse Genetics: Altering the Genome by Gene A second mouse embryo is donated and is injected with a cultured embryonic stem cell. This embryo is transplanted into a surrogate mouse, which gives birth to a chimera (a mouse with cells from both parent embryos). Mating two chimeras gives rise to offspring with the desired traits. Targeting, Trends in Genetics 5:70, 1989

PAGE 90

Chapter 3Emerging Animal Technologies l 85 result in increased endogenous somatotropin secretion and, presumably, increased milk production and more efficient growth. If successful, this technology could be used in lieu of administering bST exogenously to increase milk production. The genes controlling the production of inhibin, the ovarian hormone that reduces ovulation rate, provide another example of potential tin-gets for deactivation. The ability to inactivate genes also provides a powerful research tool for the study of the function of genes in vivo. Stem cells have been isolated in mice and hamsters and possibly rabbits. There are reports that stem cells have also been isolated for swine ( 18). Progress is being made in isolating stem cells in sheep, and much research is being conducted to isolate bokvine stem cells, but to date, this has not been accomplished. There has been no documentation of embryonic stem cells being isolated from poulty However. in a similar type of procedure. 1-day-old embryonic cells from chickens have been isolated and introduced into immtaure embryos of other chickens. About I I percent of the resulting embryos were chimeric, and one embryo developed to hatching (24). Stern cells have not been isolated in fish (34). Promotors and Gene Expression The expression of new genes in transgenic animals is poorly regulated. Apprprate levels of gene expression are important. because overexpression can lead to impaired health in the transgenic animal. Better understanding is needed of how to turn genes on and off when desired; of how to regulate the level of gene expression: and of how to direct the expression of the gene to specific tissues at different stages of development. At the present time the factors that cause genes to have tissue and developmental specificity are not well understood. Currently, fewer than 10 promotors or regulatory sefquences have been used to direct gene expression in transgenic live stock. Most of these promotors are derived from mice or viruses. The most commonly used promotor is the mouse metallothionein promoter, which is responsive to dietary stimulation by heavy metals such as zinc. Three promotors are being examined for their abilite to direct gene expression in mammary glands. A fourth promotor directs expression primarily to the liver. It may be desirable to use promotors derved from the same species that is receiving the new gene. Evidencc euggests, for example, that using a mouse promotor sequence in pigs results in somewhat different gene expression than use of the same promotor in a mouse (18. 36). Levels of gene expression do not aways correlate with the number of gene copies incorporated into the chromosome of a transgenic animal. This suggests that the site of the incorporate ion of the new gene in the host chromosome also affects gene expression. Given that embryonic stem cell procedures still require considerable development before directed insertion can occur. some researchers are examining methods to control gene expression independently of the site of integration. Research is focusing on regulatory elements that allow the new genes to provide their own environment for expression. 15 Transgenic Poultry Research emphasis has been given to improving growth and disease resistance. Bovine somatotropin has been transferred to chickens and increased the mass of the chicken. The envelope gene of avian leukemia virus has also been transferred to chickens and the cellss that expressed this gene have been shown to be resistant to subsequent infection with the same strain of vrirus (24). Research is being conducted by USDA Agricultural Research Service and universities in the United States. as well as by a limited number of private firms. It is interesting to note that most of the funding for transgenic poultry research conducted in the United States is being supplied by other countries (mainly Canada and France). Commercial availability will take 7 to 12 years after the production of an adequate number of transgenic fonder male chickens. Transgenic Swine Several genes have been successfully transferred into pigs, including those for somatotropin. human growth hormone releasing factor (hGRF). human insulin-like growth factor--l (hIGF-I ), mouse MX ( to investigate resistance to respiratory discuses), mouse whey acidic protein (WAP) (to investigate mammaryspecific expression, and light and heavy beta chains for antibodies to produce specific immunoglobulns (36). With swine. as with other livestock species. researchers are focusing on improving growth. increasing disease resistance. and producing highvalue pharmaceutical products.

PAGE 91

86 l A New Technological Era for American Agriculture Photo credit: U.S. Department of Agrculture, Agricultural Research Service. Rooster on left was injected with genes of avian Ieukosis virus when it was a 1-day-old embryo. Roosters in center and on right are of two succeeding generations which directly inherited those virus genes. Somatotropin transferred to pigs has been shown to increase feed efficiency, enhance meat quality, reduce carcass fat, and increase the rate of gain. When fed a high-protein diet, transgenic pigs containing somatotropin genes gained weight nearly 17 percent taster than controls. and showed up to 18 percent greater feed efficiency. Backfat was significantly reduced and meat was leaner (36). Transgenic pigs that expressed the somatotropin gene passed that expression on to their offspring. Offspring that contain the somatotropin and who were fathered by boars that expressed the gene also expressed the sornatotropin gene. The offspring containing somatotropin genes who were sired by boars that did not express the somatotropin gene, also did not express the gene. This suggests that the stability and functioning of the gene are the same in the parent and offspring (36). Pigs that continuously expressed high levels of somatotropin experienced significant health problems including lameness, susceptibility to stress. peptic ulcers, and reproductive problems. Animals that incorporated the somatotropin gene but did not express it, or that expressed it at low levels did not display these health problems (36). Researchers are interested in improving disease resistance. Genes that confer resistance have not been isolated. Attempts to transfer genes that code for antibodies Photo credit: Mark Lyons Transgenic pig at DNX research facility born with the capability to make human hemoglobin. to compounds contained on the surface of selected bacteria and internal parasites are being made (28.51 ). Also, genes of the Class 1 Major Histocompatibility Complex 16 have been cloned. It may also be possible to induce immunity to specific viral diseases by transferring genes from the virus to the pig. This method has been used successfully in chickens and may also be applicable t o other livestock species (36). Attempts are being made to produce rare. medically important proteins in pigs. A U.S. firm ( DNX ) has announced that it has successfully produced human hemoglobin in pigs. Transgenic swine research is being conducted by the Agricultural Research Service, a few universities, and the private sector. The American Red Cross is also interested in the production O f blood proteins in livestock. Commercial availability of transgenic pigs is not expected before the year 2000, and it is likely that the first transgenic pigs marketed will be used to produce pharmaceutical products. Additionally, pigs have a strikingly human-like physiology, and because of this, transgenic pigs are currently being developed to serve as a model system to understand and treat gastrointestinal cancers. Transgenic Ruminants The first transgenic ruminant to be successfully produced was a lamb, followed by goats and cattle. In cre16The major hlstoComPa[iblli[y complex is ~ chr~m~S~m~l re@on that COIltaillS SCVWd gCnCS lnVO]Ved In rCgUld[lng lnlI?NInC reSPOnSC.

PAGE 92

Chapter3Emerging Animal Technologies .87 ating transgenic ruminants. greatest research emphasis has been to improve growth characteristics (i. e., rate of weight gain, feed efficiency. and carcass composition). to produce valuable pharmaceutical products, and to enhance disease resistance. Genes coding for somatotropin and somatotropin releasing factor (GRF) have been purified and transferred to sheep. While the genes have been successfully transferred and expressed. control of the level and timing of expression has not been achieved. Sormtotropin levels in sheep have varied from a low of 40 nanograms (ng)/ milliliter (ml ) to over I ().000 rig/ml (31, 37). Extreme overexpression of somatotropin can lead to serious health problems in sheep. such as diabetes (39). In the future, researchers would like to alter the composition of milk and meat for improved processing characteristics, for higher nutrition. for less fat. and to alter the types of fat contained. Another major research area involves transferring genes that code for the production of valuable pharmaceuticals. Production of blood clotting factors ( factors VIII and 1X ), tissue plasminogen activator (TPA. used to dissolve blood clots that cause heart attacks), erythropoietin (used to treat bone marrow side effects resulting from AIDS treatment). and o-l-antitrypsin (AAT. used to treat emphysema are being investagated. A U.S. firm (Genzyme). in conjunction with Tufts University. has successfully produced TPA in goats (13.14). A Scottish firm (Pharmaceutical Prteins. Inc) has produced AAT in sheep. and is conducting research to produce Factors VII and IX and crythropoietin (30. 52). Transgenic cows producing high levels of pharmaceuticals in their milk have not yet been reported, but these animals are under development in a number of public and private laboratories. For example. a joint U.S. and Dutch group (GenPharm International. Gene Pharming Europe BV, and two Dutch Universities) has successfully produced tramsgenic cattle incorporating the human lactoferrin (which has antibiotic propertics) gene in the genome (25). Attempts tire being made to identify promotors that espress gene products only in milk. Research is being conducted on whey acid protein. a protein only found in milk. to identify the promotor that directs the synthesis of this protein. The goat (3-cascin promotor is also being used (14). Once appropriate promotors are found. the high levels of U.S. milk production coupled with the ease of milk collection may make this production method more cost effectove than the cell culture systems currently used in the production of certain pharmaceutical proteins. Enhanced disease resistance is another focus of research. Diseases that may be potentially controlled by the production of transgenic organisms include progressive pneumonia in sheep. and caprine arthritis-encephalitis in goats. The introduction of preformed antibodies have been shown to provide resistance to specific infections in mice and the antibody gene antiphosphoryrlcholine has been inserted in sheep (28). Researchers are also attempting to insert viral envelope genes that could possibly lead to enhanced resistance to viral infections. Researchers in Australia are attempting to increase wool production in sheep. Currently, wool production is limited by the amount of cysteine contained in and absorbed from the diet. Researchers are transferring bacterial genes that code for enzymes that produce cysteinc from sulfur in the diet (37). Research to produce transgenic ruminants is limited due to the high cost of the research. Research is conducted primarily in the United States by the Agricultural Research Service. a handful of universities. and a few private sector firms. and in Austrulia, Great Britain, and the Netherlands. It is not expected that transgenic ruminants will be commercially available before the turn of the century. Transgenic Fish Several species of transgenic fish have been produced. including rainbow trout, salmon, common carp. loach. catfish, tilapia, goldfish, zebrafish. and medaka. Several genes have been transferred to fish. including human. bovine, and trout somatotropin; genes that confer antibiotic resistance; and fish antifreeze protein genes (34). Transgenic fish containing the trout somatotropin gene grew 22 percent more than controls. and transmitted this increased growth rate to their offspring (34). Some species of fish produce a novel set of proteins that allo W them to withstand extremely cold water without freezing. These antifreeze proteins are produced year round by fish living in polar regions, and during the winter in fish living in temperate regions. The antifreeze genes in several species have been purified. Antifreeze protein genes from winter flounder have been transferred to salmon. Express ion levels of the gene were low. however. and protection against freezing was not achicved (34). Research Needs While significant advances in transgenic animal production have been made, it is unlikely that transgenic animals will be commercially available before the end

PAGE 93

88. A New Technological Era for American Agriculture Photo credit: Thomas Chen, University of Maryland Resultant transgenic carp with trout somatotropin incorporated into some but not all of their cells. The PI (middle) and F1 (top) transgenic carp are on average, 22 percent larger than their nontransgenic siblings (bottom). of the 1990s at the earliest. The ability to produce transgenic livestock possessing traits of economic value is currently limited by the absence of embryo stem cell technology, the lack of appropriate gene expression promotors, and the lack of knowledge about the physiological consequences of specific gene expressions. While the techniques for isolating and sequencing animal genes are relatively well developed, understanding of the functions of the genes has lagged. Analysis of gene function is complicated by the fact that many traits are controlled by multiple genes. Thus, manipulation of such traits will require detailed understanding of these genes and of their interactions. Ultimately, identification and understanding the physiology of the major genes controlling growth and lactation, reproduction, and disease and stress resistance in animals is needed. An active genome mapping program could enhance these developments. ANIMAL HEALTH TECHNOLOGIES Improvements in animal health will provide considerable cost savings to the livestock industry. Biotechnology is rapidly acquiring a prominent place in veterinary medical research. New vaccines and diagnostic kits are being developed to detect and prevent a variety of major livestock diseases. Vaccines Vaccines are agents that stimulate an effective immune response without causing disease. Traditional methods of vaccine development have involved killing or modifying pathogenic organisms to reduce the potential for disease while preserving that pathogens ability to induce an immune response. Biotechnology is being used to create new vaccines. Approaches used include deleting or inactivating the genes in a pathogen that cause disease, and inserting into a vector genes that cause an immune response to a pathogen. Synthetic peptides are also being produced that stimulate the immune response. Gene Deletion Vaccines Gene deletion techniques have been used to develop both viral and bacterial vaccines. The first gene deletion viral vaccine to be approved and released for commercial use was the pseudorabies virus vaccine for swine. initially, the removal of a single gene reduced the virulence of the virus. Since then, other genes have been deleted with a continuing reduction of virulence. Chickens that have been inoculated with recombinant avian leukosis virus (ALV) developed antibodies to the virus without developing the disease. Methods to decrease the virulity of live viruses lead to more effective vaccines because live virus vaccines stimulate the immune response more effectively than do killed virus vaccines (32). Bacterial vaccines have also been produced. Escherichia coli that lack certain genes. for example, have been shown to provide protection against gram-negative bacterial infections in cattle and swine. Live Salmonella modified to prevent reproduction in vivo have also proven to be an effective vaccine for cattle (32). Most gene deletion viral vaccines will not be available before 1995 with the exception of the pseudorabies vaccine, which is already available, and possibly the rabies and rinderpest vaccines, which are currently undergoing field trials.

PAGE 94

Photo credit U.S. Department of Agriculture, Agricultral Research Service Molecular biologists analyze DNA sequence reactions of a gene detection vaccine made from a modified bacterium. Vectored Vaccines New vaccines are also being created using vectors. Development involves deleting disease-causing genes from the vector if it is a pathogenic organism, or using a nonpathogenic vector. Genes that code for protective antigens produced by pathogens can be inserted into a vector. Inoculation of the animal with the recombinant vector stimulates an immune response to the inserted genes and confers protection against the pathogen. Pathogen surface protein genes are most commonly inserted into the vector. Inoculation of the animal stimulates production of antibodies to these surface proteins. When an animal is infected with the pathogen, it already recognizes that pathogen and produces antibodies against it. As an example, recombinant vaccines have been developed against the coat protein of a bacterial pathogen of the genus Vibrio, in fish. The most commonly used vector is the Vaccinia virus. Vaccinia viruses are used because they are easy and relatively cheap to manufacture, large enough to accommodate the insertion of many new genes (1). and stable without refrigeration. A single inoculation can induce immunity, and the recipient produces the bulk of the vaccine, eliminating the need for large vaccine factories. Vaccinia viruses also stimulate more than one type of immune response (i. e., they stimulate both B and T lymphocytes). However, there are disadvantages to using vaccinia virises: they have a wide host range (including humans), and could infect species other than target species; it is possible that they can revert to a virulent form; they cannot be administered orally; and they may pose a risk to immunosuppressed recipients. Vaccinia hosts have been used to produce vaccines against rinderpest (cattle), rift valley fever (sheep), Venezuelan equine encephalitis, bovine leukemia, rabies (cattle). vesicular stomatitis (cattle), avian influenza, avian infectious bronchitis, and respiratory syncytial disease ( 1, 32). Fowlpox virus is also being used as a vaccine vector. This virus cannot replicate in humans and is being used as a carrier for genes of pathogens that cause the poultry diseases of Newcastle disease, Mareks disease, bursal disease, coccidiosis, avian influenza, and avian infectious bronchitis. Raccoon poxvirus is being developed as a carrier for rabies. In fish, vaccines to control infectious haematopoietic necrosis virus (IHNV), a devastating viral disease of trout and salmon, are being developed by inserting coat protein genes into vectors. Other genetically engineered virus vectors that are in the early stages of development include avirulent adenoviruses, herpesviruses, murine and avian retroviruses, and bovine papillomavirus (1, 32). Bacterial vectors are also being developed. Escherichia coli and Bacillus subtilis are being used to produce antigenic proteins. They can be used to produce antibodies to Theileria annulata (a tick-borne parasite of cattle and sheep), coccidia in poultry, anaplasma (a parasite of cattle), and cysticercosis (a tapeworm in ruminants and swine). Pili genes from Bacteroides nodosus, the cause of foot rot in sheep, have been cloned into Pseudomonas aeruginosa, and have been shown to be an effective vaccine for foot rot (1, 32). Natural and Synthetic Peptides A number of animal species are known to produce small peptides associated with white blood cells and that are effective in destroying bacteria, fungi, and enveloped viruses. Such peptides, referred to as antimicrobial peptides, include defensins in mammals, bovine nubopeptides in cattle, magainins from frogs, and cecropins from moths. Some of the smaller peptides have been synthe-

PAGE 95

90 l A New Technological Era for American Agriculture sized and appear to have biologic activity similar to that of the natural peptides, and could be used in a manner similar to antibiotics. The genetically engineered protein lysostaphin, which kills Staphylococcus aureus, has reportedly achieved cure rates as high as 80 percent for mastitis in some clinical trials (1). Commercial development will take 5 to 10 years. Synthetic peptides can be constructed to stimulate an immune response in animals. Small fragments of proteins that are homologous to proteins coded for by the foot and mouth disease virus have been used to stimulate an immune response to that disease in cattle and pigs. Synthetic peptides have been used to inhibit critical functions of lentiviruses in sheep. Administration of a viral surface protein elicited production of an antibody and provided protection in fish. Commercial availability is not likely until the end of the decade. Monoclinal Antibodies To Confer Passive Immunity Monoclinal antibodies can be used to provide passive immunity to disease-causing microorganisms. They generally act not by stimulating the immune response of the animal itself, but rather by providing exogenous antibodies to the pathogen. Because monoclinal antibodies are specific to one antigen, they may provide only weak immunity to pathogens that have more than one immunogenic region of their surfaces. Certain strains of the bacteria Escherichia coli cause diarrhea in newborn calves. For diarrhea to occur, the bacteria must attach to the walls of the intestines. Attachments occur via cilia-like projections, called pili, that cover the surface of the bacteria. Monoclinal antibodies specific to the attachment proteins on the pili prevent attachment of the bacteria to the intestinal wall and prevent calves from getting diarrhea. A product currently on the market for diarrhea prevention in calves is Genecol-99 (50). Monoclinal antibodies specific for bluetongue also have been shown to protect sheep from this virus in trials. In addition to monoclonal antibodies, antisense agents can also provide passive immunity. Antisense agents can be synthesized and used as drugs, or used to block viral genes. They are very sensitive, but are susceptible to enzymatic degradation A delivery is a problem ( I). Immunomodulators Immunomodulators are hormone-like molecules that play a role in coordinating immune defenses to infectious agents, cancer, and autoimmune diseases. They act to boost or accentuate the immune response. Some of these molecules, the lymphokines, for example, are produced by white blood cells. Other immunomodulators, the cytokines, for example, are produced by other body cells. Two classes of lymphokines, the interleukins and the interferon, have been the focus of research attention. Interleukins are compounds that transmit signals between white blood cells. These signals help to stimulate the proliferation of disease-fighting white blood cells and the production of antibodies. Interferon induce the expression of class 11 histocompatibility antigens (define) and enhance their activity. Several interleukins and interferon have been identified in mammals, and the genes encoding some of these compounds have been isolated and cloned into bacteria (e.g., bovine alpha, beta and gamma interferon, bovine interleukin-2) (32). Lymphokines are being tested as adjuvants to boost immune responses to poorly immunogenic vaccines. For example, interleukin genes and genes for compounds that cause immune responses in animals (antigens) are being inserted together into viral or bacterial vaccines. This combination may enhance the immune response of the animal and lead to increased protection against the antigen. Recombinant interleukins produced in bacteria or other expression vectors may also be used therapeutically to assist in overcoming certain infections. For example, recombinantly produced interleukin-2 is being tested as a control for shipping fever and mastitis in cows. Mechanisms by which these regulatory proteins modulate immune response are now being investigated in domestic animals. Biotechnology is being used to identify and repilicate these compounds so that their function can be investigated. Diagnostics Safe, accurate, rapid, inexpensive, and easy-to-use diagnostic procedures are critical to the livestock industry at virtually all points in the production process. Examples of diagnostic tests include pregnancy tests and assays for pathogenic organisms. Many currently used diagnostic tests are costly, time consuming. and labor intensive, and some still require the use of aninml assay systems. Monoclinal antibodies and nucleic acid hybridization probes can be used to produce simpler, easily automated. and highly sensitive and specific diagnostic procedures. Antibodies are proteins produced by the body in response to foreign chemical substances. Monoclonal antibodies are produced by a cell line expressing only a single antibody type. They are the primary tools for bio-

PAGE 96

Chapter3Emerging Animal Technologies l 91 technology-based diagnostics. At least 15 different rapid diagnostic tests based on monoclinal antibodies are on the market or will be soon (table 3-4). These tests are highly specific and most lend themselves to automation, potentially allowing their application in mass screening systems for disease surveillance and control. Some of the tests have been adapted to field use and can be used by veterinarians or producers. The rapid commercialization of these products is having a significant impact on animal health management and disease control. Monoclinal antibodies are also being used in enzymelinked-immunoabsorbent-assay (ELISA) systems to provide sensitive, quantitative blood assays of toxins, hormones, chemicals (e. g., pesticide and antibiotic residues), and a variety of antigens including microbial agents. Many of these tests are commercially available. In some instances monoclinal antibody diagnostics have been used to replace bioassays such as mouse inoculation tests. The high specificity of monoclinal antibodies has generally been felt to make them less useful than polyclonal antibodies in initial screenings for diseases that have many serotypes. However, an ELISA kit containing just two monoclinal antibodies was able to detect 800 different Salmonella strains, so it may be possible that diagnostic kits containing just a few monoclinal antibodies could be useful for initial screening of pathogens ( l). Nucleic acid hybridization can also be used to diagnose the presence of microbes and parasites (table 3-5). Such assays rely on the bonding of a specific DNA or RNA segments (the probe) to complementary RNA or DNA fragments in a test sample. The probe is attached to (labeled by) a radioactive compound or to a color compound to allow for detection. DNA probes are most comTable 3-4Diagnostic Monoclinal Antibody Kits Avian Ieukosis Avian reovirus Bluetongue Bovine virus diarrhea Canine parvovirus Coccidiosis Episotic hemorrhagic disease Equine infectious anemia Feline infectious peritonitis Feline Ieukemia a Feline T-lymphotropic Ientivirus Feline T-lymphotropic lentivirus ( Feline leukemia Mastitis Pseudorabies a Rotavirus gastroenteritis Trichinosis a More than one company has a kit on the market SOURCE: Office of Technology Assessment, 1992. men. The development of RNA probes is very recent, and they are used to detect RNA viruses. The major limitation of nucleic acid hybridization is inadequate signal strength. The amount of target nucleic acid present in some samples may be too small to emit a signal the probe can detect. The polymerase chain reaction technique (PCR) (see ch. 2) can be used to amplify the amount of target DNA present and improve the ability of the probe to detect its presence. Similarly, bacteriophage replicase systems can be used to amplify the RNA present in a sample. Currently, the most reliable probes are those that are radioactively labeled. Use of these probes requires expensive equipment and trained technicians. thus precluding their use in the field. Alternative calorimetric techniques currently in development will replace the radioactively labeled probes and make the use of this technology more commercially attractive (32). The advantage that nucleic acid probes have over traditional diagnostic techniques is speed. Conventional tests for anaplasmosis and Johnes disease (an intestinal disease in ruminants), for example, require about 6 and 14 Table 3-5Pathogens for Which Diagnostic Kits Using Nucleic Acid Probes Are Available Viruses Bluetongue Bovine coronavirus Bovine Ieukosls Bovine virus diarrhea Equine encephalosis Foot and mouth disease Infectious bovine rhinotracheitis Porcine coronavirus Porcine parvovirus Rabies Rotavirus Bacteria Anaplasma marginale Campy lobacter Enterotoxigenic Escherichia coli Leptospira Mycobacterium Mycoplasma Salmonella Shigella Parasites Babesia bovis Eimeria tenella Eperythrozoon suis Hammondia hammondi Theileria parva Toxoplasma gondii Tritrichomonas foetus Trypanosoma brucei brucei Trypanosoma congolense SOURCE: Office of Technology Assessment, 1992.

PAGE 97

92 l A New Technological Era for American Agriculture weeks, respectively, to confirm the presence of the pathogen. This much time allows for interim spread of disease. With DNA probes the presence of these pathogens can be confirmed within a few hours. Restriction Fragment Length Polymorphism maps (RFLPs) can also be used for diagnostic purposes. This procedure has been used to distinguish different strains of African swine fever virus and has shown that equine herpesvirus1 can infect and cause abortion in cows under natural conditions. Research to develop diagnostic kits using biotechnology is being conducted in both the private and public sector. Currently, several diagnostics kits are commercially available. Development time to bring new diagnostic kits to market ranges from 2 to 5 years. Generally, less time is required to develop monoclinal antibody kits than nucleic acid probes. FOOD PROCESSING APPLICATIONS The processing of animal products into foods also will be affected by biotechnology developments. Americans consume many meat and dairy products that are fermented; genetically engineered fermentation starter cultures are being developed for these products. Starter cultures are living microorganisms used to produce fermented products such as cheese, yogurt, butter, buttermilk, sour cream, salami, and sausages. Culture organisms have been safely consumed by humans for centuries and serve as ideal hosts for the production of these natural foods. The metabolic properties of these organisms directly affect the properties of the food product, including flavor and nutritional content. In order to improve various properties of food products, food microbiologists attempt to manipulate the traits of the microorganisms, primarily through mutation and selection. The cloning and gene transfer systems developed in the 1980s are being used to construct strains with improved metabolic properties more rapidly and precisely than is possible with traditional methods. The development in this decade of new strains with precise biochemical traits will have an impact on several aspects of fermentation, including production economics, shelf-life, safety, nutritional content, consumer acceptance, and waste management (19). Although much of the current work to develop new strains of microorganisms has focused on the use of E. coli and other nonfood microorganisms, there are distinct advantages to engineering starter cultures for producing high-value foods. For example, construction of cultures resistant to attack by viral infection will impact processing costs by eliminating waste. Cloning of the genes responsible for ripening of aged cheeses can decrease storage costs by accelerating ripening. Production of natural preservatives, such as nisin (effective in inhibiting foodborne pathogens and spoilage organisms), will help ensure the safety and extend the shelf life of fermented meat and dairy products. Starter strains engineered to mimic the function of nitrates could reduce the use of these compounds in cured meats. Cloning of the gene(s) responsible for enzymatic reduction of cholesterol or modification of the degree of saturation of meat and milk fat will improve the nutritional quality of fermented products. The ability to engineer strains capable of producing enhanced flavors or natural stabilizers will influence consumer acceptance of fermented dairy foods. Enzymes, which are added to the curd to accelerate ripening, or to produce dairy products acceptable for digestion by lactose-intolerant individuals, will also be produced more economically by engineered microorganisms (19). A genetically engineered version of the enzyme preparation rennet, which is normally extracted from the forestomach of calves, has recently been approved by FDA for use in cheese manufacturing (See ch. 10). Processing of animal products generates many wastes such as blood, bone, collagen, shells, fish parts, and milk whey. Bacteria and yeast strains engineered to convert these waste products into useful products could decrease the cost and problems associated with their disposal. For example, engineered yeast strains are capable of fermenting the lactose in whey to value-added products, such as vitamin C, biofuels such as ethanol and methanol, or pharmaceuticals. Whey protein could potentially be used to produce specialty chemicals with biotechnology. Biotechnology products can be used to monitor animal products for food safety. DNA probes and monoclinal antibodies can be used to analyze raw materials, ingredients, and finished products for pathogenic organisms, bacterial or fungal toxins, chemical contaminants (i.e., pesticides, heavy metals), and biological contaminants (i.e., hormones, enzymes) (figure 3-6). Detection kits are commercially available. For example, kits are available to monitor several pesticides and antibiotics. Kits are also available to detect Salmonella. Animal cell cultures may partially replace whole animal systems to test for acute toxicity. Biosensors may be used to monitor food processing, packaging, transportation, and storage (19).

PAGE 98

Chapter 3Emerging Animal Technologies l 93 Figure 3-6Basic Steps in a DNA-Probe Hybridization Assay Isolate organism Disrupt organism to obtain dsDNA Convert dsDNA to ssDNA Add labeled probe and bind to solid support Hybridize probe Wash and detect signal Organisms present in a food product are trapped on filters and disrupted to obtain double-stranded DNA. Following denaturation of the DNA to single strands, the labeled probe is allowed to hybridize with target DNA. Hybridization can be detected by a number of methods. SOURCE: Journal of Food Protection 54(4):387, 1991 SUMMARY Biotechnology will offer many new opportunities to alter the manner in which livestock is produced in the United States. New products are being developed to enhance feed efficiency, improve livestock reproductive performance, and enhance herd health management. Producers, food processors, and consumers all potentially may benefit from these new products. Several new products are under development to enhance the feed efficiency and growth of meat-producing animals, and to increase milk yields in lactating animals. Increased feed efficiency could significantly decrease the cost of producing livestock. New growth promotants result in meat that is far leaner than that which is produced naturally, a benefit to consumers who desire less fat in their diets. Three new products (bST, pST, and betaagonists) currently are undergoing FDA review for use in livestock production. Additionally, traditional growth promotants, such as steroids and antimicrobial agents, continue to be improved. New reproductive technologies offer producers the opportunity to rapidly upgrade herd quality by selecting and incorporating desired traits at a faster rate than could be accomplished with traditional breeding. It is now possible to induce superior females to shed large numbers of eggs, and then to fertilize those eggs in vitro with the sperm of superior males. The embryos may be implanted into surrogate mothers whose estrus cycle has been synchronized to accept the embryo. Cloned embryos are currently marketed, and more efficient methods of embryo production are being developed. Advances in embryo and sperm sexing will allow livestock producers to choose the sex of the progeny and to breed for animals of highest value (e. g., females in dairy, males in beef production). Eventually, transgenic livestock will be commercially available. Efforts are under way to produce transgenic livestock with improved production characteristics such as enhanced disease resistance, leaner carcasses. and faster growth. However, the first transgenic livestock will most likely be animals that produce high-value pharmaceuticals in their milk. Several firms have successfully produced such transgenic animals; however, commercialization is not likely to occur before the end of this decade. New vaccines, therapeutics, and diagnostic kits will improve the ability of livestock producers to manage herd health. Several vaccines and diagnostic kits are commercially available, and more are under development. The food processing industry will also be affected. New enzymes and starter cultures for cheese and dairy manufacturing, and meat processing are being produced with biotechnology. One genetically modified enzyme preparation, chymosin. has been approved as generally regarded as safe (GRAS) by FDA for use in cheese making. Biotechnology can be used to improve the safety of food products through the development of nucleic acid probes and monoclinal antibodies to detect the presence of microorganisms, chemicals, heavy metals, and other contaminants in food products. Additionally, new methods to manage processing waste products. such as whey, are under development.

PAGE 99

94. A New Technological Era for American Agriculture Despite the potential opportunities offered by biotechnology, these technologies are not without controversy. Concerns have been raised about the effects of these technologies on farm survival and structure, food safety, animal welfare, and the environment. Additionally, many of these technologies will place a premium on farm management skills, and thus may not be appropriate for all farmers. These issues are discussed in more detail in the following chapters. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. CHAPTER 3 REFERENCES Bachrach, H.L. Genetic Engineering in Animal Agriculture, commissioned background paper prepared for the Office of Technology Assessment, 1991. Baile, C.A. et al., Effect of Somatotropin Treatment in Sows During Late Gestation on Birthweight and Performance of Pigs, Journal of Animal Science, vol. 67( Supp. 2), 1989, p. 67. Bauman, D. E., Bovine Somatotropin: Review of an Emerging Animal Technology, commissioned background paper prepared for the Office of Technology Assessment, 1991. Beermann, D.H. et al., Abomasal Casein Infusion and Exogenous Somatotropin Enhance Nitrogen Utilization by Growing Lambs, Journal of Nutrition, vol. 121, 1991, pp. 2020. Brinster, R.L. et al., Introns Increase Transcriptional Efficiency in Transgenic Mice, Proceedings of the National Academy of Science, U. S. A., vol. 85, 1988, pp. 836-840. Caperna, T.J. et al., Growth Response and Hormone Profiles of Growth Treated Pigs Fed Varying Levels of Dietary Protein, Journal of Animal Science. vol. 67( Supp. 1), 1989, p. 210. Choi, T. et al., A Generic Intron Increases Gene Expression in Transgenic Mice, Molecular Cell Biology, vol. 11, 1991, pp. 3070-3074. Crenshaw, T.D. et al., Exogenous Procine Prolactin and Somatotropin Injections Did Not Alter Sow Lactation Performance, Journal of Animal Science, vol. 67( Supp. 2), 1989, p. 258. Cromwell, G.L. et al., Recombinant Porcine Somatotropin for Lactating Sows, Journal of Animal Science, vol. 67( Supp. 1), 1989, p. 257. Cromwell, G.L. et al., Weekly Administration of Recombinant Porcine Somatotropin to Lactating Sows, Journal of Animal Science, vol. 67( Supp. 1), 1989, p. 258. Cromwell, G.L. et al., Recombinant Procine Somatotropin for Sows During Late Gestation and Throughout Lactation, Journal of Animal Science, 1991. Cromwell, G.L. and Dawson, K. A., Antibiotic Growth Promotants, commissioned background 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. paper prepared for the Office of Technology Assessment, 1991. Denman, J. et al., Transgenic Expression of a Variant of Human Tissue-Type Plasminogen Activator in Goat Milk: Purification and Characterization of the Recombinant Enzyme, Biotechnology, vol. 9, September 1991, pp. 839-843. Ebert, K.M. et al., Transgenic Production of a Variant of Human Tissue-Type Plasminogen Activator in Goat Milk: Generation of Transgenic Goats and Analysis of Expression, Biotechnology, vol. 9, September 1991, pp. 835-838. Enright, W. J., Effects of Administration of Somatotropin on Growth, Feed Efficiency, and Carcass Composition of Ruminants: A Review, Use of Somatotropin in Livestock Production, K. Sejrsen, M. Vestergaard, and A. Neimann-Sorensen (eds. ), Elsevier Applied Science, London, 1989, pp. 132 156. Etherton, T. D., Porcine Somatotropin: Review of an Emerging Technology, commissioned background paper prepared for the Office of Technology Assessment, 1991. Food Chemical News (Washington, DC: Jun. 3, 1991). Hansel, W., Reproduction and Embryo Transfer, commissioned background paper prepared for the Office of Technology Assessment, 1991. Harlander, S., Biotechnology in Food Processing in the 1990s, commissioned background paper prepared for the Office of Technology Assessment, 1991. Hocquette, J.F. et al., The Human Liver Growth Hormone Receptor, Endocrinology, vol. 125, 1989, pp. 2167 -2174. Houseknecht, K.L. et al., Effect of Abomasal Casein Infusion on Nitrogen Retention of Growing Steers Treated with Exogenous Bovine Somatotropin (bST), Journal of Animal Science, vol. 68( Supp. 1), 1990, p. 272. Johnson, L. A., Flook, J. P., and Hawk, H. W., Sex Pre-Selection in Rabbits: Live Births from X and Y Sperm Separated by DNA and Cell Sorting, Biol. Reprod., vol. 41, 1989, pp. 199-203. Juskevich, J.C. and Guyer, C. G., Bovine Growth Hormone: Human Food Safety Evaluation, Science, vol. 249, pp. 875, 1990. Kopchick, J. J., Transgenic Poultry, commissioned background paper prepared for the Office of Technology Assessment, 1991. Krimpenfort, P. et al., Generation of Transgenic Dairy Cattle Using In Vitro Embryo Production, Biotechnology, vol. 9, September 1991, pp. 844847. Lacy, E. et al., A Foreign B-Globin Gene in Transgenic Mice: Integration at Abnormal Chromosome Positions and Expression, Cell, vol. 34, 1983, pp. 343-358.

PAGE 100

Chapter 3Emerging Animal Technologies l 95 27. Lavitrano, M. et al., Sperm Cells and Vectors for Introducing Foreign DNA into Eggs: Genetic Transformation of Mice, Cell, vol. 57, 1989, pp. 717723. 28. Lo, D. et al., Expression of Mouse IgA by Transgenic Mice, Pigs, and Sheep, European Journal of Immunology, vol. 21, 1991, pp. 1001 1006. 29. Massey, J. M., Animal Production Industry in the Year 2000, Genetic Engineering of Animals, J. Reprod. Fert. Suppl. 41, W. Hansel and B.J. Weif (eds. ), 1990. 30. Moffat, A., Transgenic Animals May Be Down on the Pharm, Science, vol. 254, Oct. 4, 1991. 31. Nancarrow, C.D. et al., Expression and Physiology of Performance Regulating Genes in Transgenic Sheep, Journal of Reprod. Fertil, vol. 43( Supp. ), 1991, pp. 277 32. Osburn, B. I., Animal Health Technologies, commissioned background paper prepared for the Office of Technology Assessment, 1991. 33. Palmiter, R.D. et al., Heterologous Introns Can Enhance Expression of Trangenic Mice, Proceedings of the National Academy of Science, USA, VOI. 88, 1991, pp. 478-482. 34. Powers, D.A. and Chen, T. T., Transgenic Fish, commissioned background paper prepared for the Office of Technology Assessment, 1991. 35. Preston, R. L., Steroid-Like Anabolic Growth Promotants, commissioned background paper prepared for the Office of Technology Assessment, 1991. 36. Pursel, V. G., Prospects for Genetic Engineering of Swine, commissioned background paper prepared for the Office of Technology Assessment, 1991. 37. Rexroad, C. E., Transgenic Ruminants, commissioned background paper prepared for the Office of Technology Assessment, 1991. 38. Rexroad, C.E. Jr., Production of Sheep Transgenic for Growth Hormone Genes, Transgenic Animals, N. First and F.P. Haseltine (eds. ), Butterworth-Heinemann, Boston, MA, 1991, p. 280. 39. Rexroad, C.E. Jr. et al., Transferrinand AlbuminDirected Expression of Growth-Related Peptides in Transgenic Sheep, Journal of Animal Science, vol. 69, 1991, pp. 2995-3004. 40. Scanes, C. G., Poultry Somatotropin, commissioned background paper prepared for the Office of Technology Assessment, 1991. 41. Seidel, G. E., Sexing Mammalian Sperm and Embryos, Proceedings of the 1lth International Congress of Animal Reproduction and Artificial Insemination, vol. 5, 1988, pp. 136. 42. Smith, V.G. et al., Pig Weaning Weight and Changes in Hematology and Blood Chemistry of Sows Injected With Recombinant Porcine Somatotropin During Lactation, Journal of Animal Science, vol. 69, 1991, p. 3501. 43. Spence, C.A. et al., Effects of Exogenous Porcine Growth Hormone on Metabolic and Endogenous Patterns in Sows During Late Gestation and Lactation, Journal of Animal Science, vol 59( Supp. 1), 1984, p. 254. 44. Spence, C.A. et al., Effect of Exogenous Growth Hormone on Fetal Energy Storage and Lactation Performance in Sows, Journal Animal Science, vol. 59(Supp. 1), 1984, p. 246. 45. U.S. Congress, Office of Technology Assessment, Commercial Biotechnology?: An International Analysis, OTA-BA-218 (Springfield, VA: National Technical Information Service, 1984). 46. U.S. Congress, Office of Technology Assessment, New Developments in Biotechnology>: Patenting Life Special Report, OTA-BA-370 (Washington, DC: U.S. Government Printing Office, April 1989). 47. U.S. Congress, Office of Technology Assessment, U.S. Dairy Industry at a Crossroad: Biotechnology and Policy Choices-Special Report, OTA-F-470 (Washington, DC: U.S. Government Printing Office, May 1991). 48. Veenhuizen, E.L. and Anderson, D. B., An Assessment of the Effects of Beta-Adrenergic Agonists on the Food Animal Industry, commissioned background paper prepared for the Office of Technology Assessment, 1991. 49. Wallis, M., The Molecular Evolution of Pituitary Hormones, Biol. Rev., vol. 50, 1975, pp. 35. 50. Walsh, M.E. and Sundquist, W. B., A Case Study of Genecol-99: Possible Implication For Other Single-Use Agricultural Biotechnology Products, North Central Journal Agricultural Economics, vol. 10, No. 1, January, 1988, pp. 25. 51. Weidle, U. H., Lenz, H., and Brem, G., Genes Encoding a Mouse Monoclinal Antibody are Expressed in Transgenic Mice, Rabbits, and Pigs, Gene, vol. 98. 1991, pp. 185-191. 52. Wright, G. et al., High Level Expression of Active Human Alpha-1-Antitrypsin in the Milk of Transgenic Sheep, Bio/Technology, vol. 9, September 1991, pp. 830-834.

PAGE 101

Chapter 4 Advanced Computer Technology Photo credit: U.S. Department of Agriculture, Agricultural Research Service

PAGE 102

Contents Page INTRODUCTION . . . . . . . . . . . . . . . 99 SPECIFIC COMPUTER TECHNOLOGIES . . . . . . . . . 99 Knowledge-Based Systems . . . . . . . . . . . . . 99 Interfacing Technologies . . . . . . . . . . . . . . 109 Other Computer Technologies . . . . . . . . . . . . 115 SUMMARY/PROGNOSIS . . . . . . . . . . . . . 123 The Current State . . . . . . . . . . . . . . . 123 Mid-1990s . . . . . . . . . . . . . . . . . 123 2000 . . . . . . . . . . . . . . . . . . 124 CHAPTER PREFERENCES . . . . . . . . . . . . . 124 Boxes Box Page 4-A. An Example Rule for an Expert System . . . . . . . . . 101 4-B. An Example of an Explanation provided by an Expert System . . . . 101 4-C. An Example Recommendation from FLAYER . . . . . . . . 104 4-D. An Example Query and Answer to a Natural-Language Interface . . . . 112 Figures Figure Page 4-1. Trends in Semiconductor RAM Density . . . . . . . . . 100 4-2. Trends in Microprocessor and Mainframe CPU Performance Growth . . 100 4-3. Effect of Quality of Management on Milk Response of Dairy Cows Receiving bat . . . . . . . . . . . . . . . 100 44. Structure of an Expert System . . . . . . . . . . . 100 4-5. System Structure for an Object-Oriented Simulation System . . . . 106 4-6. Functional Components of the Crop Production Expert Advisor System . . 115 4-7. Topology of BITNET Commections in the United States . . . . . 116 4-8. States with Participants in DAIRY-L . . . . . . . . . . 117 4-9. Volume of DAIRY-L Messages . . . . . . . . . . . 117 4-10. Volume of DAIRY-L Requests for Remote Retrieval of Text Files and Software . . . . . . . . . . . . . . . 118 Tables Table Page 4-1. A Partial Catalog of Research Applications of Robots in Agriculture.. . . 119 4-2. A Partial Catalog of Research Applications of Sensors in Agriculture . . . 121

PAGE 103

Chapter 4 Advanced Computer Technologies INTRODUCTION Since the industrial revolution, agricultural systems have intensified, and agricultural productivity has significantly increased along with farm size. Labor-saving devices on farms have increased output per worker several-fold. and advances in understanding and application of biological principles have significantly boosted agricultural yields. With greater production per acre and animal, however, farm management becomes correspondingly more challenging and complex. In general, methods for making management decisions have failed to meet this challenge. As a result, many decisions are uninformed and many agricultural systems poorly managed. The application of advanced computer technologies to agricultural management can help remedy this situation. improved access to information will allow farmers to more effectively monitor progress toward optimal performance. Computer technologies of potential use to agricultural managers are advancing at a tremendous rate. The performance of computers has increased several-fold with each new generation of computer chip (figures 4-1 and 4-2). In the last decade. microcomputers have evolved from 64-kilobyte machines with a 320-kilobyte floppy drive to machines with several megabytes of memory and several hundred megabytes of permanent storage; such machines approach the performance of mainframe computers (25 54) and can store massive amounts of information. Advances are also occurring in software technologies, allowing improved utilization of stored information. Decision support systems, for example, provide enterprisespecific, expert recommendations to decisionmakers. Several other types of information technologies allow for rapid access to the latest information. These advances will provide the tools to improve farm management. For example. close monitoring of animal performance will allow early detection of diseases and can help reduce stress in animals. Overall, advanced computer technologies can provide managers with the ability to systematically determine the best decision rather than arrive at decisions in an ad hoc fashion. Optimal decisionmaking requires a holistic view of a farm enterprise, factors that affect it, and the probable consequences of management decisions. Thus, a farmer deciding whether to plant a specific crop on a specific field should weigh the profitability of the crop as well as overall farm needs (i. e., nutrition requirements if it is an animal enterprise). The decision will impact land sustainability and the need to use certain pesticides and herbicides or other pest-control methodologies. Computer technologies, by providing the capability of taking these multiple factors into account, can help producers arrive at the best possible decisions and management strategies. The quality of management, in turn, will influence productivity as well as the future impact of some biotechnologies. For example, the response of milk cows to bST is directly related to management. Poorly managed dairy herds have a lower response to bST than wellmanaged herds (figure 4-3). SPECIFIC COMPUTER TECHNOLOGIES Computer technology is changing at an unprecedented rate on three different fronts, causing a "three-dimensional" information revolution. Rapid advancements in traditional database and computational programs: in symbolic computing and artificial intelligence: and in systems that improve access to information constitute the three dimensions of the information revolution. Knowledge-Based Systems Traditional database and computational programs, which are largely numeric and follow established algorithms. are invaluable resources, but they cannot easily deal with symbolic data or mimic an experts reasoning process. The so-called knowledge-based systems in the category of symbolic computing and artificial intelligence have these capabilities. American agriculture is just now beginning to capitalize on these resources. Essentially, knowledge-based systems present expert knowledge in a form that can be used to solve problems. In addition to expert knowledge, such systems require situation-specific databases. For systems that operate in real-time, sensors may play an important part in collecting data for knowledge-based systems (40). General uses 1. 2. -99of knowledge-based systems include: recommending solutions for problems (e. g., diagnosis), monitoring the status of a system to determine significant deviations (i. e.. management-by-exception). and

PAGE 104

. 100 l A New Technological Era for American Agriculture Figure 4-lTrends in Semiconductor RAM Density lo+@ I 1985 1975 1985 1995 2065 Year SOURCE: J L Hennessy and N P Jouppi, Computer Technology and Architecture: An Evolving Interaction, IEEE Computer September:18. 1991 Figure 4-3Effect of Quality of Management on Milk Response of Dairy Cows Receiving bST Excellent management Good management Average management I Somatotropin dose SOURCE: D.E. Bauman, Bovine Somatotropin: The Cornell Experience. Proceedings of the National Invitational Workshop on Bovine Somatotropin, USDA Extension Service, Washington, DC, pp. 46. Figure 4-2Trends in Microprocessor and Mainframe CPU Performance Growth Figure 4-4Structure of an Expert System 1985 1970 1975 1980 1985 1990 Year SOURCE: J L. Hennessy and N.P. Jouppi, Computer Technology and Architecture: An Evolving Interaction, IEEE Computer September:18, 1991 3. forecasting the behavior of a system (i. e.. simulation). Expert Systems Expert systems are the most popular knowledge-based technology in agriculture. The main benefit of expert systems is that they emphasize knowledge acquisition. not programming. User lnterface Interface engine Knowledge base control strategy domain knowledge rule facts How to do it what to do SOURCE: Office of Technology Assessment, 1992 Expert systems are distinguished by a unique structure that separates What to do from How to do it (figure 4-4). The knowledge base tells the program what to do. It contains the expertise for solving the problem without the control structure found in traditional programs. The second component of an expert system is an inference engine that, in effect, shows the program how to do the task at hand. The inference engine contains the control strategy that determines how to combine domain knowledge to solve the problem. Domain knowledge can be represented in the knowledge base in several different forms, the most common

PAGE 105

Chapter 4Advanced Computer Technologies .101 of which is rules (e. g., If the leaves are brown, then apply insecticide X see box 4-A). Rules correspond closely to the natural reasoning of experts, are modular, and are easy to maintain. As a result. expert systems are easy to develop and to support. The knowledge in an expert system tends to be symbolic instead of numeric. This feature allows rules to be heuristic in nature, akin to "rules-of-thumb. When exact algorithms do not exist, the rules represent the experts best guess (94). Another interesting feature of expert systems is their capability of incorporating uncertainty into rules. For example, the rule "If the leaves are brown, then apply insecticide X; 0.3 means that there is a 30-percent certainty or confidence in the conclusion. Strategies have been developed for combining the uncertainty of rules to give a confidence value for each recommendation (7, 76). Therefore, the expert system is able to make recommendations even when the circumstances of the problem are uncertain. This ability mimics the reasoning of an expert. Expert systems have the added capability of explaining the reasoning used to derive a solution (see Box 4-AAn Example Rule for an Expert System IF l you are willing to speculate for higher prices AND l the price trend is up AN D l the basis trend is weakening AND l the basis trend is not expected to reverse soon AND l the timing is harvest AND l the type of available storage is farm OR l the type of available storage is commercial AND l you need downside price insurance AND l Storage revenues are greater than storage costs, THEN l forward contract your grain and buy call options. SOURCE: R.H. Thieme et al., Expert System Techniques Applied to Grain Marketing Analysis, Computers and Electronics in Agriculture 1:299, 1987. box 4-B), much as an expert might. The explanation is a map of the rules chained together by the inference engine ( 102). Because expert systems separate the inference engine and knowledge base, it is easy to remove the knowledge from the expert system. leaving a shell that can be reused in other applications. The shell contains the inference engine, user interface. and other domain-independent modules. The first expert system shell was EMYCIN, which resulted when the knowledge base was removed from MYCIN, an expert system that diagnosed human blood diseases (89). Expert system shells have become saleable products, and several are commercially available for use in agriculture ( 14). There are numerous examples of expert system applications in agriculture. These systems have tended to be diagnostic systems for addressing relatively narrow problems. Large-scale, broad-based expert systems have not been developed in agriculture. The following overview of agricultural expert systems includes systems developed for business decisions. animal production, and crop production. Farm and Area-Wide Managementeffective decisions regarding the planning, organization, and control of a farm enterprise are essential to agriculture. The legislative, economic, and environmental demands placed on farmers and government agencies that implement agricultural policy create a need for tools that help make sound farm-level and regional policy decisions in agriculture. Tools that help with agricultural problems at a watershed or farm will become increasingly important in the future. This will involve integration of expert systems with geographic information systems, area-wide monitoring systems (78 ), and remote sensing. The financial difficulties of the 1980s vividly document the cost of poor decisionmaking in the business sector. However, a major obstacle impedes the adoption of expert systems in making business decisions. Business decisions. unlike production decisions, are generally inBox 4-BAn Example of an Explanation Provided by an Expert System I conclude that the patient has dreaded lurgy caused by bug-eyed germs with a certainty factor of .76. WHY: I concluded that the patients dreaded lurgy was caused by bug-eyed germs because tight abdomen (E3) and acid saliva (E4) indicate swelling of the spleen (E), which taken together with yellowish skin cast (D) provide evidence (CF = .76) that bug-eyed germs are the cause of the dreaded lurgy. SOURCE: M. Van Horn, Understanding Expert Systems, Bantam Books. New York, NY, 1986.

PAGE 106

102 l A New Technological Era for American Agriculture fluenced by values, goals, and risk attitudes. Thus, two experts with the same knowledge and expertise may select different courses of actions (91). Only a small number of expert system applications is available for farm decisionmaking. Most existing expert systems in this area relate to design, planning, and control. Unfortunately, such functions are considered relatively unimportant by farm managers. Expert systems dealing with data acquisition and interpretation, prediction, and monitoring have not been developed. This may indicate that expert system development efforts are focusing on applications not in the area of greatest need for farm managers (91), Farm-level planning and financial analysis are active areas of expert system development. Several prototype systems are under development. One effort at farm-level planning directed at farmers needs is the Crop Rotation Planning System (CROPS) developed at Virginia Tech (6). This system uses a map-based interface to let farmers enter data about their land (soil type, topography, landuse, and field sizes) and their farming enterprise. Based on these data, CROPS provides farm-level or field-level environmental risk evaluations for soil erosion, and nutrient and pesticide leaching and runoff. It then uses Al planning and scheduling techniques to generate a wholefarm production plan so that the overall farming operation can meet user-defined yield and/or acreage targets, economic return goals, while also reducing potential environmental risks to acceptable levels. The system runs on Apple Macintosh 11 systems and is adapted for use by the Soil Conservation Service and the Virginia Department of Conservation and Recreation in their farm planning activities. The best known farm financial system is the Agricultural Financial Analysis Expert System (AFAES) from Texas A&M University (63). AFAES consists of a spreadsheet to prepare operating-year and multiyear financial statements; a program that calculates financial ratios and trends from the spreadsheet; and two expert systems that develop a performance operating-year analysis and multiyear analysis, respectively. This expert system operates on an IBM-compatible microcomputer and is marketed through the Texas Agricultural Experiment Station at a variety of prices based on the type of user making the purchase. Other agricultural expert systems have been developed for specific business decisions. One example is the Grain Market program developed at Purdue University (98). This system provides advice for marketing storable commodities (e. g., crops). An example rule from this expert system is shown in Box 4-A. The machinery selection process is aided by the Farm-level intelligent Decision Support system (FINDS) (49). This system integrates a linear program (REPFARM), a database management system, and an expert system. The expert system is used to form the link between the linear program and the user and to interpret the output of the linear model. The linear program component operates on a minicomputer, but the other components operate on a microcomputer. A decision support system for planning of land use and forage supply for a dairy farm has been developed in Denmark (34). The main components of the system are a knowledge base, a linear programming model, and a PASCAL program connecting the knowledge base, model, and interface. The model integrates the varied business activities of a dairy farm, such as crop production, storing feeds, milk production, and utilization of manure. Interactions between feeding and production of milk and meat are established by use of knowledge sets. The user interface allows for consequent analysis and can function as a tool for calculation and optimization planning. In addition to agriculture-specific expert systems for business decisions, nonfarm business systems will impact agriculture (91 ). For example, Dologite (24) developed the Strategic Planning Advisor to provide strategic planning advice. This system provides recommendations such as: l l l l l Get out of a business. Hold current position. Focus on one market niche. Invest selectively. Invest aggressively. Animal ProductionExpert systems for animal production deal with the management of farm animals and generally focus on disease diagnosis and suboptimum performance identification based on technical expertise. Most expert system activity in the area of animal production focuses on the dairy industry. There are at least two reasons for this. First, the dairy industry has a national data recording system (i. e., Dairy Herd Improvement, DHI), that provides centralized databases from which expert systems can be built (99). A second reason is that dairy animals are generally housed in confinement, and they produce a product (i.e., milk) that can be routinely monitored on an individual animal basis. This is conducive to intensive management. Spahr et al., (92) outlined several potential applications of expert systems for dairy herd management.

PAGE 107

Chapter 4Advanced Computer Technologies l 103 Some of the earliest dairy expert systems were developed by Extension Specialists at the University of Minnesota. Their first system (DMGTSCOR) ranks dairyherd management strengths and best opportunities for improvement using DHI management measures ( 16). Management action is suggested for the three best opportunities for improvement. A second system, SCCXPERT, was developed to diagnose herd mastitis problems using DHI somatic cell data and to recommend corrective actions. Another system, BLKTNKCL, provides interpretation and information about bulk tank culture data for primary mastitis causing organisms. A fourth system, MLKSYS, provides expertise to troubleshoot operational and design problems with a milking system ( 15). Two other systems have recently been developed to assist in manure management and to provide an overall analysis of the production and financial status of a dairy farm. All of these systems were developed in the Level 5 expert system shell; as a result an effort is underway to integrate them into a single system to allow data sharing among the programs. These expert systems are distributed by the Dairy Extension office at the University of Minnesota freely to extension personnel and commercially for $75 ( 17). Tomaszewski and others at Texas A&M University have developed a Dairy Herd Lactation Expert System (DHLES) to analyzes DHI milk production data and to provide recommendations for improving milk production ( 106). DHLES contains a separate module (LacCurv) to graphically display lactation curves. This system was developed in PROLOG and operates on an IBMcompatible computer. It is marketed through Texas Dairy Herd Improvement Association for $99 ( 100). Several expert system projects are under development for the dairy industry. Kalter and coworkers (45) are developing a comprehensive expert system (Dairy Pert) to evaluate dairy-herd management. The impetus behind this effort is the possible future adoption of bovine somatotropin (bST), but the system has general applicability. This system currently contains over 320 rules in the Nexpert expert system shell, a spreadsheet-based nutrition model, and entry and advice routines based on Foxs database management software. DairyPert does not utilize DHI data because of inconsistencies among the nine national Dairy Record Processing Centers. DairyPert is funded by and will be distributed to the private sector through a large pharmaceutical company. Cornell University will distribute the system to public agencies and institutions. Oltenacu et al. (73) are developing a reproduction expert system that will analyze DHI reproductive records and determine weaknesses in the reproductive program. This system utilizes LISP on an IBM workstation. Allore and Jones (42) are developing an expert system to evaluate DHI somatic cell counts that will identify areas of management that predispose cows to mastitis. This system is being developed in CLIPS and will operate on an IBM-compatible microcomputer. Oltjen et al. (74) have developed a prototype expert system that recommends whether to keep or cull commercial beef cows. The rules contain knowledge relating to the cows age, body condition score, calving difficulty, structural correctness, health, and previous reproductive performance. The expert system was integrated with a simulation model to calculate net present value for each animal. This expert system was developed in the CALEX expert system shell. An expert system to assist in the management of a sheep enterprise has been developed in Scotland ( 104). This system was developed without the aid of an expert system shell. Once a working prototype that could be delivered to an agricultural unit was developed, this project was halted as a research project. Expert systems for the management of sheep flocks are also under development in Australia. CHESS is a Dutch decision-support system designed to analyze individual swine breeding herds within an economic framework (22). It determines strengths and weaknesses in the management of a pig enterprise. CHESS consists of a decision-support system and three expert systems. The decision-support system identifies and assesses the importance of relevant deviations between performance and standards. The expert systems combine and evaluate deviations to identify management strengths and weaknesses. XLAYER (84) is a management expert system for the poultry industry and is one of the most comprehensive expert systems in animal production. XLAYER is designed to diagnose and estimate economic and associated losses as well as recommend remedial management actions for over 80 individual production management problems significantly affecting a flocks profitability. An example output is shown in box 4-C. This system contains over 400 production rules and was developed in the M1 expert system shell. Crop ProductionAll commercial crop production systems are potential candidates for expert system applications. In particular, expert systems should be considered for integrated crop management decisions that would encompass irrigation, nutrition, fertilization, weed

PAGE 108

104 l A New Technological Era for American Agriculture Box 4-CAn Example Recommendation From XLAYER You are experiencing an economic loss of about $725 per week because of a sudden change in the grain portion of your layer ration. Reformulate the ration and phase in new grains gradually, even if the cost per pound is higher. Production losses amounting to some $500 per week are being experienced because temperature in your layer house is exceeding 29.4 degrees Celsius. Use artificial cooling systems in regions where hot weather is expected to continue. If layer barn has no cooling system, construct a partial budget to evaluate alternative pooling systems such as evaporative cooling pads, roof sprinklers, high pressure misting and other forms of cooling, Water intake is very low. Check watering systems to make sure that birds are getting adequate fresh, clean water. Equipment repair costs are running $100 per week higher than normal. Check management practices related to the routine servicing of mechanical equipment. If repair and maintenance costs are consistently high, construct a partial budget to evaluate the replacement of old or poor functioning equipment. SOURCE: E. Schmisseur and J. Pankratz, XLAYER: An Expert System for Layer Management, Poultry Science 88:1047, 1989. control-cultivation, herbicide application, insect control, and insecticide and/or nematicide application (64). The first expert systems developed in agriculture were PLANT/ds (65), a program developed at the University of Illinois that identified diseases of soybeans in Illinois. and POMME(81 ), developed at Virginia Tech to identify diseases of apple orchards. Both were written by computer scientists who were using agriculture as a novel domain. Michalski, for example, was primarily interested in machine learning. Of the major crops, cotton has received the most attention to date, with at least three expert systems and one simulation-based management model now available to the public (94). COMAX (COtton MAnagement eXpert), the expert system component of GOSSYM/ COMAX was developed by the U.S. Department of Agriculture, Agricultural Research Service (USDA/ARS) in Mississippi (56). 1 Users of this system purchase a weather station linked to a personal computer running the program. The GOSSYM component is a simulation model of cotton production that uses weather data collected from the weather station. The COMAX expert system uses the model to project when to irrigate and fertilize to achieve optimal agronomic goals. The entire GOSSYM/COMAX system including the weather station and computer costs several thousand dollars. Despite the high price tag, it is used by as many as 500 cotton farms in 15 States. COTFLEX is an integrated expert system and database package developed at Texas A&M and released to the public through the Cooperative Extension Service (93). Photo credit: U.S. Department of Agriculture, Agricultural Research Service, Farmer and consultant examine data from COMAX (Cotton MAnagement eXpert) computer program. The overall system will eventually include a whole-farm economic analysis module that lets farmers evaluate whether or not to participate in Federal farm programs or to purchase Federal crop insurance. The component released to the public, however, is devoted to insect pest management of the three major insect pests of cotton in Texas. CALEX/Cotton is another integrated cotton expert system and database management tool (79). CALEX was developed as an expert system shell, and cotton was the 1 GOSSYM is a hybrid term formed by combining (1).w]piwn. the scientific name for cotton and the word simul;ition.

PAGE 109

Chapter 4Advanced Computer Technologies .105 Photo credit U.S. Depatment of Agriculture, Agricultural Research Service. Farmer and engineer check automated weather station that feeds daily weather information into the COMAX system to update its prediction for cotton yield and harvest dates. first application area. The system was supported through Californias statewide integrated pest management program and delivered to farmers for testing and use. It is one of the best-documented attempts at delivering expert systems to farmers for use in crop production (31). Because the program was developed with State support, no revenue has been collected from its users and the project continues to depend on State support. Pennsylvania State University supports a laboratory devoted to the development of expert systems and their delivery through the Cooperative Extension Service. The University has developed several expert systems using an expert system shell (PENN-Shell) developed in-house. One of these expert systems, GRAPES, recommends pest control options for insects and diseases in vineyards (83). Penn States expert systems all run on Apple Macintosh computers, and the University supports a statewide computer network for these machines. USDA-ARS researchers (28) developed a knowledgebased system for management of insect pests in stored wheat. The system determines whether insects will become a problem and helps select the most appropriate prophylactic or remedial actions. Simulation models of all five major insect pests in wheat have been developed; the models output feeds the expert system. Evans and coworkers (26) at the University of Manitoba have developed an expert system to serve as a Fertilization Selection Adviser. The current system considers only one type of crop (wheat), four different moisture regimes (arid, dry, moist, and irrigation), one soil nutrient (nitrogen), and four different fertilizer compounds (urea, ammonium nitrate, urea ammonium nitrate, and ammonia). It provides return on investment information; a risk analysis module is under development. This system was developed in the LISP programming language for the Macintosh: however, work has already begun to develop a similar system using the C programming language cm an IBM-compatible microcomputer. In general, one can find expert system applications for crop production for virtually all the major crops in this country and in many countries around the world. Insect pest management, weed control, and disease identification are the most common domains. Other systems that have received wide recognition in crop systems include: l l l l l EasyMacs, an expert system and database program developed at Cornell University for recommending pest management strategies for apple production; SOYBUG, an expert system developed in Florida that helps farmers with insect pest control in soybeans (2); SIRATAC, an expert system and simulation model developed in Australia for helping cotton farmers with pest management decisions that has since been marketed internationally (36); TOM, an expert system for diagnosing tomato diseases developed in France (5); and WHAM, a wheat modeling expert system developed at the University of Melbourne, Australia (3). Research Needs Development of commercial expert system shells is being driven by forces outside agriculture

PAGE 110

106 l A New Technological Era for American Agriculture and is proceeding at a relatively rapid rate. However, agriculture applications generally will require expert system shells to operate in a microcomputer environment whereas industrial applications often reside on workstations or minicomputers. Since this is a domain-independent problem, it may be best addressed by computer scientists outside of the agricultural sector. The main limitation to development of expert systems is adoption of computer technology. To promote this area will require more trained personnel and incentives to develop and deliver computer systems. Object-Oriented Simulation Systems In addition to expert systems, another type of knowledge-based system that is useful for planning is object oriented simulation. Traditional simulation systems model the behavior of a system by explicitly simulating individual processes. The structure of the real system usually is implicit in the model. Object-oriented simulation models have an inverse structure; they explicitly model the Figure 4-5System Structure for an Field (a parent object) structure of the real system, and the behavior of the system is implicit in that structure. Each component of the real-world system is represented in the simulation as an object. Objects are units that consist of self-descriptive data and procedures for manipulating that data. Objects can be represented in a hierarchy such that they inherit properties from more general categories (i.e., their parents). For example, an object-oriented simulation of a farm (figure 4-5) would contain a general FIELD parent object that describes the general features of all fields (e.g., a method to calculate the area of the field). Individual fields (e.g., field 23) would be represented as unit objects that inherit the properties of the parent FIELD object and may also contain some information specific to themselves (e. g., current crop planted in the field). Objects in object-oriented systems communicate by exchanging messages. For example, if field 23 is to be harvested, a HARVEST message is sent to the field 23 object. The field 23 object handles the details (internally resetting its own values) and returns the amount of crop harvested. This return message can Object-Oriented Simulation System Silo (a parent object) PROCEDURE area length x width/43,560 PROCEDURE harvest yield x area reset yield to zero message: area / I / I I response: 98.7 acres PROCEDURE harvest volume = voiume + harvest SOURCE: Office of Technology Assessment, 1992.

PAGE 111

Chapter 4Advanced Computer Technologies .107 be sent to a particular silo object which knows how to add the crop to its inventory. Once the object-oriented system is developed, the simulation sends messages to appropriate objects in a fashion similar to farm managers giving orders to their employees. There are two main advantages to this type of simulation. First, the model closely corresponds in structure to a real system. This facilitates maintaining and expanding the model. Second. procedures in an object can be represented in a symbolic fashion similar to expert systems. Thus, objectoriented simulation can be used to model processes that may not be quantitatively well defined. Object-oriented simulations have been under development since the early 1980s. An early object-oriented simulation language (ROSS) was developed in the LISP programming language by the Rand Corp. for the Air Force (62). This language has been used in military applications. Two early examples are SWIRL, an objectoriented air battle simulator, (47) and TWIRL, an objectoriented simulation for modeling ground combat between two opposing tactical forces (48). Object-oriented simulations are powerful tools for modeling the behavior of biological systems that are otherwise difficult to describe mathematically. Output from these systems can be used in planning and to determine impacts of changing management procedures. However, most existing object-oriented simulation models cannot easily be transferred to production agriculture. Several object-oriented simulation projects have been developed specifically for agriculture. Researchers at Texas A&M University developed an object-oriented model to simulate animal/habitat interactions (82). The simulation was specifically used to study the damage caused by moose migrating through forest plantations. This system was developed on a Symbolics workstation using LISP. Another agricultural simulation was developed by USDAARS to model insect disease dynamics in a rangeland ecosystem (9). This model is primarily a research tool for studying the relationship between grasshoppers and their pathogens to assist in integrated pest management programs. This system was also developed on a Symbolics workstation using the FLAVORS objectoriented programming language. Another LISP-based system was the host-parasite model developed by Makela et al. (58) to study the interaction between the tobacco budworm and one of its parasites in cotton fields. More recently, Crosby and Clapham ( 18) used the Smalltalk language to simulate nitrogen dynamics in plants; Stone (95) used an object-oriented model of a mite predator-prey system to show that chaotic dynamics rather than stable or predictable cycles, might be the norm in agricultural systems; and Sequeira et al. (87) developed an object-oriented cotton plant model for use in studying the interaction between localized pest feeding and cotton lint yield and quality. Another object-oriented simulation project is under development by Chang and Jones at Cornell University for use in agriculture ( 10). This project uses a LISP-based, object-oriented programming language (Bobject, Kessler, University of Utah) to model the operation of a milking parlor. When completed, this model will be useful to dairyfarm managers and their consultants for parlor configurations and for identifying changes in performance when changes in parlor operation are made. Research NeedsThe general paradigm of objectoriented programming is being incorporated into several traditional programming languages (e. g., C, PASCAL), but few inexpensive commercial shells exist in which to develop object-oriented simulations. Smalltalk is a good example. It is a language and a development environment in one, and it generally comes complete with many predefine object classes developed specifically for simulation. Other expert system shells like KEE, Goldworks, NExpert-Object, and Level-V Object include the objectoriented paradigm and can be used for simulation. LISP offers many advantages for prototype systems such as the parlor project. However, LISP is not a language in which final products should be delivered, since it requires too much memory and is too slow for agricultural applications. More research is needed to determine the potential value of object-oriented simulation for agriculture. Knowledge-Acquisition Knowledge-based systems are powerful computer tools because they contain and apply a significant amount of expert knowledge to problem-solving; however, this also constrains systems development. Knowledge acquisition is a slow and tedious process, and problem-solving rules and procedures are often hard to articulate. Artificial intelligence can help automate one type of knowledge acquisition (21, 66), that of rule formation. Machine learning, for example, is an artificial intelligence technique for automatically generating rules from a set of examples. This is sometimes called learning from examples. It can be used to assist experts to develop rules or fill in where experts do not exist. For instance, rules for a crop disease diagnostic expert system can be generated using a machine learning system with a database of plant descriptions and associated diseases.

