The Future of Agricultural Biotechnology
The application of biological sciences in agriculture has become increasingly prominent in the past decade. Genes were first inserted into corn using molecular techniques in 1989, and by the late 1990s farmers were growing millions of acres of transgenic corn. Clearly, the science of biotechnology for agriculture is in its infancy, yet it shows an influence beyond its years.
The previous chapters review what is known about the environmental impact of commercialized transgenic crops and approaches for monitoring that might be adapted to screen for their unanticipated effects. One key finding is that particular phenotypic characteristics of a given transgenic plant determine its likely environmental interactions; the fact that recombinant DNA methods were used in its development only indirectly affects these interactions by influencing the phenotypic characteristics of the transgenic plant. Indeed, the significance of biotechnology for environmental risk resides primarily in the fact that a much broader array of phenotypic traits can be incorporated into crop plants than was possible about a decade ago. As such, our experience with the few herbicide-tolerant and insect- and disease-resistant varieties that have been commercialized to date provides a very limited basis for predicting questions needed to be asked when future plants with very different phenotypic traits are assessed for environmental risks.
This chapter is divided into three major sections. The first includes an overview discussion of some new kinds of transgenic crops and a selective discussion of some environmental risk issues that may be associated
with the next generation of transgenic crops. The second section focuses on policy, beginning with a general discussion of the context in which environmental risk from transgenic crops should be framed and then moving on to specific topics that may arise as policies for the next generation of transgenic crops evolve. The final section is a discussion of some research needs to address future issues.
THE NEXT TRANSGENIC CROPS
This section describes anticipated future transgenic crops, including some expected to be commercialized in the next couple of years, others that may reach commercial status on a midterm horizon sometime during the next decade, and others that are mere twinkles of ideas for transgenic crops that will require research breakthroughs before they can reach fruition. As discussed in Chapter 2, it is not possible to characterize the environmental hazards that may be associated with all such crops in advance of knowledge about their phenotypic characteristics and the agricultural ecology of the settings in which they will be grown. However, the second part of this section offers a preliminary discussion of some representative environmental risk issues that may be associated with these new transgenic crops.
An Inventory of New Transgenic Crops
The first commercially produced transgenic crops were based on single-gene traits. Among these was the “Flavr-Savr” tomato, which used gene silencing to inhibit the expression of an enzyme involved in fruit ripening (Kramer and Redenbaugh 1994). The Flavr-Savr tomato was not a commercial success, but the technology was effective because the fruit not only had a slow rate of ripening but also was less susceptible to pathogen infection. Other early transgenic products were based on traits influencing agronomic performance (i.e., pathogen, insect, and herbicide resistance). The rapid and broad use by the American farmer of glyphosate-resistant soybeans and Bt-expressing cotton and corn attests to the commercial success of these transgenic crops (James 1998, USDA-NASS 2001).
Based on the successes of these initial transgenic crops, research laboratories throughout the world are now studying a wide variety of traits/ genes that could greatly expand the spectrum of products from such plants. As was true of the first genetically engineered crops, the rate at which new transgenic traits can be expected to appear in the future depends largely on the number of genes encoding them. So traits controlled by single genes, or traits that can be reduced or eliminated by the loss of
expression through gene silencing of a single gene or group of related genes, are likely to become the first available. The next logical step for expansion is the integration of single-gene traits. Crops with a set of single genes for a number of distinct traits already are in use. We can only speculate about the number of genes controlling the development and architecture of plants and the various physiological processes impacting yield. It seems safe to assume, though, that genetically complex traits will require additional years of research to understand, let alone express and regulate in a genetically engineered crop species. Nevertheless, complex traits, including those controlling adaptation to abiotic stresses, such as drought and salinity, flowering and reproduction, and hybrid vigor, are being actively investigated, and it would not be surprising if some of these could be regulated in crop plants by genetic engineering within the next 5 to 10 years.
These products have the potential to not only improve agronomic performance but also increase the nutritional value of grains consumed by humans and livestock, eliminate allergens and antinutritional factors, improve the shelf life of fruits and vegetables, and increase the concentration of vitamins and micronutrients found in seeds, creating healthier foods (Abelson and Hines 1999). While there are many different kinds of transgenic crops under development, most aim to address one of four broad social needs: improved agricultural characteristics, greater adaptation to postharvest processing practices, improved food quality and other uses of value for humans, and better mitigation of environmental pollution. Indeed, so many traits currently under investigation could become incorporated into transgenic plants that space limits consideration here to only a few.
Improved Agricultural Characteristics
Among the transgenic traits near commercial release are new Bt genes that provide protection against additional types of insect pests. One of these is a gene that protects corn against corn rootworm damage (Kishore and Shewmaker 1999). It is estimated that damage to corn roots by this pest result in losses approaching $1 billion annually. By reducing the impact of this pest, it is expected that not only will there be better corn yields but also better drought tolerance and fertilizer utilization due to the healthier root system. Research is also being done to genetically engineer tree crops to make them resistant to insects and herbicides and to increase their rate of growth. For example, a Bt gene has been inserted into hybrid poplars to protect them against defoliation by a leaf beetle. Acreage of hybrid poplars has increased because of their good wood pulp characteristics, but they have been susceptible to insect attack, which has
prompted applications of insecticides. Many of these new traits for improving agricultural production on farm are ones that could have environmental impacts that are similar in kind to the present generation of transgenic crops. Their value accrues directly to the farmer and the seed company and only indirectly to other sectors of society. Their potential risks, however, are borne by a wider segment of society. Thus, risk analysis of this next generation of traits is likely to resemble present discussions and debates about biotechnology. The evaluation of these risks is likely to become more complicated and difficult as the range of transgenic crops expands from the major grain crops to the more wild and perennial plants, such as pines and poplar.
Over the long term, new knowledge regarding the physiology and development of plants and their interaction with microorganisms could eventually provide the foundation to modify plant structure and reproduction. It may become possible to genetically engineer crop plants that are more tolerant to drought, salinity, and other abiotic stresses (see below); that are able to grow more efficiently in the acidic, aluminum-containing soils found in tropical areas (Herrera-Estrella 1999); that can compete more effectively with weeds; that can reproduce in a shorter time; and that can potentially fix their own nitrogen.
Improved Postharvest Processing
Transgenic technology is also being applied to several commercially important tree species, including poplar, eucalyptus, aspen, sweet gum, white spruce, walnut, and apple (Kais 2001). The global demand for wood and wood products is growing along with the human population. To reduce pressure on existing forests, forest plantations that grow transgenic trees are expected to play an increasingly important role in meeting the demand for tree products (Tzfira et al. 1998). As mentioned above, the first traits being genetically engineered into trees are herbicide tolerance and insect resistance, which are useful for establishing and maintaining young trees. Several traits are under development to better adapt trees to postharvest processing, and these may become commercially available in the near future. For example, there is research under way to modify the lignin content of certain tree species, in order to improve pulping, the process by which wood fibers are separated to make paper. Reduced lignin may improve the efficiency of paper production and may reduce environmental pollution from the paper production process.
To restrict the transfer of transgenic traits to wild forest and orchard tree populations, it is generally considered essential to simultaneously genetically engineer reproductive sterility. Methods currently exist to do this in crop plants (Mariani et al. 1992, Williams 1995), so this technology
is available. Besides restricting gene flow, sterility is expected to cause the trees to grow faster and produce more wood, since energy would not be wasted producing flowers or fruit. It is likely that this issue will be investigated by a future National Research Council (NRC) committee.
There is also interest in using genetic engineering technology to turn annual crop plants into factories that produce valuable chemicals (for a recent review, see Somerville and Bonetta 2001) and antibodies (Daniell et al. 2001). Plants have the capacity to synthesize a variety of complex molecules, given the simple inputs of a few minerals, carbon dioxide, water, and sunshine. It is widely thought that plants could provide a “green” renewable source of chemicals to replace those currently obtained from petroleum. This could also be a mechanism to create new markets for plant products as well as utilize excess production of agricultural commodities. The feasibility of producing a plastic precursor, polyhydroxybutyrate, in plants, was demonstrated several years ago (Poirier et al. 1992), but this was not found to be an economically viable process. Nevertheless, there is excellent potential for mass producing a variety of fatty acids in plants that serve as precursors for valuable polymers, such as nylon. The properties of plastics that incorporate starch could also be significantly improved as more knowledge is learned about the biochemistry of starch synthesis.
Research is also directed toward reducing pollution in postharvest production. For example, there is ongoing research to reduce the content of phytic acid in corn. Phytic acid stores phosphorus in the developing seed (Raboy 1997). Much of this mineral complex is not digested by livestock, so it ends up in the waste stream. Ultimately, it is released into ponds and lakes, where the phosphorus can create algal blooms (Tilman 1999b). Several molecular approaches, including genetic engineering, are being applied to reduce the phytate content of corn.
Some of these new plant products raise several risk issues that the first generation of products did not. If any of these applications displace crops for food and feed production, what are the marginal and aggregate effects on national and global food and feed supplies? If some applications lead to a more vertically integrated farm-to-product production system, what are the environmental impacts? These and other research questions may be important in the near future.
Improved Food Quality and Novel Products for Human Use
Corn and soybeans are two of the most important food and feed commodities in the United States and worldwide. Most (65 to 70%) of the 9 billion to 10 billion bushels of corn produced annually in this country are used for livestock feed; about 25% is exported, and the remaining 10%
is processed into food ingredients, nonfood coatings and adhesives, and ethanol (Corn Refiners Association 1999). In addition, approximately 20% of the dietary calories (of people) in the United States come from lipids obtained from plant seeds, with soybean oil accounting for about one-third of the total (www.soygrowers.wegov2.com/file_depot/0-10000000/0-10000/735/folder/4944/2000SoyStats.pdf). By altering the lipid, protein, and carbohydrate composition of these seeds, it may be possible to create more nutritious food and obtain byproducts with improved functional characteristics. Several of the transgenic products discussed below are in field trials or will soon be available for production.