PAGE 112

108 l A New Technological Era for American Agriculture Michalski and others (65) compared rules derived by experts and those generated by a machine learning algorithm (AQ11 ) when developing an expert system for soybean disease diagnosis (PLANT/ds). The database consisted of 630 examples based on 35 plant and environmental descriptors for 15 soybean diseases. One rule was generated for each disease. When tested in an expert system, the machine-generated rules outperformed those generated by experts. The machine rules properly diagnosed 98 percent of the test cases while the expert derived rules diagnosed 72 percent correctly. A microcomputer-based machine learning system has been developed for agricultural problems (27). This system was first used to generate rules for a grass identification system (WEEDER). Other generic machinelearning algorithms are available as commercial products (e. g., Classification and Regression Trees, California Statistical Software, Inc, Lafayette, CA; ID3, Knowledge Garden, Naussau, NY). Due to the nature of rules generated from machine learning (i.e., the rules indicate which variables are important for describing certain results), machine learning can also be used as a data analysis tool. Liepins et al. (57) investigated the use of three machine learning algorithms for analyzing natural resource data. They studied the effect of storm damage on lake acidification using a data set generated after a major storm stuck the Adirondack Park in upstate New York. Application of machine learning to these data provided no new information but reinforced many of the discoveries made using traditional statistics. Dill (23) also used a machine-learning algorithm to analyze the sale price of cattle sold at public auction. The data set contained all information available to a buyer on sale day and the price for which the animal was sold. Using machine learning. Dill was able to determine which variables influence the buyers decision and now will be able to generate an automated appraisal system from these results. Research NeedsThere are several problems associated with machine learning. One concerns data that contain random errors (i. e., noisy data). Some machine-learning algorithms are unable to handle this type of data while others perform poorly (57). Much of the data in agriculture is noisy. Another problem is that many of the machine-learning algorithms require discrete data (e.g., classification-based) while agricultural data is mainly continuous (e. g.. numeric). A third problem is that machine learning requires a complete database with associated outcomes from which to operate. Few of these databases exist in agriculture. Despite these limitations, machine learning can be a very valuable knowledge acquisition tool in certain situations. With continued development, these limitations will likely be overcome. Knowledge-Based Report Generation One of the initial goals in artificial intelligence was to develop systems capable of translating documents from one computer language to another ( I I ). An integral component of machine translation is developing a knowledge representation of the original document such that text can be generated in another language. Though machine translation will not have a major impact on American agriculture, systems that are able to generate knowledgebased reports from a database will. Farmers receive large volumes of production data with little or no interpretation; hence, they may be unable to convert these data into useful information. Knowledgebased report generation is an emerging technology that can provide them with interpretive reports to better support management decisions. In many respects, programs for knowledge-based report generation are similar to expert systems. Report generation programs contain four components: 1. a domain-independent knowledge base of linguistic and grammar rules, 2. a domain database from which the report is to be generated, 3. a domain knowledge base for interpreting the data structure, and 4. the text planning component for deciding what to say and how to say it (69). Once a system is complete, the domain knowledge can be removed to create a shell that can be used in another domain. Report generation is still largely in the research stages and commercial shells have not been made available. CoGenTex, Inc. has developed a proprietary linguistic shell for knowledge-based report generation. This shell has been used to generate weather forecasts in both English and French for the Canadian Government. A USDA Small Business Innovation Research proposal has been submitted to study the suitability of this approach for generating knowledge-based reports that interpret DHIA records for dairy farmers (46). Research Needs To date, there have been no applications of knowledge-based report generation in agriculture. Research should be directed at investigating the potential benefit of this technology to American agriculture. Once the preliminary investigations are com-

PAGE 113

Chapter 4Advanced Computer Tehnologies l 109 pleted, a better understanding of needs and benefits will be established. Interfacing Technologies Farmers have been slow to adopt personal computers. Recent surveys indicate that only 15 to 27 percent of farm managers utilize computers in management (1, 55). Two factors that may have contributed to this slow adoption rate are the lack of high quality management software (71 ) and a computer phobia on the part of some farm managers. Farm managers have available to them only a limited selection of computer programs. most of which perform similar functions. The computer phobia is caused by a lack of exposure to computers but is exacerbated by the type of user interfaces (both hardware and software) employed by most agricultural computer programs. Hardware Issues Currently, most microcomputer systems use a keyboard as the major input device. Keyboard entry is clumsy for agricultural software as many farm managers are slow typists. Even for programs that require little input. a hunt and peck typing ability can frustrate the user to the point of not using the system. Another problem with keyboard entry is impaired dexterity from excessive physical labor or injury that severely impairs the farm managers ability to type. Consequently, software should be developed allowing the use of alternative input devices. Two relatively common input devices are the mouse and the light pen. However, neither of these capture the users natural pointing instincts (77 ). A more intuitive input device is the touch-sensitive screen. Another alternative input device is speech. Touch-sensitive screens are computer displays in which portions of the display may be used as an input device. This technology has been available since the mid-1960s (41 ). Touch-sensitive screens are easy to learn, very durable, and require no additional work space. At the same time they have the disadvantage of increased cost, increased development complexity, lack of software to take advantage of touch-sensitive screens, arm fatigue. and screen smudging. A major complaint of touch-sensitive screen users is the lack of precision; however. highprecision screens have recently been developed (86). Due to their durability and user-friendliness, touch-sensitive screens have been used in specialized applications such as kiosk information systems in shopping malls and airports and for order processing in restaurants. Both of these applications have been developed to allow control of a computer systems by nontechnical users. A second area of research aimed at improving the physical link between the computer and user is speech recognition. This research has been glamorized by science fiction movies such as 200 l: A Space Odyssey, in which computers carry on a dialogue with the user. Though this is the goal of research efforts, it is not the current state-of-the-art (52). A prominent researcher has predicted that totally spontaneous, unrestricted speech recognition is still as much as 30 years from fruition ( 105). However, speech recognition appears to be suitable for applications with restricted discourses. Agriculture is one such application. Speech recognition is based on the ability to distinguish between words and on natural-language processing whereby natural language input is transformed into a form that the computer can utilize. In a common method for speech recognition, template matching, each spoken word is matched against a predetermined lexicon. The lexicon must be trained to recognize a users voice. thereby resulting in a user-specific system (52). High-performance, speaker-independent, continuous-speech recognition systems use another approach. that of statistical modeling. Commercial speech recognition systems range from speaker-dependent, single-word recognition (64-word vocabulary units) to speaker-independent, continuousword recognition (40.000-word vocabulary units ) (75). Speech recognition is not a perfect function. Most literature values for recognition accuracy range from 95 to 99 percent (97); some articles report 8 to 12 percent error rates (61 ). Several factors affect the error rate; these include presence of background noise. phonetic similarity of words, and mood of the user as he/she alters voice quality (52). Furthermore, lack of a one-to-one correspondence of sounds to words distinguishes speech from other inputs. For instance, when a key is pressed on the keyboard, the output is unambiguous. With speech recognition, the output is the most likely output which corresponds to the input. Consequently, the performance of current systems degrades (in both time and accuracy) as the vocabulary increases. When speech input was compared to traditional input methods, it was found to require the same amount of time as mouse input, 80 percent as much time as a single key stroke and 48 percent as much time as full-word typed commands (61). A commercial speech recognition system recently was added to a medical diagnostic system for clinical data entry (88). The system was an isolated-word, speakerdependent system capable of recognizing eight continuous syllables. Utterances required a half a second to

PAGE 114

110 l A New Technological Era for American Agriculture take effect and 90 percent of all utterances were recognized correctly. For this application, speech recognition proved an effective interface for improving the acceptance of the diagnostic system. Advances in hardware input devices to improve the usability of computers are being driven by multiple nonagricultural sources. For example, speech recognition is a goal of the Department of Defense ( 105) and of research aimed at providing more environmental control to the physically disabled (20). Since this technology is domain independent, advances in other domains should also greatly facilitate the use of speech recognition in agriculture during the next decade. Software Issues The software design of the user interface is the main factor determining the effort required both to learn and to use a computer program. The most important function of the user interface is to match the needs of the user. Novice users need interfaces that are easy to learn while advanced users prefer interfaces that are easy to use. Most easy to learn systems are not convenient to use. Thus, no one interface will meet the needs of all computer users (33). In general, agricultural software has not been distinguished by sophisticated user interface designs. This partly reflects the fact that most agricultural software is written by people who understand agriculture. yet have little or no training in user interface design. Currently. there are nearly a dozen different interface designs that can be used with computer programs. These range from command languages to natural language. Two common user interface designs in agriculture are command and question/answer systems. A commanddriven user interface is similar to the DOS system where a series of commands and arguments have to be known by the user. For example. in the Cornell Remote Management System, which is used to ascess DHI data. a command such as AIM 1-S1-DH1MO094 is used to run a report. This type of user interstice is easy for an expert to use, but because it is not intuitive. it is difficult to learn. Another type of command-driven user interface can be designed by mapping commands to special keys. This interface is used by WordPerfect ( WordPerfect Corp.. Orem. UT) which uses multiple combinations of the SHIFT. ALT. and CTRL keys with function keys forspecific commands. Question/answer systems require the user to enter a response. If the type of response is unambiguous, this design can be easy to use but also tedious. This type of user interface should be limited to responses which are Yes-No (e. g.. Y/N) or numeric. A type of computer interface that is more intuitive to use than command and question/answer systems is natural language. With this type of interface, commands are given in normal spoken or written language instead of a formal command language. An example of a natural language user interface is one that converts natural-language commands to DOS commands. For example, the natural-language command show me the files on drive b: is converted to the command dir b:*.* (53). Another example of a natural-language interface is one that was developed for signal processing (68). This system allows users who are knowledgeable about signal processing but ignorant of any programming languages to manipulate wave forms using English commands oriented toward mathematical operations. However. the most common use of natural-language interface has been in database querying systems. Natural language is attractive to the casual user and to the user who is unwilling to learn a formal command structure. However, natural-language user interfaces require more typing than command-language interfaces. As discussed previously, typing requirements are an important consideration for agricultural software. Therefore, natural-language is probably not a desirable user interface for systems that can be driven with a limited set of commands (e. g.. DOS). Another popular user interface design is the menu system. In the simplest form. a menu is a list of choices. The user selects one choice by entering a number or letter. Another version includes a light bar that can be positioned over the menu using the keyboard. A more sophisticated menu design, known as the graphic user interface (GUI), is the icon and mouse system. This type of system represents menu selections using a picture that is "clickedon with a mouse. The icon system was first developed for the Xerox Star workstation (90) to reduce the learning time of the user interface. The user is expected immediately to know which icon is appropriate. Thus. the icon must be unambiguous and realistic. Distinguishable and meaningful icons may be difficult to develop for several similar items (96). Accompanying text is often added to clarify the meaning of possibly ambiguous icons. Another major factor of the user interface is data entry. For this factor. interfaces called form-filling designs have been developed. The user is presented with a series of fields in which data are entered. The display relates to a written form and allows the user to see all of the fields together. Often. form-filling interfaces have data

PAGE 115

Chapter 4Advanced Computer Technologies l 111 validation and editing capabilities. For more complex data entry needs, multiple forms arranged as overlaid windows can be used. As data are entered into a field, it actuates the next form which displays with the appropriate related fields. This type of user interface is rapid, easy to use, and easy to learn (96). Design of sophisticated user interfaces has advanced to a point where they should now be considered for all agricultural software. Proper attention to user interface design issues can result in agricultural software that is more acceptable to use. For example, adaptive interfaces are aimed at satisfying the differing needs of both novice users and experienced users. An adaptive user interface determines the skills of the user and changes the interface to meet those skills. In general, novice users are provided with menus and question-answer systems, while advanced users are given the option to use command languages and special key strokes. A prototype adaptive interface has recently been developed (SAUCI); (101) for processing UNIX commands. Using the adaptive interface, users made about half as many errors and required less time to perform tasks. Research in adaptive interfaces should result in systems that are more intuitive to use and easier to learn. Information Retrieval Systems Information retrieval systems are a set of advanced computer technologies for accessing stored information. These technologies differ from decision support systems in that they offer no recommendations. Three technologies are emerging that may have a role in American agriculture in the next decade. These are natural-language interfaces, full-text retrieval systems, and hypertext systems. Natural-Language Inter-acesMaintaining a complete set of production records is a critical component of farm management. More important is the ability to rapidly and flexibly access information for management decisions. The best method of accessing production records has been through database management systems; however, these systems generally have inflexible retrieval facilities based on menus that present options of data to retrieve or predefine reports to run. Traditional systems require the user to learn the hierarchical structure of the menu system and limit the type of reports available. A natural-language interface for querying a database can offer a more flexible retrieval system (43). The current generation of natural-language interfaces was made possible by a set of linguistic theories developed by Chomsky ( 12). These theories were first implemented in an efficient algorithm in a natural-language interface for retrieving information about lunar rock brought back from the Apollo space missions (LUNAR) (107). LUNAR is based on a three-compartment model of data retrieval. The first compartment is syntax analysis, which determines the grammatical structure of the sentence. The second compartment of LUNAR is the semantic module, which is responsible for determining the meaning of the syntactic structures. The meaning is translated to a formal query language in this module. The third module of LUNAR is the retrieval component. This module executes the formal query language, based on the semantic analysis, to retrieve data from the appropriate database. When LUNAR was tested, it answered 78 percent of the questions presented to it ( 107). The purpose of developing LUNAR was to assist scientists in retrieving data on lunar rocks. Its users were primarily interested in specific data as that data related to other scientific information that had been collected. However, this style of data retrieval is not appropriate for production agriculture where management decisions need to be made. A natural-language interface for retrieval of data for decisionmaking should put the data in the proper context so that an informed decision can be made. Consequently, a knowledge-based, natural-language interface was developed to formulate more complete, intelligent answers to users questions from an agricultural database (IDEA) (44). IDEA is based on the LUNAR three-compartment model but utilizes a new approach for semantic representation. Unlike the formal query language used in LUNAR, IDEA represents the query through a set of domain concepts, which contain expert information. IDEA has the capability of responding to a query and offering additional pertinent information. An example of a query and answer is shown in box 4-D. IDEA was developed for a dairy database to assist farm managers in decisionmaking. It is capable of responding to several different types of queries. The simplest query is about a single cow (e.g., When is 5000 due to calve? or, simply, Is 5000 pregnant?). More complicated questions can be asked about subgroups of cows (e.g., Which daughters of Thor are bred to Bell?): averages (e. g., What is the average calving interval for cows in the north barn?); and counts (e. g., How many heifers are due to calve in June?). Replies are designed to contain important information that the user may not have known was in the database or may not have even asked for.

PAGE 116

112 l A New Technological Era for American Agriculture Just as generic, domain-independent shells have given expert systems widespread use; for natural-language interfaces to be successfully used in agriculture, a generic natural-language shell capable of being transported to other databases is needed. However, unlike expert systems, development of a generic shell for natural-language interfaces has proven difficult. Hendrix and Walter (37) point out that full synchronization is needed between the database management system and the natural-language interface. This is difficult to achieve when the naturallanguage interface is added as an afterthought. For example, in the dairy database that IDEA accesses, the reproductive status of a cow is given as a numeric value (e. g., O means not bred ). To access multiple databases, the natural-language interface must be able to translate all their representations. TEAM is a natural-language interface developed at SRI and designed to be transportable between databases (32). However, a database expert still is needed to adapt the system to each new database. The expert supplies information describing the database and domain-specific words, a process that TEAM has automated through a series of questions. Another source of difficulty in making a natural-language interface transportable is associating meaning to phrases. For example, consider the phrases bred to Bell and bred in May. The first refers to the service sire while the second refers to the breeding date. Most natural-language interfaces cannot handle these kinds of subtleties. Thus, for a natural-language interface to be successfully transported to a new database, a database expert and a linguistic expert are needed ( 19). Because of the problems in developing generic shells, natural language commercially lags behind many other artificial intelligence technologies (70). One of the few generic natural-language interfaces available commercially is the Intelligent Assistant interface for Q&A (Symantic Corp., Cupertino. CA), which was introduced in 1985. This system differs from most transportable natural-language interfaces in that it has its own database system. Users build their application directly in Q&A. This system also uses synonyms for acquiring new words. Box 4-DAn Example Query and Answer to a Natural-Language Interface >> which cows are due to calve next week? 4897 was bred to STARMAN on 12/15/1987 and is due to calve on 09/18/1988 Projected calving interval: 347 days >> when did 5281 calve? 5281 calved on 05/26/1988 with a heifer calf #5535 The calf was in good condition Gestation length = 278 days 5281 had a retained placenta >> is 5239 pregnant? YES 5239 was bred to TOPBRASS on 03/20/1988 and is due to calve on 12/19/1988 Projected calving interval: 466 days >> is 5449 pregnant? NO 5449 calved on 12/1 1/1987 with a heifer calf #5478 The calf was in good condition Gestation length = 283 days 5449 is 282 days in milk 5449 was bred to LEVI on 02/21/1988 5449 was pregnancy checked on 03/30/1988 and was open SOURCE: L.R. Jones and S.L. Spahr, IDEA: Intelligent Data Retrieval in English for Agriculture, A/ Applications in Natural Resource Management 5(1)56, 1991.

PAGE 117

Chapter 4Advanced Computer Technologies l 113 An attractive feature of this system for agriculture is that it operates on standard IBM-compatible microcomputers. Another commercial natural-language interface is Natural Language (Natural Language, Inc., Berkeley, CA). This system interfaces with any database that supports Structured Query Language (i.e.. SQL). Full-Text Retrieval SystemsA relatively new area of human-computer interfaces that holds great promise in making information more accessible is full-text retrieval. The goal of a full-text retrieval system is to search a collection of documents to find relevant information for the user (4). These systems can be particularly useful for accessing a collection of documents that are authored by several different people who potentially use different words to express the same thing. Such a collection of documents, including most Agricultural Extension publications, is unedited and generally not indexed. Blair and Maron (4) evaluated the effectiveness of STAIRS (STorage And Information Retrieval System). a full-text retrieval system developed by IBM. They found it to retrieve less than 20 percent of documents relevant to a particular search when the database contained roughly 350,000 pages of text. They identified several pitfalls that need to be considered in developing full-text retrieval systems. STAIRS was efficient at retrieving documents that exactly matched the wording of the request, but it performed poorly in retrieving documents that contained misspelled words, and words that were synonymous with those in the request. For example, the word gauge was spelled 9 guage in an original document, preventing its retrieval. Full-text retrieval systems must be able to account for such situations and retrieve relevant documents whose text may not match the exact wording of the request. A simple key-word search or an indexing scheme thus does not meet the needs for full-text retrieval. A full-text retrieval system developed by Gauch and Smith (30) contains an expert system and a thesaurus. The thesaurus contains domain-specific information for words, a list of synonyms for each word. its parent word(s), and a list of children words. This structure allows a particular search to be generalized or narrowed. Decisions as to the search pattern are made by the expert system. If the recall is low, it will broaden the search. If the precision is low (i.e., too many irrelevant passages are retrieved) the expert system will use a more specific search. The query is formed by the user and then passed to a full-text retrieval system that has immediate access to any passage in the text. The retrieval system requires that the text undergo two stages of preprocessing. In the first stage, the text is formatted for enhanced display. Formatting includes insertion of format marks (line. tab. italics, line, label ) and context information (section. paragraph, sentence. item). In the second stage of preprocessing. the file is converted to fixed-length records totfast access. Consequently. the system does not operate on the original documents. This is an undesirable feature as it precludes searching subsets of documents and requires additional storage. A full-text retrieval system now commercially available (Metamorph; Thunderstone, Chesterland. OH) should have wide application in agriculture. Metamorph operates on standard ASCII files using natural-language queries to search and find relevant passages in documents. The natural-language input undergoes morphological analysis to normalize each word. The normalization process converts words to morphemesthe smallest meaningful unit of a word. A set of morphemes that are related to, but not necessarily synonymous with. the original morpheme is generated. Metamorph then correlates these equivalence sets to textual passages to determine passages that relate to the natural-language query. At the first level of search, an equivalence must be present in the passage for its retrieval. If this is unsuccessful, Metamorph will broaden the search. Another important feature of the correlation procedure is that it utilizes an approximate match to account for minor discrepancies in spelling. These features fulfill the conditions Blair and Maron (4) identified as necessary for a full-text retrieval system. Numerous applications of full-text retrieval are possible. A recent project used a commercial full-text retrieval system to assist users in querying a specific DHI computer manual (29). Additionally, with the advent of mass storage systems for microcomputers (e. g., CDROM), full-text retrieval systems can play a significant role in providing expert information (e. g., extension bulletins) to county extension offices and directly to farm managers. An effort is underway to develop a national dairy database (39) consisting of full-text documents covering major dairy-management areas. This full-text database is expected to be delivered on a CD-ROM and accessed using a full-text retrieval system. Hypertext-Hypertext is a method of connecting related passages of text, graphics, animation, or computer programs in a multidimensional {i. e., hypercube) fashion such that they can be accessed in a nonlinear fashion. Each node can be connected to any number of other nodes that provide additional related information. Hypertext systems are analogous to footnotes or references in a

PAGE 118

114. A New Technological Era for American Agriculture document. For example, a footnote contains additional information related to the text. The reader determines when or if the footnote is to be read. Computerized hypertext systems are based on the same principle. Hypertext systems are relatively easy to implement but are difficult to build. They require the locations of the related text to be stored with the location of the original text. This is essentially a database management problem. The difficult part of a hypertext system is to establish the appropriate links between and among nodes. This usually requires a domain expert, but the process can be automated through full-text retrieval tools. As Extension documents begin to be disseminated in electronic form, hypertext should be considered as a method of increasing access to related subject matter. For example, an extension bulletin that describes the use of lactation curves for herd management should be linked to other bulletins describing the use of butterfat and protein curves. To demonstrate the benefits of hypertext in an agricultural setting, Rauscher and Johnson (80) delivered the six feature papers contained in an issue of Al Applications: Natural Resources, Agriculture, and Environmental Sciences in hypertext form. Integrated Systems Management of an agricultural enterprise requires a variety of decisions and, hence, a variety of decisionsupport tools. Long-range research in the area of humancomputer interface will be directed at integrating various decisionsupport programs into a single system. Current research is aimed at integrating autonomous systems, developing intelligent user-interface managers, and integrating systems through a common representation shared by an intelligent dialogue manager. An overall controlling software system that allows the user to access different decision-support tools yet maintains operational independence of tools themselves represents the lowest level of systems integration. The genera operating system of a computer is an example in that it allows the user to access multiple programs in the same environment. More advanced integrated systems assist the user in choosing the decision-support tool and provide logical links between tools. This type of integration can also be used to develop multimedia applications such as full-color, full-screen graphics; full-color, full-screen video; aural delivery of speech or music; and animation (50). An example of an advanced multimedia system for integrating several different decision-support tools is the Whole Earth Decision Support System (WEDS; reference (51. The WEDS project combines textual databases, expert systems, simulation models, traditional programs and laser-video images within the agricultural domain into a single integrated system. Each module is developed independently and inserted into WEDS. For example, an expert system for lactation curve analysis developed independently from WEDS can be incorporated and linked with other components dealing with lactation curves (e. g., documents in the textual database). In this system, the user moves between the different modules guided by logical connections. Systems such as WEDS should be able to provide a complete information resource to extension agents, agri-service personnel, and farm managers for solving problems and formulating management decisions. The multimedia approach utilized in the WEDS project should be encouraged for systems developed in the 1990s since people remember more if they combine seeing, hearing, and doing during the learning process (60). A more tightly coupled method of integrating software is to link different systems through a user-interface manager. The user-interface manager controls all user-interface functions for a set of application software (96) and validates all inputs for the application software. Screen displays, including error messages and on-line help, are also controlled by the user-interface manager. There are two major advantages to integrating software in this fashion. First, a system does not need to be redeveloped for each piece of application software. Second, the user is always presented with a consistent interface; thus, as the user moves from one application to another, the user interface remains the same. This is important for acceptability of software by laymen. Development of a generic user-interface manager awaits further research; however, several fourth-generation languages include facilities that can assist in development of generic user interfaces (%). A more advanced method of integrating software is through an intelligent user interface; such an interface allows problems to be formulated and appropriate application software selected using natural language. A prototype system for integrating crop production decisionsupport systems is under development (see figure 4-6); (59). It uses an intelligent dialogue manager (IDM) with unrestricted natural-language communication to develop a problem description. The IDM parses input into a semantic representation using knowledge of the types of queries that can be asked and the lexical entities that can be discussed. The IDM also utilizes a model for inferring the goal of the users input and relating it to the context of the overall dialogue. The semantic representation is passed from the IDM to an expertise module dispatcher

PAGE 119

Chapter 4Advanced Computer Technologies l 115 (EMD), which selects the application to respond to the query and formulates the appropriate control structure for the application software. The EMD is an expert system with knowledge of the problem-solving abilities of each application software module. This system can provide the user with a variety of problem-solving tools. Furthermore, the user does not need to know the nature of the software, the details for using it, or the situations for which it is appropriate. Other Computer Technologies Three other emerging computer-oriented technologies will impact American agriculture in the 1990s. The first involves dispersal of information to those who need it in different geographic localities. The second. robotics, will impact the labor problems associated with agriculture. The third area is sensor technology. Networks and Telecommunications American agriculture is decentralized and widely distributed, making information dissemination problematic. However, electronics can be used to provide mass distribution of information. Electronic information can be transmitted essential y at the speed of light and duplicated at minimal cost. Two electronic forms of information delivery will dominate in the 1990s: a satellite-based system and a wide-area computer network. Satellite transmission of data has become a commonday occurrence for telephone and other communications. A geosynchronous satellite receives a transmission from Earth and rebroadcasts that message back to Earth over a wide area. Different frequencies are used to send multiple simultaneous messages. Two common modes of transmission are the Ku and C bands. Interest in delivering agricultural information via satellite is growing. Several distance-learning programs have been developed at the University of Utah for delivery in Ecuador ( 13). Their developers are also preparing an undergraduate animal breeding and genetics class to be delivered over the national AG*SAT satellite instructional network, which routinely carries Extension programs. An Extension series of interactive dairy programs has been developed and delivered by the University of Washington (8) as well as by the University of Wisconsin (35). The American Farm Bureau also maintains a satellite link to 46 States and 573 of their county offices (72). This satellite link is used to transmit data as well as instructional programs. Satellites not only make possible mass distribution of information, they do so in a way that makes this information easily accessible to end users. They only need a satellite reception disk and a television. However, development of satellite-based instruction programs can be expensive. Poor planning may also reduce attendance. Other problems include limited audience interaction and low motivation on the part of the end user to view the program. The importance of in-person interactions with the live speaker should not be underestimated. However, Figure 4-6Functional Components of the Crop Production Expert Advisor System Deep reasoning User IDM I State Simulation representation Management Expert system modules problem situations : domain-specific The problem solving system component SOURCE: L.R. Maran, CPEAS: The Crop Production Expert Advisor System, Knowledge Based Systems Research Laboratory, Department of Agronomy, University of Illinois, Urbana-Champaign, 1989.

PAGE 120

116 l A New Technological Era for American Agriculture if funds for education continue to dwindle, this may remain the only feasible means to conduct an Extension program. Another method of rapidly delivering information is through a wide-area computer network. Much of the western world currently is criss-crossed with multiple computer networks. Two of the original computer networks are BITNET (figure 4-7) and ARPANET. BITNET was initiated at the City [University of New York and was used to connect major educational institutions. ARPANET was initiated by the Department of Defense. Today there are national computer networks for the government. commercial companies, and educational institutions. A number of regional networks have also been developed. These include networks such as Clemson University Forestry and Agricultural Network, CNET (Cornell University). and PENpages (Pennsylvania State University). Most of-these networks interface through the national Internet system so that messages can be sent from one network to another. Internet is funded by several government agencies and numerous companies (50). The main benefit of wide-area computer networks is the ability to rapidly share information and expertise. for instance, an industry situation report can be posted on the network and broadcast to all interested readers with access to the network. County Extension agents on the network can send and receive files in electronic format. In this way. interdisciplinary work can be conducted over long distances. Varner and Cady (103) have established a bulletin-board type system, called DAIRY-L, through which dairy professionals can request and receive information. DA IRY-L is only one of hundreds of bulletinboard systems, but a pioneer in the use of networking for Extension education. DAIRY-L, which resides on the University of Maryland mainframe computer. was initiated early in 1990. Since that time subscription has grown to 150 subscribers from 37 states and 20 foreign countries (figure 4-8). Message traffic also has increased, approaching an average of 15 messages per month (figure 4-9). Messages are submitted to a list server which in turn transmits them to all participants of DAIRY-L; therefore. all subFigure 4-7Topology of BITNET Connections in the United States SOURCE: J.R. Lambert, (Networks, Telecommunications and Multimedia Information Bases for Agricultural Decision Support, commissioned background paper prepared for the Office of Technology Assessment, Washington, DC, 1990.

PAGE 121

Chapter 4Advanced Computer Technologies .117 Figure 4-8States with Participants in DAIRY-L. New Hampshire I SOURCE Mark Varner, University of Maryland (M.A. Varner and R A Cady. Dairy Science 74(Supp 1): 201, 1991 Figure 4-9Volume of DAIRY-L Messages. 50 I SOURCE Mark Varner, University of Maryland (M A Varner and R A Cady. Dairy-L A New Concept in Technology Transfer for Extension, Journal of Dairy Science 74(Supp. 1 ) 201, 1991 Dairy-L A New Concept in Technology Transfer for Extension, Journal of DA IRY-L has proven extremely useful to extension specialists needing knowledge in areas outside their institutions expertise. Because all members see all messages, DA IRY-L is also a powerful ediucational tool. Information exchange through wide-area computer networks makes efficient use of personnel and resources. Therefore. a high priority should be given to maintaining and enhancing the backbone systems (i. e.. satellites and wide-area computer networks ) that provide rapid dissemination of information. Since these systems are national in scope, this initiative should occur at the Federal level with USDA-ES providing the leadership in agriculture. Robotics Robotics are machines that can be programmed to perform a variety of labor intensive tasks in agriculture. Since 1968, when strew Dutch companies proposed mechanisms similar to robotics for harvesting citrus, researchers have proceeded though the poposal stage and currently are testing Laboratory and field prototypes for fruit harvesting. transplanting, tissue culture propagation. and machine guidance (67) (table 4I ).

PAGE 122

118 l A New Technological Era for American Agriculture Figure 4-10Volume of DAIRY-L Requests for Remote Retrieval of Text Files and Software Number of requests Date posted File/software o 10 20 3/21/91 3/22/91 4/22/91 4/22/91 4/22/91 5/09/91 6/19/91 6/25/91 6/25/91 6/25/91 7/18/91 8/29/91 8/29/91 8/29/91 &29/91 Cow price spreadsheet Poison plant fact sheet TMR fact sheet #1 TMR fact sheet #2 TMR fact sheet #3 FTP instructions water coliform fact sheet Heat stress fact s. #1 Heat stress fact S. #2 Heat stress fact s. #3 Staff design fact sheet Dairy producer survey Antibiotic residue fact sheet Somatic cell count fact sheet Somatic cell count fact sheet SOURCE: Mark Varner, University of Maryland (M.A. Varner and R A Cady, Dairy-L A New Concept in Technology Transfer for Extension, Journal of Dairy Science 74(Supp. 1): 201, 1991. Most robotic applications under development are foreign-based. The United States is noticeably lacking in development efforts. Japan and Europe have much stronger programs and are likely to capitalize on this technology much sooner. Agricultural robotics research is proceeding in two directions. One involves sensor technology (see following section) and machine vision. This is because, unlike production line robots, agricultural robots will operate in environments where interferences will be encountered. For example, a fruit-harvesting robot must be able to locate irregularly shaped fruit despite the obscuring effect of leaves and stems. A second research concern is robot end-effecters (i.e., grippers). These are the mechanisms through which robots conduct their work. Again, unlike industrial operations, agricultural robots will generally be working with fragile products (e.g., bedding plants and fruit). Touch and force feedback are necessary to avoid bruising or damaging plants, fruits, or animal products. Three other areas of research are important for robot development but are not specific to agriculture. Manipulators are the physical linkages that move the end-effectors. Breakthroughs in speed and cost of manipulators are necessary. Agricultural robots will likely require less precision than industrial robots and will not require curvilinear motion, thus reducing the cost, Easily adopted robot components from nonagricultural applications would reduce the engineering costs of agricultural robots. A second research area is the development of computer algorithms for robot control. Significant advances in the miniaturization and integration of control hardware are needed. Integral feedback of the robots position is essential. More powerful integrated circuit chips to interface sensors and to control the manipulators are also needed. New artificial intelligence approaches to task selection will be important facets of robot control research. A final area of research, systems simulation, allows evaluation of alternative robot configurations through animated computer simulations. Advances in computer simulation would reduce the development cost and time required in engineering a robot. One major use of robots in agriculture will be for laborintensive tasks. For example, there are two Dutch companies developing robots to milk dairy cows; one prototype is operating at the University of Maryland. Laborsaving robots will enable American farmers to remain competitive in world markets despite higher labor costs and a shortage of part-time, seasonal labor. They will also help to stem the flow of young, struggling industries such as ornamental horticulture, bedding plants, and plant tissue cultures to countries with low-priced labor. If robotics can help these industries survive, they will create or maintain jobs which would otherwise be lost. Another major use of robots will be to micromanage crops. For example, a robot with an image sensor to detect weeds could be used to spot-spray herbicides. This

PAGE 123

Chapter 4Advanced Computer Technologies .119 Table 4-lA Partial Catalog of Research Applications of Robots in Agriculture Application Location Notes Fruit harvesting Apple harvesting France Able to harvest 500/. of fruit Citrus fruit harvesting University of Florida 1 fruit every 3 seconds, able to harvest a fruit on 750/0 of its attempts Tomato harvesting Kyoto University 20 seconds per fruit Cucumbers harvesting Japan In a laboratory study, the hand successfully completed the harvesting motion for 42 of 53 cucumbers. Muskmelon harvesting Purdue University 5 seconds per fruit Volcani Institute, Israel Plant material sensing and handling Transplanting l pepper plants Louisiana State University Transplanting rates as low as 1 plant l marigolds and tomatoes Purdue University every 3 seconds have been achieved l move plugs from one flat to another Rutgers University with a 95/0 success rate. Automated tissue propagation University of Georgia, Operations include retrieving the University of Florida, cuttings from a conveyor, trimming to University of Illinois, size, stripping selected petioles, New Zealand, Europe, applying rooting hormones, and Israel, Japan, Switzerland sticking the finished product into a plug flat cell. Mushroom harvester England Uses a vision system to locate and size mushrooms and guide a selective robot harvester. Forest thinning Performs automatically selective felling within the tree ranks, bunching the harvested trees and carrying them to a process zone. Animal Robot milkers Netherlands Sheep shearing Australia Egg handling University of California. Facilitated candle inspection, Davis Pork protein sensing Purdue University Robot moves an electro-magnetic scanner over a carcass. Pork carcass sectioning Sweden Oyster shucking University of Maryland Machine vision application to locate oyster hinges. Machine guidance Automated guided vehicles Michigan State University Based on machine vision sensing. Texas A&M University Plowing robot France Rice combine Japan Used edge-following to guide the machine around a rectangular field. Direct spot spraying Purdue University Machine vision application to recognize Corn detasseling plants. Purdue University Machine vision application to recognize plants SOURCE Office of Technology Assessment. 1992 would encourage farmers to adopt conservation tillage robots to perform their tasks. Reliable sensors coupled and post-emergence spray programs. with knowledge-basccl decision support systems will provide important managenment tools. Sensor Technology

PAGE 124

120. A New Technological Era for American Agriculture Photo credits: Norman Pruitt, Maryland Agricultural Experiment Stat/on. This research prototype automated milking system, developed in the Netherlands, allows scientists to study system automation and robotics that can benefit dairy farms.

PAGE 125

Chapter 4Advanced Computer Technologies l 121 or that require more vigorous sensing than we can provide. Sensor technology provides information the human senses cannot access. There generally are six classes of sensors. The newest is machine vision which processes images (e. g., camera input ) to detect patterns. Nuclear magnetic resonance (NMR) is a noninvasive technique of resonating highfrequency electromagnetic radiation in the presence of hydrogen nuclei. This technology is widely used for diagnosis in the medical field, but it is costly and difficult to apply in field situations. Neur-infrared (N/R) spectroscopy is another noninvasive technique that measures the reflectance of NIR radiation by a substance. Because organic compounds absorb and reflect NIR radiation differently this is a quantitative sensor. Acoustical measurements provide another class of sensors for measuring the density of substances. Biosensors are sensors that incorporate a biologically sensitive material (e. g.. immobilized enzyme). Electrical sensors can monitor the electrical properties (e. g., conductance) of a substance. Considerable work has been done in environmental sensing (i. e., crops, weather), somewhat less in animal sensing (i. e., estrus detection) (40). A partial list of research efforts in sensor technology is presented in table 4-2. Animal sensors are difficult to engineer due to biocompatability problems and animal welfare constraints. Photo credit: U.S. Department of Agriculture, Agricultural Research Service. Drawing of pig (left) shows where cross section was made by magnetic resonance (MR) imaging. Spine, loin muscles, and kidneys are visible in upper part of MR image (right). Scientists can measure fat development under the skin quickly without injury to the pig. Table 4-2A Partial Catalog of Research Applications of Sensors in Agriculture Application Type of sensor Electronic navigation system Used the Global Position Satellite System Automated plowing system Photodetectors sensed the furrow edge Tractor guidance Computer vision Monitor organic matter in soil Light and NIR reflectance Application of spray material Electronic surface grid Monitor gaseous ammonia NIR spectroscopy Moisture sensors for irrigation Electrical resistance Plant stress Infrared leaf temperature sensor Crop growth Spectral reflectance Weed identification Machine vision Identification of plant embryo shapes Machine vision Animal digestive system Radionuclide imaging Estrus detection Electrical conductivity Sex determination of baby chicks Machine vision SOURCE: Off Ice of Technology Assessment, 1992

PAGE 126

122 l A New Technological Era for American Agriculture Research on sensors for use in crop production generally focuses on the following objectives: l l l l l Improving operations in crop production by machine guidance systems. Applying pesticide and fertilizer chemicals. Improving the management of irrigation water to conserve the resource and reduce production costs. Developing methods of monitoring crop growth to incorporate with computer models for improving day-to-day crop management and strategic planning. Developing sensors for assessing crop maturity and fruit location as basis for mechanical harvesting. There remain numerous agricultural areas where sensors need to be developed (40). Doing so will require a multidisciplinary approach with input from professionals who understand the biology of the system in question as well as professionals who understand sensor technology (e. g., engineers and physical scientists). Some of the areas that need to be addressed include: l l l l Accurate three-dimensional fruit location sensor for crop canopies. This will facilitate robotic fruit harvesting. High-resolution navigation for field machines. Ability to program machine locations within inches, not several feet, is needed. A chemical drift sensor to monitor fertilizer and pesticide application and production of air polluting gases from animal units. Irrigation demand sensors that are not affected by soil properties and climatic factors. Photo credit: Gerald Isaacs, University of Florida An experimental fruit picking robot uses a machine vision sensor and a computer to locate individual fruit for detachment. Approximately 3 seconds per fruit are required. l l l Animal stress sensors that can remotely detect early animal health problems. A fruit-ripeness sensor that can determine optimum harvest times and detect early stages of fruit and vegetable deterioration. Microbial sensors that can detect early development of spoilage or bacterial contamination in fresh meats, including poultry and seafood. Photo credit: U.S. Department of Agriculture, Agricultural Research Service. Animal physiologists test a sensor that will detect when this cow is ready to give birth. An important component of the use of sensors in animal agriculture is telemetric data transfer and electronic identification of animals. For sensors that are to be implanted (e.g., tissue conductivity for estrus detection), telemetric data transfer must be accomplished within the size constraints which make implantation feasible. This remains a research issue. Implantable electronic identification systems have been developed and are currently under review by the Food and Drug Administration. Concern centers on the possibility that implantable sensors or identification units can enter the food chain. The development of sensors will facilitate more forms of automatic control over various aspects of agricultural production. The development of robots is closely tied to success in the area of sensor technologies. A broader implication of sensor technology may be to provide a data acquisition system and a database from which decision support systems can operate. This should result in tighter controls for management and higher profitability for the enterprise. Another important impact of sensor technology will be in the food safety arena. Sensors to detect food spoilage or contamination will greatly increase the safety of the American food supply.

PAGE 127

SUMMARY/PROGNOSIS Computer technologies change at such a rapid pace that it is difficult to foresee their application in the next decade accurately. Irrespective of agricultural policics. computer technologies will continue to advance to support the needs of other industries. Meanwhile. a number of impendiments exist that are likely to slow adoption of these technologies in agriculture. These impediments can be removed through changes in policy. Most projections of agricultural application of computer technologies have been overly optimistic. For example, Holts futuristic view of the application of computer technoolgies for farm management (38 ) is still 20 years from fruition. OTA has developed a scenario for the application of computer technologies in agriculture assuming that new technologies have a 5-year development phase. That is to say that once a research project begins it takes 5 years before that technology is applied. It was also assumed that incentives to bring new computer technologies out of the research laboratory and into production agriculture would exist. There are almost no incentives to do so today Thus. American agriculture will not be affected by these technologies in a major way for at least10 years. The Current State By and large. computers have had little impact on production agriculture to date. Predictions that every farmer would own a computer by 1990 have not come true. Fe W farmers have computers and those who do use them primarily for bookkeeping and routine calculations (e. g.. ration balancing ). Computers have had somewhat more impact on agriculture support industries. Using computer networks and tracking systems. equipment dealers are better able to provide faster service and feed dealers are better able to manage feed inventories. Most of these advances have come from directly adopting general business software with little or no input from the agricultural academic community. Another technology that currently is being adopted by farmers is fax machines. This allows for rapid exchange of printed material. An example of the use of this technology is in ration balancing. A nutritionist can receive the results of a feed analysis by fax from the laboratory. formulate a ration. and fax that to the farmer all with-in a few minutes. There is limited use of networks for exchange of information among Extension personnel ( i .e Dairy-L) and among protoype full-text databases (i. e., National Diary Database). Mid-1990s Within the next few years. many technologies currently under development should find their way into application. By the mid 1990s, the performance of microcomputers will likely double, eroding some of the current constraints to farmer adoption of computer tech nology. However. it still is unlikely that a high proportion of farmers will own a personal computer by that time. The primary application of advanced computer technology in the mid-1990s will be in the form of ad hoc expert systems to solve well-defined problems. These will be primarily problem diagnosis expert systems that are currently under development. Farmers will have a cadre of expert systems at their disposal to diagnose diseases and to evaluate animal and crop performance. These systems will generally not be integrated with each other and each will condisider one aspect of a problem. Integrated systems that solve producton problems while considering economic consequences will not become available until later in the decade. The primary use of expert systems within the next 5 years may be by agribusiness personnel, as they will be able to leverage the cost of adopting these technologies across more farms. Using expert systems to provide additional service to farmers may cause a shift in the role of some professionals. For example. expert systems help veterinarians take an epidemiological approach to solving problems (85 ). It will also cause some diversification in services provided. For example. nutritionists may be more likely to become involved in consulting for the crop program when armed with an expertsystem. Sensors will see limited application for collecting realtime data for expert systems. The primary use of sensors will be for monitoring weather and field conditions for crop management. Expert systems will help farmers to interpret these data and suggest appropriate management strategies such as irrigation, fertilization, or pesticide treatment. Another technology likely to see application within the next 5 years is full-text retrieval systems. It will be possible for farmers and Extension personnel to have a CDROM with all of the latest publications at their fingertips. Using a full-text retrieval system they will be able to retrieve pertinent information that will help them make better decisions. For example. when a farm experiences a corn mycotoxin problem, the manager can access an information base to find relevant literature. Large information bases, such as the national dairy database, will likey be developed and delivered by 1995.

PAGE 128

124 l A New Technological Era for American Agriculture Robots for highly specialized, labor-intensive tasks will begin to be applied to agriculture in the late 1990s. This would include robot transplanting of seedlings and pork carcass sectioning. Robots for milking cows could reach application by the mid1990s. 2000 The turn of the century should bring with it significant new applications of computer technologies in American agriculture. Ten years will provide sufficient time for the acceptance by farmers of computer technologies as a valid management tool and for the development of integrated management programs. It will also allow time for universities to become comfortable with these technologies and for personnel to be properly trained in developing these technologies. By 2000, whole-farm advisors, or integrated managemen t workstations, should be developed. A management workstation will consist of integrated decision support tools with a multimedia presentation of information. The workstation can thus serve as a diagnostic tool, an information source, an advisor, and a planning system. The expert systems will consider the holistic view of an enterprise when making recommendations. The systems will also share data so that information used in one system will be available to other systems. This generation of expert systems should operate as monitors that can alert producers to potential problems, as opposed to current expert systems which are situation-driven: that is, the producer must perceive a problem and decide to execute the system. The management workstation will also contain an advanced user interface consisting of speech recognition and touch-sensitive screens. The future dairy management workstation might contain decision support systems that monitor the financial records, the herd production records and the crop production records. Cropping decisions would be integrated with the dairy needs, the financial situation, and the land resources available. Currently, these decisions are all made independently. When the farmer is alerted to a problem (e. g., pest infestation). he or she can use the multimedia features of the workstation to retrieve video segments to learn how to identify the pest and the proper techniques for applying a pesticide. Robots for harvesting fruits and vegetables and for automatically guided vehicles should become available by 2000. Their application will depend on the cost associated with using human labor for the same job. 1, 2, 3 4 4. 5. 6. 7. 8. 9. 1 (). I 1. 12. 13. 14. 15. 16. CHAPTER 4 REFERENCES Batte, M. T., Jones. E.. and Schnitkey, G. D., b Computer Use by Ohio Commercial Farmers, American Agricultural E(onomic.v Asso(iation 72:935, 1990. Beck, H. W., Jones, P., and Jones, J. W., SOYBUG: An Expert System foe Soybean Insect Pest Management, Agrit,u[[[ir(ll S>,j[(,mif 30:269_~~6, 1989. Berger, R.. WHAM: A Wheat Modeling Expert System User Manual, Technical Report Department of Computer Science, University of Melbourne, Australia. 1987. Blair. D.C. and Maron, M. E., b An Evaluation of Retrieval Effectiveness for a Full-Text Docun~entRetrieval System. C(Jttlt~ll(tli(~ltiotl.~ (!/ Ihe ACI14 28(3):289. 1985. Blancard, D., Bonnet. A., and Coleno. A., TOM: Un Systeme Expert en Maladies des Tomatoes. P.H.M. RCIIIC Holti(olc. 261 :7 13. 1986. Buick, R.D. et al.. CROPS, A Crop Rotation Planning System for Management ot Whole-Farnl Resources Implementing Low-Input Sustainable Agriculture. Al Appli(wti~m.~ 6: 1992. Buxton, R., Modelling Uncertainty in Expert Systems. ltlt(~.}t(ltiot~[[l J()[{rnal l)t~(itl-~(1(.llitl(~ Studies 3 I:4 15, 1989. Cady. R.A. et al., The Role of Televideoconferences in Extension Programs. JoI/rnal (f L)t/ir\ Stienfc 74( Supp. 1 ):290. 1991. Carruthers, R.1.. Larkin, T. S.. and Soper. R. S., bSimulation of Insect Disease Dynamics: An Application of SERB to a Rangeland Ecosystem. Sim[(l[ltiotl 5 l: 101, 1988. Chang. W. and Jones. L. R., An Object-0 -iented Approach for Modeling Milking Parlor Performance, Journal ojDai\> Scien(c 74( Supp. 1 ):236. I 991. Charniak, E. and McDernlott. D.. Introduction to Artificial lntell igence (Reading. MA: Addison Wesley Publishing Co., 1985 ). Chomsky, N., Syntactic structures. Mouton, The Hague, 1957. Christensen, C. et al.. Preparing and 1nlplementing Distance Learning by Utah State University, Journal ~?jDait:} Siictlie 74( Supp. I ):289, 199 I Citrenbaum. R.. Geissmtm. J. R., and Shultz, R. Selecting u Shell, Al Appli(wtion.s in N[lt[iral Re.~our(c klanagctnen~ 2( 1 ):3. 1988. Conlin. B.J. et al.. Machine XPERT. Proceedings of the National Mastitis Council. 1989. Conlin. B .J., Stcuernagel. G.R. and Peters R.R. ,Design of an Expert System for rr(~ublc Shooting Dairy Hml Mtinugernent. J(wt-tra/ {J/ L)(/it-\ .Stietl(c 72( Supp. I ):461. 1989.

PAGE 129

17. 18. 19. 20. 21. 7? --23. 24. 25. 26. ?7. 28. 29. 30. 31. Conhn. personal conlnlunications. Sept. 17. 1991. Crosby. C.J. tind Cluphwn. W. M.. A Simultition Modeling Tool for Ni(rogcn D~mmlics Using Object-oriented Programming. Al App/ittl(i(~ll.\ 4(2):94100. 1990. Dtimerau. F. J.. Prospects for Knomlcdgc-Bused Custt~nlizati~~n ot Natural Language Query S~Istenls. 111/i~rttl{[ti{jtl Pro~c.~.ii\l(q & kl~llllltqelllcllt 24:65 1. 1988. Damper. R.].. Voice-input Aids For The Ph~sicully Disublcd. l~lt[~l}~[iti~~}l~~l JtMIrIItIl {J/ kfLlIIM{[t}linc .%tl{dic.v 21 :54 1. 1984. Dicttcrich, T. G.. b Mtichine Learn ing. AIII?IItI/ Rc\ic\\ ~J/ C(wIp[IttIr .$~iril~c 4:255, 1990. Diikhuizcn, A.A. and Huirrw. R. B. M.. CHESS An Integrated Decision Support Sjstcm T() Antilyr/c Individual Ct~\*-Herd Pcrt\mnancc. 1lll(.$tctlt.~ in t!,ql-icl~ltl{t-t-.$ll[tv.~sfi[l Pr[i~tital Appli~wiio\t.~. Ikutschc Landsw i rt hsc h afts-Ge se 11 sch~ fts ( cd. ). Fran k fu rt, Germany. 1990. p. 221. Dill. D. E.. An Electronic Model for Determining Value and Predicting Stile Price of Holstein Cattle Sold at Public Auction. Ph. D. Thesis. University of Illinois. Urbtina. 1990. Dologite. D. G.. Developing a Knowrledyc-Based System on u Personal Computer Using an Expert S~stenl Shell. J(H{I-HIII (f .$>st(ws M(iml.itcnl.$ i)! Agri(l{l tl{i-[~Pl().~]~tct.~ ,fiw Appli(wtim. Jleutsche Landwifischafts-Gesellschaft (cd. ). Frtinkfurt. Gemumyz. 1988. p. 252. Flinn, P.W. and Hagstrum, D.W. ,Stored Grai n Advisor: A Knowledge-Based System for Mtinagement of Insect Pests of Stored Grain. Al Applitwti~)]t.~ ill Natliral Re.WH(r(aC kfanllgo~~l~)llt ~(~ ):~~. 1990. Fourdraine, R. H.. Tomaszewski. M. A.. and Taylor. J. F.. Using a Natural Language Interface to Assist DH1 Supervisors, J~)[{rnal (!/D{lil-> .~~ic}l(c 73(supp. 1 ):209, 1990. Gauch, S. and Smith. J. B.. Intelligent Search of Full-Text Databases, TR87-035, Department of Computer Science. University of North Carolina. Chapel Hill, 1987. Goodell et al., CALEX Cotton: An Integrated Expert System for Crop Production and Management in California Cotton, Cal{/iu-nitl A,qriculfut-e 44:(5 ): I 8. 1990. ~-) -. 33. M. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46 47. 48. Grosz. B.J. et al.. TEAM: An Experiment in the Design of Transportable Nuturtil-Ltinguugc lntm_faces,. Arti/i~i[ll lntclli.stt~ttl.s ij~ A
PAGE 130

126 l A New Technological Era for American Agriculture 49. 50. 51, ~? -. 53. 5-I. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. Kline. D.E. et u]., Machincr} Selection Using Expert Systems and linear Programming. CcJnlp[(ter,j (iItd Elc(.troili(.s ill A,qri([(lt[tre 3 :45. 1988. Lambert. J. I?.. Networks, Tclcc~~l~lnlunic;ltit~ns iind Multi nwdi~l lnt~)rnlatit~n Buses for Agriculturtil Decision Support cwmm issit~ncd background paper prcpwui for the OIficc 01 Tcchnolog) Assessment. W:ishingtt~n. DC. 1991 L~inlbcrt, J. Il. :Ind Deter. R. S.. Whole Earth Decision Supp(wt Sjstenl. J{)l(I.IILII ofl>air] .Scicltcc 7Nsuppl I ):255. 1990. Ltinduu. J .A.. Norwich. K. H.. :md E\ri.ins. S. J., Autom;itic Speech Rcc(~gnition-Can It ]mprokc the Man-M:lchinc lntcrt:icc in Mcdiciil Expert Systcms! lltt[.lttltit~}~ill J(Jl(rn~Il ot[)iol]lttli~[ll COIIIpl{[it?,q 24: I I I I 989. Lunc. A.. DOS in English.. BYTE Dccenlbcr:261. 1987. Luyer. tl. A., Microcomputer History and Prchistory-An Archueologictil Beginning, Anncl/.Y ()/ tht~ Hi.~i(jr> ()/ Cl)\ttpl{ti~l(q 1 l; 127. 1989. Luzw-us, W.F. and Smith, T. I-l.. .A&~ption of Computers tind Consulting Su-\ices by New York DtiiD F:mmxs. J(mrIILIl (!/D~~i\;\ .hiemv 7 I; I(j(j7, 1988. Lcmmon, H .E., bCOMAX: An Expert System for Crop Mantigcment. S<-iellce 223:29-33. 1986. Licpins, G.. Gocltz. R.. tind Rush. R.. b Machine Lc~irning Techniques t-or Nuturti] Resource Ihtu .Anu]ysis, Al Applit [lti(~n.~ ill Natl(ral Rc.~ol(rcc Ivl(ln(i,qmen( 4(3):9. 1990. Makclu. M. E.. Vins(ln. S. B.. tmd Stone. N. D.. f+t)st-Pw-asitoid Pt)pulation Dynitmius in a Heterogeneous Enlir(Jnnlcnt. in Pr(M. .$(x. [~)IHputcr Si\ttl{l[iti~)\t ,Mllltit{~\l/i~l.ttl~[ oil Art~/i~i[il lit(clli,qctt~c (Ind .~inllll~iti(m: TiIc [~i~vr.~it> (){ Ap plitatio~t.$, Sun Diego. C,A. Feb. 3-5. 1988. pp n+~j~ Mm-an, L. R., CPEAS: The Crop Production Expert Acl\iw}r S}stcm. Knowltxlge Bused Sjstenl\ Resciirch l.uborut~~r), Dcp:u-tment of Agronoml, Univcr\it) t~tlllim~is. Llrb:illu-Chai~lptiign, 1989. Mw-cus. A. :md vun D:lnl. A., bUser lntcrfucc Dciclopmcnts tt)r the Nineties.. IEEE C(N~/pI(tcr Scptcmber:49. 1991. Martin. G. L., The Utilit~r of Spetch Input in UscrComputcr ln[crt-:lccs.. ltltt~)-ltil~iolltll J(NII-IW! @M~(t~Machi}lc .7tl({iie~ 30:355, 1989. h4cArthur. D., Kl~lhr, P.. and Narain, S., ROSS: An Object-oriented Language for (lmstructing Sinlul~it(ws. RLmd Ct~rp. R-3160-AF: 1984. ,M~.Grunn. J .iM.. Kurkt}sh. K., and F:dconcr. L.. A~ri~u]tur~l] Financiill Anulysis Expert S!stcnls. Texas A~!riuu]tur:l] Experiment Stution, l-CXUS A&h4 (-,lni~~r~it~ ((~]lcgc St;ltion. TX. 1989. McKinitm, J.ltl :Ind l.enlmtm. H. E.. Expert S~\65. 66. 67. 68. 69. 70. 71. 7~ 73. 74. 75. 76. 77. 78. 79. 80. terns for Agriculture, Computers und Electronics in Agriculture I :3 1. 1985. Michalski, R. S., bLem-ning by Being Told und Learning From Examples: An Experimental Conlparison of the TWO Methods of Knowledge .Acquisition in the Context ot Dcvelt~ping tin Expert System for Soybeun Disease Diagnosis. /ntermltiont{l J{~lIrnLIl ()/ Poli~~ Anal~.ji.s tIIId ln/i~rljI~IIion S~.\tem.\ 4: 125. 1980. Michalski. R. S.. Curbonel]. J. G.. and Mitchell. T. M.. kla~.hinc Ltvirnincq. AI1 Arti/i~i~~l lntclli .Y~icn~.c 74(supp. I ):290. 1991. Oltenticu. PA., Fcrguson, J.Il., and Lcdnor, A. J., A Data-Driven Expert System to Evaluate Management. En\rironnlentiil and Cow Fuctor lntluencing Reproducti\c Efficiency in Dairy Iierds. Jol~rIlttl (!fDait;\ .~cietl~c 7.3( Supp. l): 145, 1990. Oltlen, J. W. et al.. .9 Integrated Expert System for Culling M:intigement t~t Beet COW S, Cfvlrplit~~r.\ attd Electr~~ni~.s ill A,qri~l{lturr 4:333, 1990. Peacocke. R.D. and Grtif. D.H. An Introduction to Speech and Spei.ikcr Recognition, /EEL C()/Ilputer August:26. 1990. Petirl, J., Rcast~ning Under Unccrttiinty.. AnnlIIIl Re\ic\\ of Cotttp[lter S~ictt~c 4:37, 1990. Pickering, J .A.. Touch-Scnsiti\e Scrccns: The Tcchnolt~g} und Their Application. //?/i/-/(ifJi?t/?t// J(HII.IILI1 {!fMt~tl-hl[itltittt~ Stl{dic.s 25:249. 1986. [~lckerlng. J. et ~11 ., RAIN: A Novel Approuch to Conlputer-Aided Decisitmmaking in Agriculture und Forestr}.. Cot)lp[(tcr.~ atd Ele~tr~)ni~j in A,qritl[ltllr( 4:275285. 1990. Plunt. R. E., An Integrated Expert System Decision Support S}stem for Agricultural System. A,y ricl!ltltr{ll .5)YttIIn.~ 29;49. 1989. R:iuscher. 14 .N4. tind Jc~hns(~n. S., .Autht)ring Hy

PAGE 131

Chapter 4Ad}7anced Computer Technologies l 127 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. ()? 93. pcrtext Documents Using HypcrWriter. Al Appli~[itit)tt,s in N(itiirtil Rc.~~)iIr~e IWLi II[itqCmCIIt 5(2): I 16. 1991. Rotich. J. W. et al.. POMME: A Computer-Btised Consultation S}rstcm for Apple orchard Mwwgement Using PROLOG. E.rpcr[ .S)stct)is 2(2 ):56 69. 1985. Swu-tmmtia. H. et al.. ..An Artiticitil lntclligcncc Mt~delling Approach to Simultiting Animul Htibitut lntertiutitms... E~olo(qi~[ilM() ~lt~llillstcnl tor Lti~m Managcmcnt... Poi{/ti-\ Sticllte 68: 10447. 1989. Suhrcinemtikers. J.F. et al.. The Intr(xiuction of Expert S}stcm\ in Animul Husbumir}. The \rc/crilifit~] Qii{irterl] i 0( -1 ):28 i. I 988. Scars. A. and Shnciticrnum. B. ..liigh Precision Touchscrceris: Design Strategies und Compurist~n\ With A Mouw. lilt~.iititi~)il~il .Ioi{rtitil of IVl~iII M(i(lli)te .Sti((lic.j 34:593. i 99 i Sequcir:i. R.A. et al.. .Ob,lcct-oriented Simuliiti(~n: Pl:int Growth A Discrctc Org:in to (lytin Intcrtict it~ns. Ectjl. Jl(dcllilt,q 58:55-89. i 99 i Shift-man. S. ct Lil.. .13uilding :1 Spaxh intcrf:iuc to u Mcdiciil DiugntJstic S)stcm.. /hEL E.I~wrl Fcbruar)(:4 i. 1991. Sht~rtlittc, E.. 11.. Computer-ll:iwi Mtxiic.:il C~msult;iti(ms: MYCIN. Amcricun Elsmricr. Nc\\ Y(wk. i 976. Smith. D.C. ct :ii.. .Dc\igning the SIAil Lwr intcrtacc.. BYTE 7:?-$2. 1982. Sonk:i, S. T.. .Expcrt S)\tcms t~~r Busincis Dcc ision M :ik in:.. ct~mnlisii(~nc~i b:ickgrt)lin(i p~ipcr prepared t(~r the (Nticc (~t Tcchnd(lgy ASwS\mcnt. I 99 i Spiihr. S. 1.., Jones. L. R.. :imi i~ill. D. E.. E\pcrt Sy\tcm-Their Uw in [liiir~ 1 lcrci Nliin:igcmcnt .Joi{rii[il of I)(iir> S(i(}1(() 7 I :879. 1988. Storw. N. 1). ct :il.. .C~~tf]cx: A hf(xili[:i[ E\pcrt S}stcnl thiit S~nthc\i/c\ Bi~~l(~yic:il iin(i Ec(~nt)nliC:II Aniilj\is: l-hc fcst hl LIn:igcnlcn( Ad\ iv)r A\ An Exiinlplc. i n Pr(jcct~[litig.$ (~/ IIic B1ltliqii{() < olloIi Pro[ii{~ti(j\i (iti(l fic.~c~ircII {(~Ii/crctitc. N:it i(~n;tl Cott(m Council of Amcric:i. hlcmphi~. TN. 19X7. 94. 95. %. 97. 98 99 i(M). Ioi I 02. 103. 104. i05. 106. 107. Stone. N.1).. Knowiedge-Based S>stcms tor Crops commissioned background paper prepared for the otfice of Technology Assessment. U.S. Congress, Washington. DC. i99i. Stone, N. D.. bChat~s in an lndi\idual-Lc\ci Predator-Prey Model .. N(iti(r(il Rr.\oiircc M{)tieli}t.~[cIIi.~. Banttim i30t~hs. Ncu Yoi+k. NY. i 986 Viirncr. M .A. iind C:iCi) R. A., b.l)tiirj -L: ,4 New Concept in Technology Trtin\t-cr i(~r l-lxtcnsit~n... .Ioi(rn(il ()/ ll[iir~ .%~ic}i~c 74( Supp. I ):20 i I 99 I Wtiin, N.. Iklilicr. C. D. F.. :imi lli~ri~. R. l}.. A RuIc-B;IvA Inlkrencc Sjstcm l~)r Anim:il f%cduction Miinwymcnt. (.oi)tpi(lci.$ (in(l [lltclrolli~~ ili Acqri~i~lti{rc 2;277. i 988. W;ildrop. M.M A l.:incim:irk in Speech Rcut~gn it ion. .Sticii(c 2-I(): i 615. 1988. Whittakcr. A.D. et :11.. Iliir) 1 {crci NLitritioniil An~il~sis using Kmm Iccigc Sjitcms rCUhIll~LIC\. A,qrici{ltiirtil .$].~tcjtl.~ 3 I :83. i 989. W(N)cis. W.A. .Lun:ir R(~ch\ in N:itulti] Engliih: Explorations in N:itur:ii i.linyu:i~c ~Licsti(m Ans ~+ c r I n g. L illtq iii.~ t i~ .51ii(~ti(rt f>rf)ci~.~iliy. A Z~impt~lli (Mi. ) ( Ncm Yt~rk. NY: North-Hollund p~l[l[lihlll,~ C(). i 977). p. 52 I .-

PAGE 132

Part II Implications of New Technologies for Agricultural Production

PAGE 133

Chapter 5 Productivity Implications of New Technologies

PAGE 134

Contents Page TECHNOLOGY ADOPTION AND PRODUCTIVITY IMPACTS: NEW PROJECTIONS . 133 Timing of Commercial Introduction . . . . . . . . . . . . . . 133 Primary Impacts . . . . . . . . . . . . . . . . . . . 134 Adoption Profiles . . . . . . . . . . . . . . . . . . 134 Projection of Animal and Crop Production Efficiencies . . . . . . . . . 137 IMPACTS OF NEW TECHNOLOGIES ON THE STRUCTURE OF CROP AGRICULTURE 139 Large-Acreage Volume Crops . . . . . . . . . . . . . . . 139 Small Acreage Specialty Crops . . . . . . . . . . . . . . . 139 New Crops and New Uses of Existing Crops . . . . . . . . . . . 140 IMPACTS OF NEW TECHNOLOGIES ON THE STRUCTURE OF ANIMAL AGRICULTURE . . . . . . . . . . . . . . . . . . 140 Case Studies . . . . . . . . . . . . . . . . . . . . 140 New Animal Products . . . . . . . . . . . . . . . . . 144 IMPACT OF NEW TECHNOLOGIES ON AGRIBUSINESS, LABOR, AND RURAL COMMUNITIES . . . . . . . . . . . . . . . . . . 144 Agribusiness . . . Farm Labor . . . Rural Communities . POLICY ISSUES . . . Moratoriums on Agricultural Technology . . . . . . . . . . . . . . . . . . . 144 . . . . . . . . . . . . . . . . 146 . . . . . . . . . . . . . . . . 147 . . . . . . . . . . . . . . . . 148 Research or n he Implementation of New Agriculture . . . . . . . . . . . . . . . . 148 impacts of Emerging Technologies on Farm Size and Managerial Skill Requirements ..,.., 149 Displaced Farm Operators and Workers . . . . . . . . . . . . . 149 Adjusting (o Change . . . . . . . . . . . . . . . . . . 149 Figure Figure Page 5-1. Logistic Adoption Curves for Corn, Package A . . . . . . . . . . 137 Tables Table Page 5-1. Alternative Technology Scenarios . . . . . . . . . . . . . . 134 5-2. Timing of Commercial Introduction of Advancing Animal Technologies . . . . 135 5-3. Timing of Commercial introduction of Advancing Crop Technologies . . . . 136 5-4. Estimates of Crop Yield and Animal Production Efficiency by 2000 . . . . . 138 5-5. Projected Annual Rates of Growth (1990-2000) . . . . . . . . . . 138 5-6. Summary Characteristics of Representative Moderate-Size and Large Dairy Farms, by Region . . . . . . . . . . . . . . . . . . . . 142 5-7. Comparison of Average Annual Economic Payoffs From bST Adoption for Eight Representative Dairy Farms Under Three Alternative Dairy Policies, 1989 . . 142 5-8. impacts of bST Adoption on the Economic Viability of Moderate-Size Representative Farms, by Region, 1989 . . . . . . . . . . . . . . . 143 5-9. impacts of bST Adoption on the Economic Viability of Large Representative Farms, by Region, 1989 . . . . . . . . . . . . . . . . . . 143 5-10. Characteristics of Representative Moderate and Large Grain-Hog Farms in Missouri and Indiana . . . . . . . . . . . . . . . . . . . . 145 5-11. Average Annual Net Cash Farm Income due to pST Adoption for Representative Missouri and Indiana Hog Farms Under Alternative pST/Feed Response and Carcass Merit Premium Assumptions . . . . . . . . . . . . . . . 146

PAGE 135

Chapter 5 Productivity Implications of New Technologies Technologies discussed in the preceding chapters have the potential to increase American agricultural productivity, enhance the environment, improve food safety and food quality, and help increase U.S. agricultural competitiveness. Many of these technologies are fast approaching commercialization. Research in crop agriculture has advanced at a much faster pace than anticipated just a few years ago. Much of the research is aimed at improving crop resistance to weeds, insects and diseases: geoclimatic adaptation; and quality characteristics. In animal agriculture, new vaccines and diagnostics are on the market or soon will be. Growth promotants are going through the regulatory process. Reproduction technologies are advancing at a rapid pace and cloned embryos are currently being marketed. Transgenics are still in the future, but considerable strides are being made in the use of livestock to produce high-value pharmaceuticals. The advance of agricultural biotechnology and computer technologies will play an important role in increasing agricultural productivity and accelerating structural change in agriculture. These technologies, however, are not magica high degree of management skill will be needed to capitalize fully on their potential benefits. It will be important to develop management systems that make the most effective use of these technologies. This chapter and chapter 6 address these issues. In this chapter the technologies impacts on productivity are analyzed and implications for the agricultural industry are discussed. In the next chapter management issues will be examined. TECHNOLOGY ADOPTION AND PRODUCTIVITY IMPACTS: NEW PROJECTIONS OTA conducted two workshopsone for animal agriculture and the other for crop agriculturein part to assess the impacts of these emerging technologies on agricultural productivity. Workshop participants, carefully selected to include those with expertise in different stages of technological innovation, included physical and biological scientists. engineers, economists, extension specialists, commodity specialists. representatives from agribusiness and public interest groups, and experienced farmers. The workshop participants were provided state-of-theart papers on each technology prepared by leading scientists in the respective areas. These papers provided data on: 1 ) timing of commercial introduction for each technology area; 2) net yield increases (by commodity), expected from the technologies; and 3) number of years needed to reach various adoption rates (by commodity). The Delphi technique was used to obtain collective judgments from each workshop participant on the development and adoption of the technologies. Timing of Commercial Introduction Workshop participants were asked to estimate the probable year of commercial introduction of each technology under three alternative scenarios/environments assumed to extend to the year 2000: 1. 2. 3. Most likely scenarioa) a real rate of growth in research and extension expenditures of 2 percent per year, and b) continuation of all other forces that have shaped past adoption of new technology. More new technology scenario (relative to the most likely scenario)a) a real rate of growth in research and extension expenditures of 4 percent annually, and b) all other factors more favorable to new technology adoption than those of the most likely scenario. Less new technology scenario (relative to the most likely scenario) -a) no real rate of growth in research and extension expenditures, and b) all other factors less favorable to new technology adoption than those of the most likely scenario. Table 5-l shows in more detail the sets of assumptions made under the alternative scenarios. Table 5-2 shows workshop participants estimates of the probable years of commercial introduction of animal technologies, and table 5-3 shows the same for crop technologies under the three alternative scenarios. -133-

PAGE 136

Table 5-lAlternative Technology Scenarios More new Most likely Less new Factors technology technology technology Population growth rate U.S . . . . . . . . . World . . . . . . . GNP growth rate U.S . . . . . . . . World . . . . . . . . Trade policy . . . . . . . Tax policy . . . . . . . Rate of growth of export demand Grain . . . . . . . . Oilseeds . . . . . . . . Red meat . . . . . . . Energy price growth rate (constant dollars) Growth rate of research and extension expenditures (constant dollars) . . . Inflation rate . . . . . Regulatory environment . . . . . Consumer acceptance of new technology . 1.0% 0.7% 1.8% 1.6% 4% 3.4% 5% 3.5% Less protectionist, more Continuation of present favorable terms of trade policy More favorable toward Continuation of present technology development policy 1.8% 1.4% 2.3% 1.8% 2.0% 1.0% 5% 3% 4% 2% 8% 5% Less regulation, more faContinuation of present vorable climate for techpolicy nology development High Moderate SOURCE: Oftice of Technology Assessment 1992 These estimates range from the very near term for genetically engineered growth promotants and animal health technologies to 2000 and beyond for transgenic animals and certain crops. Participants thought that many of the advancing technologies may be available by the mid1990s. Of the 41 potentially available animal technologies, 21 were estimated to be available by 1995 under the most likely scenario. In crop agriculture, 19 of the 30 technologies examined were projected to be available for commercial introduction by 1995. Primary Impacts When technologies are adopted on farm their immediate technical impact on crop agriculture is usually increased yields, a changed product characteristic, and/or increased percentage of planted acreage harvested. For animal agriculture the impact is on feed efficiency for all animals, reproductive efficiency for beef cattle and swine, milk production for dairy cows, and the number of eggs per layer (producing hen) for poultry. To estimate the net impact of emerging technologies on agricultural production, workshop participants, using information provided about the new technologies at the meeting, projected net increases in crop yields, animal feed efficiencies, and other performance measures that 0.5% 1 .3% 3.0% 2.0% More protectionist, less favorable terms of trade Less favorable toward technology development .8% 1 .2% O.O% 1% 00/0 30/o More regulation, less favorable climate for technology development Low could be expected if the technologies were commercially available and fully adopted by farmers ( i.e.. adopted by all farmers). Since in practice most technologies would be used in combination with other technologies (including existing technologies), the individual technologies were grouped by the workshop participants according to their probable impacts on particular commodities under different scenarios. The commodities included corn. cotton, soybeans, wheat, beef cattle, dairy cattle, poultry, and swine. Through a Delphi process. OTA obtained estimates for each package of technologies on each of the commodities under the three alternative scenarios. Adoption Profiles When a new technology is introduced into the marketplace, only a small number of farms, mostly the large and innovative ones, will adopt the technology initially. This is because the possible payoff of the new technology is uncertain and because potential adopters need time to learn how to use the new technology and evaluate its worth. As early adopters benefit from using a new technology. more and more farmers are attracted to it, increasing the speed of adoption exponentially. Eventually. as most farmers who will adopt a new technology do so, the adoption rate will level off. Thus, the adoption profile follows an S-shaped curve (2).

PAGE 137

Chapter 5Productivity Implications of New Technologies 135 Table 5-2Timing of Commercial Introduction of Advancing Animal Technologies Technology scenarios More new Most likely Less new Technology technology technology technology Somatotropins Bovine: Dairy . . . . . . . . . . . . Beef . . . . . . . . . . . . Pork: pas t . . . . . . . . . . . . GRF . . . . . . . . . . . . Poultry: Broilers . . . . . . . . . . . . Turkeys . . . . . . . . . . . . 1991 1995 1991 1997 1991 2000 1995 1998 >2000 >2000 1995 1995 1995 1990 >2000 1995 >2000 1995 1995 1995 >2000 >2000 1990 1990 1990 >2000 >2000 >2000 >2000 >2000 >2000 >2000 >2000 >2000 2000 2000 >2000 1998 >2000 1991 1994 1992 1995 1998 1998 1991 2000 2000 1992 Beta-agonists . . . . . . . . . . . Reproduction and embryo transfer Control of ovarian functions . . . . . . . . Separation of X&Y bearing sperm . . . . . . In vitro fertilization . . . . . . . . . . Embryo sexing . . . . . . . . . . . Cloning and nuclear transfer . . . . . . . . Gene transfer . . . . . . . . . . . 1993 1992 1990 1998 1993 2000 1995 1995 1990 2000 1995 >2000 Animal health rDNA technology . . . . . . . . . . Gene deletion . . . . . . . . . . . Monoclinal antibodies . . . . . . . . . Peptides . . . . . . . . . . . . Immunomodulators . . . . . . . . . . 1993 1995 1995 1996 1996 1990 1991 1991 1991 1994 1994 1990 Antibiotic growth promotants . . . . . . . Steroid-like growth promotants Estrogen/androgen combinations . . . . . . . Controlled/sustained release . . . . . . . . 1990 1990 1990 1990 Transgenic Ruminants: Hormonally enhanced growth . . . . . . . Pharmaceutical production . . . . . . . . Enhanced disease resistance . . . . . . . 2000 2000 2000 >2000 >2000 2000 >2000 >2000 Poultry . . . . . . . . . . . . Swine: Improved productivity . . . . . . . . . Disease resistance . . . . . . . . . Disease immunity . . . . . . . . . . 2000 >2000 >2000 >2000 2000 2000 Fish: Rapid growth... . . . . . . . . . . Disease resistant . . . . . . . . . . 1995 1995 1992 2000 >2000 1995 Expert systems . . . . . . . . . . . Human-computer interactions Add-on systems . . . . . . . . . . Integrated systems . . . . . . . . . . 1992 1995 1995 2000 Sensor technology/robotics Reproduction . . . . . . . . . . . Health, . . . . . . . . . . . . Stress . . . . . . . . . . . . Carcass evaluation . . . . . . . . . . Milking system . . . . . . . . . . . 1995 1992 1995 1998 1992 1994 2000 >2000 1995 1995 >2000 1998 1998 >2000 2000 >2000 1996 Environment and animal behavior Optimizing environmental stimuli . . . . . . . Stress and immunity . . . . . . . . . Cognitive processes . . . . . . . . . Facilities and equipment . . . . . . . . . 1992 1993 1995 1992 1995 1995 2000 1994 SOURCE Office of Technology Assessment 1992

PAGE 138

136 l A New Technological Era for American Agriculture Table 5-3-Timing of Commercial Introduction of Advancing Crop Technologies Technology scenarios More new Most likely Less new Technology/problem area technology technology technology Pest control Pathogens for insect control: rDNA microbial insecticides . . . . . . . Introduction and colonization/rDNA . . . . . . Use of parasites/predators . . . . . . . . Genetic modification for resistance to insects: Bacteria . . . . . . . . . . . . Viruses . . . . . . . . . . . . Plants . . . . . . . . . . . . Insect and mite management . . . . . . . Weed control Biocontrol for weeds: Host specific pathogens . . . . . . . . Bioherbicides . . . . . . . . . . Anthropoids . . . . . . . . . . . Genetic modification for weed control Herbicide tolerance . . . . . . . . . Allelopathy . . . . . . . . . . . Disease control Microbial biocontrol of plant diseases: Manipulation of resident microbial communities . . . Antagonistic organisms . . . . . . . . Genetic modification for disease resistance . . . . Disease management: Crop loss assessment . . . . . . . . . Cropping system/agroecosystem interaction . . . . Plant stress Temperature and water stress: Biochemical/physiological indicators . . . . . Genetic modification . . . . . . . . . Root responses to stress . . . . . . . . Detection of stress . . . . . . . . . Information technology Knowledge-based systems for crops: Farm-level planning systems . . . . . . . Information networks . . . . . . . . . Expert systems for business decisionmaking . . . . Networks/telecommunications: Commercializing public databases . . . . . . Private databases . . . . . . . . . Commercializing public software . . . . . . Private software . . . . . . . . . . Robotics: Plant materials sensing/handling . . . . . . Machine guidance . . . . . . . . . 1993 1998 1998 1992 1993 1995 1990 1995 1991 1997 1993 >2000 1993 1993 1995 1991 1990 1995 2000 2000 1991 1991 1993 1990 1992 1992 1992 1992 1993 1994 1995 >2000 >2000 1995 1995 1998 1990 1998 1995 2000 1995 >2000 1997 1997 2000 1995 1990 2000 2000 2000 1995 1993 1995 1990 1995 1995 1995 1995 1995 1997 SOURCE: Office of Technology Assessment, 1992. >2000 >2000 >2000 >2000 >2000 >2000 1990 >2000 >2000 >2000 2000 >2000 >2000 >2000 >2000 2000 1990 >2000 >2000 >2000 2000 1998 2000 1990 2000 2000 2000 2000 1998 >2000 Many factors go into the decision to adopt a new techmarket segment is difficult to estimate, but it will probnology. A factor of growing importance is the ratio of ably support some producers who do not adopt hormones. consumer acceptance to rejection of a new technology. For example, it is likely that a portion of the population Other biotechnology products, suchas improved diswill prefer to purchase products that have been produced ease vaccines. most likely can be implemented effecwithout the use of growth hormones. The size of this tively by most producers and will have fewer new

PAGE 139

Chapter 5Prouctivity Implications of New Technologies l 137 Figure 5-lLogistic Adoption Curves for Corn, Package A 0 2 4 6 8 10 12 14 16 18 20 22 24 Years from introduction date More-new-technology -Less-new-technology soenario Soenario Most likely soenario SOURCE Off Ice of Technology Assessment management requirements than recombinant somatotropins. The extent to which such innovations are commercialized and adopted will depend on their profitability and effectiveness compared to that of other available technologies. To derive an adoption profile for each package of technologies under different scenarios, workshop participants were divided by expertise into commodity groups. There were four groups in the animal technology workshop (beef, dairy, poultry, and swine) and four in the crop technology workshop (corn, cotton, soybeans, and wheat). The participants were then asked the question, If a specific package of technologies was introduced in the market today, how long would it take for farmers to adopt it? Based on their answers, a logistic curve depicting the rate of adoption was fitted for each package of technologies applied to the eight commodities under different scenarios (see example in figure 5-1). Projection of Animal and Crop Production Efficiencies Based on information obtained from the workshops on: I ) years to commercial introduction, 2) primary impacts by technology package, and 3) adoption profile, OTA computed performance measurements for the eight commodity areas by the year 2000 under alternative scenarios. The results are presented in tables Table 5-4 and 5-5. Under the most likely scenario, feed efficiency in livestock production will increase at an annual rate of from 0.39 percent for dairy to 1.62 percent for swine. In addition. reproduction efficiency will also increase, at an annual rate ranging from 0.67 percent for beef cattle. to 1.25 percent for swine. Milk production per cow per year will increase at 3.01 percent per year. from 14,200 pounds to 19,200 pounds per cow, in the period 1990. During the same period, major crop yields are estimated to increase at rates ranging from 0.39 percent per year for soybeans to 2.02 percent for wheat. Wheat yield, for example, is projected to increase from 34.8 bushels per acre to more than 42 bushels per acre in 2000 under the most likely scenario. How do these rates of increase compare with historical trends and with OTAs last projections (8)? The most dramatic productivity increase is in milk production with a 3-percent annual rate of growth. Since 1960. the annual rate of growth has been about 2.5 percent. However, OTAs 1985 projection (24,200 pounds of milk per cow by 2000) was higher than its current one ( 19.200 pounds of milk per cow by 2000). A major reason for this discrepancy is the delay in marketing of bovine somatotropin. In 1985 it was predicted to be commercially available in 1987. As of early 1992 it has yet to be approved. In addition, the high milk yields projected in 1985 were revised downward in 1990 as more knowledge about the bST technology became available through additional research. Further increases in feed efficiency in livestock will lag behind historical trends in some cases and surpass these trends in others. Poultry feed efficiency has been increasing at about 1.2 percent per year for the past decade. This has resulted in making the chicken an extremely efficient converter of feed to meat. Further increases in feed efficiency will be difficult. Feed efficiency will continue to increase at 0.5 percent per year to 2000 under the most likely scenario. Feed efficiencies for beef and swine, on the other hand, have been static for the last decade. New technologies will increase feed efficiencies. Under the most likely scenario. feed efficiency for beef is projected to increase at an annual rate of 0.74 percent, reaching 0.154 pounds of beef per pound of feed in 2000; feed efficiency for swine will increase at the rate of 1.62 percent per year, reaching O. 18 pounds of

PAGE 140

138 l A New Technological Era for American Agriculture Table 5-4Estimates of Crop Yield and Animal Production Efficiency by 2000 Less new Most likely More new Actual technology technology technology 1990 2000 2000 2000 Crops Cornbu/acre . . . . . . . . Cotton-lb/acre . . . . . . . Soybeansbu/acre . . . . . . . Wheatbu/acre . . . . . . . Beef Lbs meat/lb feed . . . . . . . Calves/100 cows . . . . . . . Dairy Lbs milk/lb feed . . . . . . . Lbs.milk/cow/year . . . . . . . Poultry Lbs meat/lb feed . . . . . . . Eggs/layer/year. . . . . . . . Swine Lbs meat/lb feed . . . . . . . Pigs/sow/year . . . . . . . . 116,2 600.0 32.4 34.8 0.143 90.0 1.010 14,200.0 0.370 250.0 0.154 13.900 113.8 NA 32.6 37.7 0.146 93.750 1.030 17,247.200 0.373 250.500 0.174 14.420 128.5 708.0 33.7 42.6 0.154 96.221 1.050 19,191.600 0.389 258.0 0.181 15.750 141.6 NA 36.4 53.8 0.169 102.455 1.057 20,498.800 0.428 273.125 0.196 17.791 NOTE OTA expresses its appreciation to Yao-chi Lu and Phil Calling, Agriculture Research Service, U.S. Department of Agriculture for their assistance in deriving the estimates for this table. NA = Not available. SOURCE: Office of Technology Assessment, 1992. Table 5-5Projected Annual Rates of Growth (1990-2000) Less new Most likely More new technology technology technology Corn .21% 1.00% 1.97% Cotton NA 1,66 NA Soybeans 0.06 0.39 1,16 Wheat 0.80 2.02 4,36 Beef Lbs meat/feed 0.21 0.74 1.67 Calves/cow . 0.41 0.67 1.30 Dairy Lbs milk/feed . 0.20 0.39 0.46 Milk/cow/year 1.94 3.01 3.67 Poultry Lbs meat/feed.. 0.08 0.51 1.46 Eggs/lay/year 0.02 0.32 0.89 Swine Lbs meat/feed.. 1.22 1.62 2.41 Pigs/sow/year 0.37 1.25 2.47 NOTE: OTA expresses its appreciation to Yao-chi Lu and Phil Coiling, Agriculture Research Service, U.S. Department of Agriculture, for their assistance in deriving the estimates for twistable. NA = Not available. SOURCE: Office of Technology Assessment, 1992. pork per pound of feed in 2000. OTA made the same projection in 1985. Efficiencies in crop production will about match historical trends or climb slightly, and for the most part will exceed OTAs 1985 projections. This, in part, reflects the movement of many of the new technologies from the laboratory to the field at a much quicker pace than thought possible in the mid-80s. For example, in 1985 OTA projected wheat yields to increase at an annual rate of 1.2 percent under the most likely scenario. In the early 1990s they are projected to increase at a rate of 2 percent to the year 2000. Cotton was expected to increase at an annual rate of 0.7 percent in the mid-80s, but now is projected to increase at a rate of 1.66 percent to the year 2000. Soybeans are the exception. They were projected to increase at a rate of 1.2 percent in the mid-80s but now are projected to increase at the more modest rate of 0.39 percent, in part because biotechnology products are projected to become available to the soybean industry more slowly than previously thought. Note that corn is expected to decline from actual 1990 yield under the lessnew technology scenario. This is due, in part, to the anticipated loss of existing chemical technologies and a very slow rate of new biological technologies to replace them. Even though annual rates of growth in many agricultural products may accelerate during the 90s. the absolute

PAGE 141

quantity of yields will, for the most part, be lower than projected in the mid-80s. This is due, in part. to the fact that many of the early biotechnology inputs will be substitutes for chemical inputs and, hence, the absolute gain in productive efficiency will in many cases be negligible. This is expected to improve in the latter part of the decade as more is learncd about the genetic makeup of plants. IMPACTS OF NEW TECHNOLOGIES ON THE STRUCTURE OF CROP AGRICULTURE Production agricultural commodities generally fit into two categories: large-acreage volume crops. such as wheat, corn. and soybeans: and less volume small-acreage specialty crops, such as tomatoes, potatoes, and onions. There are several important distinctions between the two categories. First. there is less verticul integration of input, production, and marketing stages for large-acreage volume crops than for some small-acreage specialty crops. Second. the potential market for new technologies is much greater for large acreage crops than for specialty crops. This is an important driving force in terms of technoigical innovations. Third, biotechnology processes are already available to alter the harvestable component of some specialty crops such as tomatoes. This is due. in large part. to the fact that many specialty crops are easier to manipulate genetically than food and feed grain crops. Such developments are for the most part further away for the major food and feed grain crops (5). Large-Acreage Volume Crops As discussed in chapter 2, biotechnology i.implications such as herbicide resistant plants and biopesticides should be available in the near future. Unlike previous mechanical technologies. most biotechnologies will not, in themselves. generate significant economies of size. Also. there appears to be lttlc incentive for firms supplying seed and chemical inputs to expand vertically into crop production. Biotechnololgies that increase yield will have supply-increasing, price-dampening effects. These will adversely affect the survival of high-cost producers. which for the most part are small to moderate-size farm operations. Small-Acreage Specialty Crops As indicated in chapter 2, biotechnology already has the capability to modify the harvestable product for some Photo credit Grant He//man, Inc. Advancing technologies will have supply-increasing, price-dampening effects on large-acreage volume crops such as wheat. This will adversely affect high-cost farming operations. specialty crops. This capability will increase the extent to which processes specify product quality. It will also provide an incentive for vertical coordination between production inputs and the production and processing stages for a number of specialty crops. Thus, even though there are no obvious economies of size to be captured with biotechnology innovations, these innovations will facilitate vertical coordination in some cases. Small producers will be at a competitive disadvantage in specialty crops markets unless they have a particular market niche (5). For fruits and vegetables, biotechnologies will be important where product quality, shelf life, and taste are important characteristics. Technologies that allow for greater selectivity in specifying performance characteristics of different crop varieties will allow more rapid development of desirable cultivars and much more rapid propagation of plant stocks. Markets for tomatoes, let-

PAGE 142

140 A New Technological Era for American Agriculture tuce, and carrots are large and relatively focused on a few specific varieties. Improvements in these crops have the potential for rapid and widespread adoption to the benefit of growers, plant stock breeders, and consumers. There will be significant price differentials connected to biotechnology-based improvements and consumers can expect to pay higher prices for products more tailored to specific segments of the market. New Crops and New Uses of Existing Crops Biotechnology offers great potential for developing new crops and/or modifying existing crops for food, feed, and industrial uses. Examples include the modification of seed composition of corn and soybeans. Industrial use of corn for glucose, dextrose, starch, and alcohol has expanded rapidly, and biotechnology offers the capability to modify the protein, starch, and oil content of grain. Currently in the United States, approximately 3 percent of corn acreage is planted to special-use hybrids such as white corn for corn meal and grits, waxy corn for use as thickeners in the food industry, and hard yellow corn for snack chips. The other 97 percent is sold under the broad market classification of No. 2 yellow corn, without measurement of protein, starch, or other quality characteristics (6). For it to be economically feasible for farmers to grow products such as special-use corn hybrids, they must be able to capture price-premium incentives for these products. The current marketing system cannot easily accommodate new market channels for special varieties. It is expected that direct contracting between processors and growers will play an important role in the market development and growth of special-use products. The above example for corn hybrids suggests the likely pattern for marketing of other special-use crops. Where specialty market niches are small, incentives for a high degree of vertical integration in production and marketing will be substantial. This will limit the production opportunities for most independent producers (5). IMPACTS OF NEW TECHNOLOGIES ON THE STRUCTURE OF ANIMAL AGRICULTURE The U.S. livestock industry is divided into two components. One is increasingly space-concentrated, higher technology, and intensively managed. This component includes specialized cattle feedlots, broiler and swine production under confinement, and some large, highly specialized dry-lot dairy operations. A second component is the range livestock sector, which includes a large number of beef cow-calf operations along with a variety of small, lower technology livestock farms, many of which are operated by part-time farmers. A number of biotechnology applications is expected to have rather high adoption rates within the higher technology component of the livestock sector, compared to the lower technology, spatially dispersed sector. This is due, in large part, to the fact that increased managerial expertise is needed to use these new technologies effectively; such expertise tends to be associated with confinement systems. Growth promotants will be the first major biotechnology products to be made available to U.S. agriculture. The dairy and pork sectors will be the first to make use of these technologies. Case Studies Dairy Sector The dairy industry will most likely be the first to adopt technologies from the biotechnology era of the 1990s,

PAGE 143

Chapter 5Productivity Implications of New Technologies l 141 Photo credit: Grant Heilman, Inc. In the dairy industry the trend toward fewer and larger farms has been on-going for decades. The trend will accelerate as a result of new cost-reducing technologies and also will feel the first profound impacts of the emerging technologies. Biotechnology advances in reproductive technologies. animal health technologies, and growth promotants will make major contributions to the sector. In particular, bovine somatotropin (bST), a growth promotant, will significantly increase milk production. Bovine somatotropin is a naturally occurring hormone that increases milk yield in the dairy cow. Its effect has been known for decades but until it could be produced by rDNA procedures, it was not economically viable. This technology will increase milk yield per cow in 1 year to what it would take 10 to 20 years to achieve with current reproductive technologies (7). The economic effects of these emerging technologies can be visualized by analyzing the impacts on different sized farms in different regions. Representative farms used in the analysis are briefly described in table 5-6. Once bST becomes available, strong incentives will exist to adopt the technology. Payoffs from bST adoption are substantial, regardless of region (see table 5-7). Nonadopters of bST will have more problems surviving and will be more likely to exit the industry. Regional shifts in milk production patterns are expected for several reasons (tables 5-8 and 5-9). Upper Midwest farms have problems realizing sufficient earnand a more market-oriented dairy policy. ings to achieve a reasonable return on equity, compete. and survive. While Northeast farms fare better, they too were found to be at a disadvantage relative to Pacific and Southeast farms. In all regions, adoption of bST increases the potential to survive, especially for larger farms. Concern that bST will force many dairy farms out of the industry, especially in the traditional milk-producing region of the Upper Midwest and Northeast, has helped make this new technology the center of controversy. BST alone, however, will not force these traditional farms out of existence. The trend toward fewer total cows and larger farms has been underway for many decades. This trend is the result of a combination of emerging technology. economies of size, and policy. The trend will no doubt accelerate in the 1990s as the result of a combination of bST and other cost-reducing technologies, and a more market-oriented dairy policy. Such changes inherently put increased pressure on smaller traditional dairy farms. These pressures are accentuated by technological change but they are not new. For a more extensive discussion and analyses of these trends see the OTA report entitled U.S. Dairy Industry at a Crossroad: Biotechnology and Policy Choices.

PAGE 144

142 A New echnological Era for American Agriculture Table 5-6Summary Characteristics of Representative Moderate-Size and Large Dairy Farms, by Region Upper Midwest Northeast Southwest a Southeast Characteristic Moderate Large Moderate Large Moderate Large Moderate Large Cow numbers . . . 52 125 52 200 350 1,500 200 1,500 Output/cow (pounds) . . 16,850 16,850 17,940 17,830 18,590 19,690 15,340 15,310 Total asset value ($000) . 470 940 608 1,395 1,097 3,858 1,569 7,723 Land value ($000) . . 133 295 274 640 118 492 813 4,591 Percent of feed raised . 63 60 50 46 0 0 25 2 a lncludes farms from both the Pacific and Mountain USDA production regions SOURCE: Office of Technology Assessment, 1992. Table 5-7Comparison of Average Annual Economic Payoffs From bST Adoption for Eight Representative Dairy Farms Under Three Alternative Dairy Policies, 1989-98 a (thousand $) Policyscenarios Trigger b Fixed c Region size price support Quota d Lake States: Moderate . . 3.9 Large . . . . 10.3 Northeast: Moderate . . . 3.4 Large . . . . 15.8 Southwest: Moderate . . . 26.5 Large . . . . 90.5 Southeast: Moderate . . . 21.9 Large . . . . 166.4 4.1 10.9 3.6 16.6 26.6 91.7 22.8 166.3 2.4 7.0 1.0 8.8 18.3 61.2 17.2 132.0 a Economlc payoffs from bST are the average annual change in net cash farm income between a nonadopter and a bST adopter over the 1989 to 1998 planning horizon. The payoff is net of the cost of bST, the added transportation costs for milk, and the additional feed. b This option triggers a price support reduction each time the level of government purchases of milk products exceeds 5.0 billion pounds annually. c This option fixes the price support level at $10.60 per cwt. for all years. The quota policy is designed to maintain government purchases at or near a minimum government use target. This is accomplished by reducing the number of cows in a herd through a two-tiered pricing system or some other mechanism that provides disincentives for producing over quota levels. SOURCE: Office of Technology Assessment, 1992 Swine Sector As with the dairy industry, the swine sector will benefit from biotechnology improvements in the areas of reproduction, health, and growth promotants. Porcine somatotropin (pST), a growth promotant, will be one of the first technologies from the biotechnology era for the swine industry. Porcine somatotropin is a naturally occurring hormone in swine that accelerates the rate of growth, increases feed efficiency, and produces leaner hogs. Although the effects of pST on feeder hogs has been known for many years, it was not used commercially because of lack of availability. The ability to produce recombinant pST has heightened interest in using the product on commercial hog farms. Porcine somatotropin research has shown that it increases feed efficiency by as much as 40 percent, reduces fat by as much as 30 percent, and increases growth rate by as much as 33 percent. (See ch. 3.) The economic benefits of pST can be discussed by analyzing representative hog producers in the Midwest who adopt pST, and the costs to producers who do not adopt pST. An economic model was used to simulate the economic viability of two Missouri grain-hog farms (75 and 225 sows) and two Indiana grain-hog farms (150 and 600 sows) before and after the introduction of pST. The Missouri and Indiana hog farms represent two different types of Midwest hog farms. The Missouri farms raise fewer pigs per sow, in part. because their operations are not total confinement operations like those representative of Indiana (table 510). All the farms represent high-level management by progressive. full-time farmers intent on producing hogs efficiently with the best resources at their disposal. The farms were assumed to adopt pST on its introduction ( 1992) or not adopt it over the 6-year planning horizon (3). Two pST/feed response scenarios were evaluated. The first represented the average gains from pST, i.e., 25.1percent improvement in feed efficiency and a 12.7-percent increase in average daily gain. The second scenario assumed a more optimistic pST/feed response, a 34.8percent improvement in feed efficiency and a 33.3 percent increase in average daily gain. In recognition of the reduced fat to lean reported for pST-treated hogs, a 5percent price premium for market hogs was analyzed. This 5-percent carcass merit premium is within the range suggested in the literature. Results of the analysis indicate that farms that do not adopt pST will experience lower annual net cash farm

PAGE 145

Table 5-8Impacts of bST Adoption on the Economic Viability of Moderate-Size Representative Farms, by Region, 1989-98 (in percent) a 52-cow 52-cow 350-COW 200-COW Upper Midwest Northeast Southwest Southeast Measure NonbST NonbST NonbST NonbST of impact adopter adopter adopter adopter adopter adopter adopter adopter Probability of survival b 580/0 740/o 100/0 100/0 95% 97% 100/0 1 000/0 Probability of earning 5percent return on equity . . . 58 74 100 100 95 97 100 100 Probability of increasing equity c . . . 0 0 3 3 60 79 13 24 Present value of ending net worth as percent of beginning net worth d 16 29 72 77 109 128 76 89 a The analysis used a trigger-price dairy policy. b Chance that the individual farm will remain solvent through 1998, i.e., maintain more than a 10-percent equity in the farm c Chance that the individual farm WiII increase its net worth m real 1989 dollars through 1998. Present value of ending net worth divided by initial net worth indicates whether the farm increased (decreased) net worth in real dollars SOURCE: Off Ice of Technology Assessment, 1992. Table 5-9impacts of bST Adoption on the Economic Viability of Large Representative Farms, by Region, 1989-98 a (in percent) 125-cow 200-COW 1 ,500-COW 1 ,500-COW Upper Midwest Northeast Southwest Southeast NonbST NonbST NonbST NonbST Measure of impact adopter adopter adopter adopter adopter adopter adopter adopter Probability of survival b 95/0 99% 100/0 100/0 100/0 100/0 100/0 100/0 Probability of earning 5percent return on equity . . . 90 95 99 100 100 100 100 100 Probability of increasing equity c . . . 8 12 43 53 100 100 88 99 Present value of ending net worth as percent of beginning net wort h d 5 7 69 92 102 195 214 129 147 a The analysis used a trigger-pnce dairy policy b Chance that the individual farm will remain solvent through 1998, I e maaintain more than a 10-percent equity in the farm. c Chance that the farm WiII increase its net worth in real 1989 dollars through 1998. Present value of ending net worth divided by initial net worth indicates whether the farm increased (decreased) net worth in real dollars. SOURCE: Office of Technology Assessment, 1992 incomes (ranging from $13 to $33 per sow) due to lower hog prices (table 5-1 l). (The lower hog prices are due to the increased supply of meat caused by the availability of pST. ) This range of lost income is about the same across the four farms analyzed because it is a direct result of lower hog prices. For pST adopters this loss is more than offset by a 5-percent carcass merit premium for a leaner carcass. Increases range from $110 to $134 per sow (table 5-1 l). Increasing the feed efficiency and average daily gain from pST to the more optimistic feed response scenario more than doubles the economic payoffs to adoption. Without the carcass merit premium, the economic payoffs for pST average $265 per sow per year, more than double the $100 spent for pST. 2 If the producers can garner a 5-percent carcass merit premium, the per sow returns to pST adoption to a total of about $370 per sow per year. The pST figure aswmcs that pST costs $6 per pig and ii tidmlnl~tcrmf w cchly for 6 weeks. The balance of the cxwt IS xfdtxf Itihor and tccd C(YJS.

PAGE 146

144 A New Technological Era for American Agriculture Photo credit: Grant Heilman, Inc. Production of lean meat with porcine somatotropin (pST) will give meat packers a strong incentive to vertically integrate or contract with farmers. Economic pressures will be strong for most swine producers to either adopt pST or to exit the industry. The economic payoffs of pST adoption are about the same regardless of farm size. For example, the moderatesize Missouri farms per-sow payoff is within 10 percent of that for the larger Indiana farm. And, the difference in payoffs between the 150-SOW Indiana farm and the 600-sow Indiana farm are within $18 per sow. These results suggest that pST could be scale neutral. Nevertheless, pST could accelerate the concentration of the U.S. swine industry. PST adoption increases the total income of large-scale farms more than that of smaller scale farms due to the sheer volume of hogs produced on the large farms. For example, pST increases average annual net cash income $232,000 for the large Indiana farm and only $57.000 for the moderate-size Indiana farm. Thus, the large farm gains an internal source of capital for future growth far in excess of what the smaller farm gains. In addition, the smaller farms may experience lower average pST/feed response due to lower management skills while the larger farm experiences a higher than average pST/feed response and a 5-percent carcass merit premium. This results in the moderate farms average annual returns to pST in the $3,300 to $18,500 per-year range while the large farm receives $232,000 or more per year. PST may therefore contribute to a significant restructuring of the swine production sector. The production of more lean meat will give meat packers a strong incentive to vertically integrate or contract with producers and possibly pST suppliers. The economic pressures will be strong for most swine producers to either adopt this new technology once it becomes available or to exit the industry. New Animal Products Biotechnology methods capable of producing transgenic animals may alter the use of these animals from food to pharmaceuticals. Attempts are being made to produce rare, medically important proteins in pigs. Production of blood-clotting factors and tissue plasminogen activator (used to dissolve blood clots that cause heart attacks) are being investigated. A private firm has announced that it has successfully produced human hemoglobin in pigs. A blood-clotting agent has been transferred to and expressed in sheep. Transgenic cows producing pharmaceuticals have not yet been reported, but these animals are under development in a number of public and private laboratories. If successful, the production of pharmaceuticals will open new markets for livestock. Incentives will be in place for pharmaceutical companies to vertically integrate or contract with farmers for the production of pharmaceuticals from livestock. Capital costs for breeding stock is most likely to be quite high indicating that successful, large farms are most likely to meet this new market demand. IMPACT OF NEW TECHNOLOGIES ON AGRIBUSINESS, LABOR, AND RURAL COMMUNITIES Agribusiness Advancing products of biotechnology and information technology will have major impacts on agribusiness (input suppliers, processors, wholesalers, etc. ). Historically. the commodity-oriented agribusiness sector has been driven by economic forces to produce at maximum efficiency and maintain low costs. This has resulted in a system that is remarkably effective at converting un-

PAGE 147

Chapter 5Productivity Implications of New Technologies 145 Table 5-10Characteristics of Representative Moderate and Large Grain-Hog Farms in Missouri and Indiana Missouri Indiana Moderate a Large Moderate Large Hog Enterprise sows . . . . . . . Boars . . . . . . . Gilts (repI.) . . . . . . Pigs raised/sow/year . . . . Gilts sold/year . . . . . Borrows sold/year . . . . Sale weight . . . . . . Lbs. feed/lb. gain . . . . . Assets ($1,000) Land . . . . . . . Buildings . . . . . . Machinery . . . . . . Livestock . . . . . . Other Assets . . . . . . Total . . . . . . Liabilities ($l,OOO) b Real estate . . . . . . Intermediate Assets . . . . Other . . . . . . . Total . . . . . . Net Worth ($l,OOO) . . . . Acreage Owned . . . . . . Leased . . . . . . Total . . . . . . . Crops produced (acres) c Corn . . . . . . . Soybeans . . . . . . Wheat . . . . . . . 75 6 32 15.68 556 588 240 3.875 232.0 70.0 86.5 34.4 0 422.9 30.2 24.2 20.8 75.2 347.7 220 110 330 144 80 76 225 10 100 15.68 ,664 ,764 240 3.787 520.0 175,0 289.1 65.7 0 1,049.8 69.5 70.9 54.8 195.2 854.3 520 500 1,020 300 333 316 150 10 90 17.00 1,185 1,275 240 3.763 630.0 120.0 280.2 49.9 0 1,080.1 75.0 66.0 70.6 211.6 868.5 280 520 800 540 175 24 600 30 245 18.00 5,155 5,400 250 3.299 2,475.0 500.0 834.3 158.6 0 3,967.9 297.5 198.6 40.6 536.7 3,431.2 1,125 1,125 2,250 1,800 400 50 a The moderate size Missouri hog farm also has 25 cows on 100 acres of pasture. b Liabilites are reported assuming the farm has 10-percent debt on real estate assets and 20-percent debt on machinery and livestock. c Acreage of crops represents actual planted acreage in 1990 after accounting for set aside. All farms except the large lndiana farm participated in the farm program SOURCE: Office of Technology Assessment 1992 differentiated commodities into relatively low cost food. Today this sector is undergoing change inspired in part by the evolution of a more demanding and differentiated food consumer. In response, retailer strategies have emerged which focus on improving service to the end consumer. Information technology has facilitated the shifting of marketing efforts toward the discovery of consumer preferences. Information technology along with legal disclosure requirements have made it easier for the consumer to see a wider range of product attributes. Where buying decisions were once made on such aspects as variety, convenience, price stability, and value, now consumers can also evaluate additional characteristics that were previously experienced only indirectly, such as product quality, nutrition, food safety, and environmental aspects (4). To respond to a more consumer-oriented environment, input suppliers may need to explore how information technology can facilitate the coordination activities needed to assure particular attributes. In the future information technologies may facilitate new business strategies by providing improved information flows and by facilitating coordination of production and marketing activities. For example, Pioneers Better Life Grains and Frito-Lays Frito Corn Chips are two companies using information technology to assure product quality. Pioneer seeks suppliers who use a specific technology to tailor-make a seed that grows product specific attributes. Producers are required to provide specific production assurances that allow the processor to label the product for a specific set of nutritional attributes. Pioneer stands behind the attributes and accepts the implicit role as the enforcer, and information technology provides the linkages. Likewise, Frito-Lay contracts with producers for specific types of corn. The processed commodity is tracked through the market channel on a bag-by-bag basis to assure product quality (4).

PAGE 148

146 l A New Technological Era for American Agriculture Table 5-nAverage Annual Net Cash Farm Income Due to PST Adoption for Representative Missouri and Indiana Hog Farms Under Alternative PST/Feed Response and Carcass Merit Premium Assumptions Representative farms Do Do adopt Do adopt not average pST/feed response optmistic pST/feed response adopt No CMP a 5 percent CMP No CMP DST 5 percent CMP (thousand $) Missouri Moderate . . . 56.73 57.70 64.98 75.59 83.19 Large . . . . 149.16 153.93 175.66 209.15 231.85 Indiana Moderate . . . 214.22 217.53 232.66 255.48 271.70 Large . . . . 818.17 838.18 898.78 979.24 1,050.98 $/sow Missouri Moderate . . . 756 769 866 1,008 1,109 Large . . . . 663 684 781 930 1,030 Indiana Moderate . . . 1,428 1,850 1,551 1,703 1,811 Large . . . . 1,364 1,397 1,498 1,632 1,752 a CMP refers to carcass merit premium. SOURCE: Office of Technology Assessment, 1992. Input suppliers have experienced more consequences of the biotechnology era than any other part of the agriculture industry to date. In anticipation of biotechnology-enhanced seed for large-acreage volume crops, seed and chemical input industries already have transformed structurally, just as the hybrid seed-corn industry developed to become a billion-dollar business after hybrid corn became a reality 50 years ago. With the expected future gains from biotechnology, multinational chemical and pharmaceutical companies have acquired almost all of the major seed companies. Only Pioneer Hi-Bred international and DeKalb remain independent firms (6). Concentration of input industries increases the potential for monopoly power, hence the potential for exploiting farmers in their purchase of improved inputs. Overdependence on a narrow set of genetic material also raises the problem of ecological vulnerability. Economies of size in process technologies also can foster concentration in the input sector. For example, a 7 million dose-per-day bST plant can supply two-thirds of the Nations dairy herd. To the extent that efficient biotechnology manufacturing requires large plant sizes, there will be economic pressures to concentrate industry structure to a small number of firms. Moreover, in some cases, there may be incentives for manufacturing firms to integrate the manufacturing and retailing of inputs. As discussed earlier. the trend toward vertical integration in agriculture and toward proprietary production processes could result in a captive market for some biotechnology products. For example, a genetically engineered seed might be produced by a large, vertically integrated chemical-seed company with specified inputs such as fertilizer, pesticides, and herbicides produced only by that company. The potential for transgenic farm animals to produce pharmaceuticals will also provide incentives for vertically integrated companies. Firms already involved in pharmaceutical research can easily move into animal agricultural biotechnologies. The increased importance of proprietary products and processes in the input-supply sector and the increased economic incentives for further industry concentration imply a challenge for small-scale firms. The survival of such firms may depend on public research in technologies that they can effectively use in their production systems; market access to these technologies; and easily acquired information on use and management of available technologies (5). Farm Labor As has been true for most past technologies, the emerging biological and information technologies will generally shift labor from farming. At the same time, new employment opportunities will be provided in the agribusiness sector supplying these new technologies. Today

PAGE 149

Photo credit: Grant He//man, Inc. Newly emerging technologies will displace less farm labor than mechinazation, but labor will have to be substantially more skilled than in the past. only about 2 percent of the U.S. population is living on farms; about 55 percent of nonmetropolitan jobs in the food and fiber system are located off-the-farm in farm input, marketing, and other service sectors. Newly emerging technologies will displace less farm labor than mechanization, but the farm labor force will have to be substantially more skilled than in the past. This will be particularly true for workers in animal agriculture. Demand for unskilled agricultural workers will fall off. Hired field workers will be limited to specialty crop ( mainly fruit and vegetable) farms. One message seems clear: implementation of the new technologies will require a broad range of specialized Skills. For example, a key requirement of the new information technology will be computer literacy. Enhanced management skills will be needed generally to succeed within a system characterized by increased technical and economic compltexity. Programs to support skill upgrading of the farm labor force will be needed to capture fully the potential benefits of new technologies (see ch. 6 for a more thorough discussion of these requirements. ) Rural Communities The number of farms and farm population continued to decline in the 1970s and 1980s. The impacts of declining farm numbers are difficult to ascertain. In general, land is bought by other farmers and continues to remain in production so that total agricultural output does not significantly decline. However, declining farm numbers negatively affect rural community employment levels. In farming-dependent communities, for every one farmer that exits the industry, up to one additional job may be lost to the community. While in most urban areas the 1980s were years of economic recovery and prosperity. this has not been the case for rural areas. The rural economic crisis was due in part to depressed conditions in export-dependent industries such as agriculture. forestry, and mining. However, even when these industries began to recover in the mid1980s, the rural-urban gap widened. This was due, in part, to the fact that rural problems run much deeper than those of agriculture alone. extending to inadequate infrastructure, poor schools. lack of access to quality medical services, and lack of leadership to solve problems that exist. While rural communities may have once been dependent on agriculture, only 23 percent of the 3,106 counties in (his country can now be described as agriculture-dependent, nonetheless, more than 75 percent of the Nations counties are nonmetropolitan. Rural communities and agriculture are no longer synonymous (1). Much of the once agriculturally dependent popultition has moved to larger trade-center communities ( many in nonmetropolitan counties), which have therefore grown in population and business volume. Growing communities in rural areas are often preferred locations for consolidated public schools, medical facilities. and other public services. Those communities left behind are suffering the consequences, and some are particularly vulnerable to the structure of agriculture. The emergence of biotechnology and computer technologies will most likely spur on the decline of many small farms and agriculturally dependent rural communities. And, where product quality is influenced strongly by biotechnologies, such as pST in pork. and where highly specialized new markets are formed, such as pharmaceuticals, increased incentives for production-marketing links via contracting and other forms of vertical integration also can be expected. At the same time, increased demand by many farmers for one-stop shopping centers for farm supplies and technical services including those involving biotechnologies and computer

PAGE 150

148 A New Technological Era for American Agriculture 1. 2. 3. A open public discussion of biotechnology research priorities; enlightened policies and procedures regarding approval, patenting and regulation of biotechnology innovations; and insistence on high-quality and timely information about biotechnology for public and private decisionmakers. POLICY ISSUES number of policy issues surround the introduction of technological innovations in U.S. agriculture and their impacts on the industry. Many are already on the policy agenda in one form or another. Several are discussed below. Photo credit: Grant Heilman, Inc. Advancing technologies will most likely spur on the decline of agriculturally dependent rural communities. These business communities will need to substitute additional nonfarm economic activities if they are to remain viable. technologiesmay reduce the viability of business enterprises in smaller rural communities. These business communities will need to substitute additional nonfarm economic activities if they are to remain economically viable (5). In the near term, biotechnologys effects on rural communities likely will be most significant in regions of concentrated livestock production. The ability of rural communities in these regions to absorb adverse changes in agricultural employment will be closely related to the availability of off-farm employment. Because rural communities have diversified their economic base and are no longer dependent on agriculture, most rural community residents have little or no personal contact with farming, except as passive observers of environmental changes. The environmental impacts of production practices can, however, become a community issue when such externalities as water quality, chemical residues, worker safety, etc., become sources of concern. Local sensitivities about the implications of novel substances employed in animal and crop production already are significant. Perceptions of risk to health, safety, and/ or environmental diversity associated with transgenic organisms may become a further source of community conflict and controversy. To ameliorate such conflict and controversy, communities should facilitate: Moratoriums on Agricultural Research or on the Implementation of New Agricultural Technology Moratoriums have already been placed on the use of bovine somatotropin in Minnesota and Wisconsin. The dairy case study discussed earlier clearly showed that regardless of farm size or region, there will be strong incentives to adopt bST. The farms in Minnesota and Wisconsin, even if they do adopt this new technology, still will have problems realizing sufficient earnings to achieve a reasonable return on equity, compete, and survive. For farms not adopting the new technology the dilemma will be even more severe. The agricultural industry of these States will be at a great disadvantage relative to those States where a moratorium does not exist if bST is approved by FDA for commercial use. In the process of economic development a maturation process occurs such that fewer human resources are required in primary industries (farming and mining) and proportionately more workers are employed in the knowledge and service industries. American agriculture has achieved its preeminence in the world by substituting knowledge for resources. This knowledge, embodied in more productive biological, chemical, and mechanical technologies and in the managerial skills of farm operators, has given the United States a world-class agricultural industry at a time when many other sectors of our economy are losing their preeminent position. For U.S. agriculture to retain its status it is necessary to enhance public and private-sector capacity for scientific research and technology development. The costs, to consumers and producers, of failure to maintain and enhance our

PAGE 151

efficiency in production would greatly exceed the adjustment costs resulting from overabundance. Impacts of Emerging Technologies on Farm Size and Managerial Skill Requirements The post World War 11 era of farm mechanization made it virtually impossible for small unmechanized production units to compete and survive with farming as the sole source of family income. Some past chemical and biological technologies such as insecticides and hybrid seed, on the other hand, have been rather scale neutral except for price discounts afforded producers who were able to purchase them in large volume. The emerging biotechnology and information industries appear to have the potential for being relatively scale neutral in their application on those farms already large enough to support mechanization technology. But two qualifying considerations are important. First, the implementation of these emerging technologies will generally require increased management skills and, for some, computer literacy. Second, at least some of these technologies will be effective and profitable only if they are integrated into rather technically complex production systems at the farm level. Some of these systems in animal agriculture may involve environmentally controlled housing and scientifically based feeding and management procedures. Thus, increased managerial skills, and, in some cases, additional capital in the form of specialized buildings and equipment will be important components of successful farming in the future. This will most likely mean increased concentration of farm production among larger units with more sophisticated technology and management capabilities. A number of persons who have moved out of farming in the past four decades did have adequate skill levels but had an inadequate resource base of land or operating capital to succeed under a highly mechanical farming regime. Future adjustments in farming will be dictated less by large capital requirements than by the educational and managerial skill requirements for farmers. This is not to suggest that the future capital requirements in farming will not be high. They will. In fact, the capital requirements per worker in farming are very high compared to most other types of employment. But recent major deflation in agricultural capital assets, particularly farm real estate, together with creative procedures by farmers for acquiring access to land and capital resources, may result in educational and managerial skill levels becoming a more limited resource than capital. One clearcut conclusion emerges. Persons who want to compete successfully in farming will need to upgrade their managerial skills. A critical role for Extension is to develop programs and opportunities for farmers to enhance their management capabilities. Displaced Farm Operators and Workers More workers have left farming since 1940 than now remain on U.S. farms. Displacement of farmers and farm workers will continue, though at a slower pace than in the past half century. Adjustment to alternative employment is most easily accomplished by young people who are just graduating from high schools, vocational schools, and colleges or universities. Thus. strong educational programs and vocational counseling for youth in farming communities are of vital importance. Selected public policies should aim at ensuring the provision of such educational support services. Other displaced farm workers will seek nonfarm employment either with or without retraining for such employment. A number of special training programs are already in place for such individuals. These retraining programs. however, need to be geographically and financially accessible and have appropriate entrance requirements for those displaced from farming. Moreover. they need to target employment training to those skill areas for which jobs are available. A number of older farm operators and other family members without new training may have to adjust to whatever fullor part-time employment opportunities exist in the local community. The availability of such employment opportunities and the general quality of life in many rural farm-dependent communities will be heavily dependent on the local farm economy. And, in some cases businesses based on newly emerging technologies, particularly those supplying farm inputs, will provide new local employment opportunities. Adjusting to Change Policies to help farmers adjust to technological change on the farm or to off-farm employment are lacking. The Food, Agriculture, Conservation, and Trade Act of 1990 and related farm policies are aimed almost exclusively at reducing the use of farm inputs (mainly land) to curtail farm output; providing a price (and income) floor for producers of selected commodities; and enhancing the position of U.S. farm commodities in world trade. A unique exception was the dairy herd buyout program in the late 1980s, which provided some dairy farmers with an opportunity to cash out their dairy herds at more attractive prices than those afforded by the free market. New or expanded public policies are needed for upgrading the managerial skill levels of some farmers to cope with technical

PAGE 152

150 A New Technological Era for American Agriculture change and for providing retraining opportunities for others to enable them to exit from farming. Strong educational programs are also needed for all rural young people whether or not they have opportunities in future high-tech farming. Expanded Federal and State assistance will be required for effective educational programming in those rural areas with an eroding local tax base. At the institutional level, public institutions need to aim policies and programs at two somewhat different types of participantsthose who will adjust by staying in farming, and those who will seek alternative employment. Both groups need to be serviced by effective public technology transfer and training programs and supporting financial services. A reorganized and revitalized public extension service could play a major role in technology transfer while public credit agencies need to focus program delivery on the special needs of the two target groups. At the farmer level, it is crucial that individuals realistically assess their opportunities in and out of agriculture. Most should make deliberate career choices and follow up with the acquisition of the managerial skills to succeed in hightech farming or the retraining required for employment off-the-farm. Future farm commodity programs are not likely to provide an umbrella of income protection adequate for any but those farm managers who can adjust effectively and quickly to technological change. CHAPTER 5 REFERENCES 1. Knutson, R. and Fisher, D., Options in Deieloping a New National Rl{ral Polic>, Texas Agricultural Ex2. 3. 4. 5. 6. 7 8 tension Service, Texas A&M University, College Station, Tx, 1989. Lu, Y., Forecasting Emerging Technologies in Agricultural Production, in Yao-chi Lu (cd.), Emerging Technologies in Agricultural Production, Cooperative State Research Service, U.S. Depm-tment of Agriculture, 1983. Richardson, J., Farm Level Impacts Of Somatotropin Introduction and Adoption on Representative GrainHo~ Farms in the Midwest, commissioned background paper prepared for the Office of Technology Assessment, 1991. Streeter. D., Sonka, S. and Hudson, M. Information Technology, Coordination, and Competitiveness in the Food and Agribusiness Sector, American Journal oj Agricultural Econotnic.s, VOI. 73, No. 5, December 1991. Sundquist, B. and Molnar, J., Emerging Biotechnologies: Impact On Producers, Related Businesses and Rural Communities, in Agricultural Biot(~(hnolog~, Purdue University, West Lafayette, IN, 1991. U.S. Congress, Office of Technology Assessment, Agricultural Commodities as Industrial Ra}~* A4ate rials, OTA-F-476 (Washington, DC: U.S. Government Printing Office, May 199 1). U.S. Congress. Office of Technology Assessment, U.S. Dairy lndust~-> at a Crossroad: Bi(~t(~<}ltl~~l(~g> tlnd Polic? Choi(es. OTA-F-470 (Washington. DC: U.S. Government Printing Office, May 1991 ). U.S. Congress, Office of Technology Assessment, Te(.hnolog>, Publit Polic>, and the Chunging Stru[ture of American Agriculture, OTA-F-285 (Springfield, VA: National Technical Information Service, March 1986).

PAGE 153

Chapter 6 Management Implications of New Technologies Photo credit: U.S. Department of Agriculture, Agricultural Research Service

PAGE 154

Contents Page INTEGRATED PEST MANAGEMENT STRATEGIES FOR CROP AGRICULTURE . . 153 Molecular Genetics as a Tool for Detecting Resistance and Tracing its Origins . . . 156 The Influence of Genetically Engineered Crops on Pest Resistance . . . . . . 157 POLICY IMPLICATIONS REGARDING THE DEVELOPMENT AND DEPLOYMENT OF ENGINEERED CROPS . . . . . . . . . . . . . . . 166 ANEW ISSUE IN ANIMAL AGRICULTURE MANAGEMENT . . . . . . . 167 Farm Animal Well-Being . . . . . . . . . . . . . . . . 167 Biotechnology and Farm Animal Well-Being . . . . . . . . . . . 172 CHAPTER PREFERENCES . . . . . . . . . . . . . . . . 173 Box Box Page 6-A. Glyphosate: A Risk to Weed Resistance? . . . . . . . . . . . . 161 Figures Figure Page 6-1. Successful Seed Mimicry by Common Vetch Weed of Lentil . . . . . . . 154 6-2. Successful Mimicry of Barnyard-Grass Seedling for Cultivated-Rice Seedling . . . 155 6-3. Number of Crop-Pest Species Resistant to Synthetic Chemical Pesticides . . . . 155

PAGE 155

Chapter 6 Management Implications of New Technologies Biotechnology holds great promise for American agriculture, but this promise may not be realized if the technologies are poorly managed. The new technologies will demand considerable management skills and a holistic or systems approach to management. Pest resistance to technologies that control pests exemplifies management problems in the past. Many chemical pesticides are ineffective today because of pest adaptation. Evidence suggests that pest adaptation could have been delayed and, in some cases, avoided if proper management strategies had been implemented. As products from the biotechnology era are used to control pests, management strategies for delaying or possibly avoiding pest adaptation need to be identified. Good management will be of paramount importance for the effective use of new biotechnologies in animal agriculture. The new technologies are not magic bullets, and will not improve animal productivity without effective management. With or without biotechnology, a growing management issue in this decade is farm animal well-being. Little scientific evidence is available on farm animal well-being in the United States; much more is available in Europe. It is important that the American animal agricultural industries begin to focus more attention and resources on this growing issue and on the impact of new technologies on farm animal well-being. This chapter focuses on these critical management issues. First, pest adaptation to various control technologies is explored for crop agriculture. Various management strategies for delaying pest adaptation are identified for the new technologies developed through biotechnology. Second, the importance of the farm animal well-being is discussed, areas of research are identified, and biotechnologys potential impacts on farm animal well-being are explored. INTEGRATED PEST MANAGEMENT STRATEGIES FOR CROP AGRICULTURE Pest infestation is a serious problem for agriculture and effective methods to control pests are needed. Of all crop pests, weeds boast the longest recorded history of adapting to agricultural practices. It is a history dotted with examples of one of natures most interesting adaptive strategies: mimicry (35). By mimicing crop seed, weed seeds can lie hidden among crop seed stored for the next seasons planting. Successful mimicry of agricultural crops requires that weeds possess a number of important characteristics. Weed seeds must ripen by harvest time; remain on their stems during harvesting; and have a shape and density similar to that of the crop seed (35). A surprising number of weeds have evolved all the characteristics required to become crop-seed mimics. An example comes from the mimicry of lentil seeds, Lens culinuaris, by the common vetch, Vicia sativa. The lentil seed has a convex shape. Normal seeds of the common vetch are much more rounded than lentil seeds (figure 6-1 ). Another example is one of rices most serious rivals, barnyard grass. Barrett ( 1 ) discovered in weedy forms of barnyard grass so many rice-like traits that they found it more difficult to differentiate barnyard grass from rice than to distinguish two variants of barnyard grass from each other (figure 6-2). In the mechanized farming systems dominant in the United States, hand weeding may be a thing of the past, but the battle between farmers and weeds continues. Chemical herbicides used to control weeds do not discriminate on the basis of appearance. The nature of the game has switched to biochemical mimicry. Agricultural chemical companies spend millions of dollars each year inventing chemical agents that kill weeds in cultivated fields without harming crops. This has put enormous selection pressure on weeds to biochemically mimic crops. It is estimated that there are at least 84 cases of weeds with resistance to at least one chemical herbicide (figure 6-3). Like weed resistance to herbicides, the resistance of plant-pathogenic fungi to synthetic fungicides is a significant problem. By the mid1980s, more than 100 species were known to be resistant to at least one fungicide (figure 6-3). 1 The real experts at resistance to synthetic chemical agents are insects. Resistance to DDT, detected shortly after its introduction as one of the first insecticides, is 1 (h the other hand, s(mw pe~ticidcs have remained effective over the long term. For example, glyphosate has been used to control weds for more thtit 17 years without any documented examplci of resistance. LIkew ise there is no evidence of cwiling moths (pests of apples) developing rcsistuncc to organophosphate~ c~cn though these pesticides were uwd mtcnwly for 20 ywrs to control the moth (34). -153 297-937 0 92 6 QL 3

PAGE 156

154 A New Technological Era for American Agriculture frequently cited as a textbook case of rapid adaptation. Since DDT, insects have been most successful at adapting to almost all insecticides. More than 500 cases of insect adaptation to insecticides have been documented (figure 6-3). Besides the growing problem of pest resistance to chemicals, there is much criticism of chemical pesticides because of their adverse environmental side effects (95). Natural control methods are often touted as safe and effective alternatives to chemical pesticides, but there is no guarantee that pests will not adapt to these methods as well. Indeed, numerous examples abound of pests overcoming a wide variety of control methods. Pests have adapted to cultivation methods as illustrated by wild vetch in lentils and barnyard grass in rice (34). Pests also have adapted to crops bred to be pest-resistant. For example, a random sample of 63 plants bred for resistance to viral pests indicated that pests had adapted in 28 cases. Only five cases showed no evidence of viral adaptation, and the rest were inconclusive (20). Insects also have adapted to crops bred for insect resistance. Hessian flies in wheat, green bugs in grain crops. and leafhoppers and planthoppers in rice are examples (22, 33). Other insects have adapted to biological control agents. For example, alfalfa weevils and the forest pest Pristophora erichsonii have adapted to parasitic enemies, and silkworms have adapted to fungal control methods (34). Some strains of insects, the diamond back moth, for example, have developed resistance to biological control with Bacillus thuringiensis (56, 80, 91), a bacterium that is toxic to many insect pests. These examples lead to three basic conclusions: 1. pests have demonstrated tremendous ability to adapt to almost any control mechanism, 2. unilateral pest suppression tactics rapidly can be rendered ineffective due to evolutionary change in pests, and 3. the assumption that natural pest control tactics are superior to synthetic methods, at least in terms of limiting pest adaptation, is false. Control of pests requires the use of many approaches, rather than reliance on one single method. A holistic program that considers all causes of plant stresspathogens, weeds, insects and other arthropods, water and nutrient excesses and deficiencies, soil pH, salinity etc., is needed. However. developing such an integrated approach will require an enormous amount of information and an understanding of the interactions among different stress-reduction strategies. Much effort will also be needed to educate farmers in taking such a multifaceted approach to pest and other stress control. Figure 6-lSuccessful Seed Mimicry by Common Vetch Weed of Lentil Photo credit: Virge Kask Success at seed mimicry has given the common vetch the ability to contaminate lentil fields. At left is the typical seed shape of the common vetch, Vicia sativa. In a lentil field near Albion, Washington, plant pathologists recently found vetch seeds that had a distinctly different shape (center) that is quite similar to the flatter shape of the lentil, Les culinaris (right). SOURCE: Richard M. Hannon, U.S. Department of Agriculture, Agricultural Research Service

PAGE 157

Chapter 6Management Implications of New Technologies l 155 6-2-Successful Mimicry of Barnyard-Grass Seedling for Cultivated-Rice Seedling Photo credit: Beverly Benner Survival in a hand-weeded field is easier for a weed that looks like a crop plant. A barnyard-grass seedling, a serious nuisance in rice fields, is easily mistaken for a cultivated-rice seedling. Left to right, the plants shown are cultivated rice, the oryzicola variety of barnyard grass, and another barnyard grass seedling. SOURCE: Spencer C H Barrett, University of Toronto Integrated Pest Management (1PM) represents an attempt at such an approach. 1PM strategies seek to create a crop management system that combines compatible production techniques and methods in a manner that maintains pest populations at levels below those causing economic crop injury. The 1PM approach is based on Figure 6-3Number of Crop-Pest Species Resistant to Synthetic Chemical Pesticides. l l l l l 1930 194 0 1950 196 0 1970 1980 199 0 Year SOURCE: N.G Green, H.M. Lebaron, and WK. Moberg, Managing Resistance to Agrochemicals: From Fundamental Research to Practical Strategies (Washington, DC: American Chemical Society, 1990). ecological principles and requires a solid understanding of the ecological system to be managed. Development and deployment of integrated strategies requires basic knowledge about target pest species and their interactions with other pest and beneficial species, as well as with the crops to be protected and other host plants (70). Knowledge of the direct and indirect effects of other crop production and protection inputs on nontarget pests and beneficial species is also essential. Because crop/pest interactions display tremendous geographical variation for the same crop and pest, pest management systems must be adapted to local conditions. The complexity of, and lack of adequate knowledge about, pest populations and agroecosystem dynamics make 1PM an unrealistic goal at this time. Limited 1PM strategies have been used in cotton and apples to control insects, rather than weeds or disease (2 I ). Presently, 1PM efforts focus on integrating cultural controls (sanitation, crop rotation, appropriate selection of planting dates, irrigation regimes, planting densities, varietal selection); naturally occurring biological control; and the application of chemical controls when pest populations or damage to the crop reaches a threatening level. These action thresholds are based on the complex and dynamic relationship between crops and pests throughout a growing season (72). Combinations of pest-control methods ideally should act synergistically to control pests; at least they should not counteract each other. Research shows that synergism exists between some moderately resistant plants and biological control agents; in other cases, such plants adversely affect the activities of naturally occurring biological control agents (32). Compatibility with biological control agents must be a significant consideration when biotechnology is used to create resistant crop varieties and to extend the range of biological control agents. Some preliminary research involving tobacco that has been genetically engineered to produce low levels of Bacillus thuringiensis (Bt), indicates that Bt does not negatively affect natural enemies of tobacco budworm. It is possible that Bt enhances the effectiveness of the natural enemy by slowing budworm growth (34). Crops that have low to moderate levels of pest resistance, generally have responded well to chemical controls. Several cases have been documented where pest suppression has improved following insecticide use on resistant crop varieties (48, 93). However, there are also examples of antagonistic interactions (53, 55).