Corn seed has a high caloric density because of its high starch and oil content, but the protein it contains is deficient in several amino acids (lysine, methionine, tryptophan) essential for swine, poultry, and human nutrition. Transgenic corn lines that contain higher than normal levels of these amino acids and/or that produce proteins with higher contents of these acids have been created, although they are not yet in commercial production. Varieties of high-oil corn have been developed through nontransgenic technology, but transgenic technology is also being used to increase the quantity and quality of corn oil. As is also true of soybean oil (see below), the stability and nutritional value of corn oil could be improved by increasing the proportion of monounsaturated fatty acid, and there are efforts under way to do so by genetic engineering.
Natural soybean oil contains a significant proportion of di- and tri-unsaturated fatty acids (linoleic and linolenic), and although these unsaturated fatty acids are generally considered healthier to eat than the saturated fatty acids found mainly in animal fat, they have a tendency to oxidize and become rancid. These unsaturated fatty acids are also liquid at room temperature, which limits their functional properties for making certain types of foods, such as margarine. The stability of soybean oil and its functional properties are improved by hydrogenating the oil. This reduces the double bonds in the unsaturated fatty acids, yielding monounsaturated trans-fatty acids. Although trans-unsaturated fatty acids have been consumed for many years, there is increasing evidence that they are unhealthy (Taubes 2001). To address this problem, soybean was genetically engineered to produce an oil that contains predominantly a cismonounsaturated fatty acid (oleic acid; Mazur et al. 1999). This was achieved through genetic engineering by silencing the genes that produce linoleic and linolenic acid from oleic acid by a desaturation reaction. The new product is soybean oil with approximately 85% monounsaturated fatty acid, which has good stability, and reduced off-flavor and is healthier to consume.
The meal recovered after soybean oil has been extracted is rich in proteins that have excellent functional characteristics for creating a vari-
ety of foods. However, several of the most abundant soy proteins are deficient in sulfur-containing amino acids (methionine and cysteine), which are essential amino acids for humans and certain livestock. Other soy proteins are antinutritional factors, and one is a common food allergen (Ogawa et al. 1993). Thus, transgenic research has been done to remove the anti-nutritional factors and allergenic proteins from soybeans and to improve its protein quality (Mazur et al. 1999).
The deficiency in the sulfur-containing essential amino acids in soy protein cannot be addressed through conventional breeding because better genes simply do not exist in the natural gene pool. However, the problem can be approached through genetic engineering by altering the activity of the enzymes that synthesize methionine and cysteine or by overproducing a protein that contains them. A complementary approach is to block expression of the major soy proteins that lack methionine and cysteine, thereby increasing the percentage of these amino acids in the remaining proteins. Both of these strategies are being explored. The feasibility of producing a methionine-rich protein in transgenic soybeans and other pulse seeds has been demonstrated (Altenbach et al. 1989, Nordlee et al. 1996), but the trait has not been commercialized. Gene silencing has been shown to be an effective way to eliminate the major soybean allergen (Jung, 2001, personal communication), and other anti-nutritional proteins have been removed by mutagenesis (Mazur et al. 1999). The promoter used to silence the fatty acid desaturase genes in the high oleic acid transgeneic soybean effectively eliminates expression of one of the major classes of soy proteins that does not contain methionine and cysteine. Consequently, the transgenic seed that was modified for high oleic acid content also has an improved protein quality.
In addition to the macronutrients, starch, protein, and oil, plant foods provide many of the micronutrients essential in human diets. There are 17 minerals and 13 vitamins required at minimum levels to prevent nutritional disorders (DellaPenna 1999), and all of these have attracted biotechnology research. Clinical and epidemiological studies show an important role in health maintenance for several minerals (iron, calcium, selenium, and iodine) and vitamins (A, B6, E, and folate), but these are typically not present in sufficient quantities in many diets throughout the world. Common reliance on rice, wheat, maize, and soybean for macronutrients limits the diets of many people to the micronutrients these seeds contain. In particular, these foods are deficient in iron, zinc, selenium, copper, riboflavin, and vitamins A and C.
More than two billion people face serious dietary problems due to inadequate quantities of micronutrients. For example, iron deficiency leads to anemia in 40% of all women and 50% of pregnant women and is thought to cause up to 40% of the half-million deaths at childbirth each
year (Welch et al. 1997). Inadequate iron in children’s diets impairs mental development. Vitamin A deficiency is considered a global epidemic (Ye et al. 2000). Annually, 250 million children suffer from vitamin A deficiency, which contributes to illness and death for some 10 million people annually. Vitamin A deficiency causes blindness in up to half a million children each year, half of whom die after losing their sight. Besides impairing vision, vitamin A deficiency can lead to protein malnutrition and poor immune system function. Folic acid deficiency increases the risk of birth defects, heart disease, and stroke.
Much remains to be learned about the uptake and accumulation of minerals and the synthesis of vitamins in plants, but significant progress is being made in some areas of research, such that transgenic plants producing increased levels of several micronutrients have been created. For example, mulled rice grains contain no beta-carotene, but it has been genetically engineered by the introduction of three genes, one from daffodil and two from bacteria, to produce significant levels of beta-carotene, which is made into vitamin A (Ye et al. 2000). Additional genetic engineering may be necessary to raise the beta-carotene level in transgenic rice such that a daily serving provides the recommended daily allowance (RDA) of vitamin A. The efficacy of this approach to reducing vitamin A deficiencies remains controversial (Nestle 2001).
Similar genetic approaches were used to increase the level of tocopherol, the lipid-soluble antioxidant known as vitamin E, in plant oils. The RDA for vitamin E is 10 to 13.4 international units (equal to about 7 to 9 mg of a-tocopherol), which is generally accessible through consumption of plant-derived dietary components, including soybean oil. However, an excess intake of vitamin E (100 to 1,000 IU/day) has been found to be associated with a reduced risk of cardiovascular disease and some cancers, improved immune function, and reduced progression of several human degenerative conditions (Traber and Sies 1996). Thus, there could be health benefits from increasing the level of vitamin E in commonly consumed foods, such as soybean (DellaPenna 1999). By overexpressing the gene responsible for the last step in vitamin E synthesis in the model plant Arabidopsis thaliana, the effective vitamin E level was increased nearly 10-fold (Shintani and DellaPenna 1998). A similar approach is now being used to create transgenic soybean and canola plants with enhanced levels of vitamin E.
Plants contribute a number of other health-promoting chemicals to our diet (DellaPenna 1999). The glucosinolates found in broccoli and related cruciferous vegetables are thought to help detoxify cancer-inducing carcinogens (Talalay and Zhang 1996). The isoflavones found in soybeans are phytoestrogens, and they appear to reduce the incidence of breast, prostate, and colon cancers; osteoporosis; and cardiovascular disease
(Kurzer and Xu 1997). Carotenoids, the red, orange, and yellow pigments in tomatoes and green leafy vegetables, appear to reduce the risk of certain cancers, cardiovascular disease, and blindness caused by macular degeneration (Charleux 1996). The genetic and metabolic factors influencing levels of glucosinolates, isoflavones, and carotenoids from various plant sources are poorly understood, as is the appropriate level at which they should be consumed for maximum health benefits. Western diets typically contain limited amounts of these chemicals, and there is growing interest in finding ways to increase their content in food and to determine their appropriate health-promoting levels.
Plant foods could also be used as edible vaccines (Langridge 2000, Walmsley and Arntzen 2000). Many seeds contain proteins that are allergenic in certain people. When these allergenic proteins are digested, small fragments derived from them are absorbed into patches of cells on the small intestine that are part of the immune system. Antibodies are produced against these proteins, and this leads to an immune response in the individual, with potentially severe consequences after subsequent exposure to the allergenic proteins. This same process can be used to create immunity against common viral and bacterial pathogens by producing antigenic proteins derived from them in edible plant parts. Of particular interest are a group of pathogens—Norwalk virus, Vibrio cholerae (the cause of cholera), and enterotoxigenic Escherichia coli (a source of “traveler’s diarrhea”)—that cause the deaths of several million children each year, mainly in developing countries. Preliminary studies indicate that uncooked plant foods, such as potatoes or bananas, can be used to produce pathogen-derived proteins (such as virus coat proteins). These foods might then be used to inoculate children and adults against a variety of common diseases. Although a great deal of research remains to be done to demonstrate the efficacy and economic viability of this approach, results from preliminary experiments are promising.
There is also interest in using plants to produce human monoclonal antibodies. Preliminary research has demonstrated that several types of plant tissues, including seeds and leaves, have the capacity to express genes encoding the protein subunits of monoclonal antibodies and assemble them into functional complexes (Daniell et al. 2001). It remains to be seen whether plants can produce these antibodies in sufficient quantities to meet therapeutic requirements. However, if it proves possible, the technology has tremendous potential because of the expense of producing monoclonal antibodies in mammalian tissue cultures. It would be essential to grow these plants in restricted locations, but the value of the products would easily be sufficient to offset the cost of growing the crop in isolation.
These various modifications to foods raise a number of issues, many that are beyond the scope of this study. One class of modifications addresses nutritional deficiencies by increasing the concentration of the deficient nutrient or precursor in the plant. As an analysis of vitamin A deficiencies suggests (Nestle 2001), alleviating such deficiencies may require more than merely increasing the nutrient in a food. A careful, considered approach to using biotechnology to address nutrient deficiencies might start by considering all alternatives and developing a systemic approach to alleviating the deficiency in the identified population. A second class of modifications aims to optimize food quality by incremental improvements. Again, it may be difficult to know that an anticipated improvement in food quality will actually lead to an improvement, such as changing the content of “health-producing” compounds in plants, rather than a decline in human well-being. Clearly there are many questions that other scientists, including nutritionists and public health specialists, need to engage to ensure that biotechnology will lead to improved human health.
All of these future products raise parallel questions related to environmental impact. There might be indirect human health risks mediated through the environment that would require new expert analysis. Nontarget risks associated with these plants with altered nutritional characteristics (both macronutrients and micronutrients), increased concentrations of “health-producing” compounds, or edible vaccines may be considerably more subtle than the direct mortality risks associated with plants producing insecticidal toxins, which are being evaluated presently.
Mitigation of Environmental Pollution
In the future, transgenic plants may be grown for reasons other than commercialization. For example, it has been proposed that transgenic plants could contribute to removing or detoxifying heavy metal pollutants in contaminated soils (“phytoremediation”). A particular problem in some locations is mercury. Plants have already been created that can accumulate mercury. Thus, growing such plants is a potential solution for cleaning up mercury pollution at despoiled sites (Pilon-Smits and Pilon 2000). This example is discussed in detail below.