PAGE 158

156 A New Technological Era for American Agriculture Crop rotation has been employed effectively to decrease pest infestation. However, continuous cropping has also lead to a decline in incidence and severity of pest infestation by providing a more stable environment for the establishment of naturally occurring antagonistic agents. For example, the severity of take-all disease in wheat has naturally declined in fields that have been continuously planted to wheat for years. The decline is due to the establishment of a bacterium that controls the disease (97). Little is known about the compatibility of genetically engineered crops and cultural practices. Currently the use of constitutive genes (i. e., genes that are expressed in all tissues at all times in the plant) leave little room for temporal flexibility. In the above examples, the compatibility of only two control mechanisms for one pest is considered. However, many other plants, animals, and microbes, some of which are beneficial and some harmful to crops, are also part of the agroecosystem. Most of these components are studied in isolation; in a truly integrated system, all control mechanisms used to control all pests should be compatible. For example, mite management of almonds cannot be discussed without considering how simultaneously to manage codling moth, navel orangeworm, and weeds (49, 101, 102). The information needed to do this currently is unavailable. As practiced currently, 1PM strategies do not eliminate but strive to decrease chemical use by improving the timing of pesticide application to achieve pest suppression with minimal nontarget effects. Improved pesticide application technologies to minimize off-target drift could also decrease amounts of pesticides used. Pesticide delivery equipment designed to directly mix pesticides at the proper rate, eliminating the need for tank mixing, could increase the efficiency of pesticide application (78, 95). Development of pest management technologies and programs does not automatically lead to their adoption. Many obstacles stand in the way of farmer acceptance of these programs. The complexity of the programs requires high levels of management skill and this is a significant deterrent to many farmers. Information and programs tailored to meet the local needs, perceptions, resources, constraints, and objectives of farmers is imperative. Many farmers will need considerable training to use these technologies. The lack of coordination among organizations, personnel. and disciplines involved in pest management at the local and regional levels inhibits educational efforts. Development of expert systems and other information technologies may help in training and in coordinating these efforts (see ch. 4) (34). The failure of growers to perceive the long-term cost advantage of integrated pest management strategies is a significant deterrent to adoption. There is a general need to demonstrate how these management strategies might reduce production costs. For example, almond producers were generally skeptical of adopting an integrated mite management program, until it was shown that this program could be effective, was compatible with pest control tactics already being used, and could result in decreased production costs of $24 to $44 per acre (47). Developers of pest management technologies generally lack the social science training needed to demonstrate cost-effectiveness to farmers. Input from social scientists is needed to successfully develop and implement any new methods. Management of pests will continue to be a major concern of agricultural producers. Successful development and adoption of more comprehensive pest management strategies will require extensive scientific research, as well as improved methods of providing readily usable information to agricultural producers. A better understanding of the interactions between crops and pests and of mechanisms of resistance development is needed. Changes in farm management practices also may be needed. The ongoing battle to stay one step ahead of pests, given their ability to adapt, will require the development of new biological control agents, improved chemical pesticides and wholly new technologies such as genetically engineered plants. Biotechnology holds great promise for providing new ways to control plant diseases, insects, and weeds. The tools of biotechnology have created the possibility of selectively engineering plants for insect, disease, and weed resistance. In addition, these new tools are expanding the knowledge base of plant resistance and the interaction of plants and pests with the rest of the ecosystem. In particular, biotechnology will be very useful in detecting resistance by pests at a much earlier time than traditional technologies and in developing strategies to slow or alleviate pest resistance. Molecular Genetics as a Tool for Detecting Resistance and Tracing its Origins Until recently, pesticide resistance could be detected only after it became a problem in the field or through laboratory bioassays in which samples from a pest population are treated with predetermined doses of the pes-

PAGE 159

ticide in question. The number of samples that can be processed in this fashion is low, especially with insects and some weeds. If an enzyme that leads to resistance has been identified. another approach to detecting resistance is development of monoclinal or polyclonal antibodies to that enzyme (see ch. 3 for explanation of how they work). Although there are certain drawbacks to this approach, there is a potential with this system to detect resistance at very low levels using kits that can be applied directly in the field. With many pests, resistance develops in a number of localized geographic areas. It often is not clear whether these localized resistant populations arise independently or whether one population becomes resistant and rare migrants invade new areas and become the dominant form in the newly invaded area. It is important to know which of these two scenarios reflects the dynamics of resistance in order to limit further progression of the resistance problem. If the resistance developed in one location and spread to another via migration, then the mutation(s) leading to resistance are probably rare. It may be advisable to attempt to quarantine the areas of resistance and to eradicate pests within these areas. On the other hand, if resistance arises independently in each area. then the mutation frequency is probably high and the above strategy would be useless. If the biological mechanisms of resistance in two areas are clearly different. it is safe to assume that resistance arose independently. However, when the mechanisms of resistance are similar it is possible that resistance had one origin. Advances in molecular genetics have allowed scientists to clone the genes responsible for some kinds of pesticide resistance. By determining the point at which a mutation in the gene occurred in a number of different populations it will be possible to more precisely determine the number of origins of resistance. Work in this field is only beginning but progress in at least one case has been astonishing. A French molecular biology group working with a Culex mosquito species was able to demonstrate that a single, initial, mutation in an esterase locus (an enzyme that accelerates the synthesis of esters) is responsible for most of the organophosphate resistance in this species worldwide (76). Their molecular analysis demonstrated that the DNA sequences adjacent to the coding region of the gene were identical in all resistant populations. The Influence of Genetically Engineered Crops on Pest Resistance Two primary questions arise about pesticide resistant crops (and about herbicide tolerance in particular): whether the level and pattern of pesticide use will be altered by such crops; and/or whether crop production patterns will be changed. Impacts that might occur as a result of these changing patterns also need to be evaluated. Impacts include environmental and food and water safety issues and continuing or increased problems with resistance. No definitive data exists on these issues, only reasonable speculation on changing patterns (but not levels) of herbicide use that might occur. There is also reasonable speculation about changing crop patterns and pesticide use that might result from insect and disease resistance. However, more data is needed to assess environmental and food safety issues. Speculations about changing crop patterns combined with knowledge of how pest resistance develops does lead to some conclusions about the type of resistance problems that might arise. It also suggests some farm and industry management strategies that might be pursued to minimize resistance. These issues are discussed below (34). Herbicide-Tolerant Crops and Weed Resistance to Herbicides Today agriculture depends to a great extent on herbicides to control weeds. Herbicide use patterns (and related pest-resistance problems) are affected by many factors, including price, the spectrum of weeds controlled. residue effects, flexibility or timing of pre or postemergence treatments, marketing strategies, and ease of use. While biotechnology may contribute to pest resistance risks in some cropping situations, it is only one of the factors involved, and its application to American agriculture must be considered holistically. Biotechnology-agrichernical companies. and seed companies as well as public universities and laboratories are using genetic engineering to develop crops resistant to herbicides. With herbicide-tolerant crops greater quantities of particular herbicides can be used to control weeds. As the name implies, herbicide-tolerant plants can grow in the presence of herbicides that harm or kill a nontolerant plant. Some plants naturally tolerate particular herbicides. Grasses, for example, naturally tolerate certain herbicides that kill broad-leaved plants. Despite this, use of herbicides to control agricultural weeds is often limited by the sensitivity of a cultivated crop to a herbicide or by the sensitivity of other crops that subsequently will

PAGE 160

158 A New Technological Era for American Agriculture be planted in the same field. Herbicide-tolerant crops remove this limitation. They are designed to tolerate higher levels or more potent doses of herbicides than non-tolerant crops. A concern is that herbicide-resistance weeds may be created by the transfer of herbicide-tolerance genes to weedy relatives of crop plants or by the change in patterns or levels of herbicide use. Herbicide-tolerant crops could lead to increased problems with weed resistance or diminish these problems depending on the types of herbicide-tolerant crops developed and the manner in which they are deployed (27, 28). We must proceed with caution in developing and deploying herbicide tolerant crops. Resistance of weeds to herbicides is a recent problem that is predicted to worsen during the next decade. As herbicide use increases (a possible consequence of herbicide-tolerant crops) so does selection pressure for resistant weeds. Furthermore, gene mutation leading to resistance to some of the newer herbicides occurs at a reasonably high rate, leaving these herbicides in a vulnerable position. Research has shown that a number of the new herbicides (e. g., sulfonylureas, imidazilinones, and triazolopyrimidines) have the same target site in the plant, the ALS enzyme (acetolactate synthase), which is essential for plant growth. These herbicides bind to a nonactive site of the ALS enzyme, change its confirmation, and thereby inactivate it. Resistance to herbicides that inhibit the ALS enzymes has been found in eight weed species, and primarily arises through a change in the nonactive site of the enzyme (57). The mutation rate for this change is quite high ( 1 in 1 million) and companies are well aware that this presents a problem. Adaptation of a weed to one herbicide moreover can render the weed resistant to a number of other herbicides, a phenomenon called cross resistance (75). Overuse of a single ALS inhibiting herbicide or a group of ALS inhibitors in one area thus could be problematic. For example, continuous use of ALS inhibitors in soybeans and corn maybe ill advised in that it may accelerate development of resistance in target weeds. In 1991, two new herbicidal products, both ALS inhibitors, were labeled for use in corn. If these are used on a substantial crop area and other ALS inhibitors are also used on soybeans in the same area, risk of weed resistance will be significantly increased. Because the spectrum of weeds that a given herbicidal product can control is limited, a single product is rarely used everywhere or all the time. The higher the diversity of ALS inhibiting compounds, the greater the acreage that is likely to be treated with an ALS inhibitor. Herbicide Use in Corn/Soybean RotationsMany herbicides fall into two groups based on their spectrum of activity: broad-leaf herbicides; and grass herbicides. This dichotomy presents a short-term agricultural problem. Broad-leaf herbicides can be used in corn (which is a grass), but could be a problem in soybeans since it is a dicot (i. e., broad-leafed plant). Conversely, a number of herbicides that can be used in soybeans could be damaging to corn (e. g., Scepter). Until this year, imidazilinone and sulfonyl urea herbicides were used only in the soybean component of corn/ soybean rotations. Care had to be taken so that residues would not carry over to and damage the next years corn crop. Recently, collaborative work between American Cyanamid and Pioneer has lead to development of corn with tolerance of the imidazilinone products, Scepter and Pursuit, both ALS inhibitors. Scepter is currently used in southern areas on the soybean component of soybean/ corn rotations and Pursuit is used similarly in more northernly areas. If corn cultivars with imidazilinone resistance were introduced to areas with corn/soybean rotations, the door would be opened for the use of more ALS inhibitors in these areas. Pioneer is currently planning to release imidazilinone-resistant corn cultivars in the early 1990s in areas that do not generally use soybean/corn rotations ( 17). Since these areas grow continuous corn this could mean continuous use of these ALS inhibitors. Such an introduction must therefore be considered carefully. If tolerant corn cultivars were also released in areas with soybean/corn rotations, more land would receive continuous control with ALS inhibitors. Biotechnology could, on the other hand, be used to diminish risks of herbicide resistance in weeds. The ALS inhibitors are being relied on increasingly as they replace older herbicides with known environmental problems or high costs. Other types of herbicides are available that affect different target sites in weeds (e. g., glyphosate, glufosinate). Some of these compounds are limited in use because specific crops lack tolerance to them. If, for example, corn cultivars were developed with glufosinate or glyphosate tolerance, it might allow farmers to alternate use of ALS inhibitors and compounds with a different mode of action. Monsanto is currently trying to develop soybeans with tolerance to glyphosate based herbicides (e.g., Roundup). If they are successful and such soybeans were introduced

PAGE 161

Chapter 6Management Implications New Technologies l 159 Photo credit: Grant Heilman, Inc. Research is ongoing to develop soybeans with tolerance to glyphosate based herbicides. if successful, the cycle of continuous use of ALS inhibitors could be broken, thus slowing the development of resistance to target weeds. into corn/soybean rotations, the cycle of continuous use of ALS inhibitors could be broken. Herbicide Use in CottonAlthough cotton is sometimes rotated with other crops such as soybeans and corn, in major cotton producing areas of Louisiana. Mississippi, and Arkansas 75 to 80 percent of the cotton lands are planted to cotton for 5 or more years in a row (6). While soybean and cotton may be grown on the same farms, the land with the highest yield potential generally is reserved for cotton. Only about 5 percent of the land in these areas is rotated between cotton and soybean. Currently, mid-south cotton generally receives three herbicide applications, one pre-emergence and two postemergence. The most commonly used post-emergence treatments involve mixtures of Monosodium Methane Arsenate (MSMA) and fluometuron (a substituted urea) for the first post-emergence treatment, and Disodium Methane Arsenate (DMSA) plus cyanazine or prometryn (triazine compounds) as the second treatment. To date, none of these has caused significant resistance in weeds or environmental problems (7), although DSMAand MSMA-resistant cocklebur has been found in North and South Carolina (58). Some of the major weeds requiring control are the morningglories, cocklebur, prickly sida, and sicklepod, but the weed complex varies geographically, and from farm to farm. At least two companies have been working on developing transgenic cotton with herbicide tolerance. Calgene has had success in engineering cotton with tolerance of bromoxynil (a benzonitrile compound), which controls broadleaf weeds (87). Bromoxynil is especially effective against lambsquarters and young morningglories but is less effective on some other weeds. Monsanto has been attempting to develop cotton with tolerance to glyphosate. The company seems to have had some success but has altered its strategy because the original approach was not leading to sufficient tolerance levels. Monsanto has isolated what it considers promising genes to insert into cotton but has not yet tested them in any plants. Even if a high-yielding cultivar of bromoxynil-tolerant cotton were readily available, it is not clear how much acreage would be treated. Bromoxynil has a limited spectrum of activity and it will probably be heavily used only when lambsquarters or morningglory is the dominant problem. Where lambsquarters is the major problem, bromoxynil could be used twice a year. Where morningglory is the problem, bromoxynil will probably only be used once, in a post-emergence spray since other compounds can be used more effectively later in the season. Adding bromoxynil to the cotton system could result in use of more diverse classes of herbicides (and mechanisms of weed toxicity) than are currently used in that system. Little concern exists that bromoxynil will decrease this diversity (7). Thus, transgenic cotton with Bromoxynil resistance is unlikely to present a problem in terms of fostering weed resistance. If Monsanto succeeds in producing cotton with glyphosate tolerance, a very different situation may arise in cotton. Glyphosate is an effective broad-spectrum herbicide that can kill broad leaf weeds as well as grasses. If cotton were tolerant of glyphosate, this compound could replace the current post-emergence herbicides in a large portion of the cotton growing areas. While current post-emergence herbicides are generally effective, they could not match glyphosate for effectiveness nor for ease of use. Monsanto feels that two applications of glyphosate could replace current post-emergence combinations ( 14). Monsanto plans to lower the price of glyphosate to make it competitive with current practices ( 14). The U.S. use patent on glyphosate has been extended until the year 2000, but outside the United States this patent will expire soon if it has not already (26). A company in Canada is already gearing up to man-

PAGE 162

160 l A New Technological Era for American Agriculture Photo credit: Grant Heilman, Inc. Scientists have had success in engineering cotton with tolerance to bromoxynil which controls broadleaf weeds. Adding bromoxynil to the cotton system could result in use of more diverse classes of herbicides and thus it is not likely to foster weed resistance. ufacture a glyphosate-based herbicide. These changes offer incentives to reduce the price of the compound to gain market share. This price reduction would tend to make the compound appealing to farmers. The potential, thus, exists for glyphosate to be used over a large area, two or more times each season. If this happens will there be a high risk of weed resistance developing? Given the information we have to date there is no simple answer to this question. Box 6-A contains a review of some points made by scientists involved in the ongoing debate about this issue. Most of the crops that have been targeted for herbicide tolerance research are large-herbicide-use crops (i.e., the money makers). Perhaps a more important need is for herbicide tolerance in limited acreage crops for which there are few herbicides available. Herbicide tolerance could open the door for use of safer herbicides in these crops. Additionally, with limited acreage crops the risk of weeds evolving herbicide resistance is probably lower than with major crops. Crop-to-Weed GeneTransfer Before the biotechnology era, resistance of weeds to herbicides evolved through mutations in the weed plants own genetic material. The possibility that herbicide tolerance genes, engineered into crops, could find their way into weedy relatives of the crop has recently received considerable attention (e.g., Bioscience, June 1990). What will be the fate of such transferred genes, and will they increase the risk of herbicide tolerance evolving in weeds? There is no answer to these questions yet but some general statements can be made. First, it is generally assumed that natural rates of mutation leading to resistant traits in weeds are one in a million or less. Thus, any introgression (the entry of a gene from one gene complex to another) between the crop and an important weed that increases this rate without lowering the fitness of the weed could be of importance. If genes that reduce the fitness in the hybrid are tightly linked to the herbicide tolerant gene(s), the latter might not remain in the weed population long enough to cause a problem. Only empirical studies will determine the likelihood that a herbicide tolerance gene would free itself from fitness-reducing, or encumbering genes and become a problem. There are at least three things that could be done by genetic engineers to lower the risk of herbicide tolerance genes finding their way from crops to weeds, and leading to resistant weed strains. First, when developing transgenic crops containing the herbicide tolerance gene, molecular geneticists could determine if certain inserts map closely with specific crop traits that would tend to lower fitness of a weed. Second, when developing the initial constructs, a second gene could be inserted that would serve as a suicide gene if expressed in a weed seed (i. e., it would kill the whole weed). A final strategy would involve engineering herbicide tolerance into plants that required two genes to be effective. If the two genes were placed on separate chromosomes the chance that both genes would segregate when they were at low frequency in the weed population would be minuscule in an outcrossing hybrid. This could dramatically slow the rate of increase in frequency of the tolerance trait. A similar result could be achieved if the tolerance trait was controlled by a single recessive gene. Crops With Resistance to Pathogens The only breakthroughs in genetic engineering that are likely to affect pathogen control practices in the near future involve virus resistance. Work on engineering plants to express viral coat protein genes and antisense genes has resulted in plants with significant protection against

PAGE 163

Chapter 6Management Implications of New Technologies l 161 Box 6-AGlyphosate: A Risk to Weed Resistance? History Glyphosate hadbeen in widespread use for at least 17 years and no cases of resistance have been documented that could be directly traced to its use. However, due to its broad spectrum of activity, glyphosate has not been used on crop fields except in cases where weeds need to be controlled in fallow rotations. Most of the weeds that it has been used to control are perennials, and these weeds are less likely than annuals to evolve rapidly resistance. In at least one situation, however, glyphosphate has been used to control annual grasses in fallow rotations every other year for a long period of time with no sign of resistance. it has also been used on orchards (14). Chemistry Although glyphosate rapidly is degraded by some soil bacteria, plants apparently lack enzymes that can degrade this compound. in screening for resistance to glyphosate, Monsanto scientists have never found a plant enzyme that could degrade glyphosate. This further suggests that weeds are unlikely to mutate such that they become resistant to glyphosate (35). Mode of Action Unlike the sulfonyl ureas and imidazilinone herbicides that bind to an inactive site of a critical plant enzyme, glyphosate binds to the active site of an essential enzyme for synthesis of certain amino acids. Crop tolerance could be engineered by interfering with glyphosphate binding to this site. Any alteration in the active site that would inhibit glyphosate binding, however, potentially could also impair the binding of the enzyme to its target molecule and diminish the fitness of the plant. Monsantos experience indicates that this is indeed the case. This has apparently been one of the factors that has made it difficult for them to engineer crops with glyphosate tolerance. While overproduction of a less efficient form of the enzyme is possible, it still could lead to decreased growth efficiency. Lack of Persistence One important characteristic of glyphosate is that it does not persist in the environment. Therefore, weed control exerted by this compound is restricted to those weeds that are actually sprayed. Concision Certainly the question of potential of weeds to adapt to glyphosate is not yet resolved. However, it seems clear that glyphosate poses less risk than some of the ALS inhibitors. The information to date would suggest proceeding with caution in developing and deploying glyphosate-tolerant cotton. SOURCE: Office of Technology Assessment, 1992. a number of viruses (2). Such plants could be used widely iment was reported on in an anecdotal fashion (2). He in developed and developing countries. They certainly have the potential to raise yields. The question is whether this increase of yield will be stable. For 28 of 63 traditionally bred virus resistant crops examined, virus strains have been positively identified that could overcome the resistance (20). In only four cases was there good evidence that there had been no adaptation. Results were equivocal for the remaining cases. It is not clear whether or not we should expect the same track record from crops with genetically engineered resistance. Only one short-term experiment attempted to look for genetic adaptation to engineered resistance. This experindicated that he had propagated a TMV virus to high levels in an attempt to induce systemic infection of resistant plants. He passed the virus through the resistant plant seven times, after which it was collected and tested for rate of disease development. This rate was unchanged. This experiment was obviously a good first step in evaluating the potential of a virus to adapt to engineered resistance. Studies using a broader base of viral isolates and conducted over a longer period of time would be advisable and very useful before any engineered germplasm is relied on to increase yields in developing countries.

PAGE 164

162 l A New Technological Era for American Agriculture Engineered Plants With Insect Resistance BackgroundThere has been a great deal of interest on the part of industry in developing plants with resistance to insects. Although most of the traditionally-bred, resistant crop cultivars owe their resistance to secondary plant compounds (e.g., alkaloids, phenolics, terpenes) and changes in physical characteristics (e.g., spines, waxy leaves, solid stems) these traits are generally controlled by many genes and are not amenable to straightforward engineering approaches. Molecular geneticists have instead taken the approach of 1 ) finding a protein from a bacterium, plant, or an animal that is toxic to insects (e. g., venoms, bacterial toxins), 2) finding the gene that codes directly for the protein, and 3) inserting that gene into a plant. Sometimes this approach works well as with the crystal protein toxins from Bacillus thuringiensis (Bt) (59). In other cases, this approach is only partially successful, probably because the proteins are digested in the insect gut before they reach their site of action. If it were simple to design toxic proteins that could withstand the gut enzymes, plants would probably do so themselves. Another successful approach to engineering insect resistance involves the proteinase inhibitors, whose site of action is the insect gut itself. Unfortunately, high levels of the proteinase inhibitors are usually needed to inhibit insect growth. Of all the potential approaches to engineering insect resistant crops, those involving the Bt crystal proteins are farthest along. Crops that have been successfully engineered to produce insect-toxic proteins include tobacco, tomato, cotton, and potato. Other crops targeted for Bt crystal protein production include but are not limited to corn, rice, soybean, cucumber, and eggplant. The mother bacteria for the Bt toxin has been used for many years as a biological insecticide by organic farmers and to a limited extent by others. Recently, there has been an increase in the use of these bacteria in conventional, production agriculture. This is in part due to increased pest resistance to conventional pesticides. For example, few insecticides are still effective against diamondback moth and the Colorado potato beetle (23). Other reasons for increased use of Bacillus thuringiensis include better formulations and increased toxicity. Both conventional breeding and genetic engineering have been used to improve the potency of the bacterium. The Mycogen company in California has taken the gene from a crystal protein and placed it in another bacterium. They have reported field results indicating that their product has slower decay in the field than normal Bt strains and therefore is more useful for the farmer (23). Ecogen, a company in Pennsylvania, has bred a strain of Bt that produces two crystal proteins, one effective against lepidoptera (caterpillars), the other effective against beetle larvae. This product offers useful control of the Colorado potato beetle and the European corn borer when they infest potato. There appear to be some good markets for Bt products, whether engineered in plants or used as biological insecticides. One very good thing about using Bt is that it is not likely to disrupt natural enemies of pests or hymenopteran pollinators found in the crop, because most natural enemies and bees are immune to the effects of Bt. This property should make the use of Bt or Bt genes compatible with biological control. Again, the major question is whether or not Bt will offer long-term solutions to pest problems or whether pest insects will adapt to Bts and nullify their utility. There has been much concern over this issue. In the mid 1980s, there was a feeling among some workers that insects would not adapt to Bt (8). Many early attempts to select for resistance failed or produced very low levels of tolerance (24). In 1985, however, McGaughey (65) found that Indian meal moths selected in the laboratory for Bt resistance became over 100-fold resistant. 2 Further work by McGaughey and his colleague led to a level of resistance in excess of 250 fold. McGaughey and Johnson (66) also found cross-resistance to a number of other Bt strains. This was considered by some scientists to be an exception, but in 1989 Monsanto scientists published work (89) indicating 20 fold resistance to a Bt toxin in one member of the cotton bollworm complex, a major target for Bt toxin production. Further work by the Monsanto group found up to 70 fold resistance of cotton bollworms to this toxin (60). Ongoing research has found resistance in this insect to a number of Bt toxins, to plants expressing the toxin, and to mixtures of Bt spores and crystals (36). However, all of the above work was done in the laboratory, and field results do not always match laboratory findings. Nonetheless, in 1988 there was a report of field failure of Bt sprays in the Philippines due to resistance ~The meaning of this term involves u ratio. For example, if it takes 200 micrograms to kill d resistant pest compared to 2 micrograms to kill a susceptible pest, the pest has a 100-fold resistance (200 divided by 2).

PAGE 165

Photo credit: Monsanto Co. The large boll of cotton on the left is the product of a transgenic plant with Bt genes. The boll on the right was grown in the same field but comes from an unprotected, nontransgenic plant. However, resistance to Bt by bollworms is a very real possibility. of the diamondback moth (56). in 1990 resistance was carefully documented in a crop field in Hawaii (9 1 ). The level of resistance in Hawaii was about 30 fold. Recent evidence of Bt resistance in Florida, Southeast Asia and Japan indicate levels as high as 400 fold in the diamondback moth (85 ). There is no longer any doubt that at least some insects are very capable of adapting to Bt and Bt toxins. Recent work on the biochemistry of resistant Indian meal moths indicates that the difference between susceptible and resistant individuals involves a change in a receptor binding site in the midgut of the caterpillars. Interestingly, a change in the receptor that leads to resistance to one Bt toxin does not necessarily lead to resistance to other Bt toxins (96). For some insects (e. g., diamondback moth. cabbage worms), scientists have found two or more distinct groups of toxins with high activity. For species like the cotton bollworm. only one group of toxins offers high activity. There is high risk of resistance to Bt in some cropping situations. If a crop or a set of crops is engineered to produce a Bt toxin and is planted widely, the potential for resistance must be considered. CottonOne of the first major crops in which Bt genes may be commercialized is cotton. Monsanto claims to have Bt toxin expression high enough to kill 100 percent of the insects placed on a cotton sample in the laboratory. Close to that level of success was achieved in the field. Monsanto intends to commercialize Bt-producing cotton in the early-to-mid 1990s. In some areas of cotton production, cotton and soybeans are grown on the same farms although not rotated on the same field. This could be helpful in limiting selection pressure on bollworms to adapt to Bt-producing cotton because some of the insects (a refuge sub-population) will feed on soybeans. The effects of insects in refuges has been described earlier and can be quite important, especially if adaptive genes are recessive. Unfortunately, large tracts of cotton acreage are planted in solid blocks. Potential for resistance in these areas will be quite high. As long as the size of the bollworm populations is large there is likely to be sufficient genetic variation to lead to resistance. While it is impossible to say for sure that the bollworms will be able to adapt to Bt in the field, laboratory results certainly support this possibility. PotatoTwo types of Bt-toxin genes have been engineered into potato. Plant Genetic Systems in Belgium has engineered a Bt toxin into potato that is active against the potato tuberworm. Monsanto has engineered a beetlespecific Bt toxin into potato and reports to have achieved high levels of Colorado potato beetle mortality. The Colorado potato beetle (CPB) is notorious for adapting to pesticides. One reason for this is that there are few refuges for this beetle. When potatoes have been heavily sprayed with insecticides. it has very few alternative plants on which to feed. However, there is only one report of CPB resistance to Bt. which comes from a laboratory study in Michigan (68). Results of this study were only briefly described but seem to indicate approximately 30-fold resistance. No field resistance has been reported. It is difficult to assess the meaning of this since Bt sprays capable of controlling CPB have only recently come to market and have not been used widely. If potato plants with Bt expression are introduced and used widely, the selection pressure for potato beetle adaptation is likely to be as strong as that exerted by insecticides. CornSuccess with transgenic corn is very recent. Therefore, it is too early to know just what levels of Bt toxin expression will be obtainable in this crop. There is no doubt, however, that one of the goals of molecular geneticists in industry is development of corn with Bt toxin levels high enough to control European cornborer. The European cornborer currently causes over 10 percent yield reduction in certain areas of the United States (54) but is rarely the target of chemical control measures. In general, chemical control is not economically profit-

PAGE 166

Photo credit: Grant Heilman, Inc. Molecular geneticists have had recent success in developing transgenic corn with Bt levels sufficiently high to control the European cornborer. In the corn belt, there would be few alternatives for cornborers so Bt resistance could be strong, especially if corn is planted in monoculture. able because of the low value of the crop (on an acreage basis) and the difficulty of controlling this insect because of its habit of feeding in crevices and within plant tissue. Bt expression in corn would be a very desirable trait from the perspective of yield. If some farmers start to use it early on, they will have at least a temporary yield advantage over their neighbors. Certain areas of the United States where cornborers cause more yield loss than in other areas would gain an advantage. This would occur because their yield increase would be greater than in other areas (54). It is possible that corn seed with Bt genes would be adopted widely if it were priced low enough. In the corn belt there would be few refuges for the cornborers, so selection pressure for Bt resistant strains would be strong. In other areas of the country where corn is not planted in huge monoculture and cornborers feed on other crops (e.g., potato, beans, cotton, peppers, etc. ), selection pressure would not be as intense. Strategies for Delaying Pest Adaptation Need for a Comprehensive ApproachFrom the farmers perspective, the history of pest control is the saga of a long struggle to stay a step ahead of pest adaptation. Some of the techniques used to combat pests have proved relatively resistance-proof, but these successes have been limited (34). The experience with synthetic chemical pesticides has been particularly disappointing. There is growing recognition among scientists that they need to maintain an arsenal of pest-control tools in anticipation of pests evolutionary responses. That arsenal contains some potentially powerful weapons, among them the novel approaches of biotechnology. Much of the discussion of resistance management for at least the past decade has centered on ways to reduce the rate at which pests adapt to conventional pesticides. Yet pests adapt not only to pesticides but also to other agricultural pressures, and they interact with other parts of the environment in important ways. Thus, management strategies must take into account the entire spectrum of pest adaptation. As discussed above, insect adaptation to Bt toxin genes is a problem today. The following discussion of management strategies to delay insect adaptation to Bt is an example of a comprehensive approach that needs to be implemented generically for pest resistance in general. Case ExampleAdaptation to BtThere exist six basic strategies for delaying insect adaptation to plants expressing Bt toxin genes (31), each of which is appropriate in a different crop/pest system. The basic strategies are: 1. 2. 3 4. 5. 6. high expression of a Bt toxin gene with no refuges, high expression of a Bt toxin gene with refuges, high expression of two or more unrelated toxin genes with refuges, low expression of a toxin gene to slow the growth and vigor of the pest to complement natural enemies of the pest, expression of toxin genes only at times and in plant parts where protection from pest damage is required, and restricting Bt use to minor crops. These strategies for delaying adaptation to Bt are based on the same general principles of population genetics that apply to resistance to conventional pesticides. The important differences between strategies for delaying resistance to Bt toxins produced by plants, and to mechanically applied pesticides derive from inherent differences in these two toxin delivery systems. The mechanical delivery systems for insecticides usually have considerable temporal flexibility. When a scout determines that the number of insect pests in a crop is

PAGE 167

reaching an economic threshold, the information can be relayed to the farmer or crop consultant who can make the decision to spray the field with the appropriate insecticide. The farmer or consultant may have a number of insecticides on hand to choose from or can purchase them quickly. The insecticide can be applied to the field within hours if weather is not a problem and equipment and labor are available. Even in problematic cases, the insecticide can generally be applied within a few days. While there is some spatial flexibility in mechanical application procedures, it is generally not feasible only, for example, to spray plants that have two or more insects cm them. Mechanical application also permits flexibility in dosage applied. Dosage can easily be adjusted to field conditions and to the species and developmental stage of pest requiring control. The only lack of flexibility is in cost: the more you apply, the more it costs. Given insecticide decay rates in the field. doses will decrease after application and must be renewed at a cost. if needed. When the plants genetic system is used as the delivery system the situation is different. The genomes of plants and other organisms are set up to turn genes on and off as they are needed to produce specific proteins. It would not be useful for a plant to turn on a gene in a root cell if that gene was involved in producing the red pigment for flower petals. A lot of work has been conducted by molecular biologists to learn how genes are turned on and off. An important component of these switches resides in DNA sequences that flank the sequences that actually code for protein production. Some flanking sequences cause a gene to be expressed everywhere continuously; others turn the gene on only in certain plant parts; still others activate the gene only when the plant experiences a specific type of stress such as drought or attack by insects. Comments from industry (37) indicate that the first set of engineered plants to be commercialized will express Bt toxins by relying on constitutive promotors, that is, flanking sequences that activate genes under almost all conditions. This means that there will be little temporal flexibility regarding when and where a toxin is produced. In contrast to traditional pesticides, which can be applied as soon as reports of insect abundance warrant, seeds with the Bt genes must be purchased weeks or months before planting. Thus, a farmer has to assess how intense pest problems will be before a crop is even in the ground. If there is even a small chance of a pest problem and Bt seed is not too expensive, the choice will not be too hard unless the farmer has an individual concern about resistant pests. Use of Bt plants thus is generally referred to as prophylactic pest control as opposed to responsive pest control where toxins are only delivered when a problem is detected. Another difference between transgenic plants and conventional insecticide-based control programs is that the dose of a conventional pesticide can be adjusted based on need; with engineered plants the "dose" of Bt delivered is predetermined. Once the seed is in the field there is no flexibility. However, there is room for spatial flexibility in the use of engineered Bt plants. One option that a farmer has with cultivars that produce Bt continuously is to mix seed from the Bt cultivar with that of a closely related cultivar that is not resistant to pests (Strategy 1 and 2). Under certain conditions such a mixture would inhibit a pest outbreak without producing strong selection for Bt resistance. A number of models have been developed to look at this resistance management strategy, and results indicate that resistance does develop more slowly, especially if the Bt genes are recessive (29, 30). As indicated above, a number of forms of Bt toxins affect different insects. In cases where two or more distinct types of Bt toxin are available for use on one pest it is possible to have both expressed in the transgenic plant (Strategy 3). Theoretical models indicate that planting seed with two or more dissimilar toxins along with 20 to 50 percent seed that was entirely susceptible to the insect pest could preserve crop resistance 20 times longer than use of the single toxin strategy in some crop/pest systems (29, 30). There has been a good deal of work done on how partial plant resistance to insect pests could work with natural enemies of the insect pest to deter an outbreak (Strategy 4) (38). Scientists have conducted field tests with engineered tobacco that produces a low level of Bt toxin that causes about 15 percent mortality of larvae and slows the growth of survivors. The Bt was found to have no negative effect on the natural enemies of the budworm and may indeed lead to more natural enemy-induced mortality of young budworms than would otherwise be the case. This may be the result of larvae growing slower or being more restless on the plant. This low dose strategy may be a good one in some cases but not in others. Two problems that can arise are 1 ) natural enemies that cause indirect selection for adaptation to the Bt, and 2) pest genes that mediate adaptation to mild (not high) Bt stress. This later problem is considered important in the medical field where it is

PAGE 168

166 l A New Technological Era for American Agriculture sometimes advised that if antibiotics are used they should be used at high levels (9, 44, 73). Rigorous testing of the basis for this advice seems to be lacking. As indicated earlier, some genes in plants are only activated in certain plant parts at certain times (Strategy 5). Molecular geneticists have been able to move the gene activity promotors from one organism to another and basically get the same pattern of gene activation. For example a promotor region from soybeans that turns on a gene only if it is in the developing seeds cells was moved to tobacco and only turned on the gene in the tobaccos developing seed (3). Promotor sequences from tomato that only turn on adjacent genes when there is pathogen or insect stress have also been moved to tobacco and operate just as they did in the tomato (82). In some crops only certain plant parts need protection from insect damage. For example, the buds of the tobacco plant must be protected against the tobacco budworm but this insect also feeds on leaves. If the buds were protected, the budworm might switch to feeding more on mature leaves. Studies indicate that the budworm is expected to develop Bt resistance more slowly if only some plant parts express the Bt genes (36). In some crops the plants only need to be protected at certain times of the season (e. g., cotton). If Bt toxin genes were only turned on at specific times in the plants developmental cycle, the insect would experience selection pressure in one instead of three generations a year. This also should slow the development of Bt resistance. Since some plant genes are turned on only when there is tissue damage, it may be possible to find promotors that would operate like an automatic pest scout and turn on Bt genes only when a threshold of damage had occurred. Such a system would turn engineered plants from a prophylactic pest control tool into a responsive pest management tool. Such a change could significantly reduce selection for Bt resistance, especially with pests that only reach outbreak numbers once every few years. As with engineering crops for herbicide tolerance, much of the work to develop insect-resistant transgenic plants has focused on the major cash crops. This makes sense because potential industry profits are higher from working with these crops than with minor crops. If profit were not the major concern, other issues might dominate the decisions about which crops to engineer. For example, pesticides protect many small-acreage vegetable crops from insect pests up to harvest. Pesticide residues in fruits are a concern. If Bt is indeed harmless to mammals it would be useful to replace the chemical pesticides with Bt. In many cases only a small percentage of an insect pest population feeds on these minor crops, so selection for resistance to Bt would be much lower than it is in cotton or corn. If use of Bt was restricted to such crops, it would be possible to achieve long-term environmentally sound pest control (Strategy 6). Weediness of Crops With Pest Resistance Most traditional crops such as corn and tobacco are unlikely to start reproducing like weeds (i. e., uncontrollably) solely because they have pest resistance. However, semi-domesticated crops are another matter. Poplars, pine trees, and many pasture grasses and legumes can already compete well in natural habitats. Pests help maintain a balance among plant species in a pasture or forest. In mixed hardwood/pine forests, insects and pathogens are important sources of tree mortality. If a gene for insect or pathogen resistance were placed in a stand of cultured pine trees, and pollen from these trees were to reach native pines there could be a problem. Or if pine trees became resistant to their insect or microbial pests but the hardwoods did not, it is reasonable to expect a significant shift in the balance of hardwoods to pines in forest. The practical and aesthetic impact of such a change in forests must be considered. POLICY IMPLICATIONS REGARDING THE DEVELOPMENT AND DEPLOYMENT OF ENGINEERED CROPS If we maintain a laissez-faire policy regarding pest control, it is likely that developed products will be those expected to sell best. For example. farmers who have not been specifically educated about Bt-producing plants are unlikely to buy seed that produces moderately resistant plants (with hopes that natural enemies can control the rest ) if seed selling on the same shelf for an equivalent or lower price produce highly resistant plants. Only if companies exert restraint in marketing their seed will there be any potential for a multifaceted approach to resistance management. For example, if only one company has a product (such as Bt in cotton) priced such that only 50 percent of the farmers in an area decide to use it, other approaches will be adopted. When two companies have the product this is less likely to happen. Even when one company controls the market, economic analyses may dictate going for the highest volume of sales.

PAGE 169

Chapter 6Management implications of New Technologies 167 In that Bt is a naturally occurring organism that has been used by organic farmers for many years, there may be potential for regulating the use of Bt products based on resistance risk, even though synthetic chemicals have not been regulated on that basis. If it can be shown that the traditional uses of Bt would not lead to evolution of resistance as rapidly as new biotechnology approaches using Bt toxins, there may be grounds for some regulation of use. This issue is not yet resolved and the Environmental Protection Agency (EPA) does not seem to be pursuing the issue. Weed resistance problems may be somewhat different than insect resistance problems. In the case of most insects, resistance is an area-wide phenomenonwhat one farmer does affects other farmers in the region. The stage is set for a tragedy of the commons with no farmer willing to comply with practices that would help others who may be cheating. Weed seed and pollen do not move as far as most insects, so resistance can become a single-farm or even a single-field phenomenon. If one farmer overuses a herbicide and winds up with a resistance problem, other farmers who hear about it may be cautious about using that herbicide too frequently, even if it is inexpensive. If glyphosate use leads to resistance in one area of the mid-south, farmers in other areas may respond by becoming more cautious in decisions to use the product. Educational programs to point out risks to farmers would be very appropriate, and could be very effective in this case, but much research is needed to bolster the information content of such educational programs. Overall, we already have enough information to formulate general policies that prescribe judicious use of engineered crops with insect and pathogen resistance and herbicide tolerance. However, if we are to make detailed rulings about the development and use of specific products of biotechnology, we will need to generate a body of empirical knowledge relevant to these products. And, we will need an educational program designed to bring these results to the farmer and the public. A NEW ISSUE IN ANIMAL AGRICULTURE MANAGEMENT The use of new animal technologies will place a premium on the management capabilities of livestock producers. Research results clearly show the extent of response achieved depends heavily on the management capability of the producer. Use of somatotropins, for example, may require altering the animals diets. Growing pigs receiving somatotropin will require diets high in protein, and with adequate levels of the necessary amino acid, lysine. Administration of somatotropin to lactating cows may require extending the reproductive cycle to 14 months instead of using the current 12-month cycle. The availability of many different types of growth promotants may result in the use of more than one at the same time. Compatibility of these promotants will be an important management issue. Thus, producer management skills are critical to the optimal use of these technologies. As important as these management issues are, a more pressing management issue is that of animal welfare with or without biotechnology as a complicating factor. Society has focused on many of the resulting impacts of technologies such as environmental quality, food safety, and decline of the small farm and rural communities. Farm animal well-being is the most recent concern to receive attention. Much of the success in increased productivity in agriculture has been the result of lowered costs through the use of confinement systemswhich some have coined factory farming. The question from an animal welfare perspective is whether we have gone too far. Farm Animal Well-Being In the decade of the nineties, the advance of new animal technologies will coincide with increasing interest in farm animal well-being. This interest is not new. It nucleated in England at the turn of the 19th Century with the formation of the Royal Society for the Prevention of Cruelty to Animals. This in turn led to the organizing of more radical groups. In America, the American Society for the Prevention of Cruelty to Animals was formed in the 1860s by a Special Act of the New York State Legislature. However, it was not until the late 1970s and early 1980s that the majority of animal welfare/rights organizations were formed. Although no specific records are kept, estimates indicate that today there are a total of 7,000 animal welfare/rights groups in the United States with a combined total budget of $50 million (81). Widespread public concern for farm animals began to develop in 1963 with the publication of Animal Machines. This book by Ruth Harrison (46) chronicled the problems in farm animal well-being in the United Kingdom that led to the Brambell commission and its report enunciating the famous Five Freedoms to lie down. stand up, turn around, stretch, and groom. Concern built steadily in Europe, and in 1979 the first European meeting on farm animal welfare was held. European governments have allocated significant public funds to research on alternative farm systems and the European

PAGE 170

168 A New Technological Era for American Agriculture Community (EC) has supported numerous symposia on the well-being of various farm animals. Legal protection for farm animals includes far-reaching laws in Sweden and Switzerland. In the United States the level of concern has grown more slowly. However, in the past few years the pressure on farmers and animal scientists to address the issue of farm animal welfare has increased steadily. The issue of farm animal welfare has provided important impetus to a movement that may eventually be considered as significant by policy makers as that for environmental and food safety concerns. Today, the issues of animal welfare/rights foster well-entrenched polar positions. The polarity between the agricultural establishment and animal well-being advocates has highlighted the extremes of each groups position. Economics, values, and institutions determine care and treatment of farm animals. These factors divide into two animal welfare paradigms: the traditional and the alternative. Which paradigm will dominate future public policy for animal welfare remains to be seen (94). The Traditional Paradigm Those who hold the traditional paradigm of animal welfare draw on the market model of free enterprise, and on Judeo-Christian ethics. The Market ModelAdvocates of the market model argue that farm animals subject to cruelty and neglect give fewer eggs and less milk, meat, or wool than welltreated and properly cared for animals. Why not, they ask, depend on profits to ensure farm animal welfare? Quantifiable variables such as feeding efficiency, rate of growth or productivity, morbidity, and mortality rates can provide proxy measures of animal welfare. Favorable values for those objective measures of humane treatment for the most part are consistent with good management and high profits. Advocates of the market model further argue that confinement systems improve some dimensions of animal welfare. Temperature, disease, and pest control are improved. Predators are kept away. Nutrition is enhanced. Modern farming systems have lowered costs and expanded utilization, allowing more animals to exist. The Judeo-Christian EthicAdvocates of the traditional paradigm hold the Judeo-Christian ethic that God created man in his own image, that man is unique in having a soul, that man has dominion over animals, and that man as husbandman and steward of Gods kingdom is not to practice cruelty to or neglect animals (77, 86). Many advocates of this position hold that no element of society has more compassion for poultry and livestock than does the farmer (45). Other than laws protecting animals from cruelty and neglect, advocates of this view consider laws, rules, and regulation on care and treatment of farm animals to be unwarranted infringement on free enterprise. This creed holds that 1 ) proprietors deserve the right to prescribe rules under which they operate; and 2) a prime function of government is to prevent anyone, including the government, from infringing on the managerial freedom of proprietors (5). Some traditionalists will admit that, despite market incentives, cruelty-neglect laws, and producers with the Judeo-Christian ethic, animal welfare falls short of the ideal. But they contend that Big Brother intrusions of an expensive and often incompetent bureaucracy into managerial prerogatives of farmers would entail more social cost than the abuses government is attempting to correct. They favor minimal policy intervention consistent with the traditional paradigm as the lesser of two evils. Alternative Paradigm An increasing number of people reject the Judeo-Christian ethic and market paradigm in favor of an alternative paradigm emphasizing animal rights or much enhanced animal welfare. As with the traditional paradigm, the alternative has economic and ethical dimensions. Market FailureAnimal welfare has public goods properties, implying that the market alone will not bring the proper level of animal welfare. Externalities are apparent: all the public benefits from seeing livestock freely grazing in a meadow. Animal rights activists contend that the market results in confinement cages allowing too little space per animal for laying hens, sows, and veal calves. The drive to reduce costs and cater to consumer demand has kept veal calves isolated, in the dark, and on low iron diets; has disfigured animals, by encouraging practices such as trimming chickens combs and beaks and pigs and lambs tails. According to activists, animals are not allowed their natures socialization, sex, exercise, nest building, nurturing of offspring, the outdoors, and a full life. However. the role of markets in shaping the way farm animals are raised cannot be denied. Market forces have raised real prices of land and labor, and reduced the relative price of capital. Rising labor and land prices have placed a premium on labor-saving and land-saving meth-

PAGE 171

Chapter 6Management Implications of New Technologies 169 ods of production. Gains in income and population along with changes in production technologies, including disease control, have interacted with prices to create economies of size and to make confinement systems feasible. Small may be beautiful but it is frequently not competitive. The small-scale poultry operator is nearly extinct; the small Wisconsin dairy has difficulty competing with the large industrial-type California dairy farm; and the family hog farm in Iowa has difficulty competing with the large confinement operations in Arkansas. Animal welfare enthusiasts view these outcomes of market forces as a disaster to farm livestock and to traditional farmers, rather than as a means toward cheaper food. more land for urban use, and higher income for the Nation. EthicsThe alternative paradigm views man as an evolutionary product of a holistic Nature. Man is one with nature and must live in harmony with plants and animals. If he has primacy, it is to be used to ensure the rights of the rest of nature. Philosopher Jeremy Benthams (4) much-quoted comment summarizes the basis for the ethical treatment of animals under the alternative paradigm: The question is not Can they reason? nor, Can they talk? but, Can they suffer? Animals that are sentient (can experience pleasure or pain) are to be afforded rights given to people. Killing an animal is murder and eating its flesh is cannibalism. Hard-core animal rights adherents have little alternative to vegetarianism. Other advocates do not go that far but insist on improving animal welfare through provision for each species nature. Animal suffering and pain is probably the most powerful rationale for the publics concern over farm animal welfare. This concern must be addressed by objective research. Research Needed To understand and fulfill agricultural animals needs, more must be learned about their fundamental psychological and behavioral processes. Researchers must be able to elucidate farm animals cognitive and motivational processes before it is possible to begin to answer such rudimentary and obvious questions about their wellbeing such as: How does this animal feel in one environment versus another? Is the animal sufferingand if so. how much? For example, when the animals farm environment is devoid of a particular feature that would characterize its natural environment, does the animal sufferand if so, how much (11, 12)? The scientific community generally has been slow to accept the notion of animal awareness and only recently has such recognition been forthcoming. Many in agriculture now acknowledge that animals are aware of themselves and their surroundings, and thus scientists are beginning to give attention to animals conscious sensations of well-being. Only recently have factors that affect conscious well-being been considered logical criteria for the design of animal accommodations. However, there exists little hard data on which to base such a design strategy. How an animal feels, some assume, depends largely on how it expects to feel. How it expects to feel in turn depends on how it thinks, remembers, and imagines. How an animal feels also depends on factors such as the predictability and controllability of its environment ( 100). Feeling, thinking, remembering, and imagining are cognitive processes. To the extent that feeling and thus, thinking, remembering, and imagining affect an animals overall well-being, and therefore its health and productivity, these cognitive processes are factors to be considered in the economic and humane production of agricultural animals. There is reason to believe that when an animal experiences a feeling of malaise, its productivity is reduced, if only slightly. However, such decrements are cumulative; and together they can reduce productivity significantly. In the chicken, for example, there is recent evidence that as many as six stressorsammonia, beak trimming, coccidiosis, electric shock, heat, and noisecan combine in additive fashion to affect feed intake, growth, and several important physiological and pathological traits (64). In addition, stressors and combinations of stressors occurring in various sequences affect productive performance of chickens in predictable, repeatable ways (52). This linear additivity of stressor effects on such a variety of traits suggests that some single phenomenon is governing the animals overall response. This could be psychological stress. The following discussion depicts some of the production practices that animals encounter and areas of research that are needed ( 12). Thermal Comfort Little is known about the perception of thermal comfort by farm animals ( 10). Animals do respond to changing conditions in their thermal environment with different thermoregulatory behaviors. But the degree to which animals suffer when experiencing heat stress or cold stress is not known. One experiment to find the answer to cold stress of farm animals is cur-

PAGE 172

170 l A New Technological Era for American Agriculture Photo credit: University of Illinois An example of a thermal comfort experiment involving pigs operating a heat switch. The sitting pig presumably because it felt the environment was too coolhas just operated the switch in the panel to engage the heater. rently underway involving pigs operating a heat switch when they feel cold. Another thermoregulatory behavior response is wallowing by swine under heat stress. Wallowing in mud compensates for the pigs absence of thermal sweating. Research has shown that sows wallow only when environmental temperature exceeds some threshold (e.g., 12 C for sows in one experiment) (83). This limited research suggests that swine wallow only to achieve thermal comfort, not because they need to wallow or enjoy wallowing as play. If the thermal environment is maintained below 12 C all the time, sows never take advantage of a mud wallow even if it is provided. Quality of Space-The richness of an environment is somehow perceived by animals because it affects how they behave and function. The behavior repertory of swine in natural settings is larger than it is in typical production environments (88). When contemporary production environments are furnished with enriching features, pigs readily make use of these features and thereby expand their behavior repertories. Nehring (71) built a maze in a pig pen. McGlone and Curtis (67) provided pigs hiding places for their heads allowing them to submit to and subsequently avoid an aggressive pen mate. Fraser provided pigs a mezzanine for use in getting away from group mates ( 19, 74). Grandin (40) enriched pig environments with suspended manipulanda (pig toys). Pigs reared in enriched environments proved easier to be moved about than pigs in traditional production environments (43). Pigs residing in pens equipped with suspended manipulanda fouled their feeder markedly less often than did those in a relatively barren environment (92). From the above, it might be inferred that animals in richer natural or artificial environments behave differently and experience an enhanced sense of well-being compared to those in more barren surroundings. But this has not been determined scientifically to be the case, and many questions persist. For example, do pigs enjoy a higher sense of well-being when able to use enriched features? Are they starved for stimulation in less rich environments? If so, does this lead to a craving for stimulation? Commercial gilts and sows often reside during pregnancy in rectangular crates that prevent them from turning around. When living in a crate shaped so as to permit her to turn around, a pregnant gilt will turn around approximately 13 times daily in a crate 61 cm wide, but only 9 times daily in a 56 cm wide crate (in which it is more difficult for the gilt to turn around) (63). Little is known about what motivates a gilt to turn around. Does she need to turn around? Does this need affect her productive performance? How an animal perceives its living space may be crucial to its sense of well-being. Sometimes space can be modified physically or rearranged so as to make it more accommodating to the animal. For example, animals in pens have a propensity to keep their heads at or to lie around the perimeter of a pen instead of in the middle (39, 90). A triangle has 28 percent more perimeter and a square 13 percent more than a circle of equal area. Thus, of the three, triangular pens maximize the ratio of perimeter to area. Should animal facilities be built with triangular pens and cages instead of rectangular ones to enhance the animals comfort? Is it necessary to have more space in a rectangular pen to engender the same feeling of well-being that an animal would experience in a square pen of equivalent perimeter? Learned Helplessness Animals often encounter frustrating situations and presumably these may decrease their well-being. For example, when anything gets in the way of an animal on its way to the feeder to eat, that animal becomes frustrated. Frustration is one of the prepathological states indicative of stress (69). Frustrating situations generally are stressful, as indicated by various physiological indicators ( 13).

PAGE 173

Chapter 6Management Implications of New Technologies 171 Farm animals may be frustrated when engaged in any strongly motivated behavior pattern, whether eating, nesting, and engaging in sexual activities, among others. Depending on the circumstances, for example, frustrated hens may show displacement behaviorbehavior patterns that occur out of context with preceding and succeeding behavior ( 16). In other settings, an animal may find that it can neither control its environment nor predict what its environment will be, and the animal may learn to act in a helpless manner. In a state of learned helplessness, an animal stops initiating behavior aimed at controlling or making use of environmental features because it has learned to expect that these features are uncontrollable and that these attempts would be futile (84). Animals residing in certain intensive production systems might well learn to expect that they have little or no control over their surroundings. It is possible that agricultural animals living in certain housing systems may develop learned helplessness ( 14, 61. 62). Learned helplessness would be another of the prepathological states indicative of stress (69). NestbuildingFemales of all domestic avian species build nests in which they lay their eggs. The domestic hen will engage in nest-building every day, even when a previous nest exists. It seems that the performance of nest-building is itself positively reinforcing to the hen (50). Most sows attempt to construct a farrowing nest beginning 12 to 16 hours prior to delivering the first pig, regardless of where they are (51). In many modern farrowing environments, there is neither the space in which to conduct nest-building behavior nor the material with which to build a nest. Sows nevertheless direct substantial amounts of time toward small amounts of material even though a nest may not result. This suggests that for the sow, as for the hen, nest-building behavior in itself is rewarding (99). Research is needed to answer such questions as: Do hens and sows need to build nests? How much frustration do they experience when they either cannot move enough material to nest-build or cannot find nesting material? How do they feel when they cannot build a nest? Does this feeling in sows result in hormonal changes that are an anathema to oxytocins actions in birth and lactation? Electro-Immobilization Animal may find certain procedures routinely performed in agricultural production to be uncomfortable or even painful. When an animal Photo credit: University of Illinois The sowin anticipation of delivering a litter of piglets within a few hoursis building a maternal nest to protect the piglets from cold and predators. actively avoids a procedure it is presumably revealing negative feelings about the procedure. Ewes having experienced restraint by electro-immobilization and by a squeeze-tilt table, when given the choice between the two in a Y-maze avoid-avoid test, chose the squeeze-tilt table 79 percent of the time, and the electro-immobilizer 13 percent (42). Questions that need answers include: What was the ewe thinking as she hesitated at the decision point, indicating by her head movements that she is vacillating? Was she actually imagining the feeling she experienced during electro-immobilization earlier? Based on the ewes reactions, when should the electro-immobilizer not be used? What behavior indicators identify the point beyond which it would be inhumane to continue subjecting the ewe to the procedure? Chicken-Harvesting Machine Animals can adapt in a matter of seconds to machines with which they are forced to interact, provided that the machines are designed with the animals nature in mind. Take, for example, the chicken harvesting machine developed in the United Kingdom. The harvesting of birds from growing houses is a monumental task. Moreover, considerable

PAGE 174

172 l A New Technological Era for American Agriculture losses are incurred in the process of harvesting and transportation, especially in the hand-catching and hand-crating processes (25). A prototype chicken-harvesting machine has been evaluated in terms of the stressfulness of the harvesting process ( 15). By means of electrocardiograms and immobility tests, it has been found that the stress from harvesting could be reduced by catching and picking up broiler chickens with a carefully designed machine, rather than by hand. Heart rate dropped back to normal more quickly and duration of tonic immobility (a phenomenon that increases with fear) was much shorter in machineharvested birds than in those caught by hand. Research questions include: What is a chicken thinking when it is manually caught by one leg and carried upside down to the crate in which it will be transported to the processing plant? How does this contrast to what it is experiencing when it is caught by the long rubber fingers of a chickenharvesting machine, moving it onto a moderately inclined conveyer belt, which it rides to the gathering stage? Double-Rail Restrainer Conveyor SystemMeans of rapidly moving large numbers of animals of all kinds are needed in the production and processing industries. The V-restrainer, in which animals are moved along and wedged between two v-angled conveyor belts, with their legs dangling, is a vast improvement over driving animals through a chute, but it gives rise to additional problems. A prototype of this system was developed in the late 1970s, and it caused little premortem stress in animals when used in a processing plant (98). The system was further developed for applications ranging from veal, lamb, and swine slaughter lines to feedlot cattle processing. When designed specifically for the species and size range to be handled, the animals apparently find the conveyer belt comfortable to ride. Adjustable sides prevent the animal from leaning sideways which is important because tilting sideways seems to frighten the animal. As the above discussion illustrates, there are many questions to be answered regarding animal welfare. Of particular importance is the effect of animal well-being on the animals performance. Some research seems to indicate that the amount of psychological stress an animal experiences determines how the pituitary-adrenal axis responds. In other words, psychological stress may be reducing the animals performance as well as the animals well-being. Much more research is needed to understand such relationships. To date, little research has been done in the United States on animal well-being. Biotechnology and Farm Animal Well-Being In the past few years, animal protection groups have begun to voice concerns about biotechnology. Their concerns are rather diffuse and it is difficult to determine precisely what could be done to address those concerns. The new techniques for manipulating genetic material strike at some deep-seated fears amongst animal protection groups. While there are few concise papers explaining animal protection concerns, a reading of the relevant literature leads to the identification of the following issues: reinforcing notions of animals as mere property to be manipulated at the whim of human owners, and animal well-being issues (81). Manipulation of Property Genetic engineering conjures up images by some in the animal protection movement of animal machines being reconstructed by ingenious scientists to meet human needs. The push to be allowed to patent animals (discussed in ch. 15) merely reinforces the idea of animals as patentable machines. At a time when the animal movement is pushing to increase the moral status of animals to, at the very least, something between persons and property, the biotechnology era and patenting seem to be a major step backwards.

PAGE 175

Chapter 6Managenwnt Implications of New Technologies l 173 Animal Well-Being Issues The impact of biotechnology on animal well-being is probably the most challenging issue genetic engineering raises. The technology is most likely impact-neutral in that one could use biotechnology to improve animal wellbeing (e.g., engineer disease resistance, eliminate detrimental genes from a population) as well as compromise it. The clearest example of compromised well-being is the "Beltsville pig (discussed in ch. 3). This pig is the result of research at the U.S. Department of Agriculture (USDA) in Beltsville that involved the insertion of extra growth hormone genes. When the extra genes were expressed, the animal grew fast but, as it gained weight, it became lame and lethargic and suffered from degenerative joint disease and a variety of other disorders (41 ). There is little doubt that the animal was under stress as a result of the genetic manipulation. Questions also have been raised about the quality of life for the oncomouse and some of the other mice that have been developed to shorten the time of standard carcinogen and mutagen tests. It is also possible, however, that some genetically engineered animals might reduce the need for research animals and hence qualify as alternatives. Among farm animals, moreover. it may be possible to use genetic engineering to eliminate the horn gene in cattle, thereby removing the welfare problems associated with dehorning (4 1 ). While some object strongly to the proposal that farmers should create breeds of microcephalic (small brained) farm animals that are quite content in close confinement (41 ), others say that as long as the animal is in a state of positive well-being, such a creation would not be morally objectionable though there may be some esthetic problems with such creatures (79). To date, there has been little discussion or debate of these questions. and about the most that can be concluded at this stage is that careful monitoring of transgenic animals to determine their state of well-being is essential. As more experience and research with transgenic animals takes place, it will be possible to develop more sensible guidelines and conclusions. Biotechnology is a priori neither good nor bad for animals. Its impact depends on what is done and its effect. If it is used judiciously to benefit humans and animals, with foreseeable risks controlled, and the welfare of the animals is kept in mind, it is morally defensible and can provide great benefits. CHAPTER 6 REFERENCES 1. Barrett, S., Crop Mimicry in Weeds, Economic Botany 37:255, 1983. 7 k. 3 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. Beachy, R.N. et al., Genetic Engineering of Plants for Protection Against Virus Diseases, Plant Resistance m Viruses. D. Ever and S. Harnett (eds. ) (New York, NY: Wiley & Sons, 1990), pp. 151158. Benfey, P.N. and Chua, N. H., Regulated Genes in Transgenic Plants. Sciencr 244: 1 74 181, 1989. Bentham, J., The Work.v of JeremJY Bet~tilam, .I. Bowring (cd.), vol. 1 (New York, NY: Russell and Russell, 1962). pp. 142-143. Brewster, J.. Societ y Values and Goals in Respect to Agriculture, Chapter 6 in Goals ~~nd VLllue.Y in Agricultural Policj (Ames. 1A: IOWa State University Press, 1961 ). Burch, T., Professor, Cooperative Extension Program, Louisiana State University. Baton Rouge, LA, personal communication, Dec. 20, 1990. Coble, H., Professor, Crop Science Department, North Carolina State University, Raleigh, NC, personal communication, Dec. 18 and 20, 1990. Comai, L. and Stalker, D. M., Impact of Genetic Engineering on Crop Protection, Crop Pro?. 3:399 408, 1984. Curtis, C.F. and Otoo, L. N., A Simple Model of the Build-Up of Resistance to Mixtures of AntiMalarial Drugs, Trans. Roy. Sot. Trop. Med. & kl~g. 80:889-892, 1986. Curtis, S.E. Perception of Thermal Comfort by Farm Animal s, Farm Animal Welfilre and Holt.Ying, S. H Baxter, M.R. Baxter, and J. A.D. MacCormack (eds. ) (Boston. MA: Mafiinus Nijhoff, 1983), pp. 59-66. Curtis, S., Environment and Animal Behavior, commissioned background paper prepared for the Office of Technology Assessment, 1991. Curtis, S. and Stricklin, R., The Importance of Animal Cognition in Agricultural Animal Production Systems: An Overview, a paper presented at the symposium, Cognition and Awareness of Animals: Do Farm Animals Perceptions Affect Their Production or Well-Being? 1991. Dantzer, R. and Mormede, P., Can Physiological Criteria be Used to Assess Welfare in Pigs? The Welftlre of Pig.~, W. Sybesma (cd. ) (Boston. MA: Martinus Nijhoff, 1981), pp. 53. Deaton, R., Monsanto Agriculture Co., St. Louis. MO, personal communication, Dec. 18, 1990. Duncan, 1.J. H. et al., Comparison of the Stressfulness of Harvesting Broiler Chickens by Machine and by Hand, Br. Poultr. Sci. 27: 109, 1986. Duncan, 1. J. H., The Welfare of Farm Animals: An Ethnological Approach, Sci. Prog. 71:317, 1987. Fincher, R., Pioneer, personal communication. Dec. 21, 1990.

PAGE 176

174 l A New Technological Era for American Agriculture 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. Fox, M. W., Farm Animals: Husbandry, Behavior, and Veterinary Practice (Baltimore, MD: University Park Press, 1984), pp. 237. Fraser, D., Phillips, P. A., and Thompson, B. K., A Test of a Free-Access Two-Level Pen for Fattening Pigs, Anim. Prod. 42:269, 1986. Fraser, R. S. S., Genetics of Plant Resistance to Viruses, Plant Resistance to Viruses, Ciba Found. Symposium /3.?, D. Evered and S. Hamett (eds. ) (New York, NY: Wiley & Sons, 1990), pp. 6-15. Frisbee, R.E. et al., Implementing 1PM in Cotton, Integrated Pest Management Systems and Cotton Production, R.E. Frisbee, K.M. E1-Zik, and L.T. Wilson (eds. ) (New York, NY: J. Wiley & Sons, 1989), pp. 389-412. Gallun, R. L., Genetic Basis of Hessian Fly Epidemics, Ann. N.Y. Acad. Sci. 287:223, 1977. Gelemter, W. D., Targeting Insecticide-Resistant Markets, Managing Resistance to Agrochemicals, ACS Symposium Series 421, M. B. Green, H.M. LeBaron, and W.K. Moberg (eds. ), Am. Chem. Soc., Washington, DC, 1990, pp. 105117. Georghiou, G. P., Overview of Insecticide Resistance Managing Resistance to Agrochernicals, ACS Symposium Series 421, M.B. Green, H.M. LeBaron, and W.K. Moberg (eds. ), Am. Chem. Soc., Washington, DC, 1990, pp. 18-41. Gerrits, A. R., de Koning, K., and Migchels, A., Catching Broilers, Poultry 1(5):20, 1985. Goldberg, R., Environmental Defense Fund, New York, NY, personal communication, December 1990. Goldberg, R. et al., Biotechnologys Bitter Harvest, Biotech. Work. Group, Rural Advancement Fund International, Pittsboro, NC, 1990. Goodman, R. M., Future Potential, Problems, and Practicalities of Herbicide-Tolerant Crops from Genetic Engineering, Weed Sci. 35:28, 1987. Gould, F. Simulation Models for Predicting Durability of Insect-Resistant Germ Plasm: A Deterministic Diploid, Two-Locus Model, Environ. Entomol. 15: 1 10, 1986a. Gould, F., Simulation Models for Predicting Durability of Insect-Resistant Germ Plasm: Hessian Fly (Diptera: Cecidomyiidae)-Resistant Winter Wheat, Environ. Entomol. 15: 11, 1986b. Gould, F., Evolutionary Biology and Genetically Engineered Crops, Bioscience 38:26, 1988. Gould, F., Managing Pest Adaptation to Genetically Engineered Crops, pp. 80, in National 1PM Coordinating Committee, 1989, Proceedings National Integrated Pest Management Symposium/ Workshop, Las Vegas, Nevada, Apr. 25-28, 1989, Communications Service, NY AES, Cornell Univ., Geneva, NY., 1989, 276 p. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44 45. 46. Gould, F., Ecological Genetics and Integrated Pest Management, Agroecolog~, C.R. Carroll, J.H. Vandermeer, and P.M. Rosset (eds. ) (New York, NY: McGraw-Hill Publishing Co., 1990), pp. 441-458. Gould, F., Evolution of Resistance to Toxic Compounds by Arthropods, Weeds and Pathogens, commissioned background paper prepared for the Office of Technology Assessment, 1991. Gould, F., The Evolutionary Potential of Crop Pests American Scientist, vol. 79, November/ December, 1991. Gould, F. and Anderson, A., Effects of Bacillus thuringiensis and HD-73 Delta-Endotoxin on Growth, Behavior, and Fitness of Susceptible and ToxinAdapted Heliothis virescens (Lepidoptera: Noctui&e) Strains, Environ. Entornol. 20:30-38, 1991. Gould, F. and Weissinger, A., Roles for Public and Private Sector Scientists in Developing PestResistant Crops, New Directions in Biocontrol, P. Dunn and R. Baker (eds. ), Proc. Univ. California Los Angeles Molecular Biology Symposium, 1990, pp. 641-648. Gould, F. et al., Feeding Behavior and Growth of Heliothis virescens Larvae (Lepidoptera: Noctuidae) on Diets Containing Bacillus thuringiensis Formulations or Endotoxins, Entomol. E.rp. Appl. 58:199-210, 1991. Grandin, T., Livestock Behavior as Related to Handling Facilities Design, lnt. J. Stud. Anim. Prob. 1 :33, 1980. Grandin, T., *Double Rail Restrainer Conveyor for Livestock Handling, J. Agric. Eng. Res. 41:327, 1988, Grandin, T., Biotechnology and Animal Welfare, Animal Biotechnology?* and Qualih of Meat Production, L.O. Fiems, B.G. Cottyn, and D.1. Demeyer (eds. ) (Amsterdam: Elsevier, 1991), pp. 145-157. Grandin, T. et al., Electro-immobilization versus Mechanical Restraint in an Avoid-Avoid Choice Test for Ewes, J. Anim. S(i. 62: 1469, 1986. Grandin, T. et al., Richness of Pigs Environment Affects Handling in Chute, J. Anim. Sci, 63( Suppl. 1): 161 (Abstr. ), 1986. Grogl, M. et al., Leishmania spp.: Development of Pentostam-Resistant Clones in vitro by Discontinuous Drug Exposure, Exp. Parasite/. 69:78 90, 1989. Guither, H. and Curtis, S., Changing Attitudes Toward Animal Welfare and Animal Rights, FS8, The Farm and Food S>stem in Transition (East Lansing, Ml: Cooperative Extension Service, Michigan State University, 1983). Harrison, R., Animal Machines (London, England: Stuart, 1964).

PAGE 177

Chapter fi-kfanagement Implications of NeMY Tt~c.ll\l~~l~~,qi(~.s l 175 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. Headley, J.C. and Hey, M. A., b Benefit/Cost Analysis of an Integrated Mite Management Program for Almonds, J. Ecwn. Entomol. 80:555559, 1987. Heinricks, E.A. et al., Susceptibility of Rice Plant Hoppers, Nilapar\wta lugens and Sogatella ji4r(i/i)ra ( Homoptera: I)elphacidae) to Insecticide as ]ntlucnced by Level of Resistance in the Host Plant. En}iron. Entomol. 19:455-458, 1984. Hey, M. A., Almonds (California): Integrated Mite Management for California Almond Orchards. Spider Mites, Their Biolog>, Natural Enemies [Jnd Control, W. Helle and M.W. Sabelis (eds. ) (Amsterdam: Elsevier, 1985), pp. 299 10. Hughes, B. O., Duncan, I. J. H., and Brown, M. F., The Performance of Nest Building by Domestic Hens: 1s It More Important Than the Construction O( a Nest? Anim. Behatt. 37:2 10, 1989. Hurnik, J. F., A Review of Periparturient Behavior in Swine, Can. J. Anim. Sci. 65:777, 1985. Johnson, R. W., Curtis, S. E., and Shanks, R. D., Effects on Chick Performance of Ammonia and Heat Stressors in Various Combination Sequences, Poulf. Sci. 70: 1132, 1991. Kennedy, G., -Trldecanone, Tomatoes and ~e_ Iio[his ceu: Potential Incompatibility of Plant Antibiosis with lhsecticidal Contro l, Entomol. Exp. & Appl. 35:305-3 I I 1984. Kennedy, G., Department of Entomology. North Carolina State University, Raleigh. NC, personal communication, August 1990 and October 1990. Kennedy, G. G.. Farrar, R. R., and Riskallah, M. R., Induced Tolerance of Neonate Heliothis zea to Host Plant Allelochemicals and Carbaryl Following Incubation of eggs on Foliage of L]copersicon hirsutum f. glabrutum, Oecologia 73:615-620, 1987. Kirsch, K. and Schmutterer, J.. Low Efficacy of a Baci//us thuringiensis (Ber. ) Formulation in Controlling the Diamondback Moth, Plutella .rylostella (L.), in the Philippines, J. Appl. Entornol. 105:249255, 1988. LeBaron, H.M. and McFarland, J., Herbicide Resistance in Weeds and Crops, Managing Resistance to Agrochemicals, ACS S>mposiurn Series 42/, M.B. Green, H.M. LeBaron, and W.K. Moberg (eds. ), Am. Chem. Soc., Washington, DC., 1990, pp. 336-352. LeBaron, H., Ciba Geigy Corp., Greensboro, NC, personal communication, March 1991. Leemans, J. et al., Insecticidal Crystal Proteins from Bacillus thuringiensis and Their Use in Transgenic Crops, New Directions in Biological Control, R.R. Baker and P.E. Dunn (eds. ), Proceedings of Univ. California Los Angeles Molecular Biology Symposia, 1990, pp. 573-588. 60. 61, 6~ 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73, 74. 75. Marrone, P.G. et al., Discovery of M icrobial Natural Products as Sources of Insecticidal Genes, Novel Synthetic Chemistry. or Fermentation Products. Bi{~t[~(ht~{~l~jg~, Biolo* and Pest Management (New York. NY: MacMillan Publishing Co., 1989). Peters, W., b The Problem of Drug Resistance in Malaria, Parasitology 90:705-715, 1985. Phillips, P. A., Thompson, B. K., and Fraser, D., Preference Tests of Ramp Design for Young Pigs, Can. J. Anim, Sci. 68:41, 1988. Powles, S.B. and Howat, P. D., Herbicide-Resistant Weeds in Australia, Weed Technol. 4: 178185, 1990.