Finding 7.1: For predicting environmental risks of future plant varieties and their novel traits, currently commercialized transgenic crops offer only limited experience and understanding.
Finding 7.2: The production of nonedible and potentially harmful compounds in crops such as cereals and legumes that have traditionally been used for food creates serious regulatory issues.
Finding 7.3: With few exceptions, the environmental risks that might accompany future novel plants cannot be predicted. Therefore, they should be evaluated on a case-by-case basis.
Finding 7.4: In the future many crops can be expected to include multiple transgenes.
Potential Environmental Impacts of Novel Traits
The Animal and Plant Health Inspection Service (APHIS) appropriately refrains from speculation on environmental risks associated with crops that may or may not reach the stage of commercialization. Full and fair discussion of such risks presupposes information that cannot be known until a functional phenotype has been developed and grown in limited field trials. However, a brief discussion of the general issues associated with some new crops currently being developed will help frame the context in which environmental risks from commercialization of the next generation of transgenic crops may be discussed.
Tolerance to Abiotic Stresses
Abiotic stresses significantly limit crop production worldwide. Cumulatively, these factors are estimated to be responsible for an average 70% reduction in agricultural production (Bresson 1999). Drought stress not only causes a reduction in the average yield for crops but also causes yield instability through high interannual variation in yield. Globally, about 35% of arable land can be classified as arid or semiarid. Of the remainder, approximately 25% consists of drought-sensitive soils. Even in nonarid regions where soils are nutrient-rich, drought stress occurs regularly for a short period or at moderate levels. Furthermore, it has been predicted that in the coming years rainfall patterns will shift and become more variable due to increased global temperatures. Thus, improved stress tolerance may improve agricultural production.
Research to create crop plants that are transformed to tolerate abiotic stresses, such as heat, drought, cold, salinity, and aluminum toxicity, is ongoing. Current research efforts in drought tolerance include the isolation of crop plant mutants to understand the molecular basis for salt responses (Borsani et al. 2001), studies on the transduction network for signaling guard cell responses and their subsequent control of carbon dioxide intake and water loss (Schroeder et al. 2001), and genetic activation and suppression screens that influence interrelationships among multiple signaling systems that control stress-adaptive responses in plants (Hasegawa et al. 2000). The actual development and field testing of novel crop plant varieties is likely to be 5 to 10 years in the future owing to the
complex genetic mechanisms and multigene systems involved in abiotic stress tolerance in plants. Accurate assessments of the environmental risks posed by any of these stress-tolerant plants will not be possible until they are actually created because the genetic mechanism of stress tolerance will greatly determine the scope of potential risks.
Despite this uncertainty regarding the nature of stress-tolerant transgenic plants, they have raised concerns about environmental risks. Abiotic conditions, such as soil nutrient levels, water, cold, heat, salt, and metal toxicity, combined with their seasonal variations have strong determining effects on plant community structure worldwide, and the geographic distribution of many plant species is influenced strongly by these factors (Whittaker 1975). Thus, when plants are transformed to better tolerate these abiotic conditions, it raises risk questions about the possibility of impacts on plant community structure and expansion of the geographic range of a plant species. While these issues have been considered in discussions of the present generation of transgenic crops (see Chapter 4), most of these traits have not been expected to alter the invasiveness or weediness of the transformed plants (Crawley et al. 2001). The environmental risks associated with such stress-tolerant crops are both complicated and subtle. To clarify their analysis, the committee focuses here on drought-tolerant crops. Drought-tolerant phenotypes could be based on higher water use efficiency (WUE), which leads to greater biomass production per unit of water, or an increased ability to extract water from the soil. Obviously the environmental effects of these mechanisms will differ. Plants with improved water extraction will still require the same amount of water to grow, so potential environmental effects may often be related to competition for sunlight or nutrients in the soil due to the plant’s metabolic needs associated with greater biomass. In contrast, ecological theory suggests that a plant with a higher WUE would be predicted to be a better competitor for water than a nontransformed plant. This hypothesized improved competitive ability is the source of some concerns about the environmental risks of drought-tolerant transgenic plants, whether it is the crop or a wild relative that might receive the transgenes by horizontal gene flow.
But increased competitive ability for water is not necessarily sufficient to cause a plant to expand its geographic range. A lupine plant in the oak savanna with better WUE may grow more luxuriantly than its conspecifics but might not expand into surrounding habitats because there is too much shade, water, or soil nitrogen, which might neutralize its advantage in the savanna. A maize variety with higher WUE may grow better in the dryland production systems of parts of Nebraska and Kansas but may still not displace spring wheat in the neighboring counties because it still needs water over a longer growing season than wheat. It is also possible
that a farmer might clear droughty land and plant such a maize variety, leading to marginal increases in the area planted to maize and marginal decreases in the area in xeric prairie remnants. It is unlikely that maize can be transformed into a plant that could grow in arid or semiarid environments without irrigation. Thus, while there is clear potential for plants transformed to tolerate drought to expand their geographic range, there is also a limit to this potential due to the inherent characteristics of the plant, and seasonal limitations of other abiotic and biotic factors that restrict it from expanding. Assessment of these risks will require attention to the plant, trait, and environment. In a similar way, impacts on plant community structure and non-target species and interactions among traits (discussed below) might occur, but their assessment will be case specific. For example, certain transgenic salt-tolerant plants can also tolerate other stresses including chilling, freezing, heat, and drought (Zhu 2001).
In the future, transgenic plants may be grown to remove or detoxify pollutants from soils (“phytoremediation”). One persistent pollution problem is mercury contamination of soils. Plants have already been transformed that can mitigate this problem in various ways. One approach to pollution mitigation is to convert the highly toxic organic mercury in soil into less toxic elemental mercury, which is volatile (Pilon-Smits and Pilon 2000). Another approach is to accumulate mercury in plant tissues (Raskin 1996), where it can be harvested, extracted, or disposed of more safely. These “phytoextraction” plants are already capable of accumulating more than 1% of their biomass as mercury, and crop improvement techniques, including genetic engineering, may be able to increase that fraction.
The environmental impacts of these mitigation strategies are probably scale dependent. While phytoremediation by volatilization could have local benefits, the large-scale consequences are less clear. If the total amount of volatile mercury created is relatively small, it will probably be of little environmental consequence. Thus, small-scale use of the volatilization strategy could be beneficial for the environment. However, if the scale of volatilization is large, so that large amounts of mercury are volatilized, levels of atmospheric mercury may rise at regional or larger spatial scales. Atmospheric mercury is returned to terrestrial, aquatic, and marine ecosystems in precipitation as rain or snow. Once deposited, it is converted into more biologically toxic forms. The deposition of atmospheric mercury released by fossil fuel burning, garbage burning, and medical waste incineration has been identified as a potentially serious environmental problem (e.g., Jackson 1997, Macdonald et al. 2000). Thus, large-scale phytoremediation based on volatilization may exacerbate what
is already a serious concern. Phytoremediation based on “phytoextraction” would not release elemental mercury into the environment. The environmental risks associated with this strategy may be less scale dependent and have the additional benefit of potential harvest for extraction of their heavy metal content.
Combinations of Transgenic Traits
Risk assessments of transgenic crops have focused on the impacts of single traits (e.g., herbicide tolerance) on the environment and human health. Already transgenic crops expressing multiple traits have been commercialized, such as cotton and corn, that have both pesticidal properties and herbicide tolerance. Many more cultivars with “stacked” genes (multiple transgenic traits) are expected to be commercialized soon, so it is important to assess if the environmental and health effects of multiple genes are simply the sum of the effects of each of the single genes or if interactions among the introduced genes quantitatively or qualitatively alter the impacts of these cultivars.
Focusing on environmental effects, there are several levels at which we can look for interactions between inserted genes. In this discussion the focus is on two levels—that of the individual plant phenotype and that of the whole-field or farming system level.
At the individual plant level, an answer to the following question is needed: If expression of each of the single genes engineered into a plant does not have unintended and/or unacceptable effects on the plant phenotype, could the interaction of these genes have such effects? A few illustrative examples may help in exploring this issue. Suppose an agency determines that adding a single gene for tolerance to herbicide A does not confer weedy characteristics to a plant and that adding a single gene for tolerance to herbicide B also does not confer weedy characteristics to a plant. Is it scientifically reasonable to assume that a plant with both genes will not have weedy characteristics? In APHIS’s assessment of petitions for nonregulated status, one “weediness” management strategy hinges on whether “volunteer” plants with the herbicide tolerance gene can be controlled by alternative herbicides. While the answer to this may be yes for each of the single-gene insertions, the answer may be no for plants with multiple herbicide resistance gene insertions, if plants with the two genes are tolerant to the entire array of registered herbicides for the crop. Under current regulatory policies, APHIS does not regulate the plant with combined herbicide tolerance if it has already deregulated each of the traits separately.
In addition, suppose a plant has both a gene for drought tolerance and a gene for insect resistance (such as Bt toxin expression). The Envi-
ronmental Protection Agency (EPA) requires most plants expressing Bt to produce a sufficiently high dose to enable effective resistance management. Is it scientifically reasonable to assume that the gene for drought tolerance, which alters plant physiology, will not also alter expression of the Bt gene? Under the statutory authority of Federal Insecticide, Fungicide and Rodenticide Act, the EPA can and may even be required to evaluate Bt toxin expression in such a drought-tolerant variety. Under the Federal Plant Pest Act and the Federal Plant Quarantine Act, however, APHIS may not be able to make such an evaluation.
At the field or farming systems level, interaction of traits in transgenic plants could involve complex indirect causal pathways. One example is how, in some cases, stacked transgenic traits might interfere with the practice of integrated pest management (IPM) and insecticide resistance management. One tenet of IPM is that pest suppression tools are only used when the level of the pest has increased above an economic threshold. The fact that Bt genes are typically expressed constitutively in itself raises concerns about their impacts on IPM.
A more serious concern was raised during the 2000 field season, when there was a shortage of transgenic cottonseed with only the single trait for herbicide tolerance to glyphosate (Roundup). The demand for glyphosate-tolerant cotton was much higher than the supply, so farmers in a number of southeastern states had to plant either a conventional nonherbicide-resistant cultivar or a transgenic cultivar with stacked Bt and herbicide tolerance genes. In 1999 approximately 22% of the cotton acreage in North Carolina was planted to Bt-expressing cultivars; in 2000, it was 54%. Cotton entomologists (Bradley, 2001, personal communication) have determined that about two-thirds of the increase was caused by farmers who wanted to use glyphosate-tolerant cotton and therefore needed to purchase a cultivar that contained the Bt gene. Similar increases in acreage planted to Bt cotton occurred in some areas of Texas and elsewhere. Such artificial increases in the area of Bt cotton can jeopardize resistance management in all target pests by creating geographic pockets where resistance could evolve rapidly. In addition, one important cotton pest, Helicoverpa zea, the cotton bollworm, is at high risk of evolving resistance, and unnecessarily high use of Bt cotton increases this risk. Stacking Bt and herbicide tolerance genes has actually caused an increase in the risk of resistance evolution.
Within the next three years, industry intends to commercialize corn cultivars with Bt genes that control corn rootworms. It is likely that these rootworm Bt genes will be stacked with the currently commercialized Bt genes that control the European corn borer in some of the new varieties. In 2000, approximately 20% of corn grown in the United States contained Bt genes that control the European corn borer (USDA-NASS 2000). The primary reason for the large proportion of nontransgenic corn is that
losses caused by the European corn borer in many areas are not high enough to cover the additional costs associated with the Bt-expressing cultivars. There are geographic (micro and macro) regions where there will be positive economic incentives to plant corn cultivars with the new corn rootworm-active Bt toxins but where there is no need for European corn borer control. There will also be geographic regions and farm-specific situations where there will be a use for Bt toxins to suppress European corn borer but not corn rootworm. With the proliferation of traits, it will become increasingly difficult for all but the largest seed companies to maintain full inventories of all combinations of the transgenic traits and agronomic characteristics. If this results in situations where farmers must purchase corn cultivars with both rootworm and corn borer Bt toxins or cultivars without either toxin, we may see a repeat of the cotton situation. To date, neither the USDA nor the EPA have publicly addressed this issue.
In the future there is expected to be a proliferation of novel transgenic traits engineered into crop cultivars involving multiple genes. The engineered traits could be as diverse as salinity tolerance and fruit flavor. It is difficult to anticipate the nature of environmental interactions that may be associated with combinations of three or more transgenic traits. Based on past experience, such interactions must be examined on a case-by-case basis. Chapter 3 includes a review of the environmental risk determinations made in connection with the U.S. Department of Agriculture/APHIS petition (97-013-01p) for determination of nonregulated status 31807 and 31808 cotton, which illustrates the kind of issues that can arise with stacked genes (that is, multiple transgenic traits in one crop variety). As noted, APHIS’s review did not consider the possibility of interaction with stacked genes, nor did it consider the impact on farming systems.
Finding 7.5: The current APHIS approach for deregulation does not assess the environmental effects of stacked genes for nonadditive or synergistic effects on the expression of individual genes, nor does it assess stacked genes for cumulative environmental effects at the field level.
Finding 7.6: There are at least two levels at which scientists and regulators must look for interactions between inserted genes with regard to environmental effects: (1) the individual plant phenotype and (2) the whole-field or farming system level.
Industrial Feedstocks and Agronomic Traits
There is significant potential to modify the biochemistry of plants to enable them to produce a number of novel biological chemicals and proteins that can be used in other industrial processes. This technology raises
several risk issues. These transgenes may code for enzymes leading to nonedible products or potential mammalian toxins. There is a clear potential risk involved in environmentally mediated movement of these transgenes so that transgene products enter the human food system (see BOX 5.1). Segregation, isolation, or sterilization methods may be essential to manage these risks, but several uncertainties remain about the cost and efficacy of these methods (see Letourneau et al. 2001, 56–57).
FUTURE POLICY ISSUES
The policy questions raised in the future by new transgenic crops and the many other factors that influence shifts in policy perspectives worldwide will differ from those addressed in detail for the first generation of transgenic crops. Just as it is not possible to predict future environmental risk issues, it is also not possible to predict what the future policy issues will be. This section will identify and discuss several policy issues that the committee believes will become increasingly more prominent in the future as transgenic crops expand globally and new transgenic crops become commercialized. The committee concentrates on four issues, which should be interpreted as a selective slice of the many policy issues that will arise in the future. The committee does not necessarily believe that all of these issues will be crucial ones in the future, but it does believe that they will be prominent and important.
First, the committee discusses how agricultural biotechnology is influencing agricultural structure in the United States and the policy and environmental risk issues that this may raise. Next, the global context for the commercialization of transgenic crops from a societal perspective is considered, and how transgenic crops may or may not address issues related to global food supply is examined. Third, the need for involving the public to democratize decision making and the increasing role that communicating environmental risk will likely play in the future development of biotechnology are discussed. Finally, the focus is shifted back to regulatory policy issues, with a comment on the precautionary principle and drawing particular attention to some of the opportunities that the 2000 Plant Protection Act provides APHIS.
Finding 7.7: The types of new transgenic crops that are developed, as well as the rate at which they appear, will be affected by the interplay of complex factors, including public funding and private financial support for research, the regulatory environment, public acceptance of the foods and other products produced from them, and the resolution of debates over need- versus profit-driven rationales for the development of transgenic crops.
Issues related to changing agricultural structure might have indirect environmental consequences. The socioeconomic considerations for evaluating transgenic crops have proven to be more volatile than was predicted only a few years ago. The desire of some consumers, both foreign and domestic, to avoid consumption of foods containing transgenes has led to the segmentation of markets for “genetically modified” (“GM”) and “GM-free” crops. In the United States the creation of marketing standards for certified organic crops that do not contain transgenes or transgene products has established a market niche for producers. These forms of market segmentation create an opportunity for U.S. farmers to capture a price premium for nontransgenic crops. At the same time, the existence of a market for nontransgenic crops creates a new kind of environmental risk from transgenics: If they cannot be environmentally isolated, pollen and residues from transgenic crops can affect/influence the economic value of nontransgenic crop production. The development of market standards and international trade rules for nontransgenic crops will have economic implications for U.S. agriculture. In the worst-case scenario, U.S. producers may effectively be excluded from these emerging world markets, but economic incentives will be affected under any scenario. Changes in producer incentives generally lead to changes in agronomic practice, with attendant environmental effects.
Such changes may have already had both economic and environmental impacts. For example, widespread use of herbicide-tolerant crops has changed patterns of herbicide use, resulting in new types of damage claims from herbicide drift, forcing farmers to adjust their farming practices accordingly. Farmers may incur legal liability for technology fees as a consequence of neighboring fields inadvertently pollinating a crop, leading to transgenic seed production. This, in turn, may create new forms of environmental nuisance lawsuits, as farmers attempt to protect themselves from complaints lodged by the owners of transgenic technology. These issues need to be evaluated and should be given careful consideration by future study groups.
Biotechnology, World Food Supply, and Environmental Risk
Any attempt to mitigate environmental risk must be mindful of the fact that avoiding one risk can inadvertently cause another greater risk (Graham and Wiener 1995). This might occur if attempts to avoid risks from commercialization of transgenic crops resulted in the adoption or continuation of commercial agricultural practices with greater negative environmental impact. While at present this is a speculative issue, it is a
key element in debates on the efficacy of biotechnology to address global food shortages. If, for example, a lower-yielding nontransgenic crop is chosen in lieu of a higher-yielding transgenic alternative, the result could be that fragile lands would need to be planted in order produce needed amounts of food. If, for example, a chemically intensive farming system is chosen over a reduced-chemical transgenic alternative, the result could be the continuation of an unnecessarily harmful chemical load on a local environment.
Unwanted environmental outcomes could also occur if regulatory oversight itself has unintended consequences. There may be some environmentally beneficial applications of recombinant DNA gene transfer that will never be attempted simply because the costs of obtaining regulatory approval exceed the profit potential. Because environmental benefits can display characteristics of public goods, private industry or individual farmers may derive little monetary benefit from environmentally beneficial farming practices. As such, transgenic crops with environmental benefit may be particularly vulnerable to real or perceived disincentives brought about by costly regulatory procedures. Again, this possibility is speculative, and only case-by-case evaluation of proposed transgenic technologies can tell whether the environmental costs of regulatory procedures outweigh their environmental benefits.
While these examples are speculative, the general point they illustrate is true. Avoiding one risk may inadvertently cause another greater risk. But its truth is more complex than it would appear initially.
The largest and most complex factor in making comparative judgments of agricultural technology is that the main goals of this technology are to provide more and better food for public consumption and to reduce the cost of production for growers. In circumstances where food supplies are inadequate, people may become more willing to accept perceived tradeoffs between environmental quality and food supply. Furthermore, changes in food production that lower prices are especially beneficial to the poor because, even in times of plenty, poor people spend a proportionally larger share of their income on food (Lipton and Longhurst 1989). The Nuffield Council on Bioethics report (1999) on ethical issues associated with biotechnology notes that there is an apparently persuasive argument for accepting environmental risks in order to increase the availability and reduce the cost of food for hungry people and poor farmers. However, this need-based argument itself must be evaluated in light of economic analyses suggesting that insufficient availability of food is seldom the cause of famine (Sen 1982) and that sweeping changes in agricultural production can have devastating effects on the poor (Dahlberg 1979). In other words, the perceived tradeoff between environmental quality and food supply may be illusory.