PAGE 178

176 l A Ne}~~ Technological Et-u-for American Agriculture 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. Raymond, M. et al., Nature 350: 15 1, 1991. Regan, T. and Singer, P., Animal Rights and Human Obligations (Englewood Cliffs, NJ: Prentice Hall, 1989). Reichard, D.L. and Ladd, Jr., T. L., Pesticide Injection and Transfer Systems for Field Sprayers Trans ASAE, 1983, 26 p Rollin, B. E., The Frankenstein Thing: The Moral Impact of Genetic Engineering of Agricultural Animals on Society and Future Science, Genetic Engineering of Animals: An Agricultural Perspective, J.W. Evans and A. Hollaender (eds. ) (New York, NY: Plenum Press, 1986), pp. 285-297. Roush, R.T. and McKenzie, J. A., Ecological Genetics of Insecticide and Acaricide Resistance, Anw. Rev. Entomol. 32:361, 1987. Rowan, A., Agricultural Animal Welfare issues commissioned background paper prepared for the Office of Technology Assessment, 1991. Ryan, A., Proteinase Inhibitor Gene Families: Strategies for Transformation to Improve Plant Defenses Against Herbivores, BioEssa}s 10:20, 1989. Sambraus, H. H., Das Suhlen von Sauen, Deutsches Tieraerztliche Wochenschrift 88:65, 1981. Seligman, M. E. P., Helplessness (San Francisco, CA: W.H. Freeman, 1975). Shelton, A., New York State Agriculture Experiment Station, Geneva, NY, personal communication, Dec. 21, 1990. Singer P., Animul Liberation (New York, NY: Avon Books, 1975). Stalker, D. M., McBride, K. E., and Malyj, L. D., Herbicide Resistance in Transgenic Plants Expressing a Bacterial Detoxification Gene, Science 242:419-423, 1988. Stolba, A., A Family System in Enriched Pens as a Novel Method of Pig Housing, in Altet-natites to lntensi~~e Husbandn S>stem.s, Univ. Fed. for Animal Welfare, Potters Bar, UK, 1981. pp. 52. Stone, T. B., Sims, S. R., and Marrone, P. G., Selection of Tobacco Budworm for Resistance to a Genetically Engineered P.selt~i{)r?l[jtl{l.y .fll!orc~.~([~t~.j Containing the Delta-Endotoxin of Bacillus thuringiensis subsp. hwrstaki, J. In}w-t. Pathoi. 53:228234, 1989. Stricklin, W. R., Graves, H. B., and Wilson, L. L., Some Theoretical and Observed Relationships of 91. 92. 93. 94. 95. 96. 97. 98. 99. I 00. 101. 102. Fixed and Portable Spacing Behavior of Animals, Appl. Anim. Ethel. 5:201, 1979. Tabashnik, B. E., Finson, N., and Johnson, M. W., Managing Resistance to Bacillus thur-ingiensis: Lessons from the Diamondback Moth (Lepidoptera: Plutellidae), J. Econ. Entornol., 1990. Taylor, 1. A., Grandin, T., and Curtis, S. E., Pig Toys: Effects on Feeder-Fouling and Dunging Pattern J. Anim. !$ci. 63( Suppl. 1): 161 (Abstr. ), 1986. Teetes, G. L., Becerra, M.1., and Peterson, G. C., Sorghum Midge (Diptera: Cecidormyiidae) Management with Resistant Sorghum and Insecticide, J. Econ. Entomo/. 79: 109 I1095, 1986. Tweeten, Luther, public Policy Decisions for Farm Animal Welfare, paper presented to International Conj?rence on Farm Animal Weffare, June 1991. U.S. Congress, Office of Technology Assessment, Beneath the Bottom Line: Agricultural Approaches to Reduce Agrichemical Contamination of Ground}t~ater, OTA-F-417 (U.S. Government Printing Office, Washington, DC: May 1990). Van Rie, J. et al.. Mechanism of Insect Resistance to the Microbial Insecticide, Bacillus thuringiensis. Science 247:72, 1990. Weller, D. M., Biological Control of Soilborne Plant Pathogens in the Rhizosphere with Bacteria, Ann. Re\. Phytopathol. 26:379-407, 1988. Westervelt, R.G. et al. l Physiological Stress Measurement During Slaughter in Calves and Lambs, J. Anim. S[i. 42:831, 1976. Widowski, T.M. and Curtis, S. E., The Influence of Straw, Cloth Tassel, or Both on the Prepartum Behavior of Sows, Appl. Anim. Behav. Sci. 27:53, 1990. Wiepkema, P. R., Behavioral Aspects of Stress, in P.R. Wiepkema and P.W. M. Adrichem (eds. ), Biolog? qf Stress in Farm Animals: An lntegrati~+e Approach (Boston, MA: Martinus Nijhoff, 1987). pp. 113-133. Wilson, L. T., Estimating the Abundance and Impact of Arthropod Natural Enemies in 1PM Systems Biological Control in Agricultural lPi14 S?stems. M.A. Hoy and D.C. Herzog (eds. ) (Orlando, FL: Academic Press, 1985), pp. 303-322. Zalom, F,G. et al., Sampling Mites in Almonds: 11. Pr(~.~etl~e-Al>.~etl(t~ Sequential Sampling .fbr Te tran>(hus Mite Species 52: 14, 1984.

PAGE 179

Part III Environmental Quality

PAGE 180

Chapter 7 Environmental Issues: Institutions and Their Regulatory Roles Photo credit Jamie Notter, OTA Staff

PAGE 181

Page INTRODUCTION . . . . . . . . . . . . . . . . . . . 181 preface . . . . . . . . . . . . . . . . . . . . . 181 Agriculture, Field Trials, and Deliberate Release of Genetically Engineered Organisms . 183 Brief Overview of Concerns . . . . . . . . . . . . . . . . 183 Evolution of Regulation and Oversight . . . . . . . . . . . . . 183 USDA . . . . . . . . . . . . . . . . . . . . . . 184 Authority for Plants . . . . . . . . . . . . . . . . . . 184 Application to Plants . . . . . . . . . . . . . . . . . . 188 Authority for Veterinary Biologics . . . . . . . . . . . . . . 191 Application to Veterinary Biologics . . . . . . . . . . . . . . 192 Authority for Animals . . . . . . . . . . . . . . . . . 193 EPA . . . . . . . . . . . . . . . . . . . . . . . 193 Authority of FIFRA . . . . . . . . . . . . . . . . . . 194 Application of FIFRA . . . . . . . . . . . . . . . . . 196 Authority of TSCA . . . . . . . . . . . . . . . . . . 197 Application of TSCA . . . . . . . . . . . . . . . . ... ....200 OTHER AGENCIES . . . . . . . . . . . . . . . . ... ......201 National institutes of Health (NIH) . . . . . . . . . . . . . . 201 Food and Drug Administration (FDA) . . . . . . . . . . . . . 201 STATE AND LOCAL GOVERNMENT . . . . . . . . . . . . . 202 Spectrum of State Approaches to Regulation . . . . . . . .. ................202 Spectrum ofLocal Approaches to Regulation . . . . . . . . . . . 203 INTERNATIONAL REGULATORY CLIMATE . . . . . . . . .. ...........204 Europe . . . . . . . . . . . . . . . . . . . . ...204 Canada . . . . . . . . . . . . . . . . . . . . . 206 Japan . . . . . . . . . . . . . . . . . . . . . ....206 Developing Countries . . . . . . . . . . . . . . . . . 207 POLICY ISSUES . . . . . . . . . . . . . . . . .. ............207 Jurisdiction and Coordination . . . . . . . . . . . . . . . 207 Coverage . . . . . . . . . . . . . . . . . . . . . 210 Potential impacts of Regulation . . . . . . . . . . . . . . ....213 Public Participation . . . . . . . . . . . . . . .. .............216 Problematic Issues . . . . . . . . . . . . . . . . . . 218 CHAPTER PREFERENCES . . . . . . . . . . . . . . . . 219 Boxes Box Page 7-A. The National Environmental Policy Act (NEPA) . . . . . . . . . . 185 7-B. The Types of Information Requested by APHIS in a Permit Application for Genetically Engineered Plants . . . . . . . . . . . . . . . . . . 188 7-C. Council on Competitiveness . . . . . . . . . . . . . . . 210 7-D. Fish RegulationsSomething To Carp About? . . . . . . .. ..............211 7-E. EPA Research and USDA Research . . . . . . . . . . . . . 215 7-F. Two Experiences With Public Response . . . . . . . . . . . . 218

PAGE 182

Chapter 7 Environmental Issues: Institutions and Their Regulatory Roles INTRODUCTION Preface Many biotechnology products, especially agricultural products, are intended for use in the environment. Examples are transgenic cows in feed lots, insect resistant crop plants in fields, microbial pesticides applied to cropland, and transgenic fish reared in outdoor aquiculture ponds. Virtually anything introduced into the environment will have an impact, whether it be concrete slabs used to construct a highway or a chemical pesticide used to control insects on cotton. The task of environmental protection legislation is to determine what types of products to be used or activities to be carried out in the environment would have adverse effects significant enough to warrant regulation. Ideally, Federal environmental protection laws and regulations would be based on complete information on all the environmental risks associated with products and activities as well as their benefits, so that decisionmakers could weigh one against the other objectively. In reality, complete information is rarely available, particularly for new products: thus, the balancing of risks and benefits is difficult and open to bias. Biotechnology has appeared on the scene during a time of intense environmental and political scrutiny of new technologies. Oversight of biotechnology thus is significantly different from that of emerging technologies in the past and may foreshadow the reception of new technologies in the future. For example, planned introductions of recombinant DNA-modified organisms will occur in a regulatory climate vastly different from that which existed as dramatically new crop varieties were introduced in the past. Key policy documents to be discussed later (e. g., 1986 Coordinated Framework statement of Federal agencies philosophy on biotechnology, and the Council on Competitiveness report on Administrative philosophy) stress the need to regulate biotechnology only on the basis of the risk of its products. not simply because it entails the new process of recombinant DNA technology. Tension exists, however, between this philosophy and operational development of oversight treatment. This tension often seems to be triggered by the technology itself, and has led to controversy over regulation of field tests. Special regulatory attention to a new agricultural technology could have implications for environmental safety and for the successful adoption of that technology and thus for U.S. economic competitiveness. Most agricultural biotechnology products intended for use in the environment are or will be regulated according to legislation enacted prior to the advent of modern biotechnology, including laws intended to protect agriculture and the environment from chemical contamination, plant pests, pathogens, and so on. Despite the unusual level of scrutiny focused on biotechnology, its oversight is meant to arise naturally from the responsibilities traditionally held by different offices or services within the Environmental Protection Agency (EPA) and the U.S. Department of Agriculture (USDA). Given the panoply of laws applicable to biotechnology. this chapter provides a road map through the confusing territory of oversight responsibilities. Figure 7-1 is the reference point used throughout the chapter. It gives a capsule overview of roles and relationships of policymaking bodies, key documents relating to designation of authority over environmental uses of biotechnology products, agencies with regulatory authority, the specific services or offices involved in regulation of biotechnology, and statutes that pertain to the use of biotechnology products in the environment. Following an introductory description of why and how regulation and oversight for biotechnology products has evolved, this chapter describes USDAs and EPAs role in these activities. The complementary roles of the Food and Drug Administration (FDA), National Institutes of Health (NIH), and State and local governments, as well as the international regulatory climate also are covered. Finally, policy issues are discussed here, among them issues of jurisdiction and coordination among agencies, scope of coverage, potential impacts of regulation on research and on agribusiness, and public participation. (See also OTA, 1988 New Developments in Biotechnologg 3, Field Testing Engineered Organisms. Genetic and Ecologic Issues (102) and 1991 Biotechnology in a Global Econony) ( 103). This chapter lays the foundation for ensuing discussion (ch. 8) of risk assessment and risk management issues related to impending large scale, commercial uses of agricultural biotechnology and biocontrol products. -181-

PAGE 183

Figure 7-lJurisdiction and Coordination of Environmental Policy for Biotechnology-Derived Agricultural Products a Natiorlaf Environmental Policy Act (NEPA) I United States Department of Agriculture (USDA) Animal and Food Office of National Plant Health Safety and Agricultural BIological Inspection Inspection Biotecnology Impact Service Service Agricultural Assessment (APHIS) (FSIS) Biotechnology Program Research (NBIAP) Advisory Committee (OAB/ABRAC0) I President I I I 1 I 1 I i Coordinated framework Scope document I I I Environmental Protection Food and Drug Agency (EPA) Administration (FDA) I Office of Biotechnology Office of Office of Center for Center for Pesticides Sciencee Advisory Toxic Biotechnology Veterinary Food Saftey Program Committee Subtances (OB) Medicine Nutrition (BSAC) (OTS) (CVM) (CFSN) I National Institutes of HeaIth (NIH) USDA statutes EPA statutes FDA statute Plant Pest Act Federal I nsecticide, Fungicide, Food, Drug, and Cosmetic Act (FDCA) Plant Quarantine Act and Rodenticide Act (FIFRA) Noxious Weed Act Toxic Substances Control Act (TSCA) Vims-Semm-Toxin Act Federal Meat Inspection Act Poultry products Inspection Act a OSTP, Council on Competitiveness, and OMB do not have direct oversight of the Federal agencies; the connections shown are those of influence through directives, key policy documents, or review. SOURCE: Office of Technology Assessment, 1992. Recombinant DNA Advisory Biotechnology Committee (NBPB ) (RA c )

PAGE 184

Agriculture, Field Trials, and Deliberate Release of Genetically Engineered Organisms Progress in agriculture traditionally has depended on selection of the best of new varieties based on field testing of cultivars. The seed industry views cultivar field testing as an essential part of cultivar development programs. The main purpose of field testing is to determine the regional environmental adaptability and market fit of the new cultivars or hybrids to know whether the items to be tested have the required disease resistance for the areas, whether they meet the needs of the industry as far as type or quality is concerned. and whether they will perform well under the environment of the region (98). Field tests also can provide evidence that the application of currently available scientific principles and information can ensure safe commercialization of new products. Genetically modified organisms, like any other organisms, must be field tested in the environment in which they would be cultivated. For example, whether the engineered trait is expressed effectively must be evaluated in condition-s representative of those the cultivated crop will encounter. Characteristics intended to confer drought tolerance to a plant. for instance, must appear and function effectively within the plant as it copes with representative drought-stressed environments. Greenhouse experiments, conducted in facilities designed to meet containment specifications, can provide only an initial screening; the field trial is an essential evaluative step. Brief Overview of Concerns As necessary and rational as field testing is, concerns have arisen over any release of genetically engineered organisms. Living creatures reproduce themselves; they may increase in numbers; and they may even exchange genes with other wild organisms. Many are worried in particular about the uncertain possible impacts that an organism with a new trait might have on other species in the local habitat. Evolution of Regulation and Oversight These concerns and uncertainties have stimulated efforts to articulate regulatory oversight; the spelling out of jurisdiction in the Coordinated Framework for the Regulation of Biotechnology [51 Federal Register (FR 2302-23393] (77) was a significant step in the organization of regulatory oversight. This fundamental document outlining the roles, responsibilities, and policies of the Federal agencies involved in biotechnology first actually appeared in the Federal Register in 1984, when the Domestic Policy Council of the White House announced the Coordinated Framework for the Regulation of Biotechnology (49 FR 50856-50907). The framework set forth certain premises, which have guided subsequent policy: l l l previously existing knowledge was regarded as pertinent, existing laws were for the most part regarded as adequate for biotechnology oversight, and different biotechnology products were regarded as falling under the mandate of different agencies (table 7-1 ). Other key points of the framework include the following: l l the products of biotechnology, not the process itself, would be regulated; and biotechnologically altered organisms are not fundamentally different from nonmodified organisms (although the introduction to the framework recognized that certain microbial products would require the establishment of additional regulatory requirements). The framework included a compilation of existing laws, regulations, and guidelines that are potentially applicable to biotechnology, policy statements from the regulatory agencies on how they intend to apply their existing regulatory authority to biotechnology, and proposed criteria for determining what should be subject to oversight. In a basic sense, agencies draw their authority to evaluate ramifications of the new technology based on their own mandates, and from the National Environmental Policy Act (NEPA). (See box 7-A. ) Since the framework was introduced, agencies have accumulated experience with deliberate releases; based on this experience, they are continuing to refine their regulatory roles. As of September 1991, USDA-APHIS (Animal and Plant Health Inspection Service), which oversees most plant-related work and animal biologics, has issued some 181 permits for field testing of genetically engineered plants or microorganisms (not including veterinary biologic). At least half of these have been issued since the beginning of 1990. (See table 7-2. ) USDA permits issued for transgenic plants with pesticidal properties have been informally reviewed by the

PAGE 185

184 l A New Technological Era for American Agriculture Table 7-lJurisdiction for Review of Planned Introductions in Research Proposed research Responsible agencies Contained research, no release in environment Federally funded . . . . . . . . . . . . . . . Nonfederally funded . . . . . . . . . . . . . . Foods and food additives, human drugs, medical devices, biologics, animal drugs Federally funded . . . . . . . . . . . . . . . Nonfederally funded . . . . . . . . . . . . . Plants, animals and animal biologics Federally funded . . . . . . . . . . . . . . . Nonfederally funded . . . . . . . . . . . . . . Pesticide microorganisms Genetically engineered Intergeneric . . . . . . . . . . . . . . . Pathogenic intrageneric . . . . . . . . . . . . . lntrageneric nonpathogen . . . . . . . . . . . . Nonengineered Nonindigenous pathogens . . . . . . . . . . . . Indigenous pathogens . . . . . . . . . . . . . Nonindigenous nonpathogen . . . . . . . . . . . . Other uses (microorganisms) released in the environment Genetically engineered Intergeneric organisms Federally funded . . . . . . . . . . . . . . Commercially funded. . . . . . . . . . . . . . Intrageneric organisms Pathogenic source organisms Federally funded . . . . . . . . . . . . . Commercially funded. . . . . . . . . . . . . Intrageneric combination Nonpathogenic source organisms . . . . . . . . . . Nonengineered . . . . . . . . . . . . . . . Funding agency, a NIH or S&E voluntary review, APHIS b FDA, c NIH guidelines and review FDA. c NIH voluntary review Funding agency, a APHIS b APHIS, b S&E voluntary review EPA, d APHIS, b S&E voluntary review EPA, d APHIS, b S&E voluntary review EPA, d S&E voluntary review EPA, d APHIS EPA d APHIS EPA d Funding agency, a APHIS, b EPA d EPA, APHIS, S&E voluntary review Funding agency, a APHIS, b EPA d APHIS, b EPA d (if nonagricultural use) EPA Report EPA Report, e APHIS b a Rewiew and approval of research protocols conducted by NIH, S&E, or NSF. b EPA jurisdiction for research on a plot greater than 10 acres. C APHIS issues permits for the importation and domestic shipment of certain plants and animals, plant pests and animal pathogens, and for the shipment or release in the environment of regulated articles. EPA reviews federally funded environmental research only when it is for commercial purposes, Designates lead agency where jurisdictions may overlap. KEY:NIH National institutes of Health; S&E = U.S. Department of Agriculture Science and Education; APHIS = Animal and Plant Health Inspection Service; EPA = Environmental Protection Agency; NSF = National Science Foundation SOURCE: 51 Fed. Reg. 23305 (Office of Technology Assessment, 1988). EPA Office of Pesticide Programs under an interagency agreement. EPA has reviewed a total of 94 notices for field tests of microorganisms since the framework was published in 1986, 74 of which were for microbial pesticides. Under an interagency agreement, EPA has in addition provided comments on approximately 100 permits submitted to USDA-APHIS for transgenic plants with pesticidal properties. (See table 7-3. ) These field tests provide the foundation of information and regulatory experience for decisions regarding fullscale agricultural use of transgenic organisms. This report comes at a critical point in the evolution of agricultural biotechnology, as it moves from the laboratory toward large-scale commercialization and use. USDA Authority for Plants Statutory Authority The Animal and Plant Health Inspection Service, APHIS, was established in 1972 as a regulatory agency within USDA with responsibilities for protection of the environment. APHIS unites the programs within USDA designed to protect American agriculture from destructive pests and diseases. APHIS activities include the development of exclusion procedures to keep pests and diseases out of the United States; and monitoring, de-

PAGE 186

Box 117-AThe National Evironmental Policy Act (NEPA) The National Environmental Policy Act (NEPA) is the sole Federal law that is broadly applicable to all agencies and departments involved in the research or regulation of biotechnology products intended for use in the environment. Enacted in 1970, NEPA is a reflection of increasing concern about environmental quality and calls for a balance between population and resource use which will permit high standards of living and a wide sharing of lifes amenities [section 101(b)(5)]. NEPA requires that any agency decision on a major Federal action significantly affecting the quality of the human environment include consideration of the environmental impact of the proposed action and alternatives to the proposed action. NEPA does not, strictly speaking, restrict or prohibit any activity that may adversely impact the environment but rather outlines procedural requirements by which Federal agencies must become aware of and consider the environmental consequences before making a decision on a proposal. The Council on Environmental Quality (CEQ) is responsible for the implementation of NEPA (CEQ Final Regulations for Implementing NEPA, 43 Fed Reg 59978, 1978), but the specific method used for compliance by individual agencies is broadly discretionary. Because EPAs mission is to consider and protect the environment through its regulatory activities, most EPA actions are considered the functional equivalent of NEPA compliance. [Warren County v. North Carolina, 528 f. Supp. 276,286 (eDNC 1981)]. Most other Federal agencies have issued their own regulations to implement NEPA. Although agencies are given broad discretion in how they evaluate and balance environmental impacts in making decisions, NEPA does open agency actions to public and judicial scrutiny. The establishment and protection of certain environmental values by NEPA gives public interest groups and private individuals standing to bring suit to ensure compliance even though they are not directly affected by an agency action. In short, NEPA has had two principal impacts on the Federal decisionmaking process: ensuring evaluation of environmental issues by Federal agencies and increasing public participation. SOURCE: Office of Technology Assessment, 1992. tection, eradication, and control programs to control the vironment) are the Federal Plant Pest Act, the Plant Quarmovement of pests and the spread of disease. APHIS operates under a myriad of legislative authorities, some dating back to 1884. Under the Coordinated Framework, APHIS is designated the lead agency responsible for the regulation of plant and animal biotechnology products. The assumption underlying this jurisdictional determination was that Agriculture and forestry products developed by biotechnology will not differ fundamentally from conventional products and that the existing regulatory framework is adequate to regulate biotechnology ( 51 Fed. Reg. 3123, p. 23302). The primary regulatory authorities available to USDA that are most applicable to biotechnology (and the enantine Act, the Noxious Weed Act. the Virus-SerumToxin Act. the Organic Act, the Federal Seed Act, the Federal Meat Inspection Act. and the Poultry Products Inspection Act. Of these statutes, two are used as the basis for the regulation of the environmental release of genetically modified organisms: the Federal Plant Pest Act, and the Plant Quarantine Act (7 CFR 340). Like the Noxious Weed 2 Act, these two acts are exclusionary statutes intended to prevent the entry into or dissemination within the United States of living organisms considered dangerous to American agriculture. These three legislative authorities traditionally have been used as the basis for inspection, quarantine, and pest eradication programs of the Division of Plant Protection and Quarantine. With the exception of the Noxious Weed Act, they now also are used by the Division of Biotechnology, Biol1 A Plant PCS( is defined ii~ my living Sttigc of: any inxcts, m itcs, nemiitodm, slugs, mails, protozoa. or other invertebrate ~nimali, b~ctcrl~, fungi. other pmmitic plmt~ m reproductive pw-ts thcrw)t. viruw~, or an} (~rgtmiw]s simi Iar to or all id w ith any of (he f{mgoing, or any infectious substances. which can ciirwtly or indirectly in jurc or cause diwaw or dtinmgc in any plant~ or parts thereof, or any processed, manufactured, or other products of plants. ~ N OXI OU S weed is defined a~ tiny I I ving stage (including but m)t I imitwl to swds and reproductive parts) of any parasitic or other plant of a kind, or subdivisi(m ot a kind. w hlch is of foreign origin, is nev to or not w idcly prmalcnt in the Unwd Sttitcs, and can directly or indirectly inlurc crops. other useful plants. 1 iwstock, or poultr} or other interests of agriculture, including irrigation. or na~igat ion or the fish or W ildl ife rcs{)urccs t)f the Unltmi States or the public health. 297-937 0 92 7 QL 3

PAGE 187

186 l A New Technological Era for American Agriculture Table 7-2Federally Approved Biotechnology Agricultural Research Field Test Applications, USDA (through September 24, 1991) Private Public 1987 . . 9 0 1988 . . 17 1 1989 . . 31 7 1990 . . 42 15 1991 . . 47 12 Total . . 146 35 b a 41 tomato; 23 cotton; 17 tobacco; 14 corn; 13 potato; 13 soybean; 10 cantaloupe/squash; 6 alfalfa; 4 clavibacter/corn; 1 clavibacter/rice; 1 TMV/ tobacco; 1 rapeseed; 1 sunflower; 1 chrysanthemum. b 11 potato; 9 tobacco; 3 cucumber; 3 rice: 2 pseudomonas; 2 walnuts; 2 xanthomonas; 1 tomato; 1 poplar; 1 alfalfa. SOURCE: APHIS BBEP Biotechnology Permits Unit, Issued Permitss List, Sept. 24, 1991. Table 7-3Federally Approved Biotechnology Agricultural Research Field Test a Applications, EPA (through April, 1991) Total Repeats Office of Toxic Substances . 20 b 7 office of Pesticide Programs . 74 34 Field tests of microorganisms reviewed by EPA since the publcation of the 1986 Coordinated Framework. b 10 Rhizobium, 8 Bradyrhizobium, 2 Pseudomonas. c Includes a variety of bacteria, fungi and viruses, both nonindigenous and genetically modified SOURCE: David Giamporcaro, Environmental Protection Agency, personal communicatlon, Oct. 18, 1991. ogics. and Environmental Protection [established in October, 1988] to regulate the movement and environmental release of genetically engineered organisms. The Noxious Weed Act has not been used to regulate genetically modified organisms. The applicability of the Noxious Weed Act to genetically modified organism is limited by the requirement that the plant be of foreign origin and the requirement that an organism be placed on the noxious weed list before it can be regulated. The Federal Plant Pest Act. the Plant Quarantine Act, and the regulations issued to implement them are not intended to present unreasonable barriers to commerce. For example, inspection at ports of entry should be expedient so as not to retard shipment of agricultural products, particularly fresh produce whose value could be diminished or destroyed if the product to be inspected is held at the inspection station too long. Agency Interpretation/Regulatory Policy USDAs overall philosophy regarding biotechnology products is articulated in the National Academy of Sciences 1987 publication, Introduction of Recombinant DNAEngineered Organisms into the Environment: Key Issues (72); and in the National Research Council 1989 publication, Field Testing Genetically Modified Organisms. Frumework for Decisionmaking (73). Consistent with U.S. Federal policy, USDA-APHIS bases its regulatory policy on certain key premises: 1. 2. A 3. 4. the products of biotechnology do not differ fundamentally from either unmodified organisms or conventional products, the product should be regulated rather than the process by which it came to be, end-use of the products and review conducted on a case-by-case basis should form the basis for regulation, and sufficient authority for regulating the products of biotechnology is provided by existing laws. Along with these premises is a commitment to the safe development of the new technology, and to a balanced, scientifically based and risk-based regulatory framework that protects agriculture as well as facilitates technology transfer (55). The USDA regulations (7 CFR 340), that pertain to genetically engineered organisms are applicable to a broad range of organisms, including Any organism which has been altered or produced through genetic engineering. if the donor organism. recipient organism. or vector or vector agent belongs to any genra or taxa designated ...and meets the definition of plant pest. or is unclassified. or any other organism or product altered or produced through genetic engineering which the Deputy Administrator deterrmines is a plant pest or has reason to believe is a plant pest. Excluded are microorganisms that are not plant pests and produced by the addition of well characterized or noncoding regulatory regions. Any person may petition to amend the list of organisms subject to regulation under 7 CFR 340. Such a petition must include the factutal grounds as to why the organism is not a plant pest and include scientific literature in support of this conclusion. Petitions should not include Confidential Business Information (CBI). The petition also should include any information known to the petitioner that would be unfavorable to the petition. APHIS then publishes a notice in the Federal Register for comment. A must respond to the petitioner within 180 days either by approving or denying the petition in whole or in part. Once an organism or class of organisms is delisted, it may move unhindered in commmerce with no

PAGE 188

reporting requirements or monitoring required by the Federal Government. If, however, new information becomes available that leads the Secretary of Agriculture to conclude that a delisted organism does, in fact, pose a plant pest risk, an interim rule can be issued, effectively bringing that organism back under regulatory authority of the Federal Plant Pest Act. It is unclear whether industry will try to petition to exempt broad classes of organisms or single. well-defined organisms. Initially some industry executives thought that they might like to delist broad classes; but some have since reevaluated this approach since the organism-byorganism delisting procedure is a market barrier to competitor-s. Broad class delisting might make it easier for some competitors to enter commerce. Furthermore, APHIS approvals provide a stamp of approval useful in acceptance by the public and by State governments. In addition, environmentalist groups might pose a legal challenge to stop a broad class delisting under the Federal Plant Pest Act. Implementation Under the APHIS regulations. anyone wishing to move or introduce an organism fitting the definition of a regulated article must receive a permit. The four kinds of permits for which applications are made are as follows: 1. a permit for release into the environment (application submitted 120 days in advance). 2. a single 1 -year permit for interstate movement of multiple regulated articles between contained facilities, 3. a single 1 -year permit for importation into the country of multiple regulated articles into contained facilities, and 4. a courtesy permit to expedite movement of organisms not subject to regulation under 7 CFR 340 (application submitted 60 days in advance) (55) Permit applications require submission of information on the biology of the donor and recipient organisms. the molecular biology of the introduced gene(s), and plans for containment during the trial and post-trial clean-up. Information is used by APHIS to prepare an Environmental Assessment (EA) and to determine whether and under what conditions to allow the release. The application process for Environmental Release permits is clearly delineated by USDA-APHIS, with process and pemmitting requirements contained in Plant Pests,. Introduction of Genetically Engineered Organisms of Products, Final R u l e (52 FR 22892 (1987). In addition, Biotechnology, Biologics and Environmental Protection (BBEP). USDA-APHIS. has developed a User's Guide for Introducing Genetically Engineered Plants and Microorganisms to provide assistance to those submitting applications for a permit under 7 CFR 340. The folloowing steps 1. 2 -. 3 4. 5 6. 7. 8. must take place: completing an application for permit under 7 CFR 340, Genetically Engineered Organisms or Products. APHIS Form 2000: assigning an accession number; preliminary pest and environmental assessment; state review/input; site inspection; issuance or denial: appeal. if permit request has been denied: and inspection of site at initiation of experinment. From day one, scientific review proceeds. The State authorities are forwarded material by day 30 and respond by day 60. At or before day 120. the biotechnology permit is issued or denied ( 104). Scientific review is based on the dutti provided in response to the APHIS permit application data requirements. Fourteen such requirements (box 7-B) include a detailed description of the organism. the location of the field test, and containment protocols. Provision is made for companies to protect Confidential Business information: they can submit both a full proposal and one for public abvailability that has CBI deleted. The APHIS Policy Statement on the Protection of Privileged or Confidential Business Information (50 FR 30561 -63) delineates data or information, such as trade secrets and confidential commercial or financial information, that can be protected from disclosure under section (b)(4) of the Freedom of Information Act (5 U.S.C.552 (b)(4). This can include production data, formulas and processes, and quality control tests and data, along with research methodology and data generated in the development of the production process. To qualify as CBI, this information must be: 1 ) commercially valuable, 2) used in ones business. and 3) maintained in secrecy. Furthermore, APHIS must be persuaded on review of information on competition that significant commercial harm would result from disclosure. BBEP explains this option to applicants, while encouraging them to be selective as to what truly calls for CBI designation (63). APHIS requires claims of Confidentiality to be substantiated at the time of submission. An Environmental Assessment (EA) is prepared by APHIS in accordance with the provisions of the National

PAGE 189

Box 7-8The 14 Types of Information Requested by APHIS in a Permit Application for Genetically Engineered Plants. 7 CFR 340 1. Information on responsible person and type of permit requested, such as movement or release. 2. All names (scientific, common, and trade) and designations necessary to identify the donor, recipient, vector, or vector agent constituents of the transgenic plant. 3. Information on the persons who developed the transgenic plant. 4. Movement of the plant. 5. The anticipated or actual expression of the altered genetic material in the plant and how the expression differs from the nonmodified plant in respect to characteristics such as morphology, physiology, number of copies of the gene, products, etc. 6. The molecular biology of the system used to produce the transgenic plantdonor, recipient, vector, or vector agent. 7. Country and Iocality where the donor, recipient, vector, or vector agent were collected, developed, and produced. 8. The purpose of the experiment and the experimental design. 9. The quantity, schedule, and number of introductions. 10. The processes, procedures, and safeguards used to prevent contamination, release, and dissemination in the production of the transgenic plant. 11. The intermediate and intended destinations of the product; the field trial site. 12. Safeguards to prevent dissemination at each site. 13. Biological material accompanying the plant, such as inoculum or soil. 14. Method of disposal of plant material after termination of the experiment, such as autoclaving or discing. SOURCE: S. McCammon and T. Medley, Certification for the Planned Introduction of Transgenic Plants in the Environment, The Molecular and Cellular Biologics of the Potato, Michael Vayda and William Park (eds.), Wallingford, U.K. (CAB. International), 1990. Environmental Policy Act (NEPA) of 1969 (42 U.S.C. Application to Plants 4332 ( 1970)). Among the components of the EA are procedural and physical precautions against risk, enviSmall-Scale Research ronrnental consequences, and background biology. The development of the EA is a process intended to assure public safety. A permit to move or introduce an organism is issued if there has been a Finding of No Significant Impact (FONSI ) from such action, and a full-scale environmental impact statement is not required. Notice of the action and the availability of the EA and FONSI is published in the Federal Register. Special additional conditions may be added to the permit that require monitoring and data collection to ensure containment. Such test data can also contribute to the information base from which future assessments can draw. The issuance of the permit constitutes certification by APHIS that no significant risk exists to the environment or to agricultural crops from the action. Recommendations for improving APHIS assessments have included making justifications for assessment conclusions more explicit, including more opportunities for gathering data on gene flow and weediness during field tests, and encouraging more timely and complete monitoring reports (110). Theory Small-scale releases in the form of field trials are experiments. Even if companies conduct them, and although field trials are the first step toward full-scale agricultural use in the environment, they are nonetheless still research rather than commercialization activity. This activity raises some concerns, but these are, to some extent, alleviated by the small scale of field trials. The first release into the environment of an organism with a novel trait can arouse concerns simply because something relatively new is happening. The regulatory policies and procedures described above represent an attempt to address such concerns. However, given the low numbers of organisms involved, the small-scale field trial is quite a carefully controlled situation. In fact, some argue that USDA requirements for most field tests exact financial, administrative, and time costs that are disproportionate, relative to any risks presented. USDA-APHIS views the small-scale field trial as playing an educational role; data compiled from these tests will provide the underpinnings for sound and rational assessment of large-scale releases in the future. As noted earlier, each permit issued for a

PAGE 190

small-scale field trial requires the submission of subsequent data. One of the players in the oversight of field trials is the Office of Agricultural Biotechnology (OAB ), which was established in 1987 under the Deputy Secretary of Agriculture and transferred to the Assistant Secretary for Science and Education in 1989. OAB is designed to ensure coordination of biotechnology activities within USDA. Within Science and Education, it is separate in many ways from APHIS. It provides staff support for the USDA Committee on Biotechnology in Agriculture (CBA ) comprised of administrators of agencies: conducts outreach programs; and provides leadership in the development of guidelines and the dissemination of information about them. For example, a handbook, Agricultural Biotechnology: Introduction to Field Testing was produced in large part to help the "users" of the regulatory system in applications for field trials ( 11 ). In line with its particular responsibility to provide guidance to researchers. the OAB staffs the Agricultural Biotechnology Research Advisory Committee (ABRAC) composed primarily of academic and industry scientists. Industry field tests. of course, are handled through APHIS. ABRAC was established in 1988 to provide advice for the Secretary of Agriculture, through the Assistant Secretary for Science and Education, on biosafety issues in the use of agricultural biotechnology and it has assisted in the development of biosafety guidelines, as well as case-by-case review of the minority of USDA-funded research projects that do not fall under other agency authorities. Its review process is modeled after that of the NIH Recombinant DNA Advisory Committee (RAC), with meetings open to the public and announced in the Federal Register. Two working groups established early in 1991 focus on the area of biotechnology risk assessment research as set out in the Farm Bill of 1990. These groups help set priorities and are developing a classification system and confinement protocols, integrating public comments received on the proposed guidelines for risk assessment research (70). The Proposed USDA Guidelines for Research Involving the Planned Introduction Into the Environment of Organisms With Deliberately Modified Hereditary Traits was published in the February 1, 199 Iissue of the Federal Register, part 3, with public comments due on April 2; a principal intent was to assist academic scientists and their institutional biosafety committees in the design of safe field trials. USDAs Cooperative State Research Service established a new program in response to recommendations in a 1985 report of the National Association of State Universities and Land Grant Colleges Committee on Biotechnology. NBIAP (National Biological Impact Assessment Program ) has a mandate to facilitate safe field testing of genetically modified organisms and. thus, safe development of agricultural biotechnology. A principal charge to the program is to facilitate the appropriate application of knowledge derived from conventional field testing in the past to biotechnology field tests today. The program supports three areas of activity related to this function: information networks: facilitation of the development of biological monitoring techniques: and supper-t for biosafety research. An information network to support the needs of publicand priate-sector researchers is being developed by NBIAP in conjunction with a number of institutions. The information network is available, over telephone lines, through an "800" number: through interlinked mainframe computers (BITNET): on floppy disks: and in printed format. An electronic bulletin board gives up-to-date information on biosafety related research activity and serves as the gateway for 14 databases. Individuals can use the network to communicate with other scientists as well. Databases include, among others: bibliographic and other listings; current literature; U.S. patents on genetically engineered species: current text of all Federal laws, regulations, and guidelines pertaining to biotechnology and biosafety; Institutional Biosafety Committee listings: and all approved applications for federally approved field test permits. licenses. and scientific reviews. A knowledge base has been designed to help researchers identify the responsible Federal agencies to which an application should be directed and to prepare applications for permits, licenses, or scientific reviews. An intelligent form generator" will actually help the investigator prepare first drafts of applications. By dissembling information from existing knowledge. the intelligent form generator provides users with access to previously written standard text, technical descriptions, test-site information, and other resources from databases. Combining information with use of extensive menus leads to a technically specific application. The first, current version of the intelligent form generator is expected to be expanded from coverage of 8 groups of organisms to 79. The intent of the intelligent form generator is to lift some of the regulatory burden from the researcher. "Hyrpertext" information on biosafety is also provided. A second function of the NBIAP is to facilitate biological monitoring of genetically modified organisms de-

PAGE 191

literately released in the environment. NBIAP is surveying field studies that have been conducted; the information gathered should help guide future regulatory decisions. NBIAP also supports biosafety research on genetically modified organisms to improve understanding of their dispersal in the real world, and their impact on human health and the environment, to improve biosafety methods and to develop useful prediction models (51. 52). Experience BaseBetween July 16, 1987 and February 27, 1991, 102 permits were granted by APHIS for field testing of genetically engineered plants and microorganisms, along with 843 permits for importation or interstate movement of organisms regulated under 7 CFR 340. Twenty-one companies were issued permits by this date, including: Agricetus, Agrigenetics, Amoco Technology Co., Biosource Genetics, Biotechnica, Calgene, Campbell Institute for R&D, Canners Seeds, Ciba-Geigy, Crop Genetics International, DeKalb, DNA Plant Technology, DuPont, Frito-Lay, Monsanto, Northrup King, Pioneer, Rogers NK Seed, Rohm & Haas, Sandoz Crop Protection Corp., and UpJohn. Twelve research institutions. two of them USDA institutes, had received permits; they are: Auburn University; Iowa State University; Louisiana State University: New York State Agricultural Experiment Station (Geneva); North Carolina State University; Pennsylvania State; USDA-ARS (Agricultural Research Service), Albany, California: USDA-ARS, Fresno, California; University of California at Davis; University of Kentucky; University of Wisconsin; and Washington State University. Field trials were approved, with the agreement of the host State, for 33 States and Puerto Rico; the 102 permits granted as of February 27, 1991, gave rise to some 140 field test sites. By April 1991, 115 permits had been granted (78). By September 1991, 181 permits for field tests had been granted. Figure 7-2 lists the new crop plants entering field trials between 1987 and early 1991, along with the novel characteristics, or genes expressed. About half of the first generation of field tests, especially the 2 I in 1988, were for herbicide tolerance in tomato and tobacco. while the rest were almost entirely for disease and insect resistance in these two crops. Many more crops showed up in the 1989 applications, including potato, soybean. alfalfa, cotton, poplar, and cucumber: new sorts of characteristics included slowed fruit ripening and improved nutritional qualities. Modified pathogenic bacteria entered the applications in 1990, along with an increased range of cultivars and modifications, particularly in two of the countrys most economically important crops, rice and corn (60). Figure 7-2Field Trials of New Crop Plants, 1987-91 1987-88 1989 1990 1991 Tobacco Alfalf a cantaloupe Rapeseed Tomato Cotton Sunflower Cucumber Rice Poplar Squash Potato Walnut soybean Genes expressed Herbicide tolerance Insect tolerance Virus tolerance Fungal tolerance Slowed fruit ripening Heavy metal sequestration Increased lysine production Antibiotic resistance SOURCE: S. McCammon, U.S. Department of Agriculture, internal memo, 1991. Biotechnica Agriculture. Inc., then a subsidiary of Biotechnica International, Inc., received in May of 1990 the first USDA approval to field test genetically engineered corn plants. The tests, to be conducted at the companys corn breeding station in Iowa. will analyze growth under field conditions and collect environmental data for future use. Biotechnica has coordinated other field tests, including one on tobacco with a gene coding for high levels of lysine expression (9). The company has applied for permission to conduct multiple field tests of corn engineered for improved nutritional quality; the gene transferred is one of several intended to improve corn for feed (4). Northrup King has begun a 3-year field test of alfalfa plants genetically engineered to be compatible with a new herbicide claimed to be highly biodegradable and environmentally safe. With Monsanto. Northrup King has planted genetically engineered cotton in Hawaii to assess its resistance to various caterpillars (71 ). An even longer term project was initiated by USDAARS researchers at the University of California-Davis. They inserted two marker genes into walnut tree embryos and will need to wait 5 years for the trees to reach maturity to assess expression brought about by the genes (76). The first field trial of genetically engineered rice was approved at Louisiana State University. The test, taking place since June 1990 in a 110 x 63 foot plot in Baton Rouge, involves a marker gene and a transposon gene (that regulates gene movement) from corn (5).

PAGE 192

USDA-ARS scientists are field testing potatoes with marker genes in Idaho, to see if the genetically engineered potatoes match the quality of conventionally bred products, under a permit issued in 1989. Some 1,000 potatoes, originally produced in a greenhouse from genetically engineered microtubers, are planted on a halfacre plot at the University of Idahos research and extension center in Aberdeen (29). Calgene successfully harvested field plots of its FLAVR SAVR tomato in the fall of 1990. Its permit for tomato plants engineered with an antisense gene for the pectalytic enzyme, or cytokinin pathway, was issued in May of 1990 (76). A complete listing of permits issued, applicants, organisms, and genes engineered along with date of issuance and location (State) is available in "Environmental Release Permits, printed by BBEP, APHIS, September 24, 1991. Large-Scale Release Theory-The USDA plans to use data from smallscale field trials to ensure the safety of large-scale releases. A variety of analyses and conferences are addressing the issue of large or commercial-scale release. For example, APHIS has organized the following three workshops to identify issues related to the large-scale use of genetically engineered crops in the environment: 1. 2. 3 Workshop On Safeguards for Planned Introductions of Transgenic Oilseed Crucifers, October 1990. Ithaca. New York; Workshop On Safeguards for Planned Introductions of Transgenic Crops: Maize and Wheat. December 1990, Keystone, Colorado; and Workshop on Biosafety Issues of Field Tests with Transgenic Potatoes, August Scotland. A fourth workshop is planned for sues for transgenic rice plants. Experience BaseNo commercial 1991. St. Andrews, 1992 on biosafety isreleases have yet occurred, nor have applications been made, although preliminary discussions have been held between company representatives and APHIS officials. Authority for Veterinary Biologics Statutory Authority Under the authority of the Virus-Serum-Toxin Act (VSTA). as amended, USDA-APHIS regulates three cat. egories of veterinay bioiogical products derived through biotechnology. The establishment of these three categories was announced by APHIS in theJune 1986Coordinated Framework policy statement (51 FR 23339, June 26, [986). Based on that frameworks premises that recombinant DNA derived products are not significantly different from more conventionally derived products and can be handled by a network of existing statutes, the three new categories were subsumed under VSTAs treatment of other biologics. APHIS supervises all experimental uses of veterinary biological products outside of containment conditions, under the provisions of the VSTA as amended by the Food Security Act of 1985. The implementing regulations (9 CFR 103.3) require approval from the Director of BBEP for shipment and describe required information for evaluating unlicensed biological products prior to granting such approval. APHIS also licenses biological products for unrestricted shipment in or from the United States under the VSTA, as amended. Agency Interpretation and Regulatory Policy The agencys policy is to balance control with flexibility in its review and approval procedures, and to adapt as necessary to new information. Products and organisms are categorized to provide practicable. reasonable procedures for review and approval: review takes place on a case-by-case basis. Category 1 is comprised of inactivated (nonviable or killed) products prepared from recombinant DNA-derived vaccines. viruses. bacterins, bacterin-toxoids. viral subunits. or bacterial subunits. Monoclinal antibodies used prophylatically, therapeuticall y, or as diagnostics are also included. These products are viewed as presenting no risks to the environment or to safety. Category 11 consists of products containing live microorganisms that have had one or more genes added (for expression of unique marker antigens or production of biochemical by-products) or deleted (i.e.. genes for virulence, oncogenicity, enzyme activity, or other biochemical functions). Such changes in genetic information must not lead to increased virulence, pathogenicity, survival advantages, or undesirable new or increased abilities to invade or survive in the animal host; and they must not compromise the safety characteristics of the organisms. Under category III fall products that uselive vectors to carry recombinant-derivecl foreign genes coding for immunizing antigens or other immune stimulants. Live vectors may carry multiple such genes and successfully can infect and immunize the host. These organisms must be completely characterized and compared with the parent virus, and environmental and human or animal safety

PAGE 193

192 l A New Technological Era for American Agriculture concerns must be addressed in an Environmental Assessment or Environmental Impact Statement. Implementation As with all other veterinary biologics, recombinant DNA products must be shown to be pure, safe, potent, and efficacious, and not worthless, contaminated, dangerous, or harmful, with assurance of lack of negative effects on the environment and human and animal health prior to licensing. Additional information (e.g., demonstration of nonpathogenicity and nonreversion to virulence, or ability of the organism to maintain itself in a livestock population) may be requested. For recombinant-derived products the manufacturer also must report the cloned nucleotide sequence coding for the product. For category 11 and 111 organisms, authorization procedures for shipping and guidelines for review of applications for field trials are done on a case-by-case basis. The categories of physical containment involved in movement of experimental products to the field are the following: 1. stringent containment conditions (level 4, isolation), 2. controlled environment (level 3). 3. Quarantined field conditions (level 2), and 4. Restricted field tests (level 1). Unrestricted geographical distribution may occur only after issuance of a license. In considering approval of these movements. APHIS requires four kinds of scientific information: human safety, ecological concerns, characterization of the vaccine virus, and animal safety. In addition, appropriate data would include: survival and reproduction of the engineered microorganism; interactions with other organisms; effects on the ecosystem if applicable; and scale, scope, and frequency of plasmid introduction. In short, an ecological risk assessment would include the biology of the phenotypic trait and of the parent organism, as well as characterization of the environment into which the introduction will be made; the product organisms host range and potential effect on other species might also be included. The review cycle includes review and approval by an Institutional BioSafety Committee (IBC). State animal health regulatory officials and, if appropriate, public health officials. For trials with a small number of animals in quarantined conditions, APHIS must prepare a Safety Factor Evaluation assessing all parameters of the trial ( 19). Application to Veterinary Biologics Small-Scale Research TheoryThe theory underlying the approach to release of veterinary biologics is consistent with the National Research Council report (73) and the Scope Principles (i.e., that products of biotechnology are not inherently more dangerous than products of other techniques; and that existing regulations can cover them). Experience BaseSome 46 licenses that have already been granted for small-scale release of veterinary biologics went through the full testing and now qualify for large-scale release. Other projects are still in the research stage (95). One of the best known small-scale field test cases is that of the genetically engineered rabies vaccine developed by the Wistar Institute, with its corporate partner Rhone-Merieux. This is a Vaccinia virus expression system, with a gene for a protein of the rabies virus, that is intended to stimulate an immune response, but that cannot cause rabies. Provisional approval was given early in 1989 by USDA; the actual distribution of 3,000 ampules of rabies vaccine in an odoriferous bait took place on uninhabited Parmmore Island, Virginia, in the fall of 1990. South Carolina had declined to have an offshore island field test take place within its boundaries. The owner of the Virginia island, the Nature Conservancy. negotiated long and hard regarding the release. The Wistar Institute had to agree to provide full insurance coverage and indemnification against any lawsuits to the Nature Conservancy. The Conservancy demanded a strong voice in field trial and animal monitoring protocols. Although Wistar researchers assert that any risk from the release is very remote, the apparent lack of control (putting bait in the wild and waiting for animals to eat it) certainly helped to arouse concerns. A similar vaccine is being tested widely in Europe, and the Wistar Institute has had discussions regarding additional sites in the midAtlantic States (25). On the basis of satisfactory results from the Virginia field trial and additional data confirming safety in other species, APHIS authorized a second field trial in Sullivan County, Pennsylvania, on June 7, 1991, with little or no adverse public comment, and New Jersey is considering a field trial as well. In contrast. early in APHIS review of animal biologics, a suit by Jeremy Rifkins organization, the Foundation for Economic Trends, resulted in a voluntary 2-week suspension of the license issued for the first recombinant-DNA derived category 11 pseudo-

PAGE 194

rabies vaccine while APHIS prepared documentation of the assessment conducted during the licensing process. Large-Scale Release for Veterinary Biologics Theory-The USDA uses information from early smallscale trials in its subsequent assessment procedure. Biological products progress from physical containment to large-scale use in the field as follows: 1. Movement from stringent containment conditions (level 4) to quarantined field conditions (level 3). 2. Restricted field tests (level 2). 3. Unrestricted geographical distribution on issuance of a license (level 1). Experience BaseAs of October 1, 1991, APHIS had approved field testing and subsequently granted licenses for 39 Category One veterinary biological products. Twenty-six of these were for diagnostic kits; five were for bacterins, and three were for monoclinal antibodies for prophylactic or therapeutic use. The first, a bacterin, was licensed in October 1983; all have been used successfully on a large scale. Seven licenses were granted for category 11 products, all of which were designed to treat pseudorabies in swine. No licenses have yet been granted in category III, but APHIS has received, evaluated, and approved an application to field test a recombinant-DNA derived live rabies vaccine (95). The following category I and 11 licenses have been issued: l l l l Salsbury Labs and Norden Labs were the first licensees for bacterins in category for genetically engineered Escherichia coli against swine disease. Molecular Genetics Inc., received licenses for category 1, therapeutic or prophylactic use, for monoclonal antibodies. Among the category 1 diagnostic test kits licensed were kits for equine infectious anemia, avian reovirus antibody, and feline leukemia and feline Tlymphotropic lentivirus. At least four companies received category 11 licenses for a modified live virus used as a pseudorabies vaccine. Authority for Animals Statutory Authority and Regulatory Policy The Federal Meat Inspection Act (FMIA) (21 U.S.C. 601 et seq. ) and the Poultry Products Inspection Act (PPIA) (21 U.S. C. 451 et. seq. ) give responsibility to USDAs Food Safety and Inspection Service (FSIS) for the safety, wholesomeness, and proper labeling of food products made from domestic poultry and livestock. FSIS inspects the organisms and cleaned products intended for use as human food. Under the slaughter of research animal provision of the FMIA and the PPIA, FSIS has developed regulations stating that no livestock or poultry used in a research investigation is to be slaughtered at an official establishment until sufficient data demonstrate to FSIS that the edible products derived from the research animals are safe for human consumption (9 CFR 309.17 and 381.75). These regulations pertain to the slaughter of transgenic animals as well as animals treated with recombinant DNAderived products. Implementation In the event of a request for slaughtering approval, FSIS would coordinate its review with the agency having jurisdiction over the experimental product (e.g., APHIS biologics, FDAdrugs, food. and feed additives, EPA pesticide chemicals. ) Usually, data gathered by each individual agency is adequate for FSIS evaluation. Once approved for slaughter, research animals are subject to the same inspection standards as nonresearch animals. If some animals derived through new technology, such as mosaics, chimeras, and some hybrids, differ significantly from currently inspected animals, the FSIS will determine on a case-by-case basis whether the animals are covered under FMIA or PPIA or if the acts need to be amended to require inspection. FSIS also has authority over substances used in processing meat and poultry products; the use must be in compliance with applicable FDA regulations and must be functional, suitable, and kept to the lowest level necessary ( I I). The FSIS has not yet had to test its interpretation or implementation process in a case involving animals modified through biotechnology ( 1 I ). EPA EPA has jurisdiction over two broad classes of products (pesticides and new chemicals) under three Federal statutesthe Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA): the Food, Drug, and Cosmetic Act (FD&C); and the Toxic Substances Control Act (TSCA). Under the authority of FIFRA, EPA regulates the manufacture, processing, distribution, and use of pesticides and sets tolerance levels for pesticides in food and feed as directed by the Food, Drug, and Cosmetic Act (discussed in the food safety chapter). Under TSCA,

PAGE 195

194 l A New Technological Era for American Agriculture EPA must screen any *new chemical before it is introduced into commerce to determine whether or not its use presents an unreasonable risk to health or the environment and is not otherwise regulated. This section reviews EPAs statutory authority under FIFRA and TSCA and discusses its application to the regulation of biotechnology products. EPA attempts to forge a coordinated and consistent approach to its biotechnology responsibilities under FIFRA and TSCA to the extent possible given the different mandates of the two statutes. Despite these different mandates, both approaches to regulation are concerned with microorganisms having: l new characteristics (intergeneric combinations of genes) that are new to the environment in which they will be released; l potential for adverse effects on other organisms; and l potential for widespread exposure because they are used in the environment. Because FIFRA regulations were already applied to microbial pesticides, an interim regulatory policy announced in the Federal Register on small-scale field trials in relation to Experimental Use Permit (EUP) regulations was the only change necessary for the new biotechnology. However, a set of regulations for microorganisms is needed under TSCA so that EPA can regulate living microorganisms more readily. The agency heretofore has dealt principally with new chemicals, although microorganisms have been included in the TSCA Inventory of Chemical Substances since its establishment. Regulations could be developed by applying the statutory provisions of TSCA and EPAs current oversight program for new chemicals to microorganisms. The delay in the development of these regulations has been noted with particular concern by the biotechnology community as likely to have caused uncertainty among applicants and would-be applicants for deliberate release. For assistance in regulating biotechnology, the EPA formed its Biotechnology Science Advisory Committee (BSAC) in 1986 to give peer review of EPA assessments of product submissions, as well as scientific advice on its biotechnology program. Among other responsibilities, the BSAC has been involved in advising on terms for regulations, on benefits and risks of the use of antibioticresistance genes as markers in field tests, and on peer reviews of some EPA assessments of field test submissions (67). Authority of FIFRA Statutory Authority As noted above, pesticides, including those produced using biotechnology, are regulated by EPA under the aegis of the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA). FIFRA was enacted June 25, 1947, to regulate the marketing of economic poisons and devices (6 I STAT. 168; 7 USC sec 135c); it has been amended multiple times in the intervening years with major substantive amendments in 1972. 1978, and 1988 and more anticipated in the early 1990s. The heart of FIFRA is the requirement that all pesticides be registered. EPA must certify that the use of a pesticide does not pose an unreasonable adverse effect in order to register a pesticide. In deciding whether a pesticide use poses any unreasonable risk to man or the environment, EPA must take into account the economic, social, and environmental costs and benefits of the use of any pesticide. EPA must also consider the impact of any regulatory action "on production, prices of agricultural commodities, retail food prices, and otherwise on the agricultural economy. Registration requires the submission by the manufacturer of extensive data on the efficacy and human and environmental effects of the pesticide. EPA uses this data in deciding whether to register the pesticide and whether to impose conditions on its manufacture, processing, distribution, and use. After registering a pesticide, EPA retains regulatory control via the reregistration, cancellation. and suspension provisions of FIFRA. Section 6 (a) of FIFRA establishes that registrations are canceled after 5 years unless EPA receives a request for a new registration, at which point EPA may request new data about the pesticide and may, on the basis of this new information, alter the conditions of the registration. EPA also has the power to cancel a registration at any time if the agency finds that the pesticide poses an unreasonable adverse effect; however, the cancellation procedure is complex and timeconsuming. If the use of a pesticide poses an imminent hazard, EPA may immediately suspend a registration. Agency Interpretation and Regulatory Policy EPAs principal experience base lies in evaluating conventional chemical pesticides where risk issues may differ significantly from those of living organisms. Nonetheless, microbes (e.g., bacteria. viruses, fungi, and protozoa) producing pesticides or pesticidal substances. as well as plants modified to produce substances to control pests, can be interpreted as falling within the statutory

PAGE 196

definition of pesticide. EPAs office of Pesticide Programs has built a group and experience in regulation of microbial pesticides since the late 1970s. EPA will register these products it it concludes that the benefits of their use outweigh the risks. More controversy has arisen over whether pest-resistant plants are equivalent to pesticides. since all plants have some pest-resistant characteristics naturally. EPA has never, for instance, regulated plant varieties. such as virus-resistant lines, classically bred to have pesticidal properties. MicroorganismsOn October 17, 1984, EPA published in the Federal Register a notice that it would "require notification prior to all small-scale field tests involving certain microbial pesticides in order to determine whether experimental use permits are required. This is in contrast to small-scale field tests of conventional chemical pesticides. An EUP is not required for the latter if under 10 acres of land or 1 acre of water is involved. The difference in policy is based on the premise that the concepts of "small scale" or small quantity arc not applicable to living organisms capable of movement and reproduction. Notifications are required for field tests involving non indigenous microorganisms. microorganisms genetically altered by traditional means, such as mutagenesis, and recombinant microorganisms. In the case of most of these notifications. no problem is perceived by EPA and no EUP is required. In a February 15, 1989, Federal Register notice. EPA announced its intent ion to amend FIFRA regulations to require notice for small-scale releases involving I ) microorganisms whose pesticidal properties have been altered by introducing intentionally manipulated genetic material; and 2 ) microbial pesticides formed by the combination of genetic material from organisms from different genera. In an attempt to maintain flexibility, EPA is currently considering a mechanism for exempting small-scale field tests of microbial pesticides from the notification requirement as increasing information and experience so justify. Only organisms with higher risk and those that arouse higher levels of public concern would remain the targets of reviews. A draft amendment to the regulations is circulating within EPA that would clarify the scope of organisms requiring notification, emphasizing only those organisms that carry significant possibility of risk or raise high levels of public concern. There has been some support for exempting nonindigenous microorganisms and microorganisms genetically altered through traditional means from notification requirements, expressed in terms of the very absence of comments received on publication of such notification in the post (67, 91 ). PlantsIn 1987, the EPA Office of Pesticide Programs (OPP) and USDAs Animal and Plant Health Inspection Service ( APHIS ) agreed to review cooperatively proposals for field tests transgenic plants that fall under the Federal Plant Pest Act. Currently, while tests are at a small scale. on an operational level APHIS takes the lead, with OPP providing comments. Under discussion is the possibility that OPP take the lead when the plants are grown on a large scale for food use. In some cases, the products of large-scale tests might be intended for food or teed USC Modifications to 40CFR 152.40 CFR 158, and 40 CFR 172 may be needed fo new data requirements and variations on Experiment Use Permits. EPA might regulate field trials of plants with pesticidal properties, or it might set tolerance levels for residues in approved food products. To gain input as it develops procedures t-or evaluating transgenic plants, EPA conducted a workshop in June. 1990 to discuss scientific issues and seek guidance on the information needed to conduct these evaluations (23). In November of 1990, EPA held a second information gathering conference, this one focusing exclusively on pesticidal transgenic plants. One topic addressed is how to adopt the agencys usual "maximum hazard". testing approach, in which artificially high concentrations of a chemical are used to evaluate the safety of plants that produce a pesticidal chemical in small amounts; if extra supplies of the chemical are generated (in bacteria) for the tests, will this material be identical to the plants chemical, so that the test is valid? Such complexities not withstanding, EPA under FIFRA has a much more focused taskregulating substances designed to harm some living systemsthan FDA, which will have to consider the much broader arena of genetically engineered plants as food. (See ch. 10 and 11. ) Other Organisrns-Microorganisms used against other insects, such as nematodes or parasitic wasps, do not fall under the purview of FIFRA. However, the demands of particular isolated cases can elicit FIFRA staff involvement. In one case, parasitic wasps were used to control infestations in certain grain elevators in Texas. The FDA inspector checking for insect parts in the food requested a tolerance level from EPA. EPA could not comply because it had never registered the wasps as a pesticide. After much interagency communication back and forth, EPA developed a memo of exemption. This was the one case to date in which EPA staff has dealt with animal

PAGE 197

196 l A New Technological Era forAmerican Agriculture microorganisms. EPA is not involved, for example, in a case of a pesticidal nematode carrying bacteria because it is seen as a microorganism system. If, however, the bacteria in the system were genetically engineered, OPP would want to take a look at it (91). Implementation Basically, EPA reviews a proposed test and decides whether to allow the test, request more information, or require an Experimental Use Permit, for which the target review time is 90 days. Companies are encouraged to hold discussions with FIFRA officials prior to the notification and EUP stages. Review is conducted on a case-by-case basis by FIFRA staff. A list of data that must be submitted with a notification is available and includes, among other components: l l l l l l l the identity of the microorganism; means and limits of detecting the microorganism in the environment; physical, chemical, and biological features influencing the growth and survival of the microorganism; information on likely survival in the environment(s) into which the microorganism will be introduced; the genetic manipulations involved. in detail; data on potential for gene transfer, detailed description of the test program, including monitoring; and any additional factual information on possible adverse effects. Aspects considered by staff include: hazard and exposure, potential problems or issues, important questions needing answers, and likelihood of risk. Staff positions are then shared for comment with intraagency workgroups, other Federal agencies if appropriate, State agencies, and, if needed, the BSAC. Although a State-FIFRA Issues Research and Evaluation Group exists, EPA does not yet seem to have tapped or developed an established, extensive system of State-level biotechnology contacts comparable to that of USDA. Public comment is regarded as important; for some proposals, several opportunities have been provided. Notice of all notifications appear in the Federal Register; significant EUPs, including all biotechnology EUP, are placed in the Federal Register as well. Companies are encouraged to inform local communities of upcoming field tests. If the analysis indicates unreasonable risks are likely, EPA can impose restrictions. Risk management can include constraints on use, disposal, and manufacture, as well as mitigation, monitoring, or other actions. As a way of checking on its evaluations, and adding to its information base for future tests, EPA has worked on the development of monitoring methods that will lead to understanding of the possible fate and dispersal of microorganisms in the environment (67). Application of FIFRA Small-Scale Release TheoryEPA under FIFRA approaches small-scale field trials on a case-by-case basis. Experience BaseFrom 1984 up to 1989, the Office of Pesticide Programs (OPP) reviewed 36 submissions (notifications and EUPs) under FIFRA. Of 25 notifications reviewed, 21 were approved with no EUP required, 1 was withdrawn, and an EUP was required for the remaining 3. Of 11 EUPs reviewed, 10 were approved, with a decision on 1 pending (96). Companies making submissions included: AGS, Mycogen, Monsanto, Ecogen, Rohm and Haas, Crop Genetics International, and Sandoz. Universities included the University of California, Montana State University, Cornell, and the University of Arkansas. Nearly half of the tests involved Bacillus thuringiensis (85). So-called pesticidal plants," transgenic plants that produce pesticidal chemicals, are reviewed in conjunction with USDA-APHIS, with EPAFIFRA staff providing comments to USDA-APHIS. Tomato plants engineered with Bacillus thuringiensis toxin genes and tobacco plants engineered with Tobacco Mosaic Virus coat protein genes are examples. Both have been explored by more than one company. Companies whose applications for transgenic pesticidal plants received informal review by EPA are: Rohm and Haas, Monsanto, Agrigenetics, Sandoz, DuPont, and Agracetus (97). The first review of an EUP application for a genetically engineered microbial pesticide (a test by Advanced Genetic Sciences, Inc., of the Ice-minus (INA) Pseudomonas syringae) took nearly 2 years from receipt of application to the field test. Two lawsuits involving Federal and State courts temporarily stopped the test; many administrative proceedings at the State and local levels caused further delays. In contrast, a later application on an EUP submitted by Crop Genetics International in December of 1987 was granted in May of 1988, less than one-half year later. The field test was begun in June and data for the test were submitted in application for an extension and ex-

PAGE 198

pansion of the EUP. This test involved the insertion of a Bacillus thuringiensis toxin gene into a plant endophytic bacteria (67). OPP considers any microbial pesticides to be biotechnological in the broadest sense; even biochemically based pheromone products are biologically active systems designed to alter the behavior of insects. At least threequarters of the Offices workload is comprised of nonrecombinant microbial pesticides; recombinant products represent only 1 to 5 percent of the number of notifications received. While numbers of new chemicals to be reviewed have plateaued over the last 5 to 6 years, microbiological/biotechnology products are increasing linearly such that they now comprise approximately onethird of the reviews. Plans exist to add biologically trained staff during the upcoming year (91). The early stages of the regulatory life cycle of a new microbial pesticide is illustrated by a planned introduction of dead recombinant organisms into the environment. In 1986, Mycogen discussed its killed recombinant bacteria with FIFRA staff who, on receiving requested data proving that the bacteria were in fact dead, told the company that it did not need to submit a notification. In 1988, the company was moving its trials into sites larger than 10 acres, the stage where an EUP was obtained from EPA. In 1989-90, field tests took place on some 5,000 acres per year. In 1991. the company had several products approved for registration as a pesticide (91). EPA has also approved field trials of live recombinant organisms by Repligen and Sandoz Research Corps. Field trials of recombinant Bacillus thuringiensis on soybeans infected with beet army worms were approved for the fall of 1990 at Sandozs Mississippi station (86). Interest in microbial pesticides is growing among large companies. Large-Scale Release Although naturally occurring, classically derived, and killed recombinant products have moved through largescale testing and commercial registration, no large-scale releases of liverecombinant organisms have as yet been approved under FIFRA. However, at least one company has had a series of discussions with EPA staff on testing design. Authority of TSCA Statutory Authority The Toxic Substances Control Act (TSCA) was enacted in 1976 to regulate the manufacture, processing. and use of chemicals that may pose an unreasonable risk to human health or the environment ( 15 USC section 2601-54) (13, 87). Because Congress intended TSCA to be gap-filling legislation. it gives EPA broad regulatory authority over a range of substances not regulated under other Federal laws. In determining the appropriate type and level of regulation to impose. EPA must consider the environmental, economic, and social impact of any action [it] takes or proposes to take" {15 USC sec. 2601 (2)]. A S with FIFRA. EPA must carry out a risk benefit analysis before imposing restrictions on the manufacture, processing, or use of any chemical. TSCA primarily is a mechanism for screening new chemicals. EPA can review new chemicals for unreasonable risk through the mechanism of manufacturers being required to submit a premanufacture notification (PMN) to EPA prior to the manufacture of any new chemical, i.e. any chemical not included on the EPA inventory of chemical substances (TSCA sec. 5, 40CFR 720.25). Under TSCA. EPA has the authority to limit or prohibit the manufacture, processing, or distribution in commerce of a new chemical substance if it determines that the chemical substance may present an unreasonable risk to health or the environment, or pending the development of sufficient data to assess whether the chemical substance presents an unreasonable risk. The burden is on EPA to establish risk rather than on the manufacturer to establish the absence of risk. If EPA ascertains that a chemical poses an unreasonable risk or that there is insufficient data to determine the effects of the chemical, EPA can require the manufacturer to test for toxic effects. TSCA subsection 8(e) requires that manufacturers and processors maintain records of "significant adverse reactions to health and the environment (40 CFR 717. 12) and requires submiss ion to EPA of any information supporting a conclusion that a chemical or microorganism presents a substantial risk to health or the environment. Under its authority to limit or prohibit use of new chemicals posing unreasonable health or environmental risks, EPA may establish conditions for the manufacture. processing, packaging, exposure. and labeling of such chemicals or ban them outright. EPA also can issue controls over chemicals through the significant new use (SNU ), reporting, and imminent hazard provisions. The SNU provision requires prior notification for a significant new use of a chemical as defined by EPA. The agency then can set conditions, limitations. or prohibitions based on a new intended use of a chemical. Finally, as with many Federal statutes, TSCA has an imminent hazard provision

PAGE 199

198 A New Technological Era for Americun Agriculture that enables EPA to take action quickly if a chemical poses a serious risk (15 USC sec. 2606). Agency Interpretation and Regulatory Policy 1986 Coordinated Framework PolicyEPA primarily uses TSCA section 5, with its requirement for a PMN prior to the manufacture of any new chemicals to deal with products of biotechnology. The Coordinated Framework (51 FR 23302-23393, June 26, 1986) (77), which designated responsibilities for various biotechnology products held by various Federal agencies, included a policy statement by EPA as to how the agency intended to use TSCA for the regulation of biotechnology; the statement described the categories and microorganisms subject to TSCA, review procedures, and types of information to be submitted for risk assessments. At the most fundamental level, living organisms are considered to be chemical substances under TSCA. Basically, EPA views certain intergeneric microorganisms (microorganisms formed by deliberate combinations of genetic material from organisms in different genera) as new chemicals and therefore under its purview. TSCA pertains to microorganisms used in commercial applications not regulated under FIFRA, FDCA, and other statutes; these applications include chemical production, waste degradation, conversion of biomass to energy, and other environmental and industrial uses. While intergeneric microorganisms are subject to review, naturally occurring microorganisms are not considered new and therefore are not subject to the prenotification requirements of section 5(a)(1) of TSCA, although they may be subject to regulation under other sections of TSCA (i. e., the significant new use rules under section 5(a)(2)). Naturally occurring organisms are implicitly considered to be on the TSCA inventory of substances available in commerce. As for all substances subject to TSCA, manufacturers, processors, or distributors of microorganisms must notify EPA immediately if they become aware of new information suggesting risk from the microorganisms to human health or the environment (section 8(e)). 1988 Draft Proposed Regulationsin general, EPAs efforts to develop regulatory policy have not met with success, and 5 years after the appearance of the Coordinated Framework there still exists no firm EPA biotechnology regulations. Two principle efforts towards developing those regulations will be discussed herethe draft proposed regulations of 1988, which did not come to be; and, in the next section, the draft proposed regulations of 1991, which are the source of current controversy. Since the issuance of the Coordinated Framework, EPA, in consultation with its Biotechnology Science Advisory Committee, worked in 1988 and 1989 to develop draft TSCA regulations for biotechnology. Under the 1986 Framework policy, small-scale biotech R&D efforts involving field tests of intergenerics were requested to submit a PMN. The 1989 draft regulations under TSCA proposed a new regulatory mechanism, the TSCA Experimental Release Application. This mechanism involved the use of Environmental BioSafety Committees (EBCs), based on the concept of Institutional BioSafety Committees (IBCs) established earlier through the NIH-Recombinant DNA Advisory Committee (RAC). This draft EPA rule was reviewed by the interagency Biotechnology Science Coordinating Committee (BSCC) at several meetings. Many concerns were reportedly raised, including the scientific basis for the draft regulations. EPA responded to some comments by sister agencies by making some modifications and then sent the draft proposed regulations to the Office of Management and Budget (OMB) for clearance. BSCC requested that OMB hold clearance until BSCC had time to review its interagency; friction ensued (92). A Request for Comment on Regulatory Approach was published by EPA in the Federal Register on February 15, 1989 (54 Federal Register 7027). Questions raised for comment included: scope of the microorganisms to be subject to EPAs review; scope of EPAs review of R&D field releases of microorganisms into the environment; breadth of definition of commercial purposes by which EPA would have authority under TSCA in educational and research facilities; definitions of release to the environment and contained facility; and to what extent review was to be performed for EPA by independent expert review groups, such as Environmental BioSafety Committees. The draft regulations did not survive. Rulemaking was delayed until EPA policy and plans for TSCA could incorporate the scope document arrived at by interagency consensus. 1991 Draft Proposed RegulationsThe most recent draft TSCA regulations, integrating some eight specific rules, appeared and were extensively reviewed in 1991. Once EPA has completed the process of responding to the recommendations of the BSAC Subcommittee regarding this draft, the regulations enter the final phase of the Agencys internal review process. Under EPAs portion of the 1986 Framework Policy, reporting by persons intending to introduce intergeneric

PAGE 200

microorganisms into the environment for R&L) purposes is voluntary. EPA is now proposing that this is no longer voluntary. However, researchers intending to introduce intergeneric microorganisms into the environment for R&D purposes would at least have the option of filing a TSCA Experimental Release Application. or TERA, as an exemption from a full 90-day notification that would otherwise be required from commercially oriented applicants. The expedited TERA review would generally be completed in 60 days. The extent of the reporting of environmental R&D required will depend on the eventual selection of an interpretation of the statutory phrase commercial purposes. In addition, EPA proposes to exempt some categories of microorganisms when introduced into the environment for R&D purposes (30). The proposed approach is different from the agencys treatment of chemicals, for which review of small-scale R&D activities is not required, presumably because, unlike microorganisms, chemicals cannot reproduce, disseminate, and transfer genetic material. In 1986, EPA had stated that it would try to derive exemptions for some organisms used in contained facilities: in the current draft, some organisms in contained facilities are exempted from review and only a short review is required for specified lists of industrys "workhorses such as Bacillus subtilis. The list is expected to grow with experience. The agency views the new document as following directly from the coordinating principles and scientific rationale of the scope document published by the Office of Science and Technology in 1990 as proposed Principles for Federal Oversight of Biotechnology: Planned Introduction in the Environment of Organisms with Modified Hereditary Traits [55 Fed. Reg. 31, 120 (1990)] (discussed later). EPA has stated that it will subject to regulatory scrutiny only those new organisms that seem likely to present risks. Definitions of new* and risk are subjects of debate. Some view the 199 I draft-proposed regulations as inconsistent with the OSTP drafts risk-based philosophy. The TSCA 1991 draft proposal effectively singles out recombinant-DNA modified microorganisms for oversight. by exempting other categories such as classical transformation systems (e. g., conjugation or chemical mutagenesis). or rearrangements, deletions, or amplifications of genetic material by recombinant techniques. These exclusions are based on EPAs view that such things couldand dooccur naturally and are thus not new. in contrast to recombinations formed with genes from different genera. EPA is proposing three alternative interpretations of commercial purposes in its current draft rule: it may draw a very big net. The first involves selection of commercial indicators that would govern whether a particular field trial would be subject to oversight. The second would apply commercial indicators to R&D conducted in laboratories and greenhouses, for example, and would consider any environmental field release as commercial. and thus subject to screening. The third would permit researchers to rebut the presumption that a field trial was for commercial purposes by showing a lack of commercial intent. Having potentially drawn so many activities into its commercial net, TSCA would defer to whatever agency would most sensibly handle that activity. TSCAs own coverage might not increase to a great extent. Academic laboratories may well fall under the scope of this con~mercial purpose if, as is so often the case today. they have some form of a relationship with a company or perhaps even if their home institutions have dealings with industryas most universities do. Another point of controversy of the proposed draft is its attempt to institutionalize good laboratory procedure and record keeping even in academic laboratories, which previously have not been considered under its jurisdiction (21). The outcome and the acceptability of the draft are not yet known. It contains controversial points and, 5 years after the appearance of the Coordinated Framework, there exists no track record for quick finalization of EPA biotechnology regulations. Implementation 3 EPA currently requests industry to comply voluntarily with the PMN (remanufacture notice) requirements for commercial R&D involving field test releases with intergeneric microorganisms. (Commercial-scale releases are subject to mandatory reporting requirements. ) Because the standard TSCA PMN form is not applicable to microbial products, the Program Development Branch of the Chemical Control Division prepared a document, Points to Consider in the Preparation and Submission of TSCA Premanufacture Notices (PMNs) for Microorganisms, in 1990. The document is intended to give guidance for contained system (fermentation) PMNs and ~ Note the preceding section, Agency Interpretation and Regulatory Policy discusws dmrclopmcnt of EPA policy, which includes proposed implementation. This section examines
PAGE 201

200 l A New Technological Era for American Agriculture environmental release PMNs. It specifies points of desired information, including: description of recipient and donor microorganisms, construction of the PMN microorganisms, characteristics of the PMN microorganisms, production process, worker and consumer exposure, environmental behavior of the PMN strain, and environmental release protocols. Manufacturers or importers of intergeneric microorganisms in and for commerce are required under TSCA section 5(a)(1) to submit a PMN at least 90 days prior to manufacture or import. Communication with a Program Manager in the Program Development Branch, Chemical Control Division, and EPAs Office of Toxic Substance is recommended prior to submission of a PMN. Submitters are encouraged to minimize information withheld as confidential; however, two versions, one with and one without CBI, can be submitted. Companies must now pay a fee of $2,500 for each PMN or consolidated PMN submitted; small businesses must remit $100 per PMN. The EPA publishes a notice on each PMN submission in the Federal Register ( 17). Review of PMNs is conducted on a case-by-case basis, and can involve both EPA scientists and outside scientific experts. Following submission of the PMN, EPA has 90 days to make a determination as to if and how to regulate. During this time a scientifically based hazard assessment and an exposure assessment are conducted. (See ch. 8 and U.S. EPA ( 1987) Toxic Substances Discussion of Premanufacture Testing Policy and Technical Issues; Request for Comment. Federal Register 44, 16243-44). Among the items of information reviewed are: l l l l l l l l l If the identity and characteristics of the source organism, the methods and genetic material used to manipulate the source organisms, the nature of any new traits or functions, purpose and intended effect of application or release, characteristics of the site of application, method and numbers involved in application, containment and mitigation methods, monitoring procedures, and data on environmental fate and effects ( 10). EPA determines during the 90-day period that a new chemical substance may present an unreasonable risk to health or the environment, EPA can prohibit or regulate the substance; if it does not do so, the submitter may proceed. An extension to a 180-day review period can occur, for good cause. Other agencies may be asked for comment, and appropriate State regulatory agencies are contacted. Visits to test sites may occur. The BSAC may review submissions and EPA evaluations. Public comment is viewed as important (67). Application Small-Scale Research TheoryEPA approaches of TSCA small-scale field trials on a case-by-case basis. Unlike commercial research involving chemicals, under the 1991 draft of proposed regulations, recombinant DNA small-scale field tests will receive no automatic exemption from the PMN requirement, although an alternative application process (the TERA) may be used. Experience Base Since 1986, EPAs Office of Toxic Substances (OTS) has reviewed 20 premanufacture notices (PMNs) for release, with the most recent review completed in April, 1990 ( 18). It has been speculated that the absence of notices over the last year may reflect an economic climate unfavorable to commercial development of environmental uses of microorganisms (economic climate seems not to have affected plant submissions); uncertainty as to EPAs regulatory role; or the evolution of the science itself. It may be that biotechnology is to some extent moving away from deliberate release of microorganisms; plants may be easier to manipulate than previously thought. Some suggest that the lack of notices received under TSCA simply reflects the fact that no company is now actively developing rhizobia or other microorganisms subject to TSCA. Bioremediation, the commercial use of microorganisms to degrade toxic waste, will probably not significantly utilize genetic engineering in the near future. It has been suggested, however, that this particular delay may be related not simply to technical reasons but also to uncertainty about regulatory interpretations. The first biotechnology application under TSCA was filed by Biotechnica in February 1987. The application was to field test, in Wisconsin, genetically engineered strains of Rhizobium meliloti to see if these increased alfalfa yields through nitrogen fixation. A Subcommittee of the Biotechnology Science Advisory Committee reviewed the field test protocols and recommended that Biotechnica provide a fuller description of the experimental methods being employed at the site in terms of plot design and monitoring of the organisms after release into the field plot. After consideration

PAGE 202

Chapter 7Environmental Issues: Institutions and Their Regulatory Roles l 201 of BSAC suggestions and other public comments, Biotechnica obtained EPA approval to conduct the field test in spring of 1988. Another early submission was a request in June of 1987 from Monsanto to field test a fluorescent microorganism genetically engineered to be more easily distinguished from other soil microorganisms under laboratory conditions. EPA completed its review in October of 1987. The field trial was held and it demonstrated the usefulness of the gene as a marker for monitoring. Monitoring of the field trials demonstrated that the organism colonized roots; that the population continued to decline; and that migration was limited (67). With the exception of Monsantos field trial, all other environmental use submissions under TSCA have been from Biotechnica. Biotechnicas tests have involved microorganisms genetically modified for improved detection in the environment (antibiotic resistance) and for enhanced nitrogen fixation resulting in potential yield increases ( 18). Commercial-Scale Release Commercial-scale release of genetically modified microorganisms has not yet occurred. EPA might be expected to follow the same case-by-case pattern for commercial-scale release as it has for small-scale release research. As a matter of interest, there have been commercial-scale uses of genetically modified microorganisms in contained systems. Reviews have been completed on 10 PMNs involving the commercial-scale use of intergeneric microorganisms in contained fermentation systems for the production of microbial enzymes. OTHER AGENCIES National Institutes of Health (NIH) The RAC, the Recombinant DNA Advisory Committee at NIH, wrote the now-classic Guidelines for research in recombinant DNA at federally funded institutions and has reviewed cases for compliance with the guidelines. The original Guidelines, issued in 1976, counted deliberate release as one of five classes of experiments not to be initiated at the present time; in the Guideline revisions of 1978, deliberate release into the environment of any organism containing recombinant DNA was prohibited, but provisions were made for waivers through the RAC and NIH; in 1982 revisions, such prohibitions became experiments that require RAC review and NIH and IBC approval before initiation (66). Deliberate release was listed as one of the triggers for RAC review (May 7, 1986, Federal Register, vol. 5 1(88), p. 16960). In fact, however, the RAC has not reviewed any cases since 1987. Since then, EPA and USDA have interpreted their authority to have purview over the vast majority of experiments involving deliberate release. RACs acquiescence to this allocation of oversight is made clear in its Talbot Amendment, stating that once approvals or other clearances have been obtained from an agency other than NIH, the experiment may proceed (Aug. 24, 1987, Federal Register 52( 163), p. 31, 849). In addition, the RAC at its February 4, 1991 meeting, voted to consider deleting planned environmental deliberate release as one of the triggers for its involvement in biotechnology regulation. After duly publishing notice and receiving public input, the RAC met on May 31, 1991 and voted to relinquish this overview. The decision now stands before the Director of NIH. NIH funds very few scientists involved in deliberate release; it also lacks qualified staff to conduct EAs. The RAC, however, intends to maintain its overview of work with transgenic plants and animals inside laboratories, animal rooms, and greenhouses. RACs relinquishing of national overview does not preclude local Institutional Biosafety Committees (IBCs) from considering planned introductions or from bringing up problems to the RAC ( 108). Food and Drug Administration (FDA) Because FDAs authority is over the final food product in interstate commerce. it does not regulate research and therefore is not involved currently in the environmental issues concerning deliberate release research. Exceptions are its jurisdiction over live attenuated vaccines and feed additives including live microorganisms. However, if FDA gives some form of approval for the commercial use of transgenic plants for food, it may have to evaluate the potential environmental consequences of that approval, under the National Environmental Policy Act (NEPA). If a company asked FDA to affirm GRAS (Generally Recognized as Safe) status for a food or to state that a particular variety of plant is acceptable as a source for food, FDA would likely have to assess the environmental consequences of the field use of the plant as part of its evaluation of the food product. An FDA advisory opinion however, might not be a major Federal action requiring an environmental assessment. In its reviews, the agency in the past has limited its environmental assessment to the manufacture and use of the petitioned-for substance. It typically has not reviewed the environmental consequences of the original produc-

PAGE 203

tion or development of the materials used at the manufacturing site. In the case of agricultural commodities, however. the plant itself might be viewed as analogous to the manufacturing facility. If USDA has evaluated the environmental consequences of the field use of the plant, FDA should be able to make use of that information in its own evaluation. It is also possible, although not necessarily likely, that FDA may be able to exclude categorically from its own environmental review those plants that have been reviewed by USDA for commercial use (24). STATE AND LOCAL GOVERNMENT Spectrum of State Approaches to Regulation Significant concern has been expressed regarding the involvement of State governments in the regulation of biotechnology. In addition to coordination with Federal agencies. discussed in a later section. a significant question is the degree to which States should take on independent review authority. On the one hand, State governments may be argued to be "closer to" the people that they are safeguarding and therefore regarded as particularly able or trustworthy as regulators. On the other hand, duplication of Federal regulatory requirements could prove to be an untenable burden on companies. Excessive. idiosyncratic requirements at the State level also might inhibit industrial development. Furthermore, a patchwork of varying State regulatory regimes across the Nation could lead to significant uncertainty on the part of industry, a shopping around for receptive States, or a simple unwillingness to move into product lines related to biotechnology. Compliance with different standards in different States could be a costly problem for industry. State legislation relevant to biotechnology in 1990 included 19 bills spread among 13 States. These fall in the areas of DNA testing (9). bST (4). R&D and economic development (3), deliberate release ( l), general regulations ( I ). and other ( I). In the same year, some 48 bills in 18 States were introduced but not enacted. These referred to bST (20), DNA testing ( I l), R&D) and economic development ( 10), deliberate release (3), general regulations (2), and other (2). Over the past several years, the nine States of Florida. Hawaii. Illinois, Maine. Minnesota. New York. North Carolina, Oklahoma, and Wisconsin have enacted statutes pertaining directly to field testing of genetically modified organisms. with Maine and New York simply creating advisory committees to study issues. In 1991, West Virginia amended its plant pest act to pertain specifically to biotechnology. Many State statutes simply require notification of field test applications to particular State agencies that are to cooperate with the Federal process. Only North Carolina and Minnesota require additional permits. Policy stances taken by various States fall into a broad spectrum. from no or very little administrative or legislative activity (approximately half the States) to moderate activity to, in a few cases, initiation of new regulatory procedures ( 16). Case study illustrations of this range of activity follow. North Carolina In June 1988 the North Carolina Department of Agriculture and the North Carolina Biotechnology Center formed an Advisory Committee to determine whether or not any State regulation was needed and, if so, to develop a suitable regulatory framework. The 27-member Committee included university and private-sector researchers, administrators, business executives, lawyers. and farmers and representatives of government, public interest, and other groups. The committees recommended regulation was passed by the North Carolina General Assembly in August of 1989 as the Genetically Engineered Organisms Act. Funds were appropriated for a staff biotechnologist in the North Carolina Department of Agriculture to administer the law, which requires a permit (either general or limited) for environmental release and for the sale of genetically engineered organisms, with public notice given (8, 16). Minnesota In response to public suggestions in 1987 for rule changes to the Minnesota environmental review regulations, the Minnesota Environmental Quality Board formed a working group on environmental release, which recommended that the EQB should be a coordinating body for genetic engineering. A Task Force was formed, and its report was implemented by legislation in 1989. A permit is required for environmental release of genetically engineered organisms. The EQB is charged with establishing an advisory committee. reviewing proposals, and adapting rules for an environmental work sheet and for a permit for releases (8, 16). Recently, issuance of resultant proposed regulations under the EQB law have caused much controversy. A process for permitting, including an environmental assessment worksheet, would be required for each release of a genetically engineered organism (defined fairly broadly. ) Legislation in 1991 created areas of specific permit authority for the Minnesota Agriculture

PAGE 204

Department (transgenic plants; genetically engineered and experimental pesticides; and genetically engineered fertilizers. soil, or plant amendments). EQB regulations would therefore cover transgenic animals and nonagricultural engineered microorganisms. Both agencies, however, must follow the same specific procedures in proposing environmental assessments (34). California The well-publicized field tests of ice-minus bacteria in Monterey County in 1983-84 (see U.S. Congress, OTA. 1988 for the full case study) ( 102) led to a recommendation that California clarify its biotechnology regulations. Thus, an Executive Order in 1985 established the California Interagency Task Force on Biotechnology. The Task Force systematically identified, evaluated, and communicated the level of regulatory control already pertaining to various biotechnology activities in California. The first product was a handbook, Biotechnology-California Permits and Regulations, published in 1986, with at least 3,000 copies distributed by the summer of 1989. The chief finding was that the current regulations were quite complete in their coverage of biotechnology. Four permit procedures were enhanced to provide for increased input from the public (8, 16). New Jersey Stimulated by the repeated introduction (without enactment) of a State legislative bill that would have regulated environmental release, and by the enactment of several local ordinances for such regulation, the New Jersey Department of Environmental Protection developed a white paper on recommendations for the development of State policy on biotechnology. Following informal discussions among agency representatives, an Interagency Committee on Biotechnology was appointed by Departmental Commissioners in the fall of 1989. with university advisors. The committee is evaluating: l l l l l l The the effectiveness of State laws to regulate biotechnology, coordination with Federal agencies, the needs of industry in complying with regulations, other States policies, the need for biotechnology education, and appropriate roles of the State and its agencies. first priority is evaluation of New Jersey statutes and coordination with Federal agencies, with the objective of compiling a California-like handbook (8). Inter-State Gatherings and Consensuses of State Regulators In recognition of the importance of State regulatory agency officials as part of the full system of regulation. the USDA hosted conferences in 1989. 1990. and 199 I on Federal and State Regulation of Biotechnology Emphasis was placed on clear communication from Federal agency representatives to State agency representatives about the details of the implementation of Federal biotechnology regulations. The 1990 meeting attracted some 130 people. the great majority from State agencies. University, private-sector. and environmentalist representatives attended as well. The third meeting, in 1991. concentrated on the issues of large-scale commercial release. In recognition of the varying degrees of unease felt by State regulators having to come to grips with biotechnology, a special workshop for State agencies, State Oversight of Biotechnology, was held in conjunction with the second Federal conference, sponsored by the University of California Systemwide Biotechnology. Research and Education Program and the New Jersey Department of Environmental Protection. Case histories of the development of various State policies were shared. Brainstorming seminars led to a consensus set of recommendations for State regulatory officials. The resulting document, Guidance for State Governments on Oversight of Biotechnology, included the following Points to Consider for States considering how to handle biotechnology oversight: 1. 2. 3. evaluation of the existing (Federal and State) oversight framework for biotechnology; organization of a task force to include representatives from multiple agencies, industry, academic and public interest groups; and activities of the task force, which should include identifying and reviewing existing State statutes and Federal agency roles; recommending needed actions. if any; delineating clear pathways for applicants to follow; working with local governments; and communicating with and involving the public (39, 59). In 1991, a follow-up workshop emphasized specific points at which coordination between State and Federal agencies could be fine-tuned. Spectrum of Local Approaches to Regulation The first local response to biotechnology occurred in Cambridge. Massachusetts. in the ordinances passed in

PAGE 205

1977. Concerns over genetic engineering research in university laboratories led to sometimes heated hearings and local regulations. Some years later, an equilibrium seems to have been reached between town and gown. Some companies find the existence of known local regulations to be positive, although others find them problematic and subject to change with newly elected local politicians. Such an open clash has been fairly unusual, although in one 1989 case the city of Burlington, Vermont and the University of Vermont clashed over the construction of a building to house much of the universitys molecular biotechnology research. The city demanded input into, if not the approval of, experiments to be conducted in a new building. The University refused, and the press attacked the Universitys stance (6). In March of 1991, a Memorandum of Understanding between the city and the University called for the establishment of a task force to discuss plans together. Like Cambridge, Burlington was not particularly concerned with deliberate release. In New Jersey, on the other hand, initial local concerns focused on perceived risks associated with deliberate release of genetically engineered organisms. When Statelevel legislation was not enacted, concerned politicians provided to municipal governments model ordinances to restrict the environmental release of genetically engineered microorganisms. By early 1990, six municipalities had adopted such ordinances. Other municipalities debated such ordinances, but decided against enactment, in part because pertinent expertise was recognized as lacking at the local level (41). To forestall negative public reactions, the AgBiotech Center of Rutgers University in New Jersey began working with the local community from the earliest moment. They formed a Citizens Advisory Committee to provide input and air public concerns over its planned field-trial facility for genetically engineered plants. Local planning boards, a homeowners association, farmers, and agricultural organizations appointed members to the committee. The committee reviews plans for the facility and applications for field trials therein. The committee also is charged with communicating information to the public (88). INTERNATIONAL REGULATORY CLIMATE Biotechnology, as a scientific endeavor and an industrial activity, is international in scope. Those concerned with U.S. economic competitiveness or with the global environment have reason to be interested in the degree to which deliberate release regulations are internationally consistent and coordinated. A brief sketch of regulatory approaches in several countries follows. Europe Status of Regulations, EC 1992 European Community (EC) directives were passed in April of 1990 concerning contained use and deliberate release of genetically modified organisms. Member States were supposed to draft national laws by October, 1991, in alignment with these minimum standard directives. Each State can, and some may well, add more restrictive measures; different member States will achieve different balances regarding restrictiveness of regulations. Pressure groups such as the Greens in Germany, for example, will attempt to counteract the voices of industry concerned with economic competitiveness. Despite the potential for some country-to-country variation in regulatory rigor, the directives are meant to provide more of a bottom line consistency among States in terms of protecting the environment than was present in the past. According to the EC Directive on the Deliberate Release of Genetically Modified Microorganisms, No. 90/ 220/EEC, releases are permitted only in countries with relevant national approval procedures. The EC hopes for an EC-wide approval procedure for releases of commercial products. This would allow free distribution of products throughout the EC. Deliberate releases will be evaluated and approved or disapproved on a case-by-case basis; hence, there may be room for flexibility in and evolution of regulations. Environmental impact assessments and consent by competent authorities are prerequisites of release. Different stages in establishing a basis for national decision making have been reached by different EC countries. Approximately one-half of the member States passed implementing legislation by the October 1991 deadline. In the United Kingdom (UK), a biotechnology regulatory framework is part of an introduced Environmental Protection Bill that is intended to form the basis of future detailed regulations. In Germany, the Gene Law was enacted in July 1990. Under pressure from some of its largest industries, Denmark retracted its extremely stringent 1986 law; deliberations as to implementation of EC directives are ongoing. In France, procedures are straightforward and nonburdensome; over 50 field trials have taken place. In the Netherlands, permits for field trials are granted by the Ministry of the Environment

PAGE 206

( 105). In France, some 67 uncontained experiments took place between 1987 and 1990 ( 12). Some analyze EC directives with a positive spirit and view the goal of developing a coordinated science-based approach to regulation as helpful to biotechnology in the long run. Science-based regulation, even if it varies among member countries, may well be preferable to idiosyncratic applications of disparate laws already on the books in different countries (47). In any event, it is not yet clear what balances will be achieved by diverse countries weighing such factors as environmentalist pressures, industry lobbying, scientific findings, and competitiveness concerns. The foundation is laid for commonality, but the likelihood is that different countries will find their own paths. True homogenization is not likely to be achieved by Europe 1992. The loathing that industry feels for regulatory uncertainty might give the United States at least a transient competitive advantage over at least some countries if regulatory uncertainty here is minimized. The Fourth Hurdle The fourth hurdle causing real worry among biotechnology advocates refers to a fourth criterion for European regulations of biotechnology. This fourth criterion would be the inclusion of socioeconomic values in the approval process. The usual three technically based hurdles for regulations generally are safety, quality, and efficacy ( 15). The fourth hurdle is controversial, and of great concern even to U.S. industry. Perhaps discussion of this hurdle has peaked already, and it may be declining in importance. However, observers believe that interest could intensify again at any moment. An attempt based on socioeconomic values to ban veterinary growth hormones was voted down late in 1990, suggesting that institutionalization of such values may be unlikely (47). Harmonization Despite differences among member states and among EC directorates, European countries and the United States are making good-faith efforts to harmonize regulations. Enlightened self-interest regarding economic competitiveness doubtless plays a role. Several forces for harmonization include: The Organization for Economic Cooperation and Development (OECD). the Office of International Epizootics (OIE), United Nations Agencies (UN), The World Bank, and bilateral discussions with the European Commission (EC). The OECD, which includes 25 industrialized countries, many but not all of which are European, has several projects related to regulation of biotechnology, including: l l l l l l OIE Good Development Practices; Guidance for the Design of Small-Scale Field Research With Genetically Modified Plants and MicroOrganisms; Good Industrial Large-Scale Practices: Monitoring of Genetically Modified Organisms Introduced into the Environment: Findings and Suggestions; Performance Evaluations for Plant Cultivar Development; and Food Safety. discussions focus on development of internationally equivalent. appropriate standards for evaluation of veterinary biological products derived through biotechnology. Within the EC, bilateral discussions have occurred through the U.S./EC Bilateral Discussions on the Environment, the High-Tech Group, and the Task Force on Biotechnology Research (57). Perhaps the most compelling example of harmonization is the development of a common document on biotechnology safety by the 23 member countries (including many European countries, as well as the United States. Canada. and Japan) of the OECD. First published inhouse as Good Developmental Practices (GDP) for SmallScale Field Research, it was reworked and released f-or public comment in 1990. GDP outlines scientific principles and conditions for proposal review and also gives guidance to researchers designing small-scale field tests of plants and microorganisms. The document may be augmented by another paper(s) as more data are compiled: the basic approach is aligned with the principles advocated in the 1989 National Academy of Sciences report on safety in field testing (73). Acceptance of this document by 23 countries has been a significant step toward international harmonization of biotechnology fieldtrial regulations. The fact that the United States was the lead country in developing the document ensures good harmonization with U.S. regulations: this. in turn, should facilitate international trade (55). Currently, the United States is the designated lead for OECD in drafting an OECD discussion paper on scientific issues associated with performance trials of plant cultivars. A principal objective of this endeavor is to enable policy bodies to make recommendations and decisions based on sound science when they consider largescale plantings of new agricultural crops. including those developed with new biotechnology techniques (24). This represents a stage beyond the small-scale research cov-

PAGE 207

206 A New Technological Era for American Agriculture ered by GDP as performance trials involve more plants and there may be no means of ensuring that plants remain confined to experimental sites. Performance trials, however, still qualify as R&D; issues associated with commercialization of plant crops are not directly addressed in this OECD paper. Canada Status of Regulations Following the lead of a Federal Government task force in 1980, Canada implemented a national biotechnology strategy in 1983 and established the Interdepartmental Committee on Biotechnology in 1985. The committee began with the premises that the product rather than the process would be regulated, building on current legislation. Additional concerns could be addressed with guidelines. Canadian regulations would harmonize with those of other countries wherever possible and practicable. A biotechnology users guide to Federal regulations has been updated recently, assisting applicants with identification of appropriate agencies, contact people, and procedure. In 1987, an ad-hoc committee was formed on environmental release. Agriculture Canada, dealing with organisms used in agriculture, and Environment Canada, along with Health and Welfare Canada, dealing with microorganisms used for nonagricultural uses, are the chief players in the regulatory arena ( 14). Currently, regulatory bodies and others in Canada are considering a draft of Proposed Notification Regulations for Biotechnology Products under the Canadian Environmental Protection Act. Developed by Environment Canada and Health and Welfare Canada. notification requirements will eventually become regulations under the new substances provisions of the Canadian Environmental Protection Act (CEPA) and will apply to new biotechnology products manufactured in or imported to Canada. Notification and assessment periods as well as information required are defined based on whether the biotechnology product will be used in contained manufacturing or released into the environment. All biotechnology products will be considered new substances under these regulations. Environment Canada is currently in the process of developing a Domestic Substances List for those biotechnology products in commercial use in Canada between 1984 and 1986. Once a microorganism is added to this list, no further notification is required by a user if the product is used for the purpose specified in the list. Guidelines are being prepared to assist those needing to submit notifications for this list. For release into the environment, notification would be required prior to importation. commercial manufacture, small-scale field trials, or large-scale field trials. Currently, information required for a field trial would include: objectives, site details, experimental design, site supervision, introduction protocols, containment procedures, monitoring procedures, termination procedures, and mitigation procedures. In the interim, while the proposed regulations are being developed, notification to Environment Canada is recommended for those with intent to manufacture or import into Canada biotechnology products (80). Harmonization Probably the closest working international relationship in the area of biotechnology regulation exists between the United States and Canada, which may not be surprising given their geographical proximity and free trade agreement. EPA officials have met with representatives of Environment Canada and have had informal contact with other relevant Canadian agencies. USDA officials have met with Agriculture Canada officials yearly for 4 years and communicated between meetings on rationale, procedure, and so on. U.S. companies can do field tests in Canada; requests that U.S. officials accept Canadian field test data are expected in the near future. Review systems similar to the U.S. biotechnology permitting system have been established by Canada, taking into account the basic principles on the safety of field testing shared by all OECD countries (60). Japan In general terms, Japans regulation of biotechnology is in line with international standards. Research guidelines are based on the early NIH guidelines, and industry guidelines are consistent with OECD. The Ministry of Agriculture, Forestry, and Fisheries (MAFF) issued the first regulations on environmental release of plants in the summer of 1989 ( 103). Government guidelines emphasize a step-by-step approach to field tests and a case-bycase basis for approval (94). USDA has worked with MAFF on how to conduct reviews and in a consultative group on monitoring. Japans environmental directorate is looking at microbiological field releases. One field test has been approved to date in Japan; Japanese companies are requesting field trials in Mexico (58). Recognition of the importance of facilitating field trials is growing in Japan. In the second half of 1990, for example, Japans MAFF announced its intention of or-

PAGE 208

ganizing an incorporated association of over 100 Japanese biotechnology-related companies. In addition to promoting biotechnology in relevant industries. the Society for Techno-innovation of Agriculture, Forestry, and Fisheries (STAFF) is expected to be involved in promoting and authorizing field trials of genetically modified organisms (45 ). In addition, Japans first isolated. openair field site for transgenic plants has been constructed in Tsukuba, Ibaraki Prefecture, by the National Institute of Agro-Environmental Science (NIAES). NIAES scientists plan to test environmental effects of tomatoes engineered to resist the tobacco mosaic virus. Developing Countries In general. developing countries have neither biotechnology regulations nor focused biotechnology staff in their regulatory agencies. One relatively unusual example of activity is the recent formtaion of the Genetic Engineering Approval Committee in India. The group regulates the production and release of genetically engineered organisms and potentially harmful microorganisms (76). A variety of efforts from the developed countries, some based on differing premises, are being made to include developing countries in current regulatory approaches. The early stages of harmonization may take place quite naturally in developing countries that have some serious interest in biotechnology. Such countries tend to send representatives to the United States to learn about approaches taken here. APHIS-BEEP for instance. has exchanged information with China, India, Mexico. Costa Rica. Brazil, Argentina, Chile, Nigeria. Kenya, Zimbabwe, Thailand. and the Philippines on regulatory philosophy, mechanisrms by which that philosophy is implemented, and ways to handle risk assessment and risk management. USDA has held a variety of conferences on related topics, which are well attended by international representatives. Various U. N. agencies are exploring different avenues through which to assist technology transfer of biotechnology to developing countries while safeguarding environmental and human health. The U. N. Industrial Development Organization (UNIDO) has developed a voluntary code of conduct to provide guidance for introducing biotechnology products into developing countries. The World Bank has hired a biotechnology advisor to consider biotechnology issues with the Consultative Group on International Agricultural Research (CGIAR). although most of the 18 CGIAR centers are not yet close to field trials. The National Research Council has published a panel report on Plant Biotechnology Research for Developing Countries. Some developing world observers question the appropriateness of automatic wholesale adoption of stringent regulations by developing countries (42). POLICY ISSUES Jurisdiction and Coordination Mechanisms of Coordination at the Federal Level The 1986 Coordinated Framework, described earlier, was a crucial step in establishing and clarifying jurisdictional authorities for a new technology with diverse applications. To further clarity jurisdiction as biotechnology matured toward products, and to help Federal agencies formulate regulations and guidelines based on existing statutory authority, the Biotechnology Science Coordinating Committee (BSCC) was established by the Office of Science and Technology Policy (OSTP). (50FR 4717447195, November 14. 1985). BSCC was charged to monitor the changing scene of biotechnology and serve as a means of identifying potential gaps in regulation in a timely fashion, making appropriate recommendations for either administrative or legislative actions. Until recently, the BSCC provided a forum for senior policy officials from USDA, EPA, FDA. NIH. and NSF as they attempted to coordinate policy. promote consistency in review procedures, and identify key issues. One outcome of this forum was the interagency funding of the 1989 National Academy of Science (NAS) report, Field Testing of Genetically Modified organisms: A Framework for Decision-Making. The BSCC also has helped to resolve jurisdictional conundrums, such as whether EPA or USDA is the lead agency in cases of dual jurisdiction. Despite such positive contributions, however, the BSCC had difficulties achieving consensus on important issues such as risk assessment and management, levels of oversight appropriate for certain organisms. definition o! deliberate release, and coherent standards for oversight (11). These difficulties arose in part because different agencies have different statutory mandates and built-in approaches to regultaion. BSCC also wa S criticized for its closed-door deliberations and for rneddiling in regulatory agency affairs. Nonetheless. the committee helped initiate formulation of broad principles for regulation (27). In the absence of agreement within the BSCC, Dr. Allen Bromley. director of the Office Of Science and

PAGE 209

208 l A New Technological Era for American Agriculture Technology Policy, decided that the identification of organisms subject to Federal oversight had policy implications beyond the jurisdiction of the BSCC [55 Fed Reg. 31,120 ( 1990)] and the issues should be addressed by the appropriate policy bodythe Presidents Council on Competitiveness. Moved under the aegis of the Working Group on Biotechnologyof the Council on July 31, 1990, a scope document pertaining to initial releases of biotechnology-derived organisms into the environment was published for public review and comment [55 FR 147, 31118 ( 1990)] by the Office of Science and Technology Policy. The document, Principles for Federal Oversight of Biotechnology: Planned Introduction into the Environment of Organisms with Modified Hereditary Traits, proposed principles for ensuring the safety of planned introductions, while still not unnecessarily inhibiting the process. Certainly, interagency disagreement has existed. It has been said, however, that the extent of collaboration on biotechnology issues among Federal agencies that took place in the drafting of the Principles is unprecedented (61 ). The scope document expands on the Coordinated Framework; its criteria for regulatory oversight are riskbased, with the objective of differentiating between organisms that do and do not require oversight at various levels of jurisdiction. Federal agencies may implement the criteria in their own ways as they categorize organisms according to the risks associated with environmental release and thus can be excluded or exempted from oversight. Some introductions may be considered similar to preceding, safe introductions; for others, risk information or current regulations make additional Federal oversight unnecessary. On the other hand, unfamiliar organisms or organisms that might present a risk not yet assessed would be subject to an assessment (62). The scope document considered all organisms with deliberately modified hereditary traits as potentially subject to oversight, regardless of the techniques used to produce them. However, exclusions from such oversight should be granted to introductions posing no risk. Examples include: plants and animals produced through natural reproduction or breeding and microorganisms modified by chemical or physical mutagenesis or the transfer of nucleic acids through physiological processes. Such exclusions are based on previous safe experience with products produced with these traditional processes. In addition, organisms produced by other processes, including recombinant DNA techniques, should be exempt from oversight if they pose no greater risk to the target environment than parental strains that are considered safe. An extremely broad class of organisms potentially is subject to oversight. In this sense, the products of new biotechnology are not singled out as inherently more risky than those resulting from nonmolecular techniques such as plant breeding (55 Fed. Reg. 147,13 1118 (1990). Nonetheless, exclusion from oversight, based as it is on criteria of familiarity, is possible for virtually all methods of modification except those using molecular or rDNA techniques. Just as operationally, regulatory examination to date has been triggered by the process of recombinant DNA, in the near future, at least, other novel techniques are equally likely to draw the attention of regulators, if only because they point to the presence of a novel product. The apparent contradiction between this reality and the scope documents attempt to focus on products, not processes, mirrors the conflicting views of those scientists and industry representatives who maintain that the products of biotechnology pose no unique risks; and those who believe that the novel characteristics of biotechnology products and scientific uncertainty about risks warrants extra caution. The product versus process debate continually resurfaces. An exceedingly fine line divides regulation of a biotechnology product and regulation of a process. USDAs approach to the balancing act between process as trigger and product as legitimate focus is to review any implications for the safety of the end-product that might arise from the technique applied. For example, clean characterization of the gene transferred is particularly important if the genetic material is taken from a plant pest, so it is clear that no unwanted genetic information is transferred. This pragmatic approach should be readily applicable to novel techniques in addition to recombinant DNA itself. Using the safety of the product as the focus for review allows regulators to take into consideration any and indeed all pertinent aspects of any techniques or processes leading to novel products, thereby avoiding gaps in coverage. Algorithms for using risk as the trigger for oversight have been and are being developed (69). Some companies. well advanced in their product development, desire regulations that effectively will end the product v. process debate so that progress can be made in bringing products to market. On February 27, 1992 the Office of Science and Technology Policy published in the Federal Register (vol. 57, No. 39) its revised scope document, describing policy on Exercise of Federal Oversight Within Scope of Statutory Authority: Planned Introductions of Biotechnology Products Into the Environment. A principal change from the draft published earlier is the elimination of a previously controversial exclusion categoryexclusion for

PAGE 210

conventional technologies. By eliminating this exclusion from oversight, some policy makers believe the new scope document is more consistent with its own premise, i.e., that no special risk is attached to the recombinant DNA modification process. Oversight of conventional and new technologies is, however, left to the regulatory agencies. Agencies are continuing to craft regulations and guidelines in response to the scope documents policy directives that existing statutes provide sufficient authority for adequate regulation and that regulation should be riskbased. EPA, for example, is crafting its regulations for biotechnology; regulations under TSCA still have not been finalized. USDAs ABRAC guidelines for research have been put out for comment. As biotechnology moves to the commercialization stage, where releases could occur on a large scale, amendments may or may not be needed. Coordination among agencies is critical, as regulatory policy evolves to avoid redundancy and delays in policymaking. Several interagency bodies will play a coordinating role, including the Office of Management and Budget (OMB), BRS (the research-oriented successor to BSCC), the National Biotechnology Policy Board. and the Presidents Council on Competitiveness (COC). The Biotechnology Research Subcommittee (BRS), of the Committee on Life Sciences and Health, is part of the Federal Coordinating Council on Science, Engineering, and Technology (FCCSET). Formed in 1990, the BRS succeeded the BSCC and focuses on issues such as research priorities, needs, and training rather than on policy issues. As an interagency body, the BRS includes the acting heads of the NIH and the FDA, with additional representatives from the State Department and its Agency for International Development. the EPA, USDA. NSF, NASA, Department of Commerce, Department of Defense, Department of Interior, Department of Energy, Office of Management and Budget, and OSTP. The Administrations final policymaking body for biotechnology, the Council on Competitiveness (COC), includes the Vice President; the Presidents Science Advisor: White House Council; the Secretaries of HHS, Commerce, Defense, Treasury, Energy, and Agriculture,; the EPA Administrator; the NSF Director; the U.S. Attorney General; and the Chairman of the Council of Economic Advisors. Biotechnology issues will be considered first by the Councils Working Group on Biotechnology. A significant action in biotechnology by the COC was the publication of its Report on National Biotechnology Policy in February, 1991. (See box 7-C. ) The thrust of the report is that biotechnology products essentially are equivalent to products developed through other procedures and that, therefore, the domestic biotechnology industry should not be burdened by excessive restrictions The report also suggested that the COC and its Biotechnology Working Group take the lead in coordinating regulation of products introduced subsequent to the 1986 Coordinated Framework. The Working Group was also charged with coordinating communication among industries; streamlining review procedures; reevaluating regulations as necessary; and dealing with inconsistencies of international. State. and local policies, regulations, and laws (28). Responses to the COC Report are predictably diverse, ranging from those of environmentalist groups, who still call for special regulatory attention to biotechnology, to industry representatives, who hope that the report will push toward clearly defined regulatory criteria. thus enabling company executives to estimate accurately the time and costs involved in winning approval for testing and marketing biotechnology products (2). The new National Biotechnology Policy Board, established by the Administration according to the instructions from the Senate Appropriations Committee in its report on the 1989 HHS budget. will play a purely advisory role. Its public members as well as voting governmental members report to the HHS Secretary ( 100). The Board will review research. nonconfidential privately funded biotechnology activities. and the development of industries and products and make recommendations to the President and Congress (84). Comparison of USDA and EPA Approaches to Biotechnology Oversight Each of the two major agencies involved in biotechnology oversight must, under its own specific mandates, attempt to provide technically sound judgments on risk, while expediting regulatory procedures and developing a foundation of experience on which to base future judgments. Types of information used by EPA and USDA to make regulatory judgments include 1) that required for the evaluation of deliberate release applications or notifications, 2) experience base, and 3) application and notification processes. By far, the largest experience base with regard to field trials is that of USDA-APHIS in working with transgenic plants. In terms of products licensed for real-world use, USDAs largest experience base is with category I (animal biologics). While EPAs Office of Pesticide Program deals

PAGE 211

210 A New Technological Era for American Agriculture I 1. 2. 3. 4. Box 7-CCouncil on Competitiveness Report: Four Principles of Regulatory Review Federal Government regulatory oversight should focus on the characteristics and risks of the biotechnology product-not the process by which it is created. For biotechnology products that require review, regulatory review should be designed to minimize regulatory burden while assuring protection of public health and welfare. Regulatory programs should be designed to accommodate the rapid advances in biotechnology. Performancebased standards are, therefore, generally preferred over design standards. In order to create opportunities for the application of innovative new biotechnology products, all regulation in environmental and health areaswhether or not they address biotechnology-shouid use performance standards rather than specifying rigid controls or specific designs for compliance. SOURCE: The Presidents Council on Competitiveness, Report on National Biotechnology Policy, 1991. with increasing numbers of microbial pesticides, the Office of Toxic Substances has had few recent applications for planned introductions of recombinant DNA modified microorganisms and the subject matter of its applications has been limited narrowly to nitrogen fixation. The time required for, and general types of steps involved in application and notification processes are roughly comparable for the two agencies. From 3 to 6 months seem to be required for these processes. APHIS, with its large body of experience, probably has the most regularized review processes today. Coordination With States A few State governments independently have promulgated deliberate release regulations (see State and Local Government). Most feel that effective coordination with the Federal agencies will suffice. The COCs Biotechnology Working Group is charged with coordinating Federal laws, regulations, and poilcies with those at the State level. As a practical matter. the task of coordination lies with the individual agencies themselves. USDA and EPA use State input in different ways. Based on its traditional network of connections with State-level agricultural departments, USDA has explicitly incorporated State review applications for field tests into its overall review process. USDA also has brought together Federal and State regulators of biotechnology in annual national meetings. EPA, on the other hand. does not have a tradition of elaborate, direct connections to State environmental departments. Recently. EPA has attempted to identity biotechnology point people in State environmental departments (68). However. many State regulators may not feel bin the loop in terms of knowing what EPA is doing in biotechnology regulations and how their State should play a part (64, 93). EPA pubilcty acknowledges the importance of receiving State input. but procedures for gathering this input are far less formalized than is the case in USDA. Still, EPAs TSCA Office did consult with State regulators for each of Biotechnicas seven field test requests (32). For the relatively few release PMNs handled, EPA-OTS has developed an informal set of steps to: l l l l l l include telephone contact with the appropriate State regulatory agency or agencies concerned with a particular submission; make available a nonconfidential version of the PMN on request; include State personnel in a site visit; make available public docket materials on request; provide opportunity for State personnel to comment on the Agencys draft risk assessment; and give State personnel a draft of the TSCA section 5(e) order. with conditions for the field test (30). Coverage Scope Possible GapsSome concern has been voiced that under the current allocation of regulatory responsibility for biotechnology, some releases might slip through the cracks. An often cited potential gap in jurisdictional authority pertains to genetically engineered plants that are neither pesticidal nor themselves plant pests. In such a case, where neither EPA-FIFRA nor USDA-APHIS has clear responsibility, the question has been raised, who would have oversight over field trials! (35) In the past, regulatory oversight for field trials largely has been allocated with traditional recombinant DNA in mind. Even newer techniques have arisen. however. such as biolistic or gene gun approach to injecting genes into organisms. How will the new techniques being de-

PAGE 212

veloped fit into the oversight structure? Should they be? Can the experience base derived from traditional recombiant DNA be applied to new techniques? As the science of biotechnology advances. it is likely that genes of more than one trait will be inserted into a plant variety being developed. This mixing of genes could lead to an overlap of authority. For example. a Bacillus thuringiensis gene for pest resistance could trigger EPA review under FIFRA in a food crop; a gene for a nutritional component could trigger FDA responsibility: while the use of a plant pest vector could trigger USDA oversight. Even though USDA-APHIS and EPA-FIFRA have a history of cooperation, some difficulties could arise in treating such situations. A company might have to submit three packages for review due to the different roles of each agency. This could comprise a regulatory burden. It also has been asserted that. apart from federally funded research, Federal oversight of genetically engineered animals is limited to selected invertebrates and animals with genetic material from plant pests. While most livestock animals would probably generate little risk to the environment if genetically engineered, aquacultural species have been cited as potentially more problematic. The possibility of escape of genetically engineered fish from outdoor aquacultural ponds to watersheds. where interbreeding with natural populations could occur. gives rise to ecological concerns (35). (See box 7-D. ) Thus, while some observers are concerned about possible limits to and gaps in Federal oversight of transgenie plants and animails. some assert that by far most cases of release of transgenic animals wuld be covered by USDA Science and Education (for research), USDAAPHIS (for plant pest invertebrates and animals carrying animal diseases ). FSIS or FDA ( for use of animals as food). FDA and APHIS (for animal drugs and biologics). and the Public Health Service (for interstate movement of etiologic agents that carry human disease. ) Only research not receiving Federal funding, in which the animal is not a plant pest. not an agent for animal or human disease, not given a drug or biologic. and not to be sold as food (92) could constitute gaps in oversight of transgenic animals. Thus, while some observers are concerned about possible limits or gaps in Federal oversight of transgenic plants and animals, others expect the natural evolution of oversight to occur. It remains to be seen whether the regulatory framework is flexible enough to catch such cases, and how, for example, the system handles genetically engineered plants that are neither engineered for pest resistance nor themselves plant pests. Current and Projected Treatment of Such Organisms and ProductsFor its part, USDA-APHIS seems to be willing to extend its range of oversight regarding genetically modified plants. Plants abilities to act as pests can be viewed in a broad context. Potential disruption of the environment by novel plants could in the broadest sense qualify a plant as a potential pest. Some environmentalists feel that USDA already is stretching its statutory scope to deal with biotechnology, and may not have the authority to extend its scope still further. Box 7-DFish Regulations: Something To Carp About? The gene that regulates growth in the rainbow trout was transferred into carp by a team of scientists from the University of Maryland, Auburn University, and Johns Hopkins University. In experiments to date, the carp have grown 20 to 40 percent larger than their unmodified relatives. Among some participants in the fish farming and research industry, enthusiasm runs high over the prospect of impacting the Nations $900 million fish farming industry and, eventually, helping to feed the hungry of the world. Others emphasize caution. The American Fisheries Society, composed of fisheries scientists, has recommended close monitoring by the Federal Government, tight control over the environmental release of a modified fish, and sterilization of the fish (75). The transgenic fish project was started in 1986; in February 1990, USDA approved the project but protests from four public groups persuaded the North Auburn Fisheries Research Unit, at Auburn University, the site of the project, to redesign the pond. The new place was approved by USDA in November 1990, pending inspections early in 1991. Current design places the fish in 10 outdoor earthen ponds, set on concrete stabilizers, surrounded by chainIink fences covered with bird netting, double and triple screened drains and ditches. Beyond these is a 17-acre lake filled with predatory fish, and then a pond with chemical and mechanical barriers before the local creek (1). SOURCE: Office of Technology Assessment, 1992.

PAGE 213

212 A New Technological Era for American Agriculture ExemptionsExemptions, as opposed to accidental gaps in coverage, are cases or classes of planned introductions deliberately excluded from regulation. Many questions underscore the dynamic, evolving nature of the regulatory situation. For example, willor shouldthe trend toward examining new products of biotechnology carry over to the products of traditional field trials. which now are exempted implicitly from review? Or, as novel techniques become more familiar, will they be less likely to serve as triggers for product review? In other words, will we learn enough to exempt certain products resulting from certain biotechnology techniques? The NIH RAC has relaxed its recombinant DNA Guidelines as an increasing experience base has indicated the appropriateness and safety of so doing. The 1989 National Research Councils report on biotechnology endorsed such an experiential approach to environmental releases: As field tests are performed, information will continue to accumulate about the organisms. their phenotypic expression, and their interactions with the environment. Eventually, as our knowledge increases, entire classes of organisms may become familiar enough to require minimal oversight (73). The 1990 draft Scope principles reinforced the idea that information-based familiarity can lead, when appropriate, to exclusion from oversight. Both EPA and USDA endorse the concept that biotech oversight can evolve on the basis of information gathered. Already, these agencies are beginning to exempt from review or expedite review of certain classes of organisms or products if certain conditions are met (65). EPA Definition of a Microorganism as a Chemical Compound The application of TSCA to biotechnology has raised some controversial and as yet unresolved issues. Paramount among these concerns is the inclusion of biotechnology products under the definition of a chemical substance, whence EPA draws its authority to regulate genetically engineered microorganisms. Although it is clear that DNA molecules can fall under the definition of chemical substances, it is less clear whether the host organism can be so defined. On the one hand, Witt writes: Calling microorganisms chemicals is tantamount to calling chemists chemicalsor regulators chemicals. On the other hand, some in industry feel strongly that microorganisms have uses that are directly connected to their chemical nature and that EPA jurisdiction is very reasonable ( 107). EPAs interpretation has on occasion been called ripe for litigation (53). In any case, it is unclear whether the scheme of regulation envisioned and currently employed for conventional chemicals is suitable for oversight of biotechnology (48 ). Regulatory approaches for chemicals may be difficult to apply to living organisms. Indeed, the fact that TSCA regulations for biotechnology products have not yet been finalized, despite having gone through various iterations, may result in part from the difficulties inherent in manipulating rules conceptualized for chemicals into rules appropriate for living organisms, although EPA has reviewed microbial PMNs under TSCA since 1986. Other problems may include technical difficulties in defining new organisms, interagency disagreements, interpretation of commercial purposes, and the small-quantities exemption. Nonetheless, the intent of Congress that TSCA serve as gap-tilling legislation seems to invite its use for some biotechnology products that would otherwise have no obvious regulatory home. From the coordinated framework, the role of TSCA in biotechnology seems to have been accepted. on at least an operational level, even it the broad definition of a chemical compound has not been universally popular. The trigger under TSCA for PMN is manufacture of a chemical, not the issue of safety. Therefore, when this traditional trigger for TSCA is applied to biotechnology, it is not consistent with the emphasis based on technical risk in the Scope Principles. It is often argued, however, that since all new chemicals must be reviewed. no implications of risk are ascribed automatically to biotechnology products falling into this net. Commercial v. Research Authority EPABecause TSCA is a commercial statute, it arguably does not apply to the deliberate release of genetically engineered microorganisms in nonindustrial settings. EPA currently requests industry to comply voluntarily with the PMN requirements for commercial R&D involving field test releases with intergeneric microorganisms. Academic researchers performing comparable releases may be seen as left out of the loop, in a regulatory limbo. Congress expressly exempted small-scale research and development from TSCA authority. Much depends on the breadth of EPAs interpretation of commercial purposes. For example, academic research may be colored by commercial intent because it maybe funded by an industry source: because patent rights are assigned to a company for commercial development; or even because a researchers home institution receives private-

PAGE 214

Chapter 7Environmental Issues: Institutions and Their Regulatory Roles l 213 sector funding. One possibility is that all field test releases will count as commercial in intent. However, problems may arise with a broad net approach. Other agencies, as well as universities, may question the validity of this approach. EPAs possible move into the R&D laboratory under a similar approach is likely to arouse fears of excessive layers of bureaucracy among laboratory researchers. USDAThe possibility of EPA penetrating further and further into the realm of research, despite its commercial mandate, has a counterpoint: USDA appears to be exploring ways to step back a pace from its review of field trials conducted as academic research. The agencys Proposed USDA Guidelines for Research Involving the Planned Introduction in the Environment of Organism With Deliberately Modified Hereditary Traits. (FR56 (22):4134 ) seems to place much of the weight of the research review process at the institutional level, with the goal of minimizing the weight of bureaucracy on researchers while still ensuring safety. The agencys Agricultural Biotechnology Research Advisory Committee (ABRAC) played a substantial role in developing these guidelines. It is important to note that, in any event. these guidelines are just for USDA-funded research; APHIS still supplies the principal regulatory coverage. Criticism of the current situation regarding research includes alleged confusion over agency jurisdictions. For example, when Biotechnica International field tested genetically engineered nitrogen-fixing bacteria. it did so under a 1989 consent order from EPA. However, when a researcher at Louisiana State University sought to do followup studies at the site, State officials, various Federal officials, and ABRAC became involved as EPA oversight and jurisdiction became less evident. EPA clarified its position with State officials, and USDA agreed that EPA would maintain jurisdiction until it chose to relinquish that jurisdiction. While the main question appears to have been over the research value of continuing to monitor the site. rather than any safety question. it demonstrates some degree of uncertainty over jurisdiction (26). Potential Impacts of Regulation Negative Impacts Questions have been raised regarding the short-and long-term impacts of the regulatory climate on research. It is frequently postulated that academic researchers do not possess the organizational whet-withal to proceed through a regulatory maze, and may find the bureaucratic and financial weight of regulatory approval procedures so burdensome that they will choose not to carry experiments through the field trial stages (74). This perception could block research at a key step, since the field trial is the stage at which the rubber meets the road, at which the predictions of the lab are tested in the real world. The impacts on research of the rulemaking process in Federal regulation of biotechnology were explored in a national survey conducted in 1989 (52, 83). Of 355 responses to the question. Have you ever been discouraged from conducting field tests with genetically modified organisrns?, 16 percent said yes. Among private-sector responders, 23 percent felt they had been constrained. Some 12 percent of responders replied that they had chosen not to proceed with a field trial even though they had a genetically modified organism ready. Legalities, uncertainties about regulation, time needed, and paperwork required were cited as reasons for the decision not to proceed (52. 83). Criticism has been leveled as to the methodologies employed in the survey. Whether or not the percentages point to a dramatic regulatory burden on research seems open to interpret at ion. Some feel that the survey captured a real reluctance among some researchers to go through the field trial. In any case, it is not clear that regulation rather than tough resource allocation decisions drives the decision to delay (or forego) field trials (83). A 1990 survey based on personal interviews of 35 researchers and regulatory affairs specialists revealed overwhelming agreement that the coordinated framework is working and that APHIS is helpful and timely in its response to permit requests, while EPA seems to be improving. Most responders, however, asserted that biosafety and biological monitoring protocols were overly cautious, with potential implications for allocation of personnel time ( 16). A third study surveyed 430 recombinant DNA scientists regarding their perceptions of the influence of activist pressures on recombinant DNA research. Some 63 percent view current safety mechanisms as adequate and 26 percent view them as overstringent; many perceive public controversy and litigation as having led to unwarranted obstacles in the regulatory arena ( 81). A premise of USDAs Proposed ABRAC Guidelines is that the local Institutional BioSafety Committees (IBCs) can provide helpful advice to academics, streamlining the regulatory procedure. According to the level of safety

PAGE 215

concern, IBC oversight ranges from simple notice to IBC review and either approval or disapproval by the IBC and the USDA. Since IBCs previously have dealt principally with laboratory-contained experiments, they may require training to play a helpful role at the field trial stage; more agriculturally and ecologically trained members will need to be added. The University of California system-wide biotechnology program has sponsored an educational meeting for institutional biosafety officers who can work with the IBCs on matters of deliberate release (43 ). Possible Positive Impact on Research Although regulations of genetically engineered organisms may possibly inhibit one line of research (field trials). it may stimulate anotherecological research. As risk assessment methodologies are being devised for evaluating releases of recombinant DNA modified organisms into the environment, ecologists and population biologists are turning their attention toward related questions. The Ecological Society of America report on deliberate release describes a pressing need for interdisciplinary research (99). The concept of deliberate release has provided a compelling focus for questions of ecological community dynamics, migration of genes into populations, evolutionary change, and other fundamental problems. Furthermore, many researchers are stimulated by the opportunity to channel their research toward a useful analysis. Such lines of work do not fall neatly into most categories of research funding; thus funding sources may need to adjust their emphases since this work has an important role to play in the evolution of agricultural biotechnology. The 1990 Farm Bill addressed this need by setting aside funds for risk assessment research, equaling 1 percent of whatever the department spends in biotechnology research. Questions pertinent to risk assessment research, as well as the relationship between ecological research and risk assessment are described at greater length in chapter 8. As guidelines are finalized and disseminated, and riskassessment research proceeds, regulatory uncertainty should be reduced for researchers. With reduced ambiguity, as well as steady increases in information and experience, researchers may well venture more boldly in greater numbers into the field trial stage. Institutional BioSafety Committees may become better versed at giving advice and assistance to researchers, as may other university offices and field trial supervisory staff. Thus, the potential negative impacts on research could prove to be short-lived. In the future, technology transfer of genetic engineering advances may be mediated through industry-sponsored, university-based field trials. Although many companies would prefer to keep work binhouse. others may place greater value on the objectivity y of university research and the capacity of university facilities. While possible contlicts of interest would have to be resolved, both parties could thus continue to contribute to field trial research. (See box 7-E. ) The positive stimulus of the regulatory climate to ecological research may be at or nearing its peak at this time: in the shortand mid-term, assessment methodology will be developed and refined. Data gathered will be synthesized. Eventually, in the long run, assessments of the results of releases may well become yet one more subfield of ecological research, one more way to approach interesting problems that exist in a real-world context. Impacts of Regulation on Agribusiness Only half of the agricultural biotechnology companies surveyed by Burrill and Lee ( 7 ) consider Federal agency jurisdiction over the testing and selling or distribution of biotechnology products clear-cut. Nonetheless. only a minority believed that they had experienced Federal regulatory delays. Some 16 percent found delays in relation to product testing; some 16 percent found delays in relation to selling and distribution (7). For the most part, at least the large agricultural companies find that the APHIS system is predictable and works well, without inhibiting industrial activity (38, 40). Moreover, even those concerned with the competitiveness of industry also acknowledge the role of regulations in shielding industry from unfortunate occurrences that could. by thus capturing public attention. slow commercial product development (79). At least one small start-up agricultural biotechnology company, Calgene. has fared well under the current regulatory structure; between November of 1987 and October of 1990, Calgene received approvals from USDA for some dozen field trials for three genetically engineered crops in five States; the average approval time of 113 days is viewed as extremely reasonable. Representatives of Biotechnica, Pioneer, and Northrup King have also testified as to the effective workings of the APHIS system for genetically engineered plants (89). It has long been alleged that the strategic business plans of some smaller companies may have been, and may continue to be, influenced by the regulatory climate, as well as by public concern over biotechnology. The company Mycogen, for instance, deliberately used killed rather than living recombinant bacteria as pesticides; Ecogen

PAGE 216

Box 7-EEPA Research and USDA Research EPA has established a research program focused on the use of microorganisms in biotechnology and intended to meet the technical needs of the regulatory program. The six areas of research are as follows: 1. development of methods for detecting, enumerating, and analyzing microorganisms in complex samples from a variety of real-world habitats; 2. development of data and predictive models related to transport or spread between the point at which release occurs and other locations; 3. determination of potential for survival, growth, or colonization of released microorganisms under various conditions and environments; 4. assessment of factors affecting stability of genetic material and likelihood of gene exchange; 5. detection of any negative environmental response; and 6. criteria and methodologies for controlling risk. Inhouse EPA scientific staff are developing a complementary extramural research program. Regular independent peer review is intended to keep the orientation of the research toward the risk assessment needs of the regulatory staff while still encouraging scientific quality and contributing basic information on microorganisms in the environment (67). The Research Office is thought to have worked very closely with the FIFRA staff, directing research towards assistance in developing evaluation procedures. The biotechnology assessment budget, however, was cut in 1991. The 1990 Farm Bill (S. 2830) contained provisions governing USDA research. In addition to promoting Federal funding for high-priority research in areas including biotechnology, the bill created a Biotechnology Risk Assessment Research Program. A competitive research grant program is authorized for environmental assessment research to the extent necessary to help address general concerns about the environmental effects of biotechnology; research is authorized that will assist regulators as they develop policies on planned release. Eligible areas of research include: biological and physical containment methods, methods of monitoring dispersal of genetically engineered organisms, and gene transfer between genetically engineered organisms and related cultivated or wild species. The Secretary of Agriculture is required to consult with APHIS, ABRAC and OAB on specific areas of research (44). SOURCE: Office of Technology Assessment, 1992. has developed products with naturally occurring or nonrecombinant organisms (33). DNA Plant Technology (DNAP). which has to consider agricultural and food regulations, has deliberately adopted a bifocal business development approach, developing products through innovative uses of nonrecombinant technologies, such as tissue culture, as well as exploring the potential of recombinant plants. This reduces their vulnerability should regulations for the commercialization of biotechnology prove untenable to them. While DNAP currently has one regulatory staff member, it foresees the likelihood of adding more (20 ). With training in use of the Intelligent Form Generator, a software program designed by the National Biotechnology Impact Assessment Program to walk scientists through the production of an application, the NBIAP program director predicts th\aut a field trial application can be generated in less than 2 hours. Without this computer aid, he estimates. completing an application could take 1 to 2 months, with a staff, and up to 6 months without a staff (3). Resolution of regulatory processes and ambiguities will be critical as companies ready themselves to move to large-scale use of recombinant plants. One point raised by the private sector is the need for clarification of EPA role under FIFRA regarding transgenic plants with pest-resistant properties. Clarification of scope of review, preparation of a guidance document on data requirements, and harmonization with APHIS are regarded as necessary to reduce regulatory uncertainty for industry (37, 109). The vast magnitude of trials necessary for the development of any new crop variety makes it particularly important to clarify regulatory roles and requirements with respect to recombinant technology. The seed company ICI Garst, for example. has compiled figures on the development of corn varieties (82). In 1990, some 350,000 plots were used for nonrecombinant plants. The following numbers demonstrate the sheer number of lincs involved in generating new varieties in 1990 and expected in 1994 (table 7--l).