On the one hand, inadequate amounts of food and malnutrition have plagued humans from the beginning of civilization, environmental degradation is a widespread concern, and both arable land and water are limited. Indeed, as recently as the 1960s, there was starvation in China, resulting in millions of deaths (Brown 1995). The world’s population has doubled in the past 40 years, and it is expected to continue to grow, reaching a maximum of 8 billion by 2020 and perhaps 10 to 11 billion by 2050 (Vasil 1998). The reasons for hunger and limits to food security in different places and at different times are clearly complex and dynamic. However, such population projections create cause for global concern. There would be less arable land to grow crops because cities would continue to expand and erosion would carry away valuable topsoil (Kishore and Shewmaker 1999, Tilman 1999b). It is also well documented that supplies of fresh water are declining worldwide (Johnson et al. 2001), and this will severely tax our ability to grow crops. Currently, 17% of the world’s croplands are irrigated, and this area produces 40% of the world’s food. It is projected that the requirement for food in the most rapidly growing areas of the world—Southeast Asia, South America, and Africa—might double by 2025 and nearly triple by 2050 (Vasil 1998). At the same time, improved economic conditions in Asia have been increasing demand for costlier food products, such as meat and poultry, which can only be obtained by producing larger amounts of feed grains (Vasil 1998, Kishore and Shewmaker 1999). Dramatic increases in population growth are not expected in many industrialized regions such as the United States and Europe, but per-capita consumption in industrialized nations is high and sustains a disproportionately high demand for diverse agricultural products. The increasing nutritional and economic demands of the less developed regions of the world are expected to impact the more economically secure countries. Some people believe that such problems will require increases in productivity that far exceed what agricultural technology has achieved in the past and that these increases cannot be achieved without the use of transgenic crop plants (McGloughlin 1999, Borlaug 2000).
Opinions regarding the severity of the problems resulting from projected rapid increases in human population and the requirements to feed and clothe the world’s population differ (Alexandratos 1999, Johnson 1999, Borlaug 2000), but these issues and those of land and capital distribution are likely to impact agriculture worldwide. Finding ways to increase food and fiber production during the next 50 years would then become a major concern for societies around the world. Also, there is increasing evidence that food quality plays an important role in human health. According to the USDA, medical expenses and lost productivity resulting from chronic illnesses cost U.S. society some $250 billion annually, and $100 billion is related to poor nutrition (Welch et al. 1997). On
the other hand, highly aggregated statistics on food production, population, and nutritional needs can present a misleading picture of the prospects for addressing these problems through agricultural production technologies, whether through transgenic crops or by other means. A wide array of national, regional and international policies have tremendous effects on incentives for food production and consumer exchange. Yield or nutrition-enhancing technological changes have variable and sometimes negative effects in a given policy context (Drèze and Sen 1989). Hence, the use of biotechnology to address growing needs for increased food production must, at a minimum, be accompanied by a call for research on economic and sociological impacts, and appropriate policy adjustments. One group of researchers has expressed concern that a concentration of economic power and intellectual property rights will skew the development of novel crops away from those that would be of most use to the neediest people (Bunders and Radder 1995, Buttel 1995). Still other researchers refute claims that research and development of novel crop traits through biotechnology are appropriate responses to present and future problems of hunger, even for less developed countries (Busch et al. 1990, Krimsky and Wrubel 1996, Altieri and Rosset 1999). They suggest instead that promotion and support of new rural development approaches and low-input technologies spearheaded by farmers and nongovernmental organizations will make a greater contribution to food security at the household, national, and regional levels in Africa, Asia, and Latin America (Pretty 1995).
The potential for commercially grown transgenic crops to increase total world food supplies and to reduce the need for utilizing marginal lands for food production is thus an appropriate but highly contested dimension of the context in which environmental risks for the next generation of crops will be debated (Kalaitzandonakes 1999). One driving factor in the debate is the suspicion that need-based arguments promoting transgenic crops are only a ruse for profit-seeking activities by the biotechnology industry. The argument for yield-enhancing agricultural research throughout the twentieth century stressed the economic viability of farmers (Rosenberg 1961, Danbom 1979). Yet analysis of the “technology treadmill” in agriculture has established that consumers and agricultural input companies are the primary economic beneficiaries of such technology (Cochrane 1979). Several early studies of biotechnology in agriculture placed a great deal of emphasis on this phenomenon while also noting that late adopters of biotechnology may experience economic losses (see Kalter 1985, Kenney 1986, Kloppenburg 1988). The result is that some farmers and farm groups take a jaundiced view of the claim that biotechnology responds to on-farm needs for greater efficiency while
nonetheless moving quickly to adopt the first generation of transgenic crops (see Rundle 2000).
Thus, an argument that begins by noting that growing global food needs will create environmental stresses on soil and water resources becomes entangled with domestic social issues that have no obvious connection to the environmental risks of transgenic crops. While environmental risks and social impacts are logically distinct, they may have common causes that reside in what biotechnology firms do to make an adequate return on their research investments. Restricting the scope of an analysis of transgenic crops to environmental impact may seem justified in light of the way disciplinary scientific expertise tends to dissociate the causal factors that contribute to environmental impact from those that affect profitability and economic access to food. Such disciplinary divisions are also reflected in the organization of regulatory authority. Even well-meaning restrictions on the scope of an evaluation of transgenic crops may be perceived as strategically motivated attempts to limit debate on transgenic crops to topics on which they will be evaluated relatively favorably (Thompson 1997a, 2000).
The context for evaluation of environmental risks must take account of the need-based argument for expanding food capacity and production. Both human and environmental dimensions of the need for increased food production are critical to any balanced evaluation of the risks from transgenic crops. As the debate moves to consider the next generation of transgenic crops, it may be possible to formulate a way of framing the need for increased amounts of food in a manner that is more sensitive to the socioeconomic complexities of hunger and deprivation (see Conway 2000). It will be increasingly important for decision makers at all points in the global food system to be well versed in both sides of this debate.
As nonfood crops are added to the product mix of agricultural biotechnology, additional and more jarring points of tension, if not outright contradiction, will emerge. Some industrial products, such as pharmaceuticals, may be produced on such a limited spatial scale that they would have a negligible impact on total food production, but that is not necessarily the case for all transgenic crops aimed at producing industrial products. For example, interest in alternative energy sources has predictably been revived in the United States with recent news of diminishing domestic petroleum-based energy sources and the concomitant price increases. One commonly cited alternative is the production of gasohol from grain, especially corn. Corn and other grains are plentiful, renewable, and inexpensive; can be grown domestically; and might reduce U.S. dependence on foreign petroleum production (Woolsey 2000). Technically, generating gasohol from grains is feasible, but there are other factors to consider. To
many the ethical implications of diverting substantial portions of foodstocks to produce an industrial chemical are themselves substantial. There may be arguments that only excess grains are used industrially and that the United States produces too much food anyway. Furthermore, gasohol could be made from the residual straw or chaff after the food grain is removed. Nevertheless, it is not difficult to see why the public might discern a contradictory message. On the one hand they are told that risks are offset because biotechnology is needed to meet increasing global food requirements. On the other hand they are told biotechnology is needed in order to allow farmers to grow something other than food.
Finding 7.8: In an era of globalization, applications of transgenic crops in developing countries will be an important component of the context in which environmental impacts from transgenic crops are evaluated.
Involving the Public and Communicating Environmental Risk
Members of the public express an interest in the environmental impacts of agriculture and a desire to be informed about the relative environmental risks and benefits of different production methods. The general public needs to know more about the environmental risks of transgenic crops in order to form opinions about whether to consume these products or to be concerned about the activities of those who promote and those who oppose biotechnology. One of the principal reasons for public concern about transgenic crops is the belief that environmental issues are not taken seriously by agricultural scientists, biotechnology companies, and farmers who utilize high-tech farming methods. The roots of this belief are not easy to trace but almost certainly include past experiences when agricultural chemicals were promoted as safe and without environmental impact, only to be regulated and withdrawn in subsequent years (Dunlap 1981, Perkins 1982). These issues linger. Some argue that the risks of chemicals have been overstated, and that scientific risk evaluations of past farming practices were neglected when chemicals were withdrawn (Fumento 1993). Others believe that substantive risks of agricultural chemicals continue to be neglected and alternatives continue to be ignored (Pimentel 1987). In either case, there is a need to do a better job of involving and informing the public about what is done to mitigate and manage environmental risks in agriculture. Issues arising in connection with the commercialization of transgenic crops are only one dimension of this general problem.
A 1996 NRC report, Understanding Risk, notes that the initial characterization of risk requires “an appropriately diverse participation or rep-
resentation of interested and affected parties, of decision makers, and of specialists in risk analysis.” The way that participants are identified and involved is critical to the success of any attempt to measure, manage, or mitigate risks associated with transgenic crops. There are different disciplinary and practical perspectives that have the potential to contribute information that will improve the modeling and measurement of hazards, exposure pathways, and the relative probability of unwanted outcomes. Also, because biotechnology is controversial, participant involvement is becoming increasingly critical to the role of risk analysis in forming the basis for authority, believability, and public confidence in regulatory decision making and in the subsequent commercialization and widespread adoption of transgenic crops.
As noted in Chapter 2, risk analysis should play at least two roles in the regulation of transgenic plants. In the past, the USDA emphasized the role of decision support and has limited participation in its environmental risk assessment procedures to those who, in the judgment of decision makers, could provide information pertinent to the anticipation and measurement of environmental hazards. However, as biotechnology has become controversial, the fact that risk assessments have been done has been cited to justify U.S. agriculture’s rapid adoption of Bt and herbicide-tolerant crops. Such justifying citations involve a shift toward using risk assessment as the basis for claiming that the USDA has faithfully exercised its decision-making authority or that the public should have confidence in recombinant DNA technologies. However, USDA officials should not presume that risk analyses done purely to support internal decision making can bear the additional weight of supporting the public’s confidence in technology or government.
Communicating the idea that risks are being taken seriously is a key element. One of the surest ways to do this is to ensure that the entire process of characterizing and measuring risk is open to an array of participants representing perspectives that reflect the various sectors of the public taking an interest in agriculture and the food system. Efforts to minimize the seriousness of risks may seem warranted in virtue of the improbability of an unwanted impact or in virtue of the relative risks posed by a transgenic technology when compared to its chemically based alternative. However, when risks are denigrated by industry, regulators, or their representatives, and people who were not present when key interpretive judgments were made are told that the risks are “not serious,” the inadvertent message is that responsibilities are not being taken seriously. This situation is aggravated when citizens who express concern through intermediary organizations think they are being excluded from key processes of risk evaluation and management. The result can be a situation in which the belief of being at risk is amplified by the very
activities that were intended to put the public’s mind at ease (Sandman et al. 1987, Kasperson 1992).