PAGE 217

Table 7-4Genetic Lines Needed for New Corn Varieties Number of Lines Stage of developement 1990 1994 New inbreds . . . . Preliminary hybrids . . . Advanced hybrids . . . Experimental hybrids . . . E hybrids . . . . . R hybrids . . . . . N hybrids . . . . . New commercials . . . 79,000 34,000 5,600 1,600 125 30 10 9 92,000 39,000 8,000 2,400 150 30 12 9 NOTE: E-Hybrids are hybrids exchanged among breeders with the company; R-Hybrids are regional uniform strip tests; N-Hybrids are national uniform strip tests. SOURCE: ICI Garst Seed Co., 1991, Obviously, were genetically engineered plants involved in such trials, and if these had to pass a complex set of regulatory requirements, agricultural companies would be forced to weigh their options very carefully. The costs of meeting regulatory requirements might prohibit them from bringing promising recombinant plants to full commercialization as new varieties. On top of the sheer numbers involved, another key point is that multitrait selection is the normal approach to plant breeding and development of improved varieties; the approach is to improve a number of traits concurrently; multiple recombinants might be combined in different trials. Furthermore, seed from the later stages of testing is sold. Agricultural practices do not separate variety from variety; all seed corn is stored in grain elevators in bulk. Clearly this is not a set-up readily amenable to special treatment for biotechnology. The restrictions governing small-scale field trials would be logistically infeasible. Developing even a conventional hybrid can cost approximately one million dollars. Although biotechnology can improve efficiency in the early research stage, by making new genes available quickly and precisely, industry emphasizes that the rigorsand the orders of magnitudeof the hybrid testing scheme will not change. Thus, the regulatory climate will have a significant impact on whether or not biotechnology is widely used as a tool in the seed industry. Assessments of the impact of regulations on industry will need to take this into account. A responsible but reasonable and clear regulatory path towards commercialization will be crucial to the successful implementation of biotechnology in agriculture. Public Participation The U.S. public today questions the use of new technologies Based in part on general environmental awareness, skepticism about science, and negative experiences with the chemical industry and the nuclear power industry, this questioning attitude is now a potent force. Today, many analysts of biotechnology sound the clarion call of public participation; if the public is to accept biotechnology, people must have access to information, and be able to play a role in debating controversies, and achieve a sense of trust in policy makers (54, 90). Federal regulatory agencies sometimes do not receive the full trust of the public. State agencies tend to be somewhat better trusted. When Federal agencies share information and involve the public, they are likely to build confidence in their procedures. FDA attempted this by publishing scientific information relevant to its decision on bST in Science. The meetings for media and other segments of the public held by USDA represent another example of public confidence-building through involvement. A positive public perception of biotechnology is obviously critical to its growth; beyond this, participation by the public can contribute to the beneficial development of biotechnology; questions raised can indeed be pertinent. Although the public has channels through which it can participate in regulations. it may not be aware of them. For example, public input into the review process for field trials is officially ensured through notifications in the Federal Register. Environmental assessments and pending approvals are so published. Clearly, however, the general public does not as a rule pore through the Federal Register. Various environmentalist and public interest groups do, however, and can bring matters to a wider audience. In some cases, such groups challenge approvals. For example. ice field tests (102) of ice-minus bacteria used to protect crop leaves from frost in 1987 were significantly delayed due to such challenges. A very narrow nongovernment subset of the public is brought into the picture when scientists external to the agencies perform scientific reviews to augment staff review in problematic cases. Public input also can arise when States receive field trial applications from the Federal agencies. Depending on an individual state's review process, representatives of the public may well participate. The 1990 Special Workshop for State Agencies, State Oversight of Biotechnology, came to consensus on the importance of a public participation component for any State biotechnology task force (39). At the institutional level, public membership is mandated for Institutional Biosafety Committees (IBCs), which seem likely to be called on more and more frequently to

PAGE 218

examine plans for field trials at an early stage within institutions. One somewhat sensitive area in terms of public participation is that of confidential business information. As discussed earlier. Federal agencies have the legal right to protect confidential information deemed critical to a companys competitiveness. In fact, companies submitting applications for field trials can submit two forms of an application. one for in-house review. under confidentiality terms. and one with confidential information deleted. for open distribution. Only the few States with legal protection for confidential business information can be sent the complete form. Of course, the more blanks that appear in an application, the more likely that proposal will be regarded with public distrust or unease. To minimize public unease, Federal officials encourage companies to keep their designated CBI to a minimum. Complaints have been voiced when information unnecessarily designated as CBI has been unavailable to the public (35). Public input into the process leading toward field trials has changed since the early and mid 80s, when court injunctions and vandalism were commonplace. Relative acceptance of the role of field trials and their safety has grown. Indeed, evidence exists that, together with an increased experience base, positive public involvement in biotechnology regulation can expedite the field trial process. (See box 7-F.) The quieting of local public opposition to biotechnology field trials seems to be evidenced quite widely. The great majority of field trials approved at the Federal and State level have met with little if any opposition by the public (106). Opposition activity now seems to be directed primarily at special cases. A current example is that of crop plants genetically engineered to withstand particular herbicides, which can then be sprayed readily over the field, as they will cause a problem only for the noncrop plants. Environmentalist spokespeople specializing in biotechnology are far from happy about this as a goal for agricultural biotechnology. In brief, despite industry protestations that this approach allows the strategic use of particularly benign herbicides, environmemtalists see this as a mechanism to excuse, if not encourage, application of environmentally hazardous chermicals. (See Goldburg et al., 1990 (36) for a thorough discussion of antiherbicide tolerance views; also Goldburg, 1989 (35). ) Early in 1991, the National Wildlife Federation (NWF) petitioned the USDA regarding Calgene Inc.'s application to field test genetically engineered cotton in 12 States. Calgenes October 1990 application to USDA proposed a 25-site test of cotton engineered to break down the herbicide bromoxynil. Whereas Calgene maintains that use of this cotton would significantly decrease herbicide use. NWF has petitioned the USDA to halt this broadscale testing until a thorough risk assessment has been conducted as to the impact on aquatic ecology and human health (22). In this case, the value question relating to herbicide tolerance begins to be tied to questions of progressively larger scale release. moving toward commercial release. The responses of public interest groups to large-scale releases may well intensify: it remains to be seen whether other components of the public will take a similar view. such that the current atmosphere of acceptance turns to opposition as commercilazation is approached. Significant factors will include: technical experience base derived f-rem small-scale tests to date, activism on the part of environmentalist groups, media attention. public confidence in the regulatory agencies. and public perception of--and education aboutbiotechnology and risk-benefit assessments. If decisionmaking is to be informed. education of the public about biotechnology risks and benefits must take place. Many advise that the evolution of biotechnology regulations benefit from the hard lessons of other industries. such as the nuclear industry, and emphasize education of and participation by the public. Thomas Jefferson has been quoted appropriately in this regard: If we think the people not enlightence enough to exercise their control with a wholesome discretion, the remedy is not to take it from them. but to inform their discretion (46). Public perception of biotechnology has been analyzed by OTA (101), and others (50). Apprehension over the novelty and power of biotechnology is mixed with a desire for the products of biotechnology. Two biotechnology trade associations (the Industrial Biotechnology Association (IBA) and the Association of Biotechnology Companies (ABC) have prepared materials and established committees related to public education. USDA ran several meetings as early as 1987 to work with the media and others toward public eductaion. Many of the Nations State and university biotechnology centers view education about biotechnology as one of their principal roles. Increasingly, high school teachers are taking courses in, and teaching. biotechnology: the media also is becoming more 297-937 0 92 8 QL 3

PAGE 219

Box 7-FTwo Experiences With Public Response In the early years of field trials, 1983-87, two sets of experiments involving ice-nucleating bacteria in California drew local public opposition as well as public interest group opposition. Suits were filed in the case of Tulelake, California, and an injunction was enforced against the University of California researchers until an environmental amendment was made; in the case of Monterey County, the County Supervisors, making use of their zoning authority, banned such experiments for 1 year, forcing Advanced Genetic Sciences (AGS) to go to the Contra Costa Countys Board of Supervisors for approval. Although a legal challenge was not upheld, many of the plants were uprooted as vandalism. (AGS had aroused particularly negative response beginning in 1985, when it had tested the bacteria in trees on its headquarters rooftop, without authorization.) Through the various vicissitudes, the University of California test was delayed from 1983 to 1987; the Advanced Genetic Sciences test was delayed from late 1985 to spring of 1987 (102). in 1988, Biotechnica International received Federal and State approval for a small-scale field test in Wisconsin of Rhizobium genetically engineered to increase alfalfa yield for which the PMNs had been filed the year before. In 1987, Biotechnica had conducted an extensive community relationship program in the county and the state where the field trial was to take place. This program involved: presubmission briefings to opinion leaders; press releases and brochures in laymans language, including a risk-benefit, Question and Answer style brochure; public meetings in the county sponsored by the company as well as attendance by company representatives at State government and legislature committee meetings; and media relations. For the first 6 months, interest was high in the community and a small group of activists opposed the trial. After the last public meeting in the summer of 1987, no further opposition emerged and, despite intense media interest, no demonstrations or protests occurred at the time of the test itself in April of 1988. For subsequent tests, the company has followed a scaled-down program of community relations, with substantially less community interest. The local comfort level with this biotechnology venture seems to have increased significantly (31). SOURCE: Office of Technology Assessment, 1992. sophisticated and therefore more able to convey acpartmental regulation of biotechnology [delegation of accurately technical and issues in biotechnology. Problematic Issues USDA Conflict of Interest? The criticism has been leveled that USDA faces an internal conflict of interest because it has a dual responsibility to promote research and to regulate in areas of biotechnology (49). USDA officials make the argument that the Department of Health and Human Services is in the same situation, but has the luxury of having its division of labor more readily perceived by the public as distinct. Within the same Department of HHS. the National Institutes of Health have responsibility for research and the Food and Drug Administration has responsibility for regulation. A comparable, but less visible or publicly understood, division exists within USDA. The Assistant Secretary for Science and Education is responsible for biotechnology research activities (including those of the Agricultural Research Service and the Cooperative State Research Service), whereas the Assistant Secretary for Marketing and Inspection Services is responsible for dethority by the Secretary of Agriculture, published July 19, 1985 (Fed. Reg. 29367 (1985).] APHIS and the Food Safety and Inspection Service (FSIS) are the USDA regulatory agencies involved (55). Coordination between the research and regulatory arms of USDA is the responsibility of the Committee on Biotechnology in Agriculture (CBA). The Office of Agricultural Biotechnology (OAB) is set up to develop policies and procedures for research in agricultural biotechnology, coordinate environmental safety review of proposed USDA-supported research with genetically y engineered organisms, provide staff support for the CBA, and provide staff support for the Agricultural Biotechnology Research Advisory Committee (ABRAC). ABRAC in turn is to review research guidelines and proposals and provide scientific advice to research and regulatory agencies in biotechnology (56). The existence of these committees demonstrates that the research and regulatory arms of USDA do interact. In fact, the agency would be criticized if there were no attempts at coordination, although the degree of coordination actually achieved has been questioned. The co-

PAGE 220

existence within one agency of NIH and FDA seem to set a relevant precedent. Conflict of interest may be avoideed within USDA by:outside Crotiques, such as advice from ABRAC and othcr external sources of review. as well as by the perception of. and loyalty to, distinct yet complementray missions on the part of APHIS and Science and Education. Burden of Proof of Safety U.S. society today desires a zero-risk society. Arising naturally from this attitude is a desire for regulatory agencies, or science, to prove safety. The agencies are attempting to build databases through small-scale field trials and, by analyzing and extrapolating from such information. to significantly reduce the probability of any risk occurring from larger scale releases. However, absolute proof of safety will never be achieved in biotechnology field releases, just as it will never be achieved in any other dimension of society. 1. 7 -. 3 4. 5. 6. 7. 8. 9. 10. CHAPTER 7 REFERENCES After Four Years Awaiting USDA Okay, Transgcnl~ Fish Gc)nna Swim, Bi{)tt~(lltl{~lt),g> Ne\\.~}twf(h, Jan. 7, 1991. p. 1. Andrews. Edmund L. Proposuls on Genetic Technology. The Ne\\v York Times, Feb. 19.1991, pp. DI D7. Artificial Expert Trims Permit Preparation From Months to Minutes, Bi~)tt~~lt\l~Jl{)(q~ Nc\\*s}~wt~h, Aug. 8, 1990, p. ~. Biotechnica Applies for Field Testing of Genetically Engineered Corn. Getw[i( Engitwcring Nc\\Is, March 1991, p. 32. bi~~~~~~.}~tt{~l{)cq). Ntw.s\~wtch, Aug. 6. 1990, p. 6. Blinded by Science? Va)zguard Pt-e.ss, Oct. 8 Iz, ]9~9, b H i g h] a nd Par-k Weighs Bill Regulating Experiments in Genetic Engineering, Dec. 4. 1988. The SundaJY Star Ledger. Burrill. G. Steve and Kenneth, B. Lee, Biofcth 9/: A C}wnging Etl\it-{jtl~t~[~tlt, Ernst and Young. 1991, pp. 57-59. Case Histories of the Development of State Bi;~technology Oversight. Conference Document, State Oversight of Biotechnology Workshop. Sacramento, CA, UC Systemwide Biotechnology Research and Education Program, New Jersey Department of Environmental Protection. 1990. Company News, AgBit~tt~(lltl~)lt~cq). N[ws. July/ August, 1990, p. 14. Cordle, Maryln and Langston, Althaea, bRegulatory Requirements, Agricultural Bi{~ti~(lltl(?l{~g>9: Introduction to Field Tc.~titlg, H. Graham Purchase Il. 12. 13. 14. 15. 16. 17. 18. 19 20. 21. and David MacKenzie ( eds. ). CMfice of Agricultural Biotechnology. USDA, 1990, pp. 26. Cordle. Maryln K.. Pa}ne. John H.. and Y(mn.g. Alvin L.. Regulation and oversight of Biotechnologica] Applications for Agriculture tind Forestry, A.we,~,vin,g E[.~d(~,qicwl Ri.vk.~ (!fl]iot~clltlolotq~%. Lc\I R. Ginzburg ( Boston. MA: Buttcr\~orth-Heinemann. 199 I ). Deshayes. A lain. b The French Rc\icm Prtxxss for Dclibcrtitc Release of Genetically Modititxi Orgunisms. BiOlO,qi~[tl A4fotlitorin ~tl ,qinect-ed Plants ~~mi klitroiw.~. D. R. Mac Kcnzic and Sw.mnc C. Henry (eds. ) ( Intcmational SYrnlposium on the Biosatctj Rc\ults t~t Field Tests of Gcneticully Modified Plants ml Microorgtinisms. Kiuwah Island, SC, Nov. 27, 1 W()) ( Bethesda. MD: Agriculture Research 1nstitutc. 1~~1 ). pp. 1o1106. Drulej, Ray M. and Ordway. Gird L.. The Tot-it S1/bstt~t~(c.s Ccmtrol A(t. The Burcuu ot Nationul Affairs. Inc., Wtishington. DC. Duke, Louise, presentation. Canaditin Fcdmd Regulations for Biotechnology. Fourth international ABC Biotechnology Meeting. Toronto, Cunada. May 23, 1990. EC Communication to the Council and the European Parliament. Pronmting the Competitive Environment !or the Industrial Activities Based on Biotechnology within the Community. Directorate-General, Internal Market and Industrial Affairs, Apr. 15, 1991. Emcr,qitl,q Bit~ti~./ltlolo,gi~).~ ill A,qril-l!ltllro: ISSUCS and Politics (Program Report IX), Di\is ion of Agriculture Committee on B iotcchnology. National Association of State Universities and Land-Grant Colleges, November 1990. EPA Office of Toxic Substances, Points to Consider in the Preparation and Submission of TSCA Premanufacturc Notices (PMNs) for Microorganisms, I 990. EPA Office of Toxic Substances, Chemical Control Division, Program Development Branch. 1. Status Report: Biotechnology Premanufacture Notification (PMNs), Oct. 1. 1991. Espeseth. David A. and Shibley, George P., Chapter 18: bRegulatory Policies for Field Testing 13xperimcntal Recombinant-Derived Vetcri n a ry B i 01 og i c a 1 Prod ucts, A d~it~ n (es it~ Bru(ello.sis Rc.\earcl?, L. Gary Adams (cd. ) (Col Iegc Station. TX: Texas A&M University Press, 1990). pp. 277. Evans, David. DNAP, personul communication, 199 I Evans, Ron, Office of Toxic Substances, EPA, personal communication, 199 I.

PAGE 221

220 l A New Technological Eru jbr American Agriculture 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39, Ezzell, Carol, b Proposed Cotton Test Comes Under Fire, BioWor/d ToduJ Feb. 1, 1991, p. 2. Final Report: Transgenic Plant Workshop, Maryland Biotechnology Institute, USEPA Office of Pesticide Programs, College Park, MD, Jun. 18 20, 1990. Flamm, Eric, FDA, personal communication. Fox, Jeffrey, U.S. Field Test Underway, Another Planned, Bio/Technolog~ (8):897, 1990. Fox, Jeffrey L., Guidelines Drafted, Still Not Released, Biotechnology (9, April), 1990, p. 326. Fox, Jeffrey, More Changes in U.S. Regulatory Bodies BiolTe(hnolog? (8) November 7:996, 1990. Fox, Jeffrey, Quayle Likes Biotech, Not Regulation, Bi~~/T(~th/z~~l~~g~ (9), Apr. 7.1991, pp. 3225. Genetically Engineered Potatoes, AgBiotechnolo~y Ne}tu (September/October, 1990), p. 7. Giamporcaro, David E., EPA, Personal communication, Oct. 18, 1991. Glass, David, Regulating Biotech: A Case Study, Forum .f{w Applied Research and Public Policy. VOI. 4(3), 1989, p. 92. Glass, David, personal communication, 1991. Glass, David, Impact of Government Regulation on Commercial Biotechnology, The Business of Biotechnology>: From the Bench to the Street, R.D. Ono (cd. ) (Stoneham, MA: Butterworth-Heinemann, 1991) pp. 169 198. Glass, David, Analysis of Proposed EQB Biotechnology Regulations, D. Glass Associates Report. 1991. Goldburg, Rebecca J., USDAs Oversight of Biotechnology in Rejorm and Imnmwtion qf Science und Education: Plunning for the 1990 Farm Bill; Printed for the Use of the Senate Committee on Agriculture, Nutrition and Forestry, U.S. Government Printing Office, 1989, pp. 137 147. Goldburg, R. et al., Bi{~tecl~\~~~l{)g?.~ Bitter Harvest.. Herbicide Tolerant Crops and the Threat to Sustainable Agriculture, Biotechnology Working Group, 1990. Goldhamrner, Alan, Regulatory Issues Related to Plant Biotechnology Products. paper prepared for the Industrial Biotechnology Association, 1990. Goldhammer, Alan, personal communication, 1991. Guidance for State Governments on Oversight of Biotechnology, Consensus Report, State Oversight of Biotechnology Workshop, Sacramento, California: University of California Systemwide Biotechnology Research and Education Program, Jul. 12 13, 1990. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54. Herrett, Richard, ICI, personal communication, 1991. Highland Park Weighs Bill Regulating Experiments in Genetic Engineering The Sunda~* Star Ledger, Dec. 4, 1988. Hodgeson, John, Appropriate Biotechnology for Africa Biotechnology (8), September 1990, p. 793. Huttner, Suzanne, University of California, personal communication, 1991. Industrial Biotechnology Association, Legislation Enacted in the 101st Congress: A Comprehensive Summary of Legislation Affecting the Biotechnology Industry, 1990. l Japan Roundup, BiolTechnolog> (8): September 1990, p. 802. Jefferson, Thomas, letter from, to W.C. Jarvis, Sep. 28, 1820, quoted in David L. Bazelton, Governing Technology: Values, Choices and Scientific Progress, Biotechnolog>j in Sociehv: Pri~ate Initiati}es and Public O\~er.~ight, Joseph Perpich (cd. ) (New York, NY: Pergamon Press, 1986), pp. 75. Knight, Pamela, U.S. and Europe Finding Common Ground, Bio/Technolog~ (9):January 1991, p. 17. Korwek, Edward, The 1989 Biotechnology Regulations Handbook, Center for Energy and Environmental Management, 1989, p. 137. Krimsky, Sheldon et al., Controlling Risks in Biotech, Technology>* Review, July 1989, p. 65. Lacy, William, Busch, Laurence, and Lacy, Laura, Public Perception of Agricultural Biotechnology, Agricultural Biotechnology?: !ssues and Choices, Bill R. Baumgardt and Marshall A. Martin (eds. ) (West Lafayette, IN: Purdue Research Foundation, 1991 ), pp. 139-162. MacKenzie, David R. and Larson, Jean, The National Biological Impact Assessment Program, Toxicological and En\ironmental Chemistn, vol. 23: 115-120, 1990. MacKenzie, D.R. and Vidaver, Anne K., U.S. BioSafety Regulations: Too Much or Not Enough? Agricultural Biotechnolog?l: issues and Choices, Bill R. Baumgardt and Marshall A. Martin (eds. ) (West Lafayette, IN: Purdue Research Foundation, 1991 ), pp. 86-88. Marchant, Gary, Modified Rules for Modified Bugs: Balancing Safety and Efficiency in the Regulation of Deliberate Release of Genetically Engineered Microorganisms, Harvard Journal of Law and Technolog~ (1, Spring): 163, 1988. Marois, James J., Grieshop, James 1., and Butler, L.J. Environmental Risks and Benefits of Agricultural Biotechnology, Agricultural Biotechnolog~~: Issues and Choices, Bill R. Baumgardt and

PAGE 222

55. 56. 57. 58. 59. 60. 61. 6~ 63. 64. 65. 66. 67. Marshall A. Martin (eds.) (West Lafayette, IN: Purdue Research Foundation, 1991), pp. 67-80. McCammon, Sally and Medley, Terry. Certification for the Planned Introduction of Transgenic Plants in the Environment. The MoltIcular and Cellular Biologi[s of the Pot~~to. Michael Vayda and William Park (eds. ) Wallingford, U.K. (CAB. International), 1990, pp. 239-244. McCammon, Sally and Medley. Terry, Certification for the Planned Field Introduction into the Environment of Transgenic Plants and Microorganisms, Biological Monitoring of Gencti(wll?% E\~,qi\lt~<)rt)[lPl[~llt.s andkli~robe.s, D.R. MacKenzie and Suzanne C. Henry (eds. ) ( lnternatiomd Symposium on the Biosafety Results of Field Tests of Genetically Modified Plants and Microorganisms, Kiawah Island, SC, Nov. 27, 1990) (Bethesda, MD: Agriculture Research Institute, 1991 ). pp. 1071 I 7. McCammon, Sally, internal memo, 1991. McCammon. Sally. personal communication, 1991. Meagher, L. R., Guidance Document for State Governments on Oversight of Biotechnology. Geneti~ En,qitleerin,g Ne\is, October 1990, p. 3. Medley, Terry L., Agricultural Products of Biotechnology: U.S. Department of Agriculture Certification. Presentation, Fourth International ABC Biotechnology Meeting, Toronto. Canada, 1990. Medley, Terry L., APHIS Seeks to Aid Tech Transfer While Ensuring Environmental Safety, Genetic E~l(qim~critz,q NeJ\s, September 1990. p. 14. Medley, Terry L., Ask the Feds. Genetic Engineering Nei~s, February 1991, p. 3. Medley. Terry, Form letter, Dear Potential Permit Applicant. USDA-APHIS-BBEP. Messerschmid(, M. Olav, Biopesticides-the Next Steps Biological Llonitoring of Gcncticall> Engineered Platlts and A4icrobes. D. R, MacKenzie and Suzanne C. Henry (eds. ) (International Symposium on the Biosafety Results of Field Tests of Genetically Modified Plants and Microorganisms, Kiawah Island, SC. Nov. 27, 1990) (Bethesda, MD: Agriculture Research Institute, 1991 ), pp. 257~6~. Milewski, Elizabeth and Anderson. Elizabeth. EPA Stresses the importance of Flexibility in Biotechnology Regulations. Gencti( En,qitteeritl ~![ Pl[l n t M itrol)c ltltt~rcl(tiotl.\. James Nakas and Charles Hagedorn 68. 69. 70. 71. 7~. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. (eds. ) (New York, NY: McGraw-Hill Publishing Co.. 1990), pp. 329-330. Milewski, Elizabeth. EPA, personal communication, 1991. Miller, H.]. et al.. Risk-Based Oversight of Experiments in the Environment, Science 250: 1990, pp. 490-491. Minutes of ABRAC Meeting, Feb. 20-21. 1991. Monsanto/NK Test Engineered Cotton, AgBiotcchnolog> Ne\ts (January lFebruary. 1990), p. 18. National Research Council, Intr{duttion of Re(wnlbinant DNA-E~tgi\lt)[)~-[)tl Org~~nisnls i~lto the En~ironment: Ke>s Issues (Washington, DC: National Academy Press, 1987). National Research Council, Field Testing of Ge twticull> A40d(fied Organisms, A Frame }\fork .fbr Dt~L.i,\ic~~l-M~lkil?g (Washington, DC: National Academy Press, 1989). National Association of State Universities and LandGrant Colleges. Division of Agriculture Committee on Biotechnology, Emerging Biot[(/~tlologi[.s in Agriculture: Issues llnd Politics. 1990. New Prospects for Gene-Altered Fish Raise Hope and Alarm, Ne}t York Times, Nov. 27, 1990. News Briefs, AgBiotL~l./ljll)lo,g> Ne\\Ys. Novem ber/December 1990, p. 18. Office of Science and Technology Policy, (51 FR 23302), Coordinated Framework for Regulation of Biotechnology: Announcement of Policy and Notice for Public Comment, Jun. 26, 1986. Permits Issued for Release into the Environment Under 7CFR 340, Printed by APHIS. Apr. 13, 1991. Presidents Council on Competitiveness. The, Report on National Biotcchnolog~ Polic>, February 1991. Proposed Notification Regulations for Biotechnology Products under the Canadian Environmental Protection Act, Environment Canada, Ottawa, Ontario, Sep. 21, 1990. Rabino, Isaac, The Impact of Activist PressuresRecombinant DNA Research, S[iencc, Tr(hnolo,q?, and Hunlcltz Va/ucs 16( I ): 199 I pp. 70. Randall, Jo, ICI-Garst Seeds, personal communication, 1991. Ratner, Mark, Survey and Opinions: Barriers to Field Testing Genetically Modified Organisms. Bic~/T~~(/?llol~~,q? (8) March 1990, pp. 196-198. Rc(mnbim~nt DNA Te(hnical Bl{l(ctin, vol. 13(3). pp. 181 184. Regulatory Mechanisms for EPA (Pesticides). Presentation, Workshop in Dealing with Field Test Regulations tind Public Acceptance of Engineered Plants and Microbes. University of Maryland, Jul. I I 1989.

PAGE 223

222 l A Ne}~* Technological Era jiw American Agriculture 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97. Repligen/Sandoz to Start BT Field Trials, Biotechnology> Ne\isi~wtch, Oct. 1, 1990, p. 2. Rodger, William H. Jr., Et~\~ir(~rlt?l~~\lt[ll LaMI: Pesticides and Tari( Substances, vol. 3, Section 5.1 to 7.5. (St. Paul, MN: West Publishing, 1988). Rutgers Scientists Plant Seeds of Trust with Local Community, The Scientist, Feb. 18, 1991, p. 6. Salquist, Roger H.. Hearing before the Subcommittee on Department Operations, Research, and Foreign Agriculture of the Committee on Agriculture, House of Representatives, One Hundred First Congress, Second Session on H.R. 5312, Serial No. 101-75, pp. 76, Also testimony of David Glass (p. 179), Warren Springer (pp. 189-192) and Rod Townsend, pp. 197-198.), Oct. 2, 1990. Sandman, P. M., Apathy Versus Hysteria: Public Perception of Risk, Public perception of biotechnology]. L.R. Batra and W. Klassen (eds. ) (Bethesda, MD: Agricultural Research Institute, 1987). Schneider, William, USEPA, personal communication, 1991. Shapiro, Sidney, Biotechnology>* und the Design of Regulation. Report for the Administrative Conference of the United States, 1989, pp. 3 10. Smith, Roger, New Jersey Department of Environmental Protection, personal communication. Soejima, Junichi, Biosafety Review and Regulatory System of rDNA Organisms in Ministry of Agriculture, Forestry and Fisheries, Biologi[ui Monitoring of Genetically>* Engineered Plants and Microbes, D.R. MacKenzie and Suzanne C. Henry (eds. ) (International Symposium on the Biosafety Results of Field Tests of Genetically Modified Plants and Microorganisms, Kiawah Island. SC, Nov. 27 30, 1990) (Bethesda, MD: Agriculture Research Institute, 1991), pp. 127. Summary of Biotechnology Regulatory Activities USDA-APHIS, 1991. Summary of Reviews Conducted Under FIFRA ( 1984-Present), 1989. Summary of Reviews Conducted Under TSCA and FIFRA, updated Jan. 31, 1991, draft. 98. 99. 100. 101. 102. I 03. 104. 105. 106. 107. 108. 109. 110. Thomas, Paul, Cultivar Testing, A Seed Industry Perspective, HortSciente 22(6), 1987. Tiedje, James M. et al., The Release of Genetically Engineered Organisms: A Perspective from the Ecological Society of America. E(olo Nen*snwtch, Oct. 15, 1990, p. 9. U.S. Congress, Office of Technology Assessment, NcM D(\*c~l~~p\?~[~\~t,s in Bi{jt[~ti~\lc)l~~g>* (2): Plibli( fer~.eption.s of Biote(.hnolog? (Washington, DC: U.S. Government Printing Oftice, 1988). U.S. Congress, Office of Technology Assessment, Ne\tY Dt~\~t~l~~p/llt~tlt.~ in Bic~t[~[.l~~~~~l{~(q> (3): Field Testing Engineered (9rganisnts: Genetic and ELYjiogical Issues (Washington, DC: U.S. Government Printing Office, 1988). U.S. Congress, Office of Technology Assessment, Biotechnology in a Global En~ironntent, OTA-BA494 (Washington, DC: U.S. Government Printing Office. October 1991 ). USDA-APHIS BBEP Fact Sheet. Ward, Mike. EC States Rush to Harmonize National Laws with Directives on Biotechnology, Geneti( Engineering Ne}\*s, November/December 1990. pp. 6-7. Wheeler, David L., Field Tests of Genetically Altered Organisms Increase as Public Opposition to the Experiments Fades, The Chronicle of Higher Education, Jul. 26. 1989. Witt, Steven C., Bi{jtt~cl~ll(jl~jg>*, A4i(.robes. {lnd the Etz\i~~~lzl?/t~izt, Center for Science Information. 1990, p. 120. Wivel, Nelson, NIH-RAC. personal communication, 1991. Workshop on the Field Releases of Genetically Modified Organisms, ABC annual meeting, Washington, DC, May 1991. Wrubel, R., Krimsky, S., and Wetzler, R., b Field Testing Transgenic Plants,* Bio.$tience. vol. 42, No. 4, 1992, pp. 280-289.

PAGE 224

Chapter 8 Scientific Issues: Risk Assessment and Risk Management Photo credit: Grant Heilman, Inc.

PAGE 225

Contents Page INTRODUCTION . . . . . . . . . . . . . . . . ... ........225 RISK ASSESSMENT . . . . . . . . . . . . . . . . ... .....225 Concerns and Postulated Environmental Risks of Biotechnology . . . . . . . 225 Major Risk Assessment Reports . . . . . . . . . . . . . . . 227 Biotechnology Ecological Risk Assessment . . . . . . . . . . . . 231 Applicability of Diverse Bodies of Knowledge to Assessments of Large-Scale Commercial Release . . . . . . . . . . . . . . . . . 235 Commercial Release Issues . . . . . . . . . . . . . . . . 240 RISK MANAGEMENT . . . . . . . . . . . . . . . . ... .....247 Design of Science-Based Regulation . . . . . . . . . . . . . . 247 Generic v. Case-by-Case Approach . . . . . . . . . . . . . ...247 Relative Risks Compared to Traditional Practices . . . . . . . . . . 248 Cost-Benefit Analyses . . . . . . . . . . . . . . . . . 248 Small-Scale v. Large-Scale Issues . . . . . . . . . . . . . . 248 SCIENTIFIC METHODS OF MANAGING RISK . . . . . . . . .. ..........249 Promoters Turned On or Off by Specific Stimuli . . . . . . . . . . 249 Suicide Genes . . . . . . . . . . . . . . . . . . . 249 Prevention of Gene Transfer . . . . . . . . . . . . . . . . 250 Combinations of Genes . . . . . . . . . . . . . . . ... ......250 AGRONOMIC METHODS OF MANAGING RISK . . . . . . . . ... .......250 Physical Barriers . . . . . . . . . . . . . . . . . . . 250 Spatial Barriers . . . . . . . . . . . . . . . . . . . 251 Temporal Barriers . . . . . . . . . . . . . . . . . . 251 SUMMARY POINTS . . . . . . . . . . . . . . . . . ... ...251 CHAPTER PREFERENCES . . . . . . . . . . . . . . . . 252 Boxes Box Page 8-A. Ecological Risk Assessment Questions . . . . . . . . . . . . 233 8-B. Learning by Doing: Successive Field Releases . . . . . . . . . . 238 8-C. Monitoring Microorganisms . . . . . . . . . . . . . . . 245 8-D. Relevant Research Fields . . . . . . . . . . . . . . . . 246 Figures Figure Page 8-1. Alternative Risk Analysis Approaches . . . . . . . . . . . . . 229 8-2. Risk Assessment Framework for Environmental Introductions . . . . . . . 232 Table Table Page 8-1. Comparison of Traditional and Developing Biotechnology . . . . . . . . 248

PAGE 226

Chapter 8 Scientific Issues: Risk Assessment and Risk Management INTRODUCTION The large-scale commercial use of agricultural biotechnology gives rise to several questions. Does the release of large numbers of genetically engineered organisms into the environment pose special risks? If so, what is the order of magnitude of these risks compared to the risks of traditional agricultural practices? What benefits offset such risks? Generally, concerns about genetic engineering focus on: possible escape of a genetically engineered organism, such that it invades new ecological niches or outcompetes naturally occurring organisms and becomes a pest; possible disruption of a delicately balanced ecosystem; possible direct risks to humans or wildlife; possible problems of gene stability and of gene transfer to unintended recipient organisms; possible impact on evolution; and the sheer newness of the technique. This chapter addresses these concerns and describes the range of scientific views on biotechnology and risk. A consensus has developed that risk assessment is desirable and feasible. Risk assessment in general is founded on principles and methodologies that can apply to biotechnology. We know what questions to ask in assessing ecological risks of planned introductions. A knowledge base already exists pertinent to these questions and risk assessment studies on this topic are proliferating. Science-based risk management builds on this technical knowledge and on our capabilities for risk assessment. Risk assessment methodologies and our technical knowledge base make it possible to conduct effective risk assessments of specific introductions and to manage risks of acceptable introductions. Science-based regulations are central to effective management of risk. A variety of scientific and agricultural methods can be used to manage risk in particular situations. RISK ASSESSMENT Concerns and Postulated Environmental Risks of Biotechnology General Concerns Questions arise concerning the impact of introduced genetically engineered organisms: What is the likelihood that such organisms will persist in the environment? What is the likelihood that they will spread, constituting an invasion into the ecological community? Will they become pests, with a deleterious effect on other species? Will the expression of the gene itself lead to an unwanted effect on the ecosystem? Other questions have to do with the recombinant gene itself (38): What avenues exist for gene transfer within and between various species in nature? How probable are such exchanges and at what rates would they occur. if at all? If introduced genes are transferred to genomes existing in nature. how well and how stablywill the functions for which they code be expressed? Finally, broader, more fundamental questions can be posed: Are we in fact dealing with a phenomenon so novel that we have no way of predicting outcomes, of Performing adequate risk assessment? Do we have a moral right to manipulate still further the species and the ecology of our planet? Are we losing an intangible, aesthetic quality to our lives by so doing? Can we at-ford to say no to the benefits that this technology can confer on agriculture Concerns About Plants Specific concerns relating to genetically engineered plants include the possibility that transgenic plants will persist and become serious agricultural weeds; that the transgenic plants will invade natural habitats and disrupt local ecological interactions; and that the pollen of transgenic plants will act as a vector. bringing the introduced genes to other species that may then themselves become problem weeds. The likelihood of such possibilities occurring remains somewhat controversial, underscoring the importance of information from field trials and research. It is noteworthy, however, that transfer of genes from conventionally bred crop plants to noncrop plants has not created obvious problems in the past, and that traditional crop plants rarely have invaded natural ecosystems ( I 4). lnvasions of plants (by seeds. fruits. or vegetatively reproducing units ) involves dispersal, persistence. and establishment: all three stages must be successful if engineered plants are to become weeds. For transgenes (introduced genes) to move from crop plants and cause or contribute to a weed problem, hybridization with a reproductively compatible species must occur. For tiny

PAGE 227

given crop species, only a small number of the wild relative species that are reproductively compatible are actually likely to present serious weed problems; however, it is theoretically possible for a plant to become a weed in a novel environment (43). One specific concern posed frequently by some environmentalists, among others (32), is that genes for herbicide tolerance might be transferred from crop plants to weeds. If this were to occur, natural selection could favor the trait in weedy neighbors of crops treated with the herbicide. With any use of herbicides, furthermore, increased selection pressure is put on wild species for any herbicide tolerance traits they might already possess. Such developments might lead eventually to increased use of chemical herbicides. A fundamental debate has arisen between industry scientists who maintain that crops can be genetically engineered to be tolerant of particularly environmentally friendly herbicides and some environmentalists who say. essentially, that no new technology should be used to favor continued use of chemicals in the environment. Concerns About Microorganisms In part because they are invisible and relatively unknowable, microorganisms tend to elicit more concerns on the part of the public than do plants. Parameters of concern related to genetically engineered microorganisms include the possibility of gene transfer and recombination. the possibility of movement into new environments, and the possibility of infection of nontarget organisms. Questions asked include: Will genetically engineered microorganisms give rise to biological risks for humans or other species? Will they give rise to environmental problems? Do we have the technical understanding to evaluate and predict any such problems? Whether bacteria, fungi. viruses, or baculoviruscs, microorganisms suffer from a bad reputation at the broadest level of public perception: they are. after all often equated withgerms One specific concern raised with regard to genetically engineered organisms is the possibility of genetic material from such organisms being transferred to human gut bacteria. The risk of infection of humans. or other deleterious effects, is clearly going to be examined for planned introductions of microorganisms. For example. among the questions raised by Monterey County staff considering the Advanced Genetic Sciences (AGS) proposal to field test Frostban R was whether or not the Pseudomonas fluorescens could sensitize or aggravate existing health conditions among sensitive human populatitons living near the proposed test site (66). To assess risk of problematic infection of humans by genetically engineered organisms. information must be available on exposure level. This hinges on such factors as bioavailability or likelihood of absorption into cells or tissues, specificity, and level of interaction possible of the microorganisms or their chemical products with nontarget (human) tissues; and potential of the microorganisms for colonization or infectivity. The degree of pathogenicity must be considered as well. Some relevant factors include virulence. Possession of toxins, host range, and relative susceptibility. Generally. risk assessment will factor in predictability of the behavior of the recombinant DNA identified microorganisms based on their parent organisms, as well as knowledge of specific recombinant techniques used (40). Scientists concerns focus less on pathogenicity and more on the possible impacts of genctically engineered microorganisms on the environment. Suggested impacts include possible influences on: indigenous population size. diversity of species. the ecological community. natural cycles, and evolution of the introduced organisms (76). Microbial environments are complex. By one estimate some 10 9 microorganisms, representing a variety of taxonomic groups, inhabit one gram of soil. Uncetanties exist as to possible consequences of sudden introductions on balanced microbial ecosystems (46). Microbial diversity in the soil is high (88). This limits the niches available to introduced mirorganisms (86). While introduced microorganisms may thus compete poorly. they may persist in low-density populations. A key issue is whether or nor an unexpected later resurgent bloom or population expansion from a low-density population can be reasonably envisioned (84). Since microorganisms can and do change location. questions of dispersaland possible subsequent reproduction in nontargeted ecological sitesalso are raised. The ability of a particular strain to transfer genes to other specics will affect the likelihood of other microorganisms being affected in new. nontarget areas. All questions bearing on survival, multiplicaton, and dispersal of genetically enginered microoganisms: on possible exchange of genes between introduced and indigenous microorganisms: and ultimately on issues of environmental and public safety, are engaging attention of academic and industrial scientists, the public, and governmental regulators alike (22). Views Held in the Scientific Community Particularly in the early days. the issue of planned introductions of genetically engineered organisms sparked

PAGE 228

a range of views on safety even among scientists (50). In the mid-eighties, microbiologist Winston Brill argued that. for centuries, traditional breeding has altered animals and plants without negative consequence: and that microorganisms, including pathogenic species, have been added to the soil in hopes 0ft beneficial impacts. also without negative consequences (7). His conclusion that these observations alone formcd a basis for risk assessments of organsms that have had one or a few genes added drew fire from a group of ecologists ( lo). These critics pointed out that mutations that increase an organisms niche range can be ecologically significant. and that some ramifications of an organisms impact on the environment are not predictable from knowledge of its introduced genes alone. Casc-by-case quantitative risk assessment for deliberate release was recommended. In 1987, Science published side-by-side articles by Frances Shw-pies (75) and Bernard Davis ( 15). Sharples. an ecologist. reaffirmed the need for casc-by-case assessments, given the complexity of any organisms interactions with the environment. Molecular biologist Davis suggested that the experience of ecologists with introductions of higher organisms is less pertinent to risk assessment of engineered microorganisms than are the insights of fields mom concerned with the specific properties of those microorganisms: population genetics. bacterial physiology. epidemiology. and the study of pathogenesis. The range of possible views on safty runs from zero risk to catastrophic risk; those who presume smallrisk, pending research occupy the middle of the spectrum. In the mid-eighties, molecular biologists tended to stress the relavance of the safty record of laboratory biotechnology and graviated toward the zero-risk end of the spectrum. Ecologists, who tended to stress the complexities of the natural environment. were less sanguine about potential risks. but stopped short of the cat astrohic-risk position taken by certain envronmentalists. An important distinction exists between ecologists and evironmentaists. The former are: l The l scientists concerned with the fundamental properties, proccsscs, and components of ecological systems latter, by definition, are concerned with various sociopolitical aspects of environmental quality and management. They may or may not be experts in understanding ecological processes and the organization of ecological systems (63). Some environmentalists, keenly aware of problems posed by past technologies, argue that the proposed user of new technologies bears the burden of proving safety. Biotechnology proponents, in contrast, argue that any risks are to date hypothetical. so that the burden of proof should rest with the doomsayer ( 5 I ). In the late 1980s and early 1990s. discussion has increasingly centered around developing appropriate riskassessment parameters and frameworks and designing regulator-y treatment according to risk. The current operational approach is in agreement with analyses in key reports that will be described in the next section (50). Presumed small risk. or risk in exceptional cases. with research or risk assessment required. is becoming more of a common theme. Arguments are tending to become more refined. revolving about such issues as legitimacy of risk-assessment parameters; the degree to which lessor-is from past field trials can be generalized: correct assessment procedures for casc-by-evaluations; development of predictive science related to these issues; science-btised regulations; and scientific mainagement of risk. Today the imminence of large-scale release is bringing all these discussions into sharp focus. Major Risk Assessment Reports Introduction to Risk Assessment Why Risk Assessment Is Needed-Society today has been sensitized to technology the public, in all its many forms, looks at past technnologiesthose of the chemical or nuclear industries for example-and sees negative outcomes that were not thoroughly considered prior to implementation of the technologies. Along with skepticism is a strong strain of environmentalism. a growing uneasiness that far too often, for our convenience. we carelessly and permantly harm the environment. Furthermore, however unrealistic it may be, a desire for zero-risk seems to underliie many responses to technology and to life in general today. For these reasons as well as to achieve the fundamental objective of promoting safety it behooves regulatrs and other responsible parties to conduct reasonable risk assessments of new technologics. Biotechnology, in particular planned introductions of recombinant DNAmodified orgainisms. is among the technologies for which risk assessment is now done. This is necessary for regulators. important to the publics sense of cconfidence, and useful to users of biotechnology. including researchers in academdia, industry. and government.

PAGE 229

Principles of Risk Assessment Risk can be defined as the potential for negative or adverse consequent to arise from an activity or an event (23). Risk also can be defined as the probability of an event occurring multiplied by the cost of its occurrence (44). Risk assessment can be viewed as the process of obtaining quantitative or qualitative measures of risk levels. including estimates of possible health effects and other consequences as well as the degree of uncertainty in those estimates (23). Risk assessment simply is an analytical tool that pulls together a great deal of diverse data in order to estimate a potential risk from an event or a process (81). Often, historical data on possible adverse consequences are difficult or impossible to obtain, making risk assessment an inexact process that attempts to characterize and quantify uncertainty, but never completely eliminates it. Nonetheless, despite the limitations and challenges, use of risk assessment principles makes it possible to organize and interpret knowledge so as to improve the prediction of possible outcomes and ultimately to manage risk (23). Risk assessment has been defined as a five-stage process: 1. 2. 3. 4. 5 Risk identification defining the nature of the risk, source, mechanism of action, and possible adverse consequences; Risk-source chracterizationcharacterizing the source of potential risk; Exposureassessment assesing the intensity. frequency and duration of human or environmental exposures to risk agents; Dose-response assessment assessing the relationship between dose of the risk agent and health or environmental consequences; and Risk estimation intergrating a risk-source characterization with an assessment of exposure and dose-response, leading to overt measures of the level of the health, safety or environmental risk involved (59, 92). Clearly these stages can be adapted to fit a variety of kinds of risks, and the entire process can take several different forms. (See figure 8-1. ) The choice of an approach to risk assessment depends in large part on the extent and quality of available knowledge, degree of expected precision, and importance attached to outcomes at a low probability. Where the knowledge base is large and little uncertainty exists, a risk or hazard may be described quite readily and a more precise deterministic consequence analysis might even be performed. On the other hand, when less knowledge is available and the level of uncertainty is high. a qualitative risk screening may be all that is possible. perhaps leading to a more quantitative probabilistic risk assessment. A much-used framework to assess risk is that developed for the evaluation of health effects associated with chemicals in the environment. This was endorsed by a National Academy of Science report (67) and refined at the Environmental Protection Agency (EPA). This chemical risk-assessment framework sometimes has been adapted for evaluation of planned introductions of recombinant DNA-modified organisms into the environment (13. 16. 30). The National Research Council (NRC) and the Ecological Society of America (ESA) (69, 85) developed in 1989 risk assessment frameworks designed for recombinant DNA-modified organisms. But they were quite different from the chemical approach. The NRC procedure takes account of the degree of familiarityof a planned introduction; the ESA uses a risk attributes categorization; both lead towards the determintaion of an appropriate level of concern. While differing somewhat in perspective. the two approaches nonetheless resemble each other in basic conclusions and therefore together provide a solid framework for risk assessment of planned introductions. Clearly, choice of framework for risk assessment will influence the kinds of data required for evaluation and for permit applications (50). The two reports described below have had significant impact on the recent framing of discussions about planned introductions. Even proponents of chemical risk-assessment procedures point out that these procedures can be used to determine whether or not a particular organism should be evaluated intensively using an analogue of a chemical risk assessment (81). National Research Council Report BackgroundIn late 1989, the National Research Council published Field Testing Genetically Modified Organisms: Framework for Decisions. This was requested by the Biotechnology Science Coordinating Committee (BSCC) on behalf of its member regulatory agencies. The report covered: l plants and microorganisms, l field-test introductions (but not large-scale commercial applications and related issues), l environmental (but not human health) effects,

PAGE 230

Figure 8-lAlternative Risk Analysis Approaches Precision of ananalysis Probabilistic risk assessment Consequence analysis with confidence bounds Hazard Qualitative description risk screening I > Level of Uncertainty SOURCE: J. Fiksel and V.T. Covello, The Suitability and Applicability of Risk Assessment Methods for Environmental Applications of Biotechnology in Biotechnology Risk Assessment: Issues and Methods for Environmental Introductions (New York, NY: Pergamon Press, 1986), pp 1scientific issues principally (but not regulatory policy). field test conditions in the conterminous United States, and general procedures for determining categories (not specific case recommendations ). fundamental principle underlying the study. and first introduced in an earlier National Academy of Science document (68), is that safety assessments of a recombinant organism should be based on the nature of the organism and the environment into which it will be introduced, not on the method by which it was modified. A related point is that no conceptual distinction exists between genetic modification of plants and microorganisms by classical methods or by molecular methods that modify DNA and transfer genes. Topics analyzed for the 1989 report include: relevant biological characteristics of genetically modified plants; experience with genetic modification and introductions of plants modified traditionally and by molecular genetic techniques; potential weediness: the features of the genetic modification in microorganisms; phenotypic characteristics of the parent organism and of its genetically modified derivatives: and relevant features of the environment into which the organism will be introduced. Findings-The report recommends that the impacts of genetic modification on the phenotype of the organism and the mobility of the altered gene be assessed. In some cases. when persistence of the modified orgtanism is not wanted or when uncertainty exists as to effects on the immediate environment. risk assessment should emphasize the phenotypic properties relating to the persistence of the organism and its modification. Questions to be considered include: fitness of the genetically modified organism; its tolerance to physicochemical stresses; its competitiveness range of available substrates; and. if applicable. pathogenicity, virulencc. and host range. The report describes the long historry of safety in the useful employment of plants and microorganisms. and underscores the need for field tests to increase the capability to assess any risks of large-scale introductions. Specific scientific conclusionss of the report pertaining to plants include: 1. 2. 3. 4. 5. 6., The current means for making evaluations of introductions of traditionally bred plants are appropriate (on the basis of experience with field tests of hundreds of millions of genotypes over decades). Crops altered by molecular and cellular techniques should pose risks no different from those posed by crops modified by traditional genetic methods for similar traits. The potential for enhanced weediness is the principal risk to the environment seen from introductions of genetically moditied plants. although the likelihood of this occurring is low. Confinement by biological. chemical spatial, physical. environmental and temporal means is the principal means of maintaining the safety of field introductions of classically modified plants. Experimental plants grown in field confinement rarely if ever escape to cause problems in the environment. Established confinement options are equally applicable to field introductions of plants modified with

PAGE 231

230 A New Technological Era for American Agriculture molecular or cellular methods and to plants modified with classical genetic methods. Conclusions concerning microorganisms included: 1. Many molecular techniques make possible genetic changes in microbial strains that can be fully characterized. 2. The molecular techniques are powerful in their capability to isolate genes and transfer them across biological barriers. 3. Field experience has given rise to a great deal of information about some microorganisms; nonetheless, less information exists on microbial ecology and less experience with planned introductions of genetically modified microorganisms than there is for plants. No adverse effects have been noted from microbial introductions to date; a field test should g O forward when sufficient information is available for its safety evaluation. 4. The probability of adverse effects can be minimized or eliminated by appropriate means of confining the microorganism to the environment into which it was introduced; one example would be the use of suicide genes. The framework for evaluating risk developed in the report is structured around the following questions: 1. 2. 3. Are we familiar with the properties of the organism and the environment into which it may be introduced? Can we confine or control the organism effectively? What are the probable effects on the environment should the introduced organism or a genetic trait persist longer than intended or spread to nontarget environments? (69) The familiarity criterion is key to this report and has reappeared consistently in risk assessment discussions since. Familiarity means having sufficient information on which to base a reasonable assessment of safety or risk. Thus, as our information base increases, so does the scope of familiarity. When the familiarity criterion is not met. the possibility of confining or controlling the organism and the potential consequences of failing to control it must be evaluated. The report is intended to provide a basis for a flexible, scientifically based. decisionrnaking process. The classification of an introduced organism into a particular risk category is made possible by the framework for evaluating field tests (69). The 1989 NRC report is often cited and has provided a conceptual framework for many approaches to risk assessment of planned introductions of genetically engineered organisms into the environment. Its level of detail made it more palatable to technical audiences than the 1987 pamphlet, which was at times criticized for making assertions without documentation ( 11, 50). The Ecological Society of America Report Another seminal assessment was published in 1989, The Planned Introduction of Genetically Engineered Organisms: Ecological Consideratons and Recommendations (85). This report was prepared for the Public Affairs Committee of the Ecological Society of America (ESA) and also has been broadly disseminated and cited. Dr. James Tiedje chaired a workshop committee in April 1988, examining ecological aspects of planned environmental introductions of genetically engineered organisms. The Workshop Committees initial draft was reviewed at great length by the ESA Public Affairs Committee, the ESA Executive Committee, and other ecologists. The report supports the use of advanced biotechnology for the development of environmentally sound products. and states that the phenotype of a transgenic organism, not the process used to produce it, is the appropriate focus of regulatory oversight. Ecological risk assessment of proposed introductions must consider the characteristics of the engineered trait, the parent organism, and the environment that will receive the introduced organism (85). Like the NRC report, the ESA report emphasizes product, rather than process, as the appropriate focus of evaluation and regulation. Thus, genetically engineered organisms should be evaluated and regulated according to their biological properties (phenotypes), rather than according to the genetic techniques used to produce them (85). Yet the report acknowledges the potential for novelty and consequent likelihood of evaluation inherent in the new techniques. The report acknowledges, however, that because many novel combinations of properties can be achieved only by molecular and cellular techniques, products of these techniques may often be subjected to greater scrutiny than the products of traditional techniques. Moreover, it recognizes that even precise genetic characterization of transgenic organisms does not necessarily allow scientists to predict all ecologically important expressions of phenotype in the environment. The ESA report emphasizes the importance of considering a variety of ecological factors in ecological ramifications of planned introductions. Among these are

PAGE 232

survival, reproduction, interactions with other organisms, and effects on ecosystem function and dynamics. Potential undesirable impacts must he weighed in evaluations. While explicitly calling attention to the complexities of ecological risk assessment, the report supports the position that ecological oversight of planned introductions should be directed at promoting effectiveness while guarding against potential problems. Thus, the authors observe that most cases will present a minimal risk to the environment and provide a set of specific scientific criteria for sealing the level of oversight to individual cases. The four categories of criteria included: 1. attributes of genetic alteration, 2. attributes of the parent organism, 3. phenotypic attributes of the engineered organism in comparison with the parent organism. and 4. attributes of the environment. Specific attributes are grouped according to level of risk presented and corresponding level of scientific risk assessment needed. Coming as it did from a group of ecologists, the ESA report is often cited as a touchstone for those wishing to balance the positive potential of biotechnology with a sensitivity to the environmental consequences of actions. Biotechnology Ecological Risk Assessment Introduction A central goal of ecological risk assessment of planned introductions of recombinant DNA-modified organisms is to make a reasonably accurate prerelease prediction of the behavior an organism is likely to exhibit in its new ecological context and given its particular genetic modification, and to be able to detect and avert potential problems before they occur (76). Most scientists seem to concur that the focus of risk assessment should be on a particular organism, with its characteristics (genetically modified or not) and the genes that code for them, in a particular environment. Experimental protocols for ecological risk assessments need to be refined to screen out potentially problematic introductions before release (l4). While scholars argue as to which risk assessment model would best apply to environmental introductions of recombinant DNA-modified organisms, all agree that the complexity of ecological factors renders biotechnology risk assessment particularly challenging. Living organisms can change locution, reproduce, and perhaps exchange genes. Once released into the environment. they will interact in a dynamic fashion with other species. They are indeed different from chemicals. Ecological risk assessment is a still young methodology, and not standardized. Some argue that directly relevant data are scarce enough, and ecological phenomena are sufficiently complex that resasoned qualitative judgments are more feasible than more precise quantitative assessments. In practice, expert review panels using good scientific judgment and common sense. along with guidelines of points to consider, achieve qualitative assessments of the riskiness of various combinations of factors. As experience is gained. codification of the principles of review should evolve for application to future cases. Augmentation of human judgment with knowledge system technology has been suggested as a means of facilitating the process (24, 66). One way of conceptuall y applying risk assessment procedures to planned introduction of recombinant DNAmodified organisms into the environment is to match the three classic risk assessment stages (A. risk-source characterization: b. exposure assessment: and c. dose-response assessment) with the five stages involved in planned introductions. (See figure 8-2. ) Information about stage one. formation of a recombinant DNA-modifiedd organism. and stage two, its deliberate release or accidental escape into the environment contributes to risk-source characterization. Exposure assessment would take into account data on stage three. proliferation of the organisms, including dispersal and possible exchange of genetic material, as well as stage tour. their establishment in an ecosystem. Stage five, human and ecological effects, relate quite directly to dose-response assessment (23). Another way of looking at risk assessment of planned introductions is to consider the defination of risk as the product of exposure and hazard. Exposure is related to the possibility of escape of the arganism, its survial. reproduction. and spretd. as well as to the gene transferred and the vector, if present. Assessment of the hazard, or potential environmental impact. depends on the ultimate fate of the introduced organism-whether it becomes extinct. establishes a balance with indigenous species. or overruns the recipient environment (53). Specific objectives of ecological risk assessment for plants. for example. include: 1. determination of the potential for crops to persist and spread in a variety of habitats, 2. discovery of the range of species that can crosspollinate with various transgenic crops.

PAGE 233

232 l A New Technological Era for American Agriculture Figure 8-2Risk Assessment Framework for Environmental Introductions 1 Release 3 5 f I 1 Establishment In ecosystem I I SOURCE: Office of Technology Assessment, 1992. 3. 4. Box investigation of the ecological performance of hybrid plants produced, and development of protocols making it possible for crop breeders to carry out ecological risk assessments on new transgenic plants in the future. 8-A illustrates the sorts of specific questions that can be asked and answered about plant introductions based on field observations, field experiments, and contained experiments ( 14). Risk assessment pertaining to genetically modified (or nonmodified) viruses used in weed biocontrol, as an additional example, would include: 1. 2. 3. 4. In information on virus attributes such as virulence, host range, vector specificity, survival, and dispersal characteristics; information on desirable and undesirable virus attributes and on the stability of these attributes; information on the virus effect on the target weeds genetic stability; and information on the release site and how a variety of ecological variables affect infection, dispersal, population dynamics, and safety (87). summary, key features of ecological risk assessments of planned introductions include properties of the Risk-source characterization Exposure Dose-response Ecological effects assessment introduced organism (not the method by which it was produced) and of the recipient environment. including the demographic characteristics of the organism, the genetic stability and likelihood of gene transfer, and the interactions between the species and the physical and biological parameters of the environment. Scale and frequency of introductions should also be factored into risk assessments. Furthermore, since recapture or recall of introduced organisms usually will not be feasible, assessments should also consider possible means of containment, monitoring, and possible mitigation if adverse consequences occur (74). Research Needs and Promise of Risk Assessment The current interest in effective risk assessment of the products of biotechnology has stimulated workshops, conferences, discussions. and articles. More and more frequently. insights from the fields of ecology. population biology, population genetics, and evolution are being recast into the language of risk assessment (31, 50, 52. 62). Additional research needs to be undertaken on a variety of fronts to facilitate risk assessment. For example, a need exists to develop models and use data from field tests to predict the rate of spread of introduced organisms in various situations (54).

PAGE 234

Box 8-AEcological Risk Assessment Questions Field observations Field experiments Contained experiments Persistence What is the survival of the What is the fate of seeds sown into How is pollen viability affected in vegetative parts of the plant under a range of plant communities, transgenic plants? a range of climatic conditions, on including other arable crops, forage How is seed dormancy affected? soils of different kinds with different crops, permanent grasslands, and categories of drainage? natural habitats? How do transgenic plants perform How is perennation affected by the What is the fate of transplanted in competition experiments with crop plants and with selected introduced genes? seedlings in different habitats? native plants? What factors influence plant What is the fate of transplanted mortality outside arable fields and mature plants (or rootstock) in how are these influenced by the different vegetation types? novel genes? How long does experimentally What is the nature of seed planted seed remain dormant but dormancy under different viable in a range of soil types? environmental conditions, and how does the introduced genetic change influence triggering, duration, and hardiness during dormancy? Spread of the vegetative plant What is the seed production of the What is the vegetative growth rate Is seed size or morphology plant when grown in a crop and in on different substrates and with different in transgenic plants, and natural vegetation? different competing species? how might this affect seed Is seed production limited by the Is the thinning rule (i.e., densitydispersal? rate of pollination? dependent plant mortality) similar Do transgenic plants present What is the germination rate of for transgenic and nontransgenic greater risks of spread by seeds in soil? plants? vegetative fragments? What is the mortality of seeds and What kind of compensatory growth seedlings in arable soils and is exhibited (e.g., gap-filling)? beneath native vegetation? What is the phenology of seedling emergence and growth? What are the natural enemies of the seedlings? What is the role of vertebrate and invertebrate herbivores in crop and noncrop habitats? What is the mechanism of seed dispersal? How far are seeds dispersed and how does this vary with environmental conditions? Do the seeds produced by plants grown outside arable fields give rise to a second generation of plants? (continued on next page)

PAGE 235

234 l A New Technological Era for American Agriculture t Box 8-AEcological Risk Assessment QuestionsConthuecf Field observations Field experiments Contained experiments If the plant were to prove invasive, at what rate would it spread and which habitats would it occupy? Which plant species (if any) are displaced when (and if) the plant is established in natural habitats? Which plant species are responsible for the competitive suppression of the plant in different natural habitats? Horizontal gene transfer through pollen How much pollen is produced? What is the fate of labeled pollen? Which plant species allow pollen What is the phenology of pollen How much pollen reaches the stiggermination on their stigmas? production and what is the phenolmas of other wild plants under difHow is pollen dispersal affected in ogy of stigma receptivity of other ferent conditions? transgenic plants? plant species growing in the neighWhich insects carry the pollen? Which plant species form viable, borhood of crops (i.e., within 500How far away from the crop can an hybrid seed and at what rate is this 1,000 m)? individual, potted crop plant be polseed produced? Over what distance is pollen disIinated and how does the rate of What is the germination rate of hypersed under different meteorologipollination fall off with distance unbrid seed? cal conditions? der a range of habitat conditions? What phenotypes are exhibited by Which is the pollen deposited, on What plants make the most effihybrid individuals? which species, and in what numbers? cient pollen barriers for the conWhat is the performance of hybrid struction of guard rows; is it Where is the pollen deposited, on plants in competition experiments nontransgenic members of the with crop plants and with selected which species, and in what numsame species or plants that form native plants? bers? physical barriers to pollen flow or to insect flight? What is the nature of perennation What is the geographic distribution and vegetative dormancy in hybrid of closely related wild plants in the and transgenic plants? vicinity of centres of crop cultivation and what is their small-scale (100s m) distribution as weeds within arable fields and on land adjoining field foundaries? What natural habitats are found within 1,000 m of arable fields, in those areas where the crops are grown, and what flora is supported by these habitats? SOURCE: Michael J. Crawley, The Ecology of Genetically Engineered Organisms: Assessing the Environmental Risks, Intorduction of Genetically Modified Organisms into the Environment, Harold A. Mooney and Giorgio Bernardi (eds.) (New York, NY: John Wiley and Sons, 1990).

PAGE 236

Achieving predictive capabilities in extrapolating from field tests to large-scale introductions is an additional goal. Along with further research, data from field tests and research then can feed into the design of future field tests and large-scale introductions. our scientific understanding pertinent to ecological risk assessment should increase exponentially over the next few years. This explosion of knowledge not only can improve safety but also the effectiveness of introduced organisms in various habitats. There seems to be general agreement, even among ecologists and environmentalists, that most biotechnology products will not be harmful. However, because uncertainty does exist, for instance, as to which applications might be harmful. reasonable caution and willingness to assess risk are appropriate (76). Risk assessment prior to introductions is a reasonable and necessary step, consensus dictates. More research can sharpen our powers of prediction and build on an already sol id foundation of information. Eventually, criteria can be developed to match individual cases with appropriate risk categories. In the meantime, as a broader knowledge base is being built, the safety of each introduction needs to be judged, basically, on a case-by-case basis (51). Understanding gained from case studies and other relevant research can be employed in the current transition to risk assessments of large-scale introductions. Applicability of Diverse Bodies of Knowledge to Assessments of Large-Scale Commercial Release Introduction In all approaches to risk assessment, the key question is predictability. Do we have sufficient information to make a reasonable prediction as to what will occur for a particular release? Can we in fact legitimately draw on knowledge gained from agricultural experience, laboratory tests, past field tests of recombinant DNA-modified organisms, and accumulated knowledge of genetics, microbiology, molecular biologics, and ecology? Are the characteristics of any individual large-scale release familiar enough that we can bring such knowledge to bear on the risk assessment? Species Introductions Those interested in the evaluation of risks from biotechnology sometimes turn to the experience base with introduced exotics, species accidentally or deliberately released in a completely new environment. Dutch elm disease is often-cited as a consequence of the accidental introduction of a fungus; kudzu vine. running rampant in the South after being brought in as a roadside ground cover, is pointed to as a deliberate introduction gone awry. One viewpoint holds that species invasions may be useful analogues of planned introductions of genetically engineered species, i.e., an invasion is an invasion. Thus. experience with analyses of key properties of successful invaders, as well as of vulnerable environments. theoretically can be brought to bear in evaluating planned introductions (63). Most scientists agree, however, that invasions by exotics have limited applicability to planned introductions of genetically modified species. For example, introduced exotic plants that have caused problems come with many traits that enhance weediness; whereas genetically modified plants, by contrast. are modified in only a few characteristics (69). The distinction between the introduction of modified genotypes of crop organisms and the introductions of totally new exotics-whether or not they are genetically engineeredis, in fact, generally regarded as an important one (14). Even so. lessons learned as to the ecological parameters of invading species and recipient environments may be useful in categorizing degrees of risk for a specific planned introduction of a recombinant DNA-modified organism. For example, comparisons can be made between the characteristics of such an organism and the characteristics often found in very successful invading species. Habitat characteristics can also be compared to help assess site for vulnerability or resistance to invasion (63). Agriculture Perhaps the oldest analogue to planned introductions of genetically modified species is agriculture itself. For much of human history, new forms of crops and domesticated animals have been introduced to the environment. Major crops have been bred by the millions for centuries; all these field tests and commercial releases provide a substantial experience base. Throughout this vast experience, no significant harm to human or animal health has occurred due to these introductions per se, nor have major crop plants become bad weeds. Normal selection procedures have eliminated plants with problems. Furthermore, recalls of crop varieties are common under the laws of supply and demand. In short, no evidence exists in the United States that plant breeding leads to ecological problems (6). The NRC reports call for l familiarity as a criterion for risk assessment makes drawing on the experience base

PAGE 237

236 l A New Technological Era for American Agriculture Photo credit: Monsanto Genetically engineered tomato plants are shown being planted by researchers at a Monsanto-leased farm in Jersey County, IL. co. of agriculture logical for most planned introductions of genetically modified agricultural organisms. A specific example of how the agricultural experience can be applied to biotechnology risk assessment is the 80 years of usage of Bt (Bacillus thuringiensis with its toxin) as a natural insecticide; its history of safe use is often regarded as evidence that transferring the gene for a Bt toxin would be environmentally safe (6). The 100-year experience base with vaccines, rhizobial bacteria, and other biological controls provides information applicable to largescale microbial introductions (20, 29, 62, 90). As a final example, corn breeders have significantly changed the corn genome and have conducted planned introductions into the environment of these modifications for the past 70 years, without negative ecological experience. Breeders have gained experience in protecting the purity of these genomes, calculating the likelihood that the modifications will spread to other plants, deploying the modified genomes, and maximizing their strengths and minimizing their weaknesses ( 18). Although there are limitations to the analogy between seed purity and gene transfer to weeds (notably, the risks associated with weed genes contaminating seed for planting crops are quite different from those associated with engineered genes getting into a weed population), this analogy does represent a useful starting point for risk assessment in controlled release. Although some observers emphasize the novelty of gene combinations that can be brought about through biotechnology, a key difference between traditional crop breeding and the new biotechnology is that changes in genomes are more precise using biotechnology. With genetic engineering, one gene is moved at a time; by contrast, huge numbers of genes are recombined in crosses that lead to new plant varieties. It is nonetheless true that ecological effects of a changed phenotype sometimes may not be predictable even with precise changes in genotype (85). Certainly, risk assessments are needed of individual cases involving particular genes. For example, forage crops such as alfalfa, which are not so dependent on cultivation practices, may have higherand perhaps problematicsurvival capabilities outside of the farm than others (6). Two of the chief concerns about planned introduction of genetically modified species have no analogs in traditional agriculture. With the exception of some introduced crops that become weeds in tropical countries, crop plants have not invaded natural habitats. Furthermore. no obvious problems have arisen due to transfer of genes from traditionally bred crops to wild plants ( 14). Laboratory Testing Results of laboratory tests have been drawn on by those interested in risk assessment of genetically engineered microorganisms in particular. Various studies of microbial genetics, as well as use of soil microcosms (or laboratory model ecosystems) that mimick the natural environment, have provided useful information. A great many reported laboratory tests involve investigations of mechanisms and likelihoods of gene transfer. For example, transformation (the uptake of naked DNA into a competent or receptive cell) is a form of gene transfer well understood in the laboratory, but not well described in natural settings. Laboratory records on transduction (the transfer of genes between bacterial strains by virus particles) have led to theoretical models predicting the possibility and frequency of transduction from an introduced genetically modified microorganism to a natural species. Another mechanism of horizontal gene transfer studied in the laboratory is conjugation, the process of genetic exchange between bacterial cells. Finally, transposition, the process by which mobile genetic sequences change positions within a genome can be associated with gene transfer. Soil microcosms, even with sterile soil, are a feasible way of assessing what kind of gene transfer mechanisms can occur in nature; they are therefore a useful tool in risk assessment (38, 70). Research has now been done using more realistic soil microcosms, with the objective of learning more about the impact of conjugation on

PAGE 238

introduced genetically modified microorganisms. For example, some experiments have been done using nonsterile soils, in an attempt to produce a closer analogue to nature. Another set of questions that laboratory tests can help address is related to population biology. Relative fitness of genetically modified microorganisms in the laboratory, for example, pertains directly to establishment and possible spread of introduced organisms in an environment; some information toward quantitative risk assessments can be gained from contained laboratory testing in chemostats (44). Laboratory tests also can help illuminate the role played by various soil environments in successful introductions (93). Of course, constraints exist on the applicability of laboratory tests, having to do with feasibility and with the impossibility of reproducing the full complexity of a natural environment. Some important parameters relevant to introductions are, for example, the relative fitness of the introduced recombinant DNA-modified organism in the new environment with its multiple dimensions of biological, chemical, and physical features, including competition with other microorganisms; microbial population density, which may vary over time and space; population dynamics; and availability of habitats (5). The dynamic complexity of many such features makes it impossible for a laboratory test to mimic reality completely. Work is beginning on testing for effects such as pathogenicity or toxicity in more realistic multispecies systems or microcosms (26). Perhaps the principal lessons learned from laboratory research have to do with the potential to work creatively with soil microcosms. The more realistic the soil microcosm used, the higher the predictive value of the laboratory tests is likely to be, particularly where extrapolation from the laboratory to the field is relatively well understood. It has been suggested that mesocosms (larger contained walk-in chambers, the environmental parameters of which can be controlled) could provide more realistic complexity than soil microcosms. This added realism might improve risk assessment (93). Small-Scale Field Tests Field tests of conventionally produced crop varieties represent part of a step-wise progression toward full-scale commercialization; the same is true of field tests of recombinant DNA-modified organisms. Initially. new varieties are assessed in a laboratory or greenhouse; then they are observed in small-scale field plots where they are evaluated according to various protocols, statistical Photo credit: Monsanto Co. Researchers begin test of tomato plants carrying the Bt toxin gene in test plant. procedures, and analytical methods. Large-scale tests and commercialization complete the process. Each stage provides information for the next stage (53 ). For the most part. principles and procedures useful in small-scale field tests are also relevant at the large-scale test and commercialization stages as well (36). Field testing and monitoring constitute real world empirical methods that are important components of risk assessment (23). Small-scale field tests can be used to elucidate characteristics that will be factored into risk assessments of possible large-scale planned introductions. For example, survival and spread of particular recombinant bacteria in a particular soil environment, as well as efficacy of function and stability of an introduced gene. can be estimated in field tests ( 1, 3, 47). Field tests also can be used to assess invasiveness of transgenic crops (73). Data from field tests can be integrated into quantitative predictive models of gene flow and gene spread (39). Field tests also provide agronomically significant information, including data on the expression or performance of the introduced gene and on the overall growth and vigor of the genetically modified plant (64). For example, 1990 field tests of insect-resistant cotton plants have allowed such agronomic traits as yield, fiber length, fiber strength, fiber quality, seed composition, and quality to be evaluated by Monsanto, which is planning for commercial introduction in 1994 or 1995 ( 28). Well-designed, well-monitored field tests of increasing scale and complexity also should allow undesirable impacts to be observed while there is still an opportunity to correct them (43). A stepwise progression in test

PAGE 239

design is seen as an approach to field trials that will reduce complexity and otherwise benefit later I urge-scale efforts (47 ). (See box 8B.) An important stage is expansion from single-site into multisite field testing, which allows sites to undergo different conditions, such as weather, and thus provides information on the variation possible in performance and impact (73). Testing over more than 1 year can provide information on the consistency of measured characteristics such as survival and efficacy. Such information will have significant implications for commercial scale planned introductions. Good. statistically sound experimental design can be important in facilitating effective transitions from the field test to commercial-scale introduction (57). For agronomic and risk assessment purposes. scale-up from field tests is a useful and informative process. There are, however. a few constraints on the applicability of small-scale field tests to large-scale tests or commercialization. An important one is the emphasis often placed on containment in small-scale field tests involving recombinant DNA-modified organisms. Containment is, of course, the antithesis of uncontained, large-scale introduction (36). Bagging plants, for example, prohibits pollination and, furthermore. would not be feasible at a large-scale (53). When a product is commercialized. it will be far more widespread in the environment than it was in the days of its field test; many more nontarget species will be exposed to it (26). As people increasingly use transgenic plants. the chance for errors will increase because some users may not follow safety procedures (43 ). Despite these limitations, field tests are providing the datai about agronomic qualities and risk assessment considerations needed for the design of large-scale tests and commercialization. Detection and monitoring techniques are improving. A step-by-step progression from individual field tests through multisite field tests to large-scale testing to commercialization is being followed for recombinant DNA-modified organisms as it has been for conventionally produced organisms. without problems. Research still needs to be done to identify important distinctions between small-scale and large-scale tests; this should improve experimental design and efficiency (53). Deliberations on Field Tests and on Large-Scale Release Over the past several years. field tests have made important contributions to risk assessments for large-scale release of DNA-modfied organisms. The data from field tests provide the most directly relevant basis for predicBox 8-BLearning by Doing: Successive Field Releases Crop Genetics International (CGI) is a company that has used a stepwise progression in test design as it has moved from an initial field test to later tests. The focus was the delivery of biopesticidal gene products by endophytic bacteria inoculated into seeds. First tested was a bacterial endophyte (Clavibacter xyli subsp. cynodontis) genetically modified to produce low levels of the delta-endotoxin of Bacillus thuringiensis (Bt) subsp. kurstaki, and inoculated into corn seed. CGI developed a strategy for multiple risk assessment studies of field releases. The focus of the field release studies was twofold: performance of plants grown from endophyte-inoculated seed; and persistence and spread of the genetically modified strain under different environmental conditions. The first two releases were used to develop a profile of the recombinant strains behavior in the environment. In 1989, the test design was extended to multiple sites in four States to examine its behavior overdiversified environmental conditions. This was the first release to take place in multiple States of a viable microorganism genetically modified to produce a biopesticide. In 1990, a new recombinant strain selected for its activity against the target pest (European corn borer) was incorporated readily into the well-established testing procedures and program, with the objective of determining efficiency. As the study progressed between 1988 and 1990, by agreement with regulators, levels of containment were gradually lowered as data on safety were obtained. In fact, the early tests were specifically designed to address risk assessment issues such that future small-scale introductions could be made with less rigid containment and such that containment requirements could be eliminated in large-scale field tests. Efficacy studies now can be done under reduced containment requirements. Multiple-site field testing of the improved strains is the next logical step toward large-scale tests and commercialization. Stepwise progression of tests is a rational strategy from a companys point of view, as well as from a regulators point of view. SOURCE: Stanley J. Kostka, The Design and Execution of Successive Field Releases of Genetically Engineered Microorganisms, Biologicd Monitoring of Genetically Engineered Plants and Microbes, D.R. MacKenzie and Suzanne C. Henry (eds.) (International Symposium on the Biosafety Results of Field Tests of Genetically Modified Plants and Microorganisms, Kiawah Island, SC, Nov. 27-30, 1990) (Bethesda, MD: Agriculture Research Institute, 1991), pp. 167-176.