The practice of sending messages that persistently minimize the significance of risks can also have the unintended effect of making those who should in fact be attentive to environmental impacts overconfident. In some cases the individuals needed to mitigate an environmental hazard may be farmers or nonscientific employees of agricultural companies. Such individuals need to be advised about environmental risks and the methods to contain them in unvarnished language (Otway 1992). Although there were no enduring food or environmental safety issues associated with the contamination of corn stocks with unapproved Cry9c protein in the autumn of 2000, one lesson is that, when key individuals—in this case farmers and grain handlers—are persistently told that risks are minimal, a general culture of laxity may come to prevail. Adequate risk management for the coming products of biotechnology depends on a culture that reinforces the seriousness with which environmental risks must be addressed.
A more inclusive approach is especially important in environmental risk assessment, as compared to human health risk assessments, because classifying (or failing to classify) a possible event as a hazard involves value judgments that are more likely to be controversial. Describing some possible event or set of events in terms of risk always implies that these events are regarded as adverse. Although we talk about the risk of dying or losing money, we do think of our chances of recovering from disease or winning the lottery as forms of risk. The tendency to overlook the value judgments made in risk assessments that emphasize adverse effects on human health arises simply because these value judgments are not controversial. In comparison to human health, the adversity of events in the environmental arena may not be at all obvious, especially to individuals lacking key background knowledge of agriculture and ecology. In addition, conflicting values and attitudes about which nature is wild, which needs protecting, and about the relative value of wild and agricultural land use will affect the way in which environmental risks from agricultural production are defined (Cranor 1997, Thompson 1997b).
The question of whether genetically engineered crops involve environmental risks is thus not wholly a scientific one. An answer depends partially on how one understands for classifying events associated with crop production as adverse. One interpretation of adversity might stress the impact on such factors as insect or disease resistance, soil fertility, and prevalence of pollinators or noxious weeds. Such factors affect the chances of successful farming. Other effects related to successful farming are indirect: herbicide-tolerant crops affect patterns of herbicide use and with that comes sources of risk associated with spray drift, for example, and
pollen from transgenic crops can substantially damage the value of crops grown for organic markets. Such events might be regarded as adverse from one farmer’s perspective but not from another. An altogether different interpretation of environmental risk might stress the impact on habitats for nonagricultural plants and animals or on genetic and species diversity.
Some of the parameters for interpreting what counts as adverse have been reasonably well articulated in the statutes on which the USDA exercises its regulatory authority. However, it is not likely that risk assessment will succeed in building the public’s confidence in biotechnology unless there is a more systematic and more public effort to acknowledge the diverse values that influence judgments on adversity. Each of the actors in the agri-food system must find ways to involve a broader array of value perspectives throughout all phases of research planning, product development, and regulatory oversight.
Finding 7.9: Adequate risk management for future biotechnology products will depend on a regulatory culture that reinforces the seriousness with which environmental risks are addressed.
Finding 7.10: Public confidence in biotechnology will require that socioeconomic impacts are evaluated along with environmental risks and that people representing diverse values have an opportunity to participate in judgments about the impact of the technology.
The preceding discussion presents the general context in which the risks of coming generations of transgenic plants should be evaluated. The types of crops developed in the second generation and the rate at which new transgenic crops become available will depend in part on specific regulatory policies, on how need-based versus profit-based technology debates play out, on continued public and private financial support for the research needed to create new transgenic crops, and finally on public acceptance of the foods and other products produced from them. Although these factors can be conceptually and logically distinguished from one another, they impact one another in complex ways. Regulatory policies affect financing and profitability of research ventures. A lack of public acceptance shapes markets for transgenic crops and creates political pressure for regulatory responses that may not be strictly justifiable based on scientific measures of health and environmental risk.
The probability of environmental damage from the next generation of transgenic crops will be determined by the specific phenotypic traits of
each crop as well as by agronomic practices and land uses specific to each particular crop that is commercialized. Throughout this report, discussion of risk characterization and risk management has focused on procedures that can and should be used to measure the likelihood and extent of such damage and to mitigate environmental risks. It is appropriate that regulators focus intently on such specifics when evaluating any particular transgenic crop. In looking toward the future, however, it is also important to ensure that the details of risk analysis do not obscure the larger context in which discussions about agriculture and the environment are situated. In this section, the discussion is focused on three related regulatory concerns from this larger perspective. The committee considers how the future of commercialized transgenic crops may challenge both U.S. and global capacity for environmental risk analysis and the ambiguities surrounding the precautionary principle and the development of its use worldwide. Then, the committee turns to the changing domestic standards for environmental regulation and some implications for regulation of any agricultural practice.
U.S. and Global Capacity for Environmental Regulation
The overall adequacy of the U.S. Coordinated Framework for the Regulation of Biotechnology has been a frequent subject of discussion throughout the brief history of transgenic crops. Krimsky (2000) finds the coordinated framework to have failed to adequately manage environmental risk, while Malinowski (2000) finds it to have been remarkably successful. It is beyond the scope of this report to evaluate the entire framework, and so instead several points are provided for consideration that may challenge the regulatory system in the near future.
As noted above, some of the coming applications of biotechnology may involve the use of plants to produce pharmaceutical products, biologics, fuels, and other substances not intended for human food use. The introduction of such transgenes poses the potential for environmentally associated risks of a wholly different order than those associated with existing transgenic crops. If such a transgene moves into food crops, either through pollen transfer or physical contamination, there could be serious human safety risks. If such a transgene moves into a wild relative, there could be widespread environmental dissemination of the pharmaceutical substance or other nonfood substances that could have impacts on wildlife as well as microbial populations.
Provisions for regulation of such environmental risks are identified under the existing coordinated framework. Environmental risks associated with pharmaceutical-producing plants, for example, could be regulated by the Food and Drug Administration under the authority it cur-
rently exercises to regulate drug production. It is not too soon to raise questions about the interagency technical capacity to review and manage such risks, the potential for duplicative regulation and new regulatory gaps, and the need to plan research to understand and adequately assess such risks. In addition, there may be a need for changes or modifications in the current regulatory approach to commercialization of transgenic plants.
A broader issue associated with global regulatory capacity relates to a potential unintended consequence of U.S. regulatory decision making. The United States has conducted environmental reviews of more transgenic crops than any other country. Many countries lack sufficient regulatory oversight of environmental impacts, especially the scientific capacity to conduct environmental risk analyses and the administrative capacity to enforce environmental decisions that are made. In these conditions there becomes a tendency to rely on previous regulatory reviews conducted in other countries, whether indirectly when an applicant submits the identical data package for risk analysis or overtly when regulatory personnel parrot the previous review. As discussed in Chapter 3, while APHIS notes some non-U.S. environmental risks of transgenic crops in some of its reviews, it does not and cannot do so in every case. Even when some global environmental impacts are noted, U.S. environmental reviews are keyed to environmental conditions and farming practices prevalent in the United States. Transgenic crop varieties that may be environmentally benign in this country might have a high risk of ecological disruption were they to be grown in some developing countries. For example, corn grown in southern Mexico and Guatemala and potatoes grown in the highlands of Peru will cross readily with wild relatives, which could lead to environmental risks. The global scientific capacity to conduct locally relevant environmental risk analyses is sorely deficient, and this needs immediate attention. The relationships among nationally-based environmental risk assessment and regulation, multinational agreements, and global environmental hazards must also be included in the broader context for future discussions of environmental risk from transgenic crops.
The Precautionary Principle
At the same time that new products and applications of gene transfer are being applied in crop production, there has been discussion of new regulatory approaches. One important component of this discussion has been the “precautionary principle.” A review of this discussion, however, reveals that there are actually several principles being proposed rather than only one. Language endorsing a precautionary principle and taking a precautionary approach in regulatory decision making is in the Euro-
pean Union’s (EU) Directive 90/220 on biotechnology and its revision in 2001, the Cartagena Protocol on Biosafety, which was finalized in Montreal during 2000, under the Convention on Biological Diversity (CBD). The CBD has been ratified by many countries but not the United States. It is an international agreement to conserve and use sustainably biological diversity for the benefit of present and future generations.
The Rio Declaration of 1992, which is a part of the CBD, includes the following language in Principle 15:
In order to protect the environment, the precautionary approach shall be widely applied by States according to their capabilities. Where there are threats of serious or irreversible damage, lack of full scientific certainty shall not be used as a reason for postponing cost-effective measures to prevent environmental degradation.
On January 20, 2000, the Cartagena Protocol on Biosafety (Biosafety Protocol) was adopted by consensus of the parties. It applies to transgenic organisms that will be intentionally introduced into the environment. Food and feed commodities, which might lead to accidental introductions, are excluded. The protocol reaffirms Principle 15 and expands it in its Articles 10.7 and 11.8.
Lack of scientific certainty due to insufficient relevant scientific information and knowledge regarding the extent of potential adverse effects of a living modified organism [some transgenic organisms] on the conservation and sustainable use of biological diversity in the Party of import, taking also into account risks to human health, shall not prevent that Party from taking a decision, as appropriate, with regard to the import of the living modified organisms in question in order to avoid or minimize such potential adverse effects.
This statement of the precautionary principle enables regulatory decisions to err on the side of precaution when there is scientific uncertainty. Most of the details on implementation are left to the discretion of the parties, and it is within these details that the scope of the precautionary principle will be established. Consequently, unless there are additional negotiations, the regulatory impact of the precautionary principle will be determined by practice and only after its use will it be possible to articulate what it actually is.
The European discussion of the precautionary principle has occurred in the context of attempts to harmonize grades, standards, and regulatory approaches among member states. These discussions have occurred during a spate of food scares and failings in government oversight, most notably associated with transmissible bovine spongiform encephalopathy (Mad Cow Disease), that have rocked public confidence in government regulation of the food system. Although debate over the food safety and
environmental risks of transgenic crops has figured prominently in Europe, European pronouncements on the precautionary principle neither offer a specific evaluation of transgenic crops nor provide language that would give regulators specific instructions on how to take a precautionary approach with respect to transgenic crops. The March 12, 2001, revision to EU Directive 90/220 states that the requirements of the Biosafety Protocol, including the precautionary principle, should be respected. The 2001 directive indicates that the precautionary principle was taken into account in the drafting of the directive and must be taken into account when implementing it. Consequently, the directive does not provide a statement of the precautionary principle but identifies regulatory processes that are precautionary. It is beyond the scope of this report to provide a detailed analysis of this new directive, but suffice it to say that the precautionary principle has been incorporated into at least the labeling and traceability standards, the monitoring standards, and the EU approval process. Much the same as in the Biosafety Protocol, the precautionary principle will be defined in Europe through its application to specific cases.
U.S. discussion of the precautionary principle has been sparked by both these European developments and the publication of papers from a 1998 working group conference on the meaning and applicability of the precautionary principle in a number of legal and public health settings (Raffensperger and Tickner 1998). This collection of essays on the precautionary principle includes a number of interpretations and applications, all of which should be regarded as supportive of the precautionary principle and the precautionary approach. There is considerable variation even among contributors to Raffensperger’s and Tickner’s book advocating implementation of the precautionary principle. Cranor (1998) focuses on the need to emphasize the minimization of type II statistical errors in science intended for regulatory application. Ozonoff (1998) argues that the precautionary principle should be interpreted as a screening device to apply “scientific evaluations to situations where the proportion of cases that are hazards is high and/or use methods with high specificity, that is, that correctly identify ‘no-hazard’ situations” (104). Arguably, these approaches simply reflect the current regulatory philosophy in place for transgenic plants under the coordinated framework. However, other contributors explicitly advocate the use of nonscientific criteria, such as political solidarity or respect for life, to evaluate the possibility and potential for environmental hazards (Bernstein 1998, M’Gonigle 1998).
In summary, a number of different formulations are given for the precautionary principle. Advocacy of a precautionary approach can be interpreted to convey a need to reconsider and possibly strengthen regulatory oversight of technology. One key is the advocacy of statistical and
scientific evidentiary decision standards that err on the side of preventing serious and irreversible health and environmental effects, even in the absence of definitive demonstration of harm to humans or non-target organisms. Given the ambiguity in its various formulations as well as opportunities to interpret the seriousness of environmental hazards in different ways, the precautionary principle does not provide a single alternative to existing regulatory policies for the U.S. government. Indeed, some interpretations of the precautionary principle would be consistent with current approaches. Nor does the principle provide unambiguous guidance for the evaluation of transgenic crops. While both advocates and critics of the precautionary approach appear to have assumed that this would result in an adverse evaluation of transgenic crops (Carr and Levidow 2000, Miller and Conko 2001), it is not necessary that such a result would follow. A more definitive evaluation of the precautionary principle must await more specific criteria for its application in agriculture (Soule 2000). For the meantime, the “precautionary principle” should be regarded as a number of potential regulatory approaches pertaining to the use of science to formulate public health and environmental policy, rather than a specific proposal for change. However, as it is applied in Europe and under the Biosafety Protocol, its meaning will become clear. As it gains clarity, there may be significant implications of its use with respect to U.S. regulatory policy.
Finding 7.11: The precautionary principle forms an important but imprecise context for the regulation of future transgenic organisms.
Raising the Regulatory Bar
It will soon become necessary to reconsider the general philosophy of regulation for environmental impact that has been applied to all forms of agricultural technology since World War II. Plant scientists have developed an array of techniques for introducing genetic novelty into crop varieties through conventional breeding methods such as mutagenesis, wide crosses, and embryo rescue. These techniques have the potential to introduce genes and genetic variations into crops that in some cases equal the novelty associated with recombinant DNA techniques. It is entirely possible that crops developed with either of these techniques could possess phenotypic traits associated with elevated levels of environmental risk.
As BOX 1.1 notes, Green Revolution varieties that increased commercial crop capacity to utilize nitrogen fertilizer shifted land-use patterns and sparked a worldwide increase in the adoption of chemical inputs. A comparison of the environmental risks from Green Revolution varieties
with those of the first, second, or third generations of transgenic crops would be a highly speculative enterprise. Yet in retrospect it is difficult to argue that Green Revolution impacts on the environment were automatically safe, benign, or even acceptable simply because they were developed using conventional breeding. With the potential for introducing additional novel traits through nontransgenic methods, it becomes increasingly difficult to defend the idea that conventional crops should automatically be excluded from scrutiny for environmental impact.
As noted earlier, NRC reports have consistently found that use of recombinant DNA technology in the development of an agricultural crop does not in itself create a new class of risks. As with conventionally bred crops, it is the phenotypic characteristics of the plant that are the source of environmental risks. This report and the 2000 NRC report (2000c) on pest-protected plants cite a number of environmental risks that should be accounted for in the regulation of both current and future crop varieties. When these observations are combined, the possibility that nontransgenic crops may also pose environmental risks requiring a regulatory response becomes logically inescapable. Yet new crop varieties posing potential hazard fail to require regulatory scrutiny under APHIS while potentially benign transgenic crops do because APHIS oversight excludes conventional crops. Those plants produced via recombinant DNA are regulated, and those produced by other methods are exempt, even if the final product has an identical phenotype and therefore presents similar potential risks.
Moving beyond the realm of crop modification, it is also clear that the bar has been raised substantially for acceptable environmental effects for novel pesticides and new agricultural practices (e.g., changes in crop rotations and changes in cultivation practices). As our perspective on the ecological interactions and interchange between agricultural and nonagricultural lands evolves (see Chapter 1), the environmental standard being set for transgenic plants may be a better overall environmental effects model for agriculture than the model developed in the early 1900s for assessing the acceptability of conventional crop varieties and agricultural practices.
Although government regulation of conventionally modified plants has been virtually nonexistent, the agricultural research establishment has not ignored the potential of genetic modification of crops to result in environmental change. Indeed, some of the work begun in the 1960s to assess the long-term impacts of Green Revolution cultivars and cropping practices demonstrated insight into the needs for examining environmental effects on long and large scales. As discussed in Chapter 1, it was only because of long-term 30-year experiments on the impact of cropping intensification that researchers were able to clearly document the effects of
the new cropping practices on soil characteristics. These long-term experiments were sponsored by nonprofit groups and local governments. Today, most efforts at examining environmental impacts of transgenic crops emphasize short-term laboratory and field plot experiments. While such experiments are useful, there is a need to examine effects that cannot be seen at such small scales. For example, short-term and small-scale experiments would not have predicted the effects of Green Revolution practices on lowering the water table in semiarid regions of India or the rise in water tables in other areas that has been accompanied by salinization of the soils in the root zones of crops.
Finding 7.12: The environmental impacts of transgenic plants and other new agricultural practices can be studied at a number of ecological scales ranging from the specific toxicological effects of a newly produced compound to the large-scale and long-term spatial and temporal effects of changes in agricultural practices induced by the introduction of a novel crop variety.
Finding 7.13: Currently APHIS environmental assessments focus on the simplest ecological scales, even though the history of environmental impacts associated with conventional breeding points to the importance of large-scale effects, as seen in the impacts of Green Revolution cultivars.
Recommendation 7.1: APHIS should include any impact on regional farming practices or systems in its deregulation assessments.
Society demands that commercial products are deemed safe for health and the environment. The public trusts government regulators to assess risks adequately, to exclude from commerce products posing unacceptable risks, and to impose appropriate risk management strategies to reduce risks. Some people want regulators to test “everything for everything” prior to commercial release. All regulatory agencies, including APHIS, must work with limited financial, temporal, and human resources, so testing everything is not feasible. Instead, prudence and fiscal reality dictate that regulators identify those products most likely to be hazardous and concentrate scrutiny on them, applying a science-based approach to risk assessment.
Currently, one of the risk-based triggers for assessment of a crop by APHIS is that it is a transgenic crop variety (see Chapter 2). This trigger captures all products of genetic engineering but excludes potentially hazardous products derived from conventional methods. This regulatory trigger is imperfect because it does not provide regulatory scrutiny for certain conventional crop plants that may have environmental risks. As discussed more thoroughly in Chapter 2, it fails to capture potentially
hazardous conventional products, such as stress-tolerant canola. At the same time, it brings under regulatory scrutiny some transgenic crop plants that it later determines to have no significant environmental risks. Indeed, an intraspecific recombinant DNA variety would be regulated even if the gene were removed and replaced back in its original location with no extraneous DNA. Thus, there are reasons for considering other or additional potential triggers for the regulation of crop varieties.
APHIS has the authority to regulate all plants that pose a plant pest risk under its present statutory authority. On May 25, 2000, the U.S. Senate and House of Representatives agreed on a conference report on a new Plant Protection Act (PPA), which was developed to improve the government’s ability to prevent and mitigate the effects of organisms that might harm agriculture and the environment. This act, signed by President Clinton, on June 20, 2000, includes products of genetic engineering but is not limited to them. The PPA offers an opportunity for APHIS to refocus regulation on those plant products that pose risk, regardless of the method of derivation. As developed in more detail in Chapter 2, however, scientifically defensible ex ante risk assessment is not yet possible, so it will not be possible to develop a science-based trigger based solely on predictions of the risks associated with particular crop varieties. In addition, it would be imprudent to bring all conventional crop varieties under regulatory oversight. Thus, careful scientific thought must be applied to this problem to avoid extraneous regulation.
Finding 7.14: The committee finds that the Plant Protection Act of 2000 can be viewed as an opportunity to clearly define the types of novel plants, regardless of method of breeding, that trigger regulatory scrutiny.
As explained in previous chapters, APHIS regulates transgenic plants under the authority of the Federal Plant Pest Act (FPPA) and the Federal Plant Quarantine Act (FPQA). While the scope of these acts is quite broad, there are some taxonomic and functional limits that make regulation of some potentially hazardous organisms problematic. For example, the FPPA does not recognize vertebrates as pests. Furthermore, transgenic plants that have been modified using Agrobacterium DNA or the DNA from certain viral species fall clearly within the regulatory authority of the FPPA because these organisms are plant pests. Although APHIS can regulate plants that have been genetically transformed without the use of DNA from a plant pest, determining which plants will be regulated is not simple. As identified early in this report, the regulatory options available to APHIS through the FPPA are limited. For example, once a plant is deregulated, APHIS has no authority to restrict its use or to even monitor it.
The broad language of the new Plant Protection Act defines “plant pest” as including all vertebrate and invertebrate animals except humans. The PPA repeals the FPPA and the FPQA, however, it provides for regulations issued under the repealed acts to be enforced until new regulations are developed. The PPA also provides the potential for developing new procedures for regulating plant products that pose risk. For example, regulations could presumably be written giving directives to APHIS on how to involve the public and external scientific experts in its review process. In addition, regulations under the PPA could be used to provide the flexibility lacking in present regulations. For example, it would be useful to allow APHIS to recall or rescind deregulated plants under appropriate circumstances. Current federal policymakers will determine to what extent the new PPA will be used as a vehicle for change.