PAGE 240

(ion as to the safety of large-scale release, particularly in cases where a small-scale field test is itself scaled-up to a large-scale introduction. Equally important, scientists in many disciplines have been gaining practice through field testing in the process of risk assessment. Now that applications for large-scale release are imminent, researchers familiar with comparable evaluations at a smallscale can begin to integrate their experience and apply it to the new assessment task at hand. Several recent conferences have helped to define approaches to the risk assessment of large-scale introductions. Commonalities arc emerging, suggesting that a state of readiness for largescale introductions is in fact being reached. Several biological principles with implications for assessment of large-scale introductions emerged from the International Symposium on the Biosafety Results of Field Tests of Genetically Modified Plants and Microorganisms (November 27, 1990. Kiawah Island. South Carolina). For example: l l l l l The integration of genes into the chromosomes of recombinant DNA-modified organisms has proven to be predictably stable. Gene transfer frequencies of recombinant DNAmodified organisms are consistent with patterns recorded for natural populations. The frequencies of transposon relocations in recombinant DNA-modified organisms are consistent with those of natural populations. Some microorganism detection methods are extremely sensitive. and this contributes to better understanding of the fate of a microorganism in the environment. Background microbial populations have been characterized as complex. and thus the release of genetically modified microbes may be insignificant by comparison.. The symposium also highlighted the strong foundation of conventional knowledge in crop improvement, microbial testing. and food processing that is available to support safe commercialization of biotechnology products. Research needs cited included: detection methods, sampling methodologies. monitoring protocols and modeling techniques, and empirical data for improved design and evaluation of experiments (53). A workshop on transgenic plants conducted by the Maryland Biotechnology Institute and the USEPA Office of Pesticide Programs (June 18, 1990) evaluated the human and environmentall impacts that could result from the widespread. full scaleuse of plants genetically modified to produce a pesticidal substance. Workgroups discussed: 1) studies and information needed for assessment; 2) scientific rationale for determining the occasional need for specialized studies; and 3) availability and test protocols for developing risk assessment information. The consensus of all groups was that such transgenic plants posed concerns and possible effects that are not unique, and risk assessment issues can be addressed through readily obtainable information on possible effects of the plant or of the pesticidal substance (89). The USDA-sponsored Workshop on Safeguards for Planned Introductions of Transgenic oilsecd Crucifers (October 9, 1990, Cornell University) was held to identify agricultural biosafety issues relevant to oilsecd rape (or canola) as soon as possible. Unlike most crops, oilseed rape has weedy relations in North America. The potential for. and possible results of, gene transfer are therefore of concern. The workshop group agreed that with mill ions of acres planted. gene transfer will occur. Therefore, an ecological map of wild species was called for, so that the location of field trials could be planned to deliberately minimize proximity and hence possibility for gene transfer. Experimental trials and research were recommended to quantify risks. as were studies of the factors influencing gene transfer potentiali .e.. travel of pollen. effective fertilization, the production of viable seed, and the plant reaching reproductive age and passing on its new set of genes. The group agreed that studies should emphasize the conditions under which transfer and expression of the transferred gene take place. and the consequenses-relative riskof such events (61). A comparable meeting was held for maize and wheat (Keystone, Colorado, December 6, 1990); another is planned for rice. Summary A long history of agriculture provides an immense bunk of data relevant to risk assessment: diverse scientific fields contribute principles and knowledge. Data from small-scale field tests of recombinant DNA-modified oganisms not only provide specifics necessary for the evaluation of large-scale counterparts. they also provide a risk assessment testing ground. Each risk assessment of a field test adds to the regulator's experience base in adopting risk assessment methodologies to planned introductions. This learning through experience is a natural part of the evolution of oversight as we move from smallscale to large-scalec introductions.

PAGE 241

240 l A New Technological Era for American Agriculture Commercial Release Issues A variety of issues relevant to planned introductions of recombinant DNA-modified organisms are receiving heightened attention as large-scale commercial releases become imminent. Principal concerns focus on the fitness of the engineered organism (defined as overall genetic contribution to future generations, usually quantified as number of offspring produced) and its potential to become established as a weed or a pest, the stability of the engineered gene, the potential for gene transfer, and impact on other organisms and the environment. Basically, these concerns are the same ones raised with regard to small-scale field tests of genetically engineered organisms. Large-scale agricultural uses involve large numbers of organisms that are usually less contained than their less numerous counterparts in field trials. Fitness and Potential to Become Established For a species to become established in a natural community, its relative fitness must be such that it competes successfully with other species. The lack of weediness on the part of most major crops illustrates a direct contrast between domestication and what is useful for survival in the wild ( 14. 43). Many traits necessary for successful weediness either have never existed in or have been deliberately bred out of crop plants to maximize productivity in a cultivated setting. One analysis showed that serious weeds tend to have on average 10 to 11 weedy characteristics; crop plants have on average only 5 of these characteristics (42). Thus, the chances of any crop plant simultaneously undergoing five to six relevant gene changes to become a weed are vanishingly small (37). Features of organisms that ecologists identify with weediness include broad ecological tolerance, ability to exploit an under-utilized resource. or readaptation to a new habitat to which the organism is well-suited and in which controlling biological agents do not exist (76). Other characteristics that help to make a plant thrive as a weed include the following: l l l l l l l l l rapid growth to a flowering stage, continuous seed production as long as growing conditions allow, high seed output. long-lived seed, pollination by wind or unspecialized insects, high competitive ability, broad environmental tolerance, seed dispersal over short and long distances, and vegetative persistence and propagation. The probability of successful establishment of a recombinant DNA-modified organism as compared to its unmodified counterpart will naturally be dependent on the nature and phenotypic expression of the specific genotypic modification made, along with the rest of the organisms phenotype, in relation to these ecological criteria. Different kinds of engineered genes will vary in the degree and nature of their impact on the phenotype of the engineered organism. Also, engineered genes may vary in terms of the conditions under which they will be expressed. For example, if a gene is only induced to be expressed under specialized conditions, then its phenotypic impact will be negligible the rest of the time. It has been well established from studies of induced mutations that most dramatic phenotypic changes in an organism result in reduced fitness (2, 14). Engineered genes that affect the growth, resource allocation, or some other aspect of an organism may convey added economic value, but may also produce a maladapted plant that is unlikely to survive outside of cultivation. On the other hand. genes that have relatively little effect on the overall phenotype, such as genes induced only on certain occasions for disease or pest resistance, might confer a real fitness advantage, even in natural populations. It is generally assumed that genes for disease resistance present a physiological cost that reduces fitness in the absence of disease, although the importance of that cost has been challenged (7 I ). However, sometimes if the gene is not expressed, such costs go down, contributing to its potential long-term persistence. Assessments of the risks of introduced organisms becoming pests must take these factors into account as well as others. For example, introducing a character into an organism whose ecological properties are otherwise wellknown, or taking a particular property associated with terrestrial bacteria and introducing it into another terrestrial bacterium, enables some prediction of how that character might respond in that target ecological setting. Thus, in assessing the potential risk associated with a particular phenotypic modification, the target environment should be considered. If a species became established as a pest, existing communities would be disrupted; fortunately, the likelihood of either a genetically modified plant or a microorganism becoming a pest is relatively low. Most crop varieties produced through conventional means do not become pests (6). Experiments to date indicate that genetically modified microorganisms in some cases may not persist at significant levels (3) and therefore may

PAGE 242

often be unlikely to proliferate and disrupt existing com munities composed of vast numbers and numerous species of microorganisms (19. 86). So for all organisms modified in any way. emphases in risk assessment of microorganisms should be placed on the specific product. Until more is known about consequences of large-scale use of genetically modified plants. a deliberate approach rather than complacency seems warranted. Gene Stability The stability of an engineered gene is important to risk assessments of planned introductions of recombinant DNAmodified organisms. A gene that has become a stable component of the transgenic organism is more predictable in its function, expression. and possible mobility than one that has not. one aspect of gene stability is persistence. An engineered gene construct usually consists of several components, all of which must be present and intact for the gene to function. In addition to the structural gene that codes for the desired gene product. a promoter gene is needed for it to be expressedto be turned on or off. Such constructs maybe broken apart by natural genetic recombination. A promoter separated from its structural gene is useless; the structural gene without the promoter remains unexpressed. The stability of a particular gene also may be directly influenced by the vector used to introduce it into the engineered organism. Bacterial plasmids or DNA-carrying bodies, are potentially the most mobile of the vectors used to insert genes. Plasmids function by inserting themselves into the bacterial chromosome. carrying an engineered gene along with them. Insertion sites for such plasm ids are nonrandom; they are specific sequences that could be recognized by other plasmids, which may pick up the inserted gene and carry it along to another organism. on the other hand, it also is often true that insertion of a particular plasmid will immunize the cell against insertion of similar plasmids. Genes directly inserted into chromosomes are more stable than genes carried by plasmids. However, chromosomes are complex structures, and the manner in which particular genes express or recombine is determined by their relative positions on chromosomes. An engineered gene inserted in some parts of the chromosome may be more exposed to recombination than genes on other parts of the chromosome. The relative stability of an engineered gene in a plant species can be increased by inserting it into portions of chromosomes subject to lower levels of recombination. To summarize. a genes stability depends on the nature of the gene itself and on the means of introducing it into the recipient organism. Either of these can be manipulated deliberately to increase stability. Gene Transfer Another appropriate focus for risk asessments of planned introduction of recombinant DNA-modified organisms is the possibility) that novel genes may become incorporated into related wild species. Such transfers, it is argued, might lead to harmful bacteria or weeds with an improved" characteristic such as resistance to pest attack: this might make them more difficult to control. Three key questions to be considered are: What is the probability that a gene will move from an agricultural organism to wild species? What can be done to lower the probability? What would be the consequences of such gene transfer on agricultural and natural communities
PAGE 243

On the bright side, a number of recently developed techniques exist that can greatly facilitate studies of bacterial interactions in natural substrates (48), including flow cytometry (a technique that involves the use of laseractivated fluorescence of stained particles) and polymerase chain reaction (PCR) (65), involving the amplification of a particular gene contained at low concentration in soil to sufficiently high concentrations that it can be detected by standard DNA analysis. PCR can be used to monitor the movements of introduced genes in natural substrates (79). This allows the population dynamics of the engineered organism to be more closely monitored, the transmission of the engineered gene to background organisms to be quantified, and potential risks to be evaluated. Also, the introduced population can be tagged with a specific but nonfunctional DNA sequence such that the growth or decline of that population in the soil can be monitored independently of the engineered gene(s). Actual probabilities of gene transfer of various kinds among microorganisms are still being researched. Although differing opinions certainly exist, one school of thought is that the order of magnitude of microorganisms present in the natural community, and the probable frequency with which they exchange genes, renders the potential impact of most recombinant genes being transferred relatively low. For higher organisms. vector-mediated transfer of engineered genes is not a major concern. For example, a widely used vector for dicotyledormus plants, Agrobacterium tumefaciens (crown gall virus) can be readily screened out of transformed organisms before they are released. Furthermore. for many important crop species, notably cereal crops, vectors for gene transfer are not used: rather ballistic incorporation of genetic material into tissue-cultured cells (using gene guns ) is the method currently in development. Using this method there is no chance of vector-mediated gene transfer. This leaves gene transfer through hybridization of crops and reproductively compatible (i. e., closely related) weeds as a possibility. In higher plants, the main risk associated with gene transfer from transgenics into surrounding populations is. in fact, that of hybridization. Modified genes potentially could be transferred from transgenic plants and incorporated into the genome of a weedy species through introgressive hybridization, whereby genes are transmitted through pollen in sexual reproduction. However, working against this possibility are limited viaility of pollen, distance and physical barriers to pollination. genetic dissimilarities (i. e., incompatible fertilization processes), and failure to produce viable, fertile offspring. Most crops grown on a large scale in temperate regions, such as corn and wheat, are grown outside of their geographic region of origin; consequently there typically are no related weed species growing in association with them. Therefore, for most crop species in the United States, pollen-mediated transfer of modified genesis only of theoretical concern. However, there are several important crop species for which closely related weed species have become introduced. Specifically. many crops in the family Brassicaceae. such as canola (oil-seed rape) and radishes, have co-occurring weedy relatives (21). Sunflowers had their center of origin in the United States and have related weedy species here as well. Most major crop species originated in what arnow regarded as developing countries. For example, corn was developed in Central America, wheat was first cultivated in the Middle East, rice in Southeast Asia. and potatoes in South America (77). Consequently, introduction of genetically engineered crops into such regions should be handled with particular attention to the probability of gene transfer into background populations. Additional concern focuses on the potential impact of introduced genes on the genetic structure of natural populations of plants related to important crop species. These populations represent the genetic heritage of the crop and are an irreplaceable reservoir of diverse genetic variation that may be needed in future development of the crop (8). If, because of a novel gene effect, one strain or lineage became a super weed it might outcompete and therefore eliminate other lineages; genetic variation potentially useful for crop development could be lost. More generally, biodiversity is intrinsically valued by many ( 12). Pollen-mediated transfer of novel genes from crops into related weeds might also result in weeds becoming similar to the crop species. A number of well-known instances exist where selection pressures exerted by traditional agronomic practices have caused weedy species to evolve to resemble the crop species. Such weeds cannot be eliminated by standard control practices (4). Thus, weeds are capable of a wide range of genetic adaptation even without the introduction of novel genes. Although there could clearly be problems associated with potential gene transfer from transgenic plants into weed populations. there is also a large experience base in agricultural and natural populations on which to draw for predictions in this area.

PAGE 244

A great deal is known about pollen transfer in plants (35) and associated Likelihoods of gene transfer. In the past few years, there has been a growing interest in tracking pollen in natural populations through paternity analysis, a technique directly analogous to human paternity analysis (58, 82). The development of such approaches provides a useful means of evaluating the potential spread of modified genes, as well as a means of testing the efficacy of various measures to prevent pollen spread into wild relatives. Gene flow in many crop species has also been studied extensively in order to determine necessary distances for genetic isolation of different plots to reduce genetic contamination of seed crops in conventional agriculture. For example, genetic contamination of seed in plantations of conifers can reach levels of 30 to 50 percent and is an extensively studied problem (78). A review of gene transfer from corn to related species concluded that the prospects for introgressive hybridization in corn were limited (17). However. it is unwise to dismiss completely consideration of gene transfer because genes transmitted at low levels could be rapidly enhanced through natural selection if they confer an advantage to their recipients. A study of hybridization among six different rice cultivars developed through conventional agriculture and the related weed red rice (Oryza sativa L.) found widely varying rates of hybridization with the different cultivars. The hybrids generally showed evidence of convergence towards the crop, thus opening the possibility of generating a particularly noxious weed that closely resembles the crop (49). For specific applications of biotechnology, it is possible to articulate potential risks of gene transfer and evaluate their probability. Furthermore, long-standing agricultural practices (e. g., isolation of crops for seed certification) can be useful in managing this risk. For the few U.S. crops with weedy relatives (i.e., canola), and for other countries where crops have multiple related species, careful risk assessment should lead to reasonable risk management. It is important to remember that successful cross hybridization is in fact a complex multistep process and does not usually lead to viable, fertile hybrids. unless the species are closely related. Evolutionary Pressures Placed on Other Organisms Evolutionary pressures on indigenous organisms can arise in several ways. Novel organisms in a biotic community may provide new levels of competitive interactions; they may impose direct selection pressures on the native organisms; they may also enhance one species at the expense of others. Thus the assessment of risks (and benefits) associated with the planned introduction of recombinant DNA-modified organisms must consider the engineered organisms probable interactions with the target biotic community. Many such interactions occur in convolution of a cultivated species and its associated pathogens. pests. and weeds. One interaction that should be beneficial in terms of controlling crop pathogens involves a pathogens response to resistance factors. Factors conferring resistance to pathogens can be conventionally bred or genetically engineered into plants. It is well established that the introduction of pathogen resistance factors imposes selection pressures on pathogens to overcome these factors by evolving greater virulence (34). Using conventional breeding methods, it can take longer to introduce a resistance factor into a crop species than it does for pathogens to respond. Genetic engineering promises greatly to reduce the time frame for introducing resistance factors. This buys the crop some lead time before the pathogen evolves a response. Strong selection pressures also are exerted on pest species to evolve counter measures to control technologies. The use of Bacillus thurigiensis ( Bt. for example, is an effective means of controlling insect pests that could become overutilized and thus rendered ineffective. The bacterium itself often is used in broadcast spray applications to control insect pests, and the gene for toxic agents in Bacillus thuringiensis has been cloned. The gene now is being incorporated into crop species in field tests. This will exert even stronger selection pressure on insect pests. Several approaches may help to diminish selection pressure and thus slow down the rate of evolution of resistance. (See ch. 6.) It may be possible, for example, to introduce the Bt gene in such a way that it is only turned on during certain stages of development, only in certain parts of the plant. or only at times of insect attack, thus decreasing its impact. Scientists from several agricultural companies have formed a Bt resistance club to discuss how to slow the evolution of resistance to Bt. Another concern is that use of genetically engineered crops for herbicide resistance may result in overuse of specific herbicides and thus impose strong selection on weeds to evolve resistance to those herbicides. For example. if even environmenttally friendly" herbicides are overused in conjunction with transgenic monoculture, weeds might evolve resistance fairly rapidly. This may lead to a desperate use of far more damaging herbi-

PAGE 245

244 l A New Technological Era for American Agriculture cides. Management strategies for slowing the development of resistance may be needed. (See ch. 6.) The convolution of a cultivated species and its associated pathogens and weeds is a quite predictable process if one genetic locus for one resistance factor is considered. The sequential introduction of resistance factors in a crop species ultimately can lead to the socalled gene for gene condition in which each gene for some resistance factor in the host is matched by a gene for virulence in the pathogen. One way to break this cycle is simultaneously to introduce multiple resistance factors, thus impeding the pests evolutionary response. Similarly, different resistance factors might be cycled from year to year so that the pest never fully responds to any one resistance factor (34). The use of genetic engineering techniques could greatly facilitate such strategies because it provides a tool for rapid generation of new lines containing different combinations of resistance factors. Monitoring Assessing the potential risks of environmental introductions of recombinant DNA-modified organisms, and evaluating how best to manage these risks, entails spatial and temporal monitoring of the organisms and of their introduced genes. Monitoring contributes to risk assessment and management in two ways. First. in a specific situation, it tracks indicators of gene transfer or spread of introduced organisms so that action can be taken if needed. Beyond this, monitoring adds to our database, so that risk assessments of subsequent introductions are even more accurate. Monitoring of field tests can provide information pertinent to subsequent field tests and to large-scale introductions. For example, presence or amount of gene transfer from transgenic crops to related or nonrelated weedy species could be estimated from monitoring species surrounding a test field containing a recombinant DNA-modified crop. These data can be used in future field tests or large-scale introductions involving similar crop/weed complexes. Monitoring also can help elucidate any spread of introduced microorganisms. As the ecology of their spread is understood more fully, risk assessments of new introductions can be improved. Thus, monitoring has an important role to play in the natural evolution of science-based. risk-based regulatory oversight. Highly sensitive monitoring techniques are developing rapidly. (See box 8-C. ) The following is an example of the kind of data that the Animal and Plant Health Inspection Service (APHIS) can require from monitoring (in this case recombinant entomocidal or insect-killing bacteria were field tested). Required monitoring provided data on: 1. 2. 3. 4. 5. 6. 7. 8. 9. plant colonization by the recombinant bacteria at 4 weeks after inoculation; colonization of all plant parts by the recombinant bacteria monthly for 4 months; dispersal, natural and mechanical, in the field of the recombinant bacteria after 60 days; presence in run-off water of the recombinant bacteria; presence in soil of recombinant bacteria populations; effect on crop yield of the recombinant bacteria; effect on crop residue decomposition of the recombinant bacteria; effect on vesicular-arbuscular mycorrhizae of the recombinant bacteria 3 and 6 weeks after planting; and effects of the recombinant bacteria on saprophytic gram-negative bacteria in the phylloplane. Other points of interest needed to be addressed through the ability to track the recombinant bacteria, as well (22). The monitoring data collected enabled APHIS to assess patterns of the spread of the recombinant bacteria on the targeted plant and its various parts, the dispersal of the bacteria in the field water and soil, effects of the bacteria on crop yield and decomposition, and the effects of the recombinant bacteria on mycorrhizae and other plant bacteria. In short, required monitoring of plants and soil contributed directly to understanding of dispersal and effects of the recombinant bacteria. Plants generally are easier to monitor than microorganisms. As techniques for monitoring improve, field test data and, soon, large-scale test data will improve our knowledge of survival and spread of recombinant DNAmodified organisms and their genes. thus aiding us in reasoned risk assessment and management. Research Needs For the past two decades, basic research in molecular biology has generated many novel scientific insights and products. As a result of strong government support for such research, we have reached a point where the planned introduction of recombinant DNA-modified organisms is a reality. However, the fields of ecology and evolutionary biology, which can provide the kind of information and expertise needed to predict the impacts of planned introductions, have enjoyed less support. Fortunately, ecologists are now taking a leading role in defining a research

PAGE 246

Chapter 8Scientific Issues: Risk Assessment and Risk Management l 245 Box 8-CMonitoring Microorganisms Detection and tracking (monitoring) of recombinant DNA-modified organisms and their genes makes possible quantification of persistence or spread. Highly sensitive new techniques, among them polymerase chain reaction (PCR) and antibodies, are being utilized to contribute to the efficacy of monitoring. Data resulting from monitoring in turn contribute to the knowledge base on which risk assessments of prospective small-scale and large-scale introductions can be based. In fact, regulatory agencies request that certain parameters be monitored in field tests allows them to fine tune upcoming assessments of large-scale applications, and to make plans for their management. The first approved environmental introduction of a living genetically modified soil-borne bacterium in the United States, in fact, had as its goal monitoring of the bacteriums population dynamics, persistence, and movement through the soil. The genes lac Z and lac Y were engineered into a root-colonizing fluorescent pseudomonas (P. aureofaciens), part of a bacterial group that often promotes plant growth and protects against some plant diseases. The added genes allow the bacterium to use lactose as a source of carbon and energy and result in readily discernible deep blue bacterial colonies on a petri dish, thus providing an excellent monitoring tool. Scientists from Clemson University and Monsanto studied bacterial spread, population dynamics, and persistence over three crop cycles (19 months) in a wheat field and found similar values for both the modified and the nonmodified strains. Both strains declined to below detectable limits 38 weeks after inoculation. Also monitored were the foliar tissue of the first winter wheat crop analyzed 3 weeks before harvest and found not to have either strain present; and native soil bacteria, to which the Iac Z and Iac Y genes were not found to have transferred. The studys multifaceted sampling design, use of new techniques such as chromosomal DNA fingerprint patterns, presence of a control in the form of a nonengineered strain, and followup over three crop cycles set good examples for thorough monitoring studies in other situations (45). This work also is noteworthy as the first study analyzing frequency of genetic exchange in the environment of genes inserted into bacterial chromosomes rather than plasmids. This success of the chromosomal approach has implications for scientific management of gene transfer in microorganisms. In future monitoring studies, the transgenic organism or the inserted gene itself might be tracked by a nucleic acid probe for a specific DNA sequence; as well as by selective media for metabolic characteristics or by antibodies to a characteristic antigen. Some tracking techniques require that bacteria be isolated and grown in the laboratory, but others are being developed that can analyze bacterial DNA as isolated from environmental samples, a capability useful in estimating the population of the introduced organisms. Still other techniques, including pulsed field electrophoresis, can be used to analyze total DNA in a simple community and possibly to then quantify different members from the sample. In communities that are more complex, higher resolution is needed and probes maybe necessary. In such cases, antibodies may give a great deal of information by tracking phenotype through detection of proteins present (19). Polymerase chain reaction methodology is an innovative technique that can be used essentially to magnify sensitivity of detection. Flow cytometry, a cell-sorting technique, may also have some application to monitoring. SOURCE: Philip C. Kearney and James M. Tiedje, t Methods Used to Track Introduced Genetically Engineered Organisms, Biiotechno/ogy for Crop Protection, Paul Hedin, Julius Menn, Robert Hollingworth (ads.) (Washington, DC: American Chemical Society, 1988). agendna to respond to a variety of social needs, including tributions to these predictive capabilities. However. funding the planned introduction of recombinant DNA-modified organisms (85 ). Since introduced species or their genes may be incorporated into natural biota, over time. a similar agenda is needed for evolutionary biology to assess the likelihood of propagation and persistence. The likelihood of an introduced organism becoming established. competing with other organisms. spreading. exchanging genes with members of other species. indirectly affecting nontarget species, or changing over evolutionary time all need to be predicted in risk assessment. A number of fields in biology are already making confor further research is needed. (See box 8-D. ) Development of mechanisms for effective coommunication between fields is critical to meeting research needs associated with the planned introduction of recombinant DNA-modified organisms. It has been noted that interdisciplinary research is critical for the development of risk assessment and risk management pertinent to planned introductions (92). In particular. the gap between ecology and molecular biology needs to be spanned. Scientists in both areas need to be trained or encouraged to be more aware of each others fields.

PAGE 247

130x 8-DRelevant Research Fields Community Ecology Community ecology is the study of interactions of populations of different species in a given habitat. Interspecific competition, predation, and other interactions are the province of this field. Modern community ecology is an experimental field; however, most experimental studies are limited in scope to consideration of two, or at most three interacting species. Larger experiments focusing on more realistically complex interactions, and desirable predictability of response to perturbation, will require more research. Ecological systems research on topics such as nutrient cycling can provide relevant information as well. Population Ecology Population ecology is the study of the dynamics and growth of populations. Such studies may emphasize properties of the species itself, such as fecundity or mortality rates, or they may emphasize effects of environmental or biotic interactions. There is a growing trend to incorporate population ecology into conservation biology. Analysis of life history can be used to determine which stages (e.g., seedling establishment versus adult survivorship) are limiting to population growth. Such analyses of sensitivity in population dynamics (9) could be useful in risk assessment of ecological impacts of recombinant DNA-modified organisms. Population Genetics Population genetics is the analytical study of properties of genes and changes in gene frequency over time. The mechanism by which genes are transmitted from one generation to the next and the relationship between particular genes and fitness are key to this field. This field is distinctive among biological fields because of its sophisticated theoretical framework. The theory enables some level of prediction about the behavior of genes in populations, but more emphasis on empirical studies is needed to generate useful predictive models of gene change. Evolutionary Biology One way to encourage empirical work in population genetics would be to place more emphasis on research in evolutionary biology. Changes over time in genetic structure and consequent phenotypes-of populations are foci of evolutionary theory. Emphasis on dynamics of change predisposes the field towards questions of relative Spread of genes and impact of phemotypes in an ecosystem over time; these are questions that are relevant to risk assessment of planned introductions. Systematic The field of systematic encompasses analysis of variation of different levels of taxonomic organization. Although the ultimate goal of such analysis is taxonomic classification, this field is increasing in importance in analysis and monitoring of biotic diversity. This field could contribute to risk assessment through analysis of species relationships and species ranges to evaluate the probabilities of hybridization. Mathamatical Modeling Mathematical modeling entails construction of a mathematical framework to describe a process and predict outcomes from that process. Modeling has been an effective approach in risk assessment and strategic planning in agriculture, For example, models have demonstrated that allowing the existence of marginal populations of pests lets them serve as reservoirs for genes that confer susceptibility to pesticides and other means of control, such populations therefore can beneficially slow the rate of evolution of resistance (34). This seemingly counterintuitive result contraindicates a straightforward program of eradication. Risk Assessment Methodologies Risk assessment involves the ranking of probable outcomes from possible events. As such, in order to rank risks, one needs to first define the risks of a given practice. Development of risk assessment methodologies is an ongoing practice, and practitioners must always be ready to adapt to new problems as they arise in different situations, such as commercialization of diverse crops in a variety of environments. SOURCE: Office of Technology Assessment, 1992.

PAGE 248

Chapter 8Scientific Issues: Risk Assessment and Risk Management .247 More communication between scientists involved in basic research and applied research is also needed. Just one example is the need for communication and interaction between plant-resistance breeders and evolutionary biologists (33). Another example would be communication between farm management systems research and ecology. Many people who work in basic research are in part motivated by applied concerns. However, it does little good to generate insights on an applied problem unless there are lines of communication whereby the results of those insights are incorporated to solve the problem. Questions regarding applied problems also need to be articulated to basic researchers. RISK MANAGEMENT Genetically modified organisms introduced into the environment do not present us with radically novel problems. Furthermore, we have a sufficient enough base of technical knowledge and risk assessment methodologies that we can make reasonable, science-based assessments of the likely impacts of individual proposed introductions. The concerns raised do not need to paralyze agricultural progress based on biotechnology. These concerns can be respected, weighed, and addressed as necessary through science-based regulations and scientific and agronomic methods of managing risk. Design of Science-Based Regulation The 1986 Coordinated Framework (5 I FR 23302-23393, 1986). the more recent scope document (55 FR 147, 3118. 1990). and other reports attempt to create a technically sound context for biotechnology oversight. (See ch. 7. ) Reviews of field trials to date have been based on technical issues of risk reduction. Technically sound evaluations of safety can provide principles for regulation and oversight. Agencies receiving proposals can add specific stipulations for risk management (66). A variety of scientific fields ranging from molecular genetics to ecology need to be brought to bear on the design or performance of oversight. As research progresses. predictability about risks and insights as to how they should be managed will improve. The imminence of large-scale introductions underscores the need for clarification of how risk will be managed in various situations. Identification of issues. development of policy, and structure for large-scale tests and commericializations, along with modifications of the approval process for small-scale field tests, are all being requested from regulatory agencies, who are themselves grappling with the issues involved (36). Generic v. Case-by-Case Approach Extrapolation of results of risk assessment from one site to another still needs refining; this has ramifications for multisite, large-scale introductions. Many believe that evaluation of risks must be specific to the particular application. However, attempts have been and doubtless will be made to associate individual cases with appropriate categories of risk and to manage them accordingly (51). One key issue in the approach to risk management in planned introductions of recombinant DNA-modified organisms is whether to use a case-by-case analysis approval process or a process built on generic categories. Some, looking at the large number of applications coming down the pipeline, advocate a shift from the current case-by-case review of experiments toward more of a generic approach. Possible strategies under this approach include categorical exemptions, licensing certain categories of tests. licensing individual scientists. or delegating authority to institutions (53). Others fully expect large-scale tests and commercialization. in particular, to be reviewed on a case-by-case basis. but they do encourage the rapid appearance of protocols or some other form of guidance so that safe and effective products can be developed (36). Advocates of a case-by-case approach point to its flexibility. As different cases arise, each can be dealt with in a manner appropriate to its nature; no one set of rules and regulations, it is argued, will cover all of the many and varied applications of biotechnologgy. Nonetheless, over time, as our experience and research base grows it is likely that some generic approaches to certain sorts of introductions in certain sorts of environments will emerge. The criteria by which these generic approaches are defined (some requiring more attention than others) will themselves change over time (74). These developments were anticipated in the ESA report (85), which made a significant step toward scaling risks. Eventually. generic categorizations of likely risk are probable, yet each case will need to be double-checked for any idiosyncratic particularity that could trigger more focused review. It is important for the successful application of biotechnology to agriculture that sufficient long-term flexibility is built into the regulatory and oversight system

PAGE 249

so that risk management can evolve based on improved understanding. Relative Risks Compared to Traditional Practices Risk management involves the weighing of costs and benefits. To put planned introductions in context, their risks could be compared to risks of traditional practices in agriculture and society. For example, risks today are associated with the widespread use of chemical pesticides; accumulation of nonbiodegradable materials; toxic wastes; agricultural practices giving rise to genetic uniformity in farm animals and crops, with loss of biological diversity; and natural biological calamities, such as the current epidemic of AIDS. Not only are risks of planned introductions put into perspective by these nonbiotechnologyrelated problems, but biotechnology itself may help to solve some of them. For example, biotechnology can provide alternatives to chemical pesticides, assist in the degradation of toxic wastes, provide alternatives to selective inbreeding, and contribute to development of diagnostics and vaccines for AIDs and other illnesses (74). On a more specific level of cost/benefit comparisons, new biotechnology techniques can be compared to those associated with traditional biotechnologies. (See table 81 for one view of such a comparison. ) Certainly controversy existsfor instance, over the relative predictability of the ecological behavior of the phenotypes of transgenic organisms even when genotype changes are precise and well-understood. On the other hand, conventional breeding changes many genes simultaneously, with consequent multiple phenotypic changes. The newer, more precise techniques may actually show up well in the comparison. Cost-Benefit Analyses Risk management includes the weighing of risks or of actual costs on the one hand against benefits on the other, and then trying to achieve a reasonable balance (74). Agricultural biotechnology has potential to create positive benefits for agriculture. horticulture, range management, and forestry in the 21st century (43) if it is not stalled in its developmental stages; on the other hand, it is to no ones best interests to proceed without attention to identifying and minimizing any likelihood of risks. An appropriate balance is necessary. In a time when the expansion potential of land for agriculture is small, when labor is expensive, and when additional use of chemicals in agriculture generally is regarded as a negative, the possible exploitation of new capabilities and new information through new technologies cannot be ignored. Thus, regulations that are not science-based could exact a very real cost, that of not introducing an innovative, promising product. Small-Scale v. Large-Scale Issues As agricultural biotechnology nears the commercialization stage, risk management must take into account a number of realities, as was mentioned in the previous chapter. For example, large plots at a number of locations are needed to test a recombinant corn line. This testing needs to be done within I to 2 years of the creation of the recombinant line for a company to stay competitive in the development of new varieties. Furthermore, many hybrids will be undergoing evaluation at the same time; several of these may contain the same recombinant gene and several recombinant genes might be examined simultaneously. In short, if the recombinant material goes Table 8-lComparison of Traditional and Developing Biotechnology Characteristics Organismal Cellular Molecular Processes Breeding Culture Cell rDNA Anther Embryo Selection Regeneration Mutation Fusion Control over changes Random Semi-random Directed, precise Primary changes Unknown Semi-known Known Number of variants needed Large Intermediate Small, in vitro selection methods Species restriction Mainly within Within & across Within & across Familiarity Very high Intermediate Low but expanding Ability to ask and answer risk questions Low Intermediate High Containment Dependent on organism and independent of method; established procedures for domesticated organisms. SOURCE: R.W.F. Hardy, (Large-Scale Field Testing and Commercialization: Thoughts on Issues, Biological Monitoring of Genetically Engineered Plants and Microbes, D.R. MacKenzie and S.C, Henry (ads. ) (Bethesda, MD: Agriculture Research Institute, 1991),

PAGE 250

Chapter 8Scientific Issues: Risk Assessment and Risk Management l 249 successfully and quickly through testing, the breeder will soon work to combine it with other useful traits, in different genetic backgrounds, as part of genetic improvement. Large numbers of new lines will emerge from the integration of recombinant genes into conventional breeding programs so that new hybrids can be tested and commercialized. Specific recommendations for risk management of the transition to large-scale could include: I ) making geographic maps of crop relatives and placing them in an accessible database, and 2) modifying the process for approving small-scale introductions based on experience base or familiarity (36). Marshaling evidence from experiences in agriculture, laboratory tests, introductions, field tests, and current, ongoing research will make possible reasoned risk assessment and management. SCIENTIFIC METHODS OF MANAGING RISK The power and precision of biotechnology can be harnessed for risk management itself. Controlling the spread of introduced genes through the manner in which they are introduced is one example. Risk management can be greatly aided by using supplementary transferred genes to ensure that the ensuing recombinant DNA-modified organism only functions on certain occasions, under certain environmental conditions, or for a finite period of time. The genetic modification can be designed to: 1 ) constrain the potential for gene transfer (increasing the containment of the gene within the organism into which it was inserted), and 2) maximize its key activity while minimizing effects in the recipient environment (60). Mechanisms for fine-tuned technical control of this sort still are being developed; a few approaches are described briefly here. In general. in addition to turning the gene on or off under certain conditions, several approaches to containment could be considered: autodestruct mechanisms (e. g., suicide genes), engineering genes such that the host has diminished survival (as through defective regulation of metabolism), and decreasing chances of horizontal gene transfer to other organisms (as through reducing the stability or ease of inheritance of the introduced genes) ( 19). Promoters Turned On or Off by Specific Stimuli One way to limit the effect of the engineered gene itself is to attach it to a promoter that only allows expression under certain conditions (83). When a gene is not being expressed, the physiological expenditure associated with expression of the gene can be allocated to other purposes. This maintains the efficiency of the organism and keeps the impact of the genes phenotype to a minimum. For example, some genes are only expressed when triggered or induced (usually through a promoter gene) by a certain chemical, such as a herbicide, or in the event of local disturbance of tissue, such as a wound response resulting from chewing by insects. A gene for some form of pest resistance attached to such an inducible promoter gene would have little phenotypic impact except in the presence of a pest. This is a realistic strategy with diverse applications, some of which already have been field tested. For example, a field test was conducted by Iowa State University to assess whether or not transgenic tobacco plants would respond to insect attack by turning on an inserted gene. Plants often can respond to insect attack by activating genes coding for defensive compounds. Such compounds may, for instance, block the digestive system of insects, reducing their leaf consumption. A marker geneone used to trace the success of the recombination experiment (chloramphenicol acetyl transferase, CAT). modified from proteinase inhibitor II genes in the tomato family, was put into tobacco to determine levels of its activation by insects under actual field conditions. Upon insect attack on foliage, the transgenic plants showed induction of the transferred proteinase inhibitor genes. This has positive implications for using the wound-inducible inhibitor promoter in biological control of insect-caused foliage damage. The potential exists for a well-managed. efficient system, in which the inserted genes function only on an as-needed basis (83). Suicide Genes When it is important that particular recombinant DNAmodified organisms not establish viable populations, a mechanism that has been proposed for their containment is to include, along with the desired gene, a "suicide" gene that will sufficiently cripple the organism that it will not survive beyond its intended use. The suicide gene may, for example. prompt a metabolic pathway resulting in death of the cell in the presence of a specific external cue (44, 70). Another approach to containment is to introduce mutations that inactivate the transgenic organisms ability to synthesize necessary aromatic amino acids or other key metabolic pathways of the cell ( 19). Alternatively. a kill gene can be inserted to be expressed constitutively all the time-unless a protection gene is turned on by the same promoter gene that causes expression of the key functional gene. That 297-937 0 92 9 QL 3

PAGE 251

250 A New Technological Era for American Agriculture promoter can be geared to respond to some signal from the environment, such as temperature or presence of a pollutant chemical. For instance, if a protection gene for a vaccine strain is only activated above temperatures of 30 C.. the vaccine organism will express the kill gene and die if it passes out of the hosts body ( 19). The advantages of a suicide strategy are straightforward. Existing experimental data indicate that genetically modified microorganisms introduced into the environment usually fail to establish viable populations unless the numbers of introduced organisms are very large. To accomplish a useful effect, as in agricultural treatments or environmental clean-up, planned introductions of microorganisms generally will require inocula of large populations. Once the goal of the planned introduction has been met, a trigger factor to set off the suicide gene can be introduced that will leave behind only a small fraction of the introduced population, which may then be at too low a frequency to sustain itself. Suicide genes are most frequently suggested for containment of microorganisms; their feasibility in plants has been questioned. With plants complicated physiology. difficulties could exist, for instance, in triggering the action of specific genes by any environmental cue other than some deliberate applied chemical, such as a herbicide (37). Overall, the potential effectiveness of suicide genes at this point is controversial (25). One key problem with the use of suicide genes is that natural selection would encourage the evolution of genetically based mechanisms counteracting the suicide effect. Prevention of Gene Transfer In the case of transgenic plants, concerns exist about the possible transfer of engineered genes to neighboring weedy populations of related species. One way to prevent gene transfer through pollen would be to shut down pollen production in the transgenic plant. This can be accomplished by introducing a male-sterility factor into the plant along with the desired trait. The use of naturally occurring male-sterility mutants has been a significant tool in traditional plant breeding. Quite recently, genes for male sterility have been cloned and reintroduced into several plant species, including canola (56). These genes were expressed in the transgenic plants and hence brought about male sterility. This strategy has a great deal of promise and currently is feasible. Its application to canola is especially pertinent because that species is among the most likel y to effect vector gene transfer to related species in North America. For leafy crops (e.g., spinach) or root crops (e. g., sugar beets), male sterility would not be problematic. In fact, it has been suggested that male sterility used in timber tree plantations would channel more of a trees resources to board feet production, in lieu of reproduction. Some crops (e. g., cereals), however, require pollination, so that mixed varietal plantings of male sterile transgenic plants and male fertile. untransformed varieties could be needed (37). Several strategies seem to have potential to decrease the risks associated with gene transfer between microorganisms. For example, a protection gene might be inserted far away from a kill gene. which itself is close to the desired gene being introduced to a host; then, if the functional gene happens to be transferred, the new recipient microorganism also would receive the kill gene, without the protection gene. Another approach might be to insert a gene for a particular active nuclease so that when a cell dies, its DNAincluding the introduced fragment-released after death will have been significantly reduced. A variety of ways of inserting defects that would disrupt the hosts mobilization and conjugation systems could also cut down significantly on horizontal gene transfer (19). Engineering changes into a chromosome rather than a plasmid may decrease the likelihood of gene transfer between microorganisms; this approach also is being explored (45, 48). Combinations of Genes As the number of genes involved in a desired effect goes up. so does the possibility that that effect will be lost in the next and subsequent generations because of natural recombination. Thus, a possible strategy for decreasing the long-term probability of establishment of an engineered genetic effect would be to have the desired effect depend on the interaction among several separate genes. AGRONOMIC METHODS OF MANAGING RISK Physical Barriers Complete containment was the preferred method of controlling risk when genetic engineering was introduced on a small scale. Examples of physical containment are boundary strips in the form of fences or hedgerows that can trap some large percentage of pollen. particularly that dispersed by wind. This might, however, be unfeasible to install or cause unwanted shade (37). Overall.

PAGE 252

Chapter 8Scietific Issues: Risk Assessment and Risk Management l 251 Photo credit: Grant Heilman, Inc. A traditional approach to isolation of plants is to spatially separate desired plants from other plants. Similar guidelines for spatial segregation have been applied to transgenic plants as well. a complete containment strategy has extremely limited applicability beyond small field tests. Once an organism has been placed in the field in the numbers required by agricultural production, it is likely to be exposed to a variety of biotic interactions beyond the control of reasonable physical barriers. Spatial Barriers The traditional approach to isolation of plants genetically improved through conventional breeding, usually for the purpose of generating a seed crop, is to isolate spatially the desired plants from other plants. Similar guidelines for spatial segregation have been applied to transgenic plants as well (64). Certainly this is feasible at the small field trial stage, and could be effective in an experimental setting to evaluate the properties of the organism as a potential pest. Some spatial separation may be feasible at the large-scale test stage, as well. Another approach to separating plants in terms of gene flow is to surround a field with flowers that will attract pollinators of the transgenic crop, so that these trap flowers rather than surrounding wild vegetation would be more likely to receive any transgenic pollen. This approach might conceivably diminish the pollinators activity in pollinating the crop itself, however. Weed control practices using herbicides or cultivation could also decrease the chance of hybridization between the crop and wild species. A straightforward mechanism is to decrease the length of the boundary of the field and thus decrease the number of opportunities for neighbors along the boundaries to exchange genes. Large, square fields minimize these opportunities (37). Temporal Barriers Many problems associated with planned release could be addressed by the timing of the release. For example, if a given engineered line is released in an area with an uncultivated relative that could incorporate the engineered genes, one could manipulate the flowering (phenology) of the engineered organisms so that the crop did not flower at the same time as the weed. For example, wild relatives need short days for flowering, bush type green beans do not (72). Similarly, one could release the introduced plant at a time of year when the weed is dormant or even engineer the crops for cold tolerance, for example, to shift its flowering and production period away from that of its wild relatives. Agricultural experience and ecological understanding will play a significant role in the development of such barriers. Some agronomic practices such as irrigation can allow crop production at a time of year unfavorable for related weeds, diminishing the possibility of cross hybridization. Crop rotation could be used to force a weed rotation. This could decrease the number of weed individuals present in the field and, therefore, the likelihood of gene transfer; it might also eliminate hybrids produced in preceding crop production periods. Crop rotation could prevent genes from being transferred to weeds outside the field for a whole season or two at a time, diminishing the chances that the gene would become established in the weed community and making it more likely to be lost due to genetic drift. The timing of harvesting could also build a barrier to cross hybridization. For some crops, such as cabbage, spinach, collards, lettuce, sugarbeets, carrots, turnips, radishes, celery, garlic, and onions, the crop product is vegetative; careful harvesting would remove the plants before their flowering, reproductive stage, thereby diminishing pollen transfer (37). SUMMARY POINTS Issues and concerns raised by planned introductions of recombinant DNA-modified organisms can be addressed by the integration of risk assessment methodologies with the currently existing knowledge base, continuously augmented by ongoing research and by additional data resulting from field tests. Risk management is therefore possible, with its chief compo-

PAGE 253

252 A New Technological Era for American Agriculture nents being science-based regulation, scientific management methods, and agronomic management methods. A natural evolution of risk management and regulatory oversight is occurring as our experience base with field tests and in performing ecological risk assessments grows. This step-by-step progression in the use of recombinant DNA-modified organisms in the environment, emphasizing science-based risk assessment strikes a balance between a laissez-faire approach and a paralysis of the use of new technology. Biotechnology has the potential to contribute significantly to agriculture; scientifically sound risk assessment and management promote its acceptance as well as its safety. CHAPTER 8 REFERENCES 1. 2. -Z -. 4. 5. 6. 7. 8. 9. Amarger, Noelle and Delgutte, Dominique, Monitoring Genetically Manipulated Rhizobiurn leguminosarum bv Viciae Released in the Field, Biological Monitoring of Genetically Engineered Plants and Microbes, D.R. MacKenzie and Suzanne C. Henry (eds. ) (International Symposium on the Biosafety Results of Field Tests of Genetically y Modified Plants and Microorganisms, Kiawah Island, SC, Nov. 2730, 1990) (Bethesda, MD: Agriculture Research institute, 1991 ), pp. 22 1. Antonovics, Janis and Grant, Michael, Experimental Evidence on the Frequency of Neutral Mutations, J. of Heredity 65: 241-242, 1974. Bakker, P. et al., Biological Monitoring of Genetically) Engineered Plants and Microbes, D. R. MacKenzie and S.C. Henry (eds. ) (International Symposium on the Biosafety Results of Field Tests of Genetically Modified Plants and Microorganisms, Kiawah Island, SC, Nov. 27, 1990) (Bethesda, MD: Agriculture Research Institute, 1991 ), pp. 201204. Barrett, S. C. H., Crop Mimicry in Weeds, Economic Botan?~ 37: 255, 1983. Berry, Duane F. and Hagedom, Charles, Soil and Groundwater Transport of Microorganisms, Assessing Ecological Risks of Biotechnolog>l, Lev R. Ginzburg (cd. ) (Boston, MA: Butterworth-Heinemann, 1991), pp. 57. Boyce Thomson Institute, Regulator}t Considerations: Genetically Engineered Plants, Center for Science Information, San Francisco, CA, 1988. Brill, W. J., Safety Concerns and Genetic Engineering in Agriculture, Science 227;381, 1985. Brown, A.H. D. et al., The Use of Plant Genetic Resources (New York, NY: Cambridge University Press, 1989). Caswell, H., Matrix Population Models, Sinauer Associates Inc., Sunderland, MA, 1989. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21, 22. Colwell, R.K. et al. Letters, Genetic Engineering in Agriculture, Science 229: 11 1 12, 1985. Colwell, R. K., Another Reading of the NAS Gene Report, Bioscience 38:421, 1988. Colwell, R. K., Natural and Unnatural History: Biological Diversity and Genetic Engineering, Scientists and their Responsibility, W.R. Shea and B. Sitter (eds. ) (Canton, MA: Watson International Publishing, 1989). Covello, V. T., and Fiksel, J.R. (eds. ), The Suitability and Applicability of Risk Assessment Methods for Environmental Applications of Biotechnology, Final report to the Office of Science and Technology Policy, Executive Office of the President, Report No. NSF/PRA 8502286, National Science Foundation, Washington, DC, 1985. Crawley, Michael J., The Ecology of Genetically Engineered Organisms: Assessing the Environmental Risks, Introduction of Genetically Modfied Organisms into the Environment, Harold A. Mooney and Giorgio Bemardi (eds. ) (New York, NY: John Wiley and Sons, 1990) pp. 133-150. Davis, B., Bacterial Domestication: Underlying Assumptions, Science 235: 1329, 1987. Dean-Ross, D., Applicability of Chemical Risk Assessment Methodologies to Risk Assessment for Genetically Engineered Microorganisms, Rec. DNA Tech. Bull. 9( 1): 16-28, 1986. Doebley, J. Molecular Evidence for Gene Flow Among Zeq Species, Bioscience 40: 443-448, 1990. Duvick, Donald N., A Case Study: Development and Release of a Corn Hybrid for the U.S. Corn Belt, talk presented at B1OTECH USA, Nov. 27 29, 1990, Washington, DC, 1990. Dwyer, Daryl F. and Timmis, Kenneth, Engineering Microbes for Function and Safety in the Environment, pp. 79, Introduction of Genetically Modified Organisms into the Environment, Harold A. Mooney and Giorgio Bemardi (eds. ) (New York, NY: John Wiley and Sons, 1990), pp. 167189. Ehler, L. E., Planned Introductions in Biological Control, Assessing Ecological Risks of Biotech nologj~, Lev R. Ginzburg (cd. ) (Boston, MA: Butterworth-Heinemann, 1991), pp. 21. Ellstrand, N.C. and Hoffman, C. A., Hybridization as an Avenue of Escape for Engineered Genes, Bioscience 40:438-442, 1990, p. 40. Faust, Robert M, and Jayaraman, Kunthala, Current Trends in the Evaluation of the Impact of Deliberate Release of Microorganisms in the Environment: A Case Study with a Bioinsecticidal Bacterium, [introduction of Genetically Modified Organisms into the Environrnen?, Harold A. Mooney and Giorgio Bemardi (eds. ) (New York, NY: John Wiley and Sons, 1990), pp. 167-189.

PAGE 254

Chupter 8Scientljlc Issues: Risk Assessment and Risk Management s 253 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. Fiksel, Joseph and Covello, Vincent T., TheSuitability and Applicability of Risk Assessment Methods for Environmental Applications o f Biotechnology, Biotechnology? Risk Assessment: Issues and Methods fijr E\l\ir{~\~t?li~tlt~~l introductions, Joseph Fiksel and Vincent T. Covello (eds. ) (Final Report to the Office of Science and Technology Policy, Executive Office of the President, Report No. NSF/PRA 8502286) (New York, NY: Pergamon Press. 1986), pp. 1. Fiksel, Joseph, Applications of Knowledge Systems for Biotechnology Risk Assessment and Management, Risk Assessment in Genetic Engineering, Morris A. Levin and Harlee S. Strauss (eds. ) (New York, NY: McGraw-Hill, Inc., 199 l), pp. 354-367. Fox. J.L. Suicide as b Painless Eco-Safeguard! BiolTech 7:412-413, 1989. Frederick, Robert J. and Pilsucki, Robert W., Nontarget Species Testing of Microbial Products Intended for Use in the Environment, Risk Assessment in Genetic Engineering. Morris A. Levin and Harlee S. Strauss (eds. ) (New York, NY: McGraw-Hill, Inc., 1991 ), pp. 32-50. Fristrom, J.W. and Clegg, M. T., Principles of Genetics, 2d ed. (New York, NY: W.H. Freeman, 1988). Fuchs. Roy L. and Serdy, Frank, S., Genetically Modified Plants: Evaluation of Field Test Biosafety Data, Biological Monitoring of Genetically? Engineered Plants and Microbes, D. R. MacKenzie and Suzanne C. Henry (eds. ) (International Symposium on the Biosafet y Results of Field Tests of Genetically Modified Plants and Microorganisms, Kiawah island, SC, Nov. 27-30, 1990) (Bethesda. MD: Agriculture Research Institute, 199 l), pp. 25. Fuxa, James R., Release and Transport of Entomopathogenic Microorganisms, Risk Assessnwn? in Genetic Engineering. Morris A. Levin and Hat-lee S. Strauss (eds. ) (New York, NY: McGraw-Hill, Inc., 1990), pp. 83 13. Gillett, J. W., Risk Assessment Methodologies for Biotechnology Impact Assessment, Potential lnzpacts of Environmental Release of Biotechnology Products: Assessment, Regulation, und Research Needs, J.W. Gillett et al. (eds. ), Report no. ERC075, Ecosystems Research Center, Cornell University, Ithaca, NY, 1985. Ginzburg, L.R. (cd.), Assessing Ecological Risks of Biotechnology (Boston, MA: Butterworth-Heinemann, 1991). Goldburg, Rebecca et al., Biotechnolog~ls Bitter Harvest: Herbicide-Tolerant Crops and the Threat to Sustainable Agriculture, Biotechnology Working Group Report, 1990. Gould, F., Genetics of Plant-Herbivore Systems: Interactions Between Applied and Basic Study, Variable Plant and Herbivores in Natural and Man34. 35. 36. 37. 38. 39. 40. 41. 42. 43. aged S}stems, R. Denno and B. McClure (eds. ) (New York, NY: Academic Press. 1983), pp. 599-653. Gould, F., Evolution of Resistance to Toxic Compounds by Arthropods, Weeds and Pathogens, commissioned background paper prepared for the Office of Technology Assessment, 1991. Handel, S. N., Pollination Ecology, Plant Population Structure, and Gene Flow, Pollination Biology, L. Real (cd. ) (New York, NY: Academic Press, 1983), pp. 163-211. Hardy, Ralph W .F., Large-Scale Field Testing and Commercialization: Thoughts on Issues, Biological Monitoring qf Genetically} Engineere, Lev R. Ginzburg (cd. ) (Boston, MA: Butterworth-Heinemann, 1991 ), pp. 123-149. Kareiva, Peter, Manasse, Robin, and Morris. William, Using Models to Integrate Data from Field Trials and Estimate Risks of Gene Escape and Gene Spread, Biological Monitoring of Geneticwllj Engineered Plant.s andMicrobes, D.R. MacKenzie and Suzanne C. Henry (eds. ) (International Symposium on the Biosafety Results of Field Tests of Genetically Modified Plants and Microorganisms, Kiawah Island, SC, November 27, 1990) (Bethesda, MD: Agriculture Research Institute, 1991), pp. 31. Katz, Lori S. and Marquis, Judith K., The Toxicology of Genetically Engineered M icroorgan isms Risk Assessment in Genetic Engineering, Morris A. Levin and Harlee S. Strauss (eds. ) (New York, NY: McGraw-Hill, Inc., 1991), pp. 51-63. Kearney, Philip C. and Tiedje, James M., Methods Used to Track Introduced Genetically Engineered Organisms, Biotechnology> for Crop Protection, Paul Hedin, Julius Menn, and Robert Hollingworth (eds. ) (Washington, DC: American Chemical Society, 1988), pp. 352. Keeler, Kathleen H., Can Genetically Engineered Crops Become Weeds? Biotechnology vol. 7: 1134 1139, 1989. Keeler, Kathleen H., Turner, Charles E., Management of Transgenic Plants in the Environment, Risk Assessment in Genetic Engineering, Morris A.

PAGE 255

254 l ANewTechrrological Erafor American Agriculture Levin and Harlee S. Strauss (eds. ) (New York, NY: McGraw-Hill, inc., 1991), pp. 189-218. 44. Kim, Hunhyong, Ginsburg, Lev R.,andDykhuizen, Daniel E., Quantifying( heRisks of Invasion by Genetically Engineered Organisms, Assessing EcologicalRisks ofBiotechnology, Lev R. Ginzburg (ed.)(Boston, MA: Butterworth-Heinemann, 1991), pp. 193-214. 45. Kluepfel, D.A. etal., FieldTesting ofa Genetically Engineered Rhizosphere Inhabiting Pseudomonas: Development of a Model System, Biological Monitoring of Genetically Engineered Planl.s and Microbes, D.R. MacKenzie and Suzanne C. Henry (eds. ) (International Symposium on the Biosafety Results of Field Tests of Genetically Modified Plants and Microorganisms, Kiawah Island, SC, Nov. 27 30, 1990) (Bethesda, MD: Agriculture Research Institute, 1991 ), pp. 189 199. 46. Kluepfel, D.A. and Tonkyn, D. W., Release of Soil-Borne Genetically Modified Bacteria: Biosafety Implications from Contained Experiments, Biological Monitoring of Genetically Engineered Plants and Microbes, D.R. MacKenzie and Suzanne C. Henry (eds. ) (International Symposium on the Biosafety Results of Field Tests of Genetically Modified Plants and Microorganisms, Kiawah Island, SC, Nov. 27, 1990) (Bethesda, MD: Agriculture Research Institute, 1991), pp. 55. 47. Kostka, Stanley J., The Design and Execution of Successive Field Releases of Genetically Engineered Microorganisms, Biological Monitoring of Geneticullv Engineered Plants and Microbes, D. R. MacKenzie and Suzanne C. Henry (eds. ) (International Symposium on the Biosafety Results of Field Tests of Genetically Modified Plants and Microorganisms, Kiawah Island, SC, Nov. 27, 1990) (Bethesda, MD: Agriculture Research Institute. 1991), pp. 167-176. 48. Lacy, George H., Stromberg. Verlyn K.. Pre-Release Microcosm Tests with a Genetically Engineered Plant Pathogen, Biological Monitoring of Genetically Engineered Plants and Microbes, D.R. MacKenzie and Suzanne C. Henry (eds. ) (international Symposium on the Biosafety Results of Field Tests of Genetically Modified Plants and Microorganisms, Kiawah Island, SC, Nov. 27-30, 1990) (Bethesda, MD: Agriculture Research Institute, 1991 ), pp. 81-98. 49. Langevin, S. A., Clay, K., and Grace, J. B.. The Incidence and Effects of Hybridization Between Cultivated Rice and Its Related Weed Red Rice (Oryza sativa L.), Evolution 44(4): 1000-1008, 1990. 50. Levin, Morris und Strauss, Harlee, Introduction: Overview of Risk Assessment and Regulation of Environmental Biotechnology, Risk Assessment in Genetic Engineering, Morris A. Levin and Harlee S. Strauss (eds. ) (New York, NY: McGraw-Hill, Inc., 1991), pp. 1-17. 51. Levin, Simon, Ecological Issues Related to the Release of Genetically Modified Organism into the Environment, Introduction of Genetically Modified Organisms into the Environment, Harold A. Mooney and Giorgio Bernardi (eds. ) (New York, NY: John Wiley and Sons, 1990), pp. 151-159. 52. Levy, Stuart B. and Miller, Robert V. (eds. ) Gene Transfer in the Environment (New York, NY: McGraw-Hill Publishing Co.), 1989. 53. MacKenzie, D.R. and Henry, Suzanne C., Towards a Consensus, Biological Monitoring of Geneticailj Engineered Plants and Microbes, D.R. MacKenzie and Suzanne C. Henry (eds. ) (international Symposium on the Biosafety Results of Field Tests of Genetically y Modified Plants and Microorganisms. Kiawah Island, SC, Nov. 27-30, 1990) (Bethesda, MD: Agriculture Research institute, 1991 ), pp. 273. 54. Manasse, Robin and Kareiva, Peter, Quantifying the Spread of Recombinant Genes and Organisms, Asses-sing Ecological Risks of Biote~hnolog>, Lev R. Ginzburg (cd. ) (Boston, MA: Butterworth-Heinemann, 1991), pp. 215. 55. Maniatis, T., Fritsch, E. F., and Sambrook, J., Molecular Cloning: A Laborator> Manual (Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press, 1982). 56. Mariani, C. et al., Induction of Male Sterility in Plants by a Chimaeric Ribonuclease Gene, Nature 347:737-741, 1990. 57. McIntosh, Marla S., Statistical Techniques for Field Testing of Genetically Engineered Microorganisms Risk Assessment in Genetic Engineering, Morris A. Levin and Hat-lee S. Strauss (eds. ) (New York, NY: McGraw-Hill, Inc., 1991 ), pp. 219-239. 58. Meagher, T. R., Analysis of Paternity Within a Natural Population of Chamae/irium luteum, 1. identification of Most-Likely Male Parents, Amer. Nat 128: 199-215. 1986. 59. Merkhofer, M. and Covello, V., Risk Assessment and Risk Assessment Methods: The State of the Art, NSF study, Washington, DC, as cited in Fiksel and Covello, 1986. 60. Miller. Robert V. and Levy, Stuart B., Horizontal Gene Transfer in Relation to Environmental Release of Genetically Engineered Microorganisms, Gene Transf&r in the En\ironment, Stuart B. Levy and Robert V. Miller (eds. ) (New York, NY: McGrawHill Publishing Co. 1989), pp. 405-420. 61. Moffat, Anne, Workshop Discusses Biosafety Aspects of Oilseed Crucifer Field Trials, Genetic Engineering News: January 1991: p. 15. 62. Mooney, Harold A. and Bernardi, Giorgio (eds. ), Introduction ofGenetically Modljied Organisms into

PAGE 256

Chapter 8-S~.ientific Issues: Risk Assessment and Risk Management l 255 63. 64. 65. 66. 67. 68. W. 70. 71. 72. 73. the Environment (New York. NY: John Wiley und Sons, 1990). Mooney, Harold A. and Drake, J. A., The Release of Genetically Designed Organisms in the Environment: Lessons from the Study of the Ecology of Biological Invasion s. Introduttion of Genetically> A40dlfied Org[itlisms into the L)t\*i)~)tl\?~t~tlt, Harold A. Mooney and Giorgio Bernardi (eds. ) (New York, NY: John Wiley and Sons, 1990). pp. I 17 129. Muench, Stephen R., Rcquircmcnts tind Considerations in Successful Field Releases of Gcnetictilly Engineered Plants, Bi(dogit[11 Monitoring (!f Ge ncticall~~ Engineered Pl[int.$ an(i Mi~robes, R. MacKenzie and S.C. Henry (eds. ) (International Symposium on the Biosatety Results of Field Tests of Gcneticul Iy Modified Plants tind M icroorgtinisn~.. Klawah Island, SC, N(JY ~7, 1990) ( Bcthe\dti< MD: Agriculture Research Institute, 199 l), pp. 37. Mullis, Kary B., The Unusual origin of the Polymcrase Chain Rctiction, .?ticntlfi~ Anwritwn. April 1990: 56-65. Naimon, Jonuthun S.. Using Expert Panel~ to Assess Risks of Envirtmmental B iotechnolog~ Applications: A Ctisc Stud) of the 1986 Fr(~stban~ Risk Assessments, Risk A.~.~t.~.~I?It~IIt ill Genetit En,qi nccrin,y. Morris A. Levin and Har]ec S. Strauss (cd\. ) (Ncw York. NY: McGraw-hill. Inc., 1991 ), pp. 31 9. National Academy of Scicncm (NAS ), National Research Counc i]. Ri\k A.~.v[.s.s\ttt~\tt ill the Federal G-lJ\crntnent. M~~\l~Igin A40dlficd Or,qanit~i?15~.. FI-{lI~I(~IIq{jI.k ,fi)r De~.i.si~~n.s ( Wush ingt(~n, DC: Nutit)nal Acudemy Press, 1989). Olson. B.H. et al., The Implicatit)ns of Horizontal Gene Transfer for the Environmental Impact of Genctical Iy Engineered Microorganisms, Risk A.\.ws.ittlcnt ill Genctic Etlcgi\lcerin,g. Morris A. Levi n tind Harlce S. Strauss (cds. ) ( Nev York, NY: McGraw-Hill. Inc., 199 1). pp. 163-188. Parker, M., The Pteiotropj Thtory for Pol~mmrphism of Disease Rc\istant Genes in Plmts.. ~)(~lutioll 44: 1872-74, 1990. Purse glove. J. W., Tr~~pi~~il (rl)p~, [licot>lt~lo)l~ (London. Englimd: Longmtin. 1974). Rces, Mark et al., An Ecolt)gictil Pm_spccti\re to Risk Assessment, Bi{dc)gittll M{)iti[(willg of (heticwlly hlcqintwed Pll[nt.$ LInd kli~-r~h),$, D.R. k9acKenzie and Suzanne C. Ha-g (cA. ) ( Intematkwil Symposium on the Bkwifkt} Re\ults t~f Field Tests ~)t Gc 74. 75. 76. 77. 78. 79. 80. 81. g~ 83. 84. 85. netically Modified Plants and Microorganisms, Kiawah Island. SC, Nov. 27-30, 1990) (Bethesda, MD: Agriculture Research Institute, 1991, pp. 9 19. SCOPE/COGENE, Joint Statement, Mooney and Bernard, 1990, pp. xviixxiv. Sharples, Frances E., Regulation of Products from Biotechnology, Science 235: 1329, 1987. Sharples, Frances E., Ecological Aspects of Hazard Identification for Environmental Uses of ~Jeneticaliy Engineered Organisms, Risk A,\.\(~.\-.\t?I(~tl/ i)? Genetic Engineering, Morris A. Levin and Harlee S. Strauss (eds. ) (New York, NY: McGraw-Hill, Inc., 1991 ), pp. 18-31. Sinlnl{~nds. N. W., E\dution of Crop P[allt.v (New York. NY: Longman, Inc., 1976). Smith, 1;. B. md Adams. W.I., measuring Pollen Contamination in Clonal Seed Orchards With the Aid of Genetic Markers, Pm. 17th South. Forest Tree Imprm. Conf. Athens, GA. 1983. pp. 69-77. Steffan, R.J. and Atlas, R. M., DNA Amplification to Enhunce Detection ot Genetically Engineered Bacteria in Environmental Samples.. App/icd ~Ind Etl\i\-(~\lt}t[~~lt~~l kli~robiolog-y 5-I: 2185-2191, 1988. Stotzky. G., Gene Transfer Among Bacteria in Soil, Gene Transfir in the Ell~i~oilltlt~\lt, Stuart B. I,evy and Robert V. Miller (eds. ) (New Yt~rk. NY: ,McGraw-Hill Publishing Co., 1989). pp. 165. Strauss, Harlee S., Lessons from Chcmical Risk ,4ssessment, Risk Assessment i