Recommendation 7.2: The committee recommends that the Plant Protection Act of 2000 be viewed as an opportunity to increase the flexibility, transparency, and rigor of the APHIS decision-making process.
Finding 7.15: Nontransgenic crop breeding techniques have the potential to introduce genes and genetic variation into crops that equal or surpass the novelty associated with recombinant DNA techniques.
It is entirely possible that crops developed using these techniques could possess phenotypic traits associated with elevated levels of environmental risk. When evaluating transgenic crops for deregulation, it becomes increasingly difficult to defend the idea that these new nontransgenic crops should automatically be excluded from scrutiny for environmental impact.
THE NEED FOR STRATEGIC PUBLIC INVESTMENT IN RESEARCH
Perhaps more than anything else, the experience with commercialization of transgenic crops has revealed gaps in the knowledge base for understanding and measuring the environmental risks of crop production, irrespective of whether recombinant DNA technologies have been applied. Crops developed through nontransgenic methods can and have resulted in avoidable environmental damage. Managing the risks of crop production in the future will depend on improving basic scientific knowledge and research capacity with respect to five key areas discussed below.
Improved Risk Analysis Methodologies and Protocols
Formal research support in the United States for the study of environmental impacts of transgenic plants has been sparse. The longest continuous funding, now almost a decade, has come from the USDA’s Biotechnology Risk Assessment Research Grants Program (BRARGP) to assist federal regulatory agencies in making science-based decisions about the safety of introducing genetically modified organisms into the environment. The program accomplishes its purpose by funding scientific research in priority areas determined in part by the input of the regulatory agencies. Research proposals submitted to this competitive grants program must address risk assessment, not risk management, and are evaluated by a peer panel of scientists. The program has allocated no more than a few million dollars for research each year. Recently, the USDA’s Initiative for Future Agriculture and Food Systems (IFAFS) program has included a competition for funding research, education, and extension on the management of environmental risks of agricultural biotechnology. Both funding programs have substantial limitations—BRARGP because its focus is only on assessment and because the total amount of funding is so low; IFAFS because the focus is only for risk management and the funding program itself is anticipated to have a short life. Neither program funds monitoring or research related to monitoring.
Research on the environmental impacts of transgenic plants can be accomplished through other funding sources if the research questions asked have general significance. For example, issues directly associated with the impacts of transgenic plants may often be associated with critical, but largely unanswered, questions in other fields. For example, whether or not the introgression of pest resistance transgenes into wild populations will result in the evolution of weediness or invasiveness is directly associated with important questions in population biology regarding the genetic and ecological causes and correlates of invasiveness (Traynor and Westwood 1999).
There are several critical areas of research related to risk analysis that would benefit from increased funding.
In the area of hazard identification and risk assessment, there is a need for improved, scientifically sound protocols to detect effects of transgenes and transgene products on non-target organisms. Present protocols are better adapted for screening for effects of toxic chemicals with broad ecological effects, rather than biologically released materials with more targeted specificity. This includes better understanding of effects on pollinators, natural enemies, species of conservation concern and soil organisms. There is also a need for improving understanding of the environmental effects of transgene
movement. Present research has concentrated on determining the conditions under which transgene movement would likely occur. It will be essential to develop a deeper theoretical and empirical understanding of the kinds of environmental effects that could result from transgene movement and the conditions under which such effects are likely to occur. This needs to be understood for transgene movement associated with pollen, viruses, and bacteria. In addition, there is a need to develop assessment protocols that can identify the circumstances under which a plant is likely to become invasive. Because deliberate introduction of plants is the most common route of establishment of invasive plants, it will be critical to understand how genetic modification affects invasiveness.
Research on the effectiveness and efficiency of present regulatory systems is needed to develop approaches to enable the regulatory systems to improve based on scientific principles and data as they acquire information and experience. This includes analysis of alternative methodologies as discussed in Chapter 2.
Research is needed to evaluate the environmental risks associated with conventionally produced crop plant varieties. This will involve understanding the kinds of risk and the circumstances under which these risks might occur. Because conventional crop breeding methods are constantly changing, this evaluation should include a dynamic assessment.
Because many novel transgenic organisms are likely to be developed in the future, it will be useful to fund research to identify and investigate possible environmental hazards associated with these new types of transgenic organisms. These efforts might best be initiated several years before any of these plants become likely to be commercialized. This research effort could be coordinated with research on transgenic methods to minimize risk, as discussed below.
Postcommercialization Validation and Monitoring
As discussed in Chapter 6, there is a need for formal postcommercialization validation testing to determine if the results of small-scale precommercialization tests are relevant at larger spatial and temporal scales. The infrastructure for such testing exists but resources must be targeted to carry out such testing in a rigorous manner. Development of long-term monitoring systems, and development and coordination of trained observer monitoring will require a major commitment to building infrastructure.
Long-term monitoring is based on the use of ecological indicators (NRC 2000a), which are intended to provide valid, reliable, and cost-
effective information about the status of important ecological systems. Repeated observations of these indicators can enable the identification of associations between changes in indicator values and the use of transgenic crops. To improve long-term monitoring of transgenic crops, research is needed to fill knowledge gaps related to ecological indicators. Important research areas include temporal behavior of ecological indicators, and the need for improvements in mechanistic understanding of the links between ecological dynamics at population and other levels and the ecosystem process measures frequently used as indicators. The latter area is particularly important to ensure that ecological indicators will be able to resolve the ecological effects of transgenics from the effects of other factors that may be associated with the use of transgenic crops. There also are important research questions about the structure and organization of scientific efforts to accomplish long-term monitoring. For example, how should feedback loops be established between validation efforts and long-term monitoring efforts so as to clarify the adequacy of available ecological indicators for monitoring of ecological effects of transgenic crops? Finally, research is needed on processes for establishing long-term monitoring indicators in relation to transgenic crops as well as research to evaluate the efficacy and effectiveness of potential long-term indicators.
Trained observer monitoring attempts to detect effects the transgenic crops that are so poorly understood—or simply unforeseen—that they cannot yet be the target of more specific monitoring efforts. The goal is to develop a network of trained observers that could detect effects of this sort. The outstanding research needs related to this kind of monitoring concern questions about what sort of practical approaches could serve this purpose, as well as the broader challenge of “posting” observers that can monitor environmental effects of technological change in agriculture generally, since it is unlikely that monitoring aimed solely at transgenic effects will be cost effective.
Specific research needs include improving understanding of the organization and facilitation of observer activities that are or could be performed by existing organizations that have a current or potential interest in performing relevant ecological monitoring (e.g., the Christmas Bird Count of the National Audubon Society). Another important issue is the structure and functioning that would provide integration, organization, and development for the trained observer network. Many questions remain about how such a network would perform its functions—for example, proactively conducting two-way communication about monitoring issues with interested parties in organizations that do monitoring, developing curricula for training sessions for observers, providing a clearinghouse for reports of notable observations from field monitoring efforts, and developing quality assurance standards.
Improved Transgenic Methods to Reduce Risks and Improve Benefits to the Environment
Some of the concerns raised about early biotechnology-derived crop varieties can be avoided in the future with our expanding knowledge base. For example, pollen flow issues could be reduced if there are advances in plastid transformation, if sterility systems are engineered into the varieties, or if gametophytic factors (i.e., genes expressed in the pollen or egg) can be used to restrict fertilization of the egg. As nontissue culture methods are developed for gene transfer, such as cocultivation of developing flowers with A. tumefaciens (see Chapter 1), gene transformation in a number of plant species might become much less genotype dependent, perhaps even to the point of using an elite cultivar for the initial transformation event, thus avoiding linkage drag as the result of subsequent breeding procedures. In addition, tissue culture itself often introduces unwanted and unpredictable genetic and epigenetic variation to transgenic plants that such methods may avoid.
Alternatively, different promoters could allow greater control of where in the plant the gene product is produced; tissue-specific promoters could preclude expression of pesticidal proteins, for example, in pollen and other tissues. Temporal and spatial regulation of gene expression also can be controlled sometimes via an exogenous inducer. The ability to identify and transfer genes from elite germplasm collections of the same species could dramatically improve as a result of knowledge gained from genomics and proteomics research. Antisense technology will improve our ability to knock out the function of expressed genes. Targeted recombination and gene replacement technologies can be expected to improve, which will allow greater precision in gene insertion following transformation. Our ability to produce primary transformation events with selectable markers that can be removed through genetic crosses or that make use of benign (more acceptable) selectable markers will be enhanced. Improvements in DNA sequencing already allow the full sequence analysis of primary transformation events, such that optimal ones with simple single-gene insertions are selected early in the process of engineering transgenic traits.
When new technologies are involved in large-scale social changes, public debate and opinion formation on value-based issues can be facilitated by research that articulates and analyzes ethical, legal, and cultural traditions as they might bear on novel questions, as well as more traditional economic research on costs, benefits, and likely effects on the prices
and availability of food. Such research helps both scientists and the public understand what is at stake and how the issues might be viewed from differing perspectives. The potential for economic and sociological research has been well established in agriculture, and there is clearly a need for more studies that focus on biotechnology. The potential for philosophical, legal, and culturally-oriented research on issues in involving genetics has been demonstrated by the Ethical, Legal, and Social Issues (ELSI) program associated with the Human Genome Initiative. This program has produced both scholarly studies and educational materials that help people go beyond initial reactions of enthusiasm or repugnance. The USDA has funded little research of this kind, and few U.S. colleges of agriculture offer training or coursework to prepare professionals for ethical decision making—coursework that is now routinely offered in medical schools. The possibility cannot be dismissed that at least some of the controversy and turmoil that have greeted genetic engineering in agriculture is due to this omission. The USDA should encourage the development of a systematic program of research and teaching on ethical, legal, and cultural dimensions of agriculture and food issues, especially as they involve genetic techniques.
Recommendation 7.3: Significant public-sector investment is called for in the following research areas: improvement in risk analysis methodologies and protocols; improvement in transgenic methods that will reduce risks and improve benefits to the environment; research to develop and improve monitoring for effects in the environment; and research on the social, economic, and value-based issues affecting environmental impacts of transgenic crops.