Are Bt crops safe?

The US EPA's analysis of Bt crops finds that they pose no significant risk to the environment or to human health.


Federal oversight of Bt crops
Using a voluntary consultation process, FDA determines whether foods and animal feeds developed from GE crops with pesticidal traits are as safe as their conventional counterparts. It does this by determining whether the companies producing them have answered all the appropriate questions about the new plant varieties, such as whether new allergens are present and whether there are increased levels of natural toxicants or perhaps reductions of important nutrients. Any changes in nutritional properties or crop processing or the presence of new allergens could require labeling to inform consumers of the important changes to the food or feed.
The USDA is responsible for protecting US agriculture against pests and diseases. All GE crops with pesticidal traits are considered plant pests until USDA concludes that the crop is not a plant pest and makes a determination of nonregulated status-that is, decides that the plant will no longer be regulated by USDA as a plant pest. Until that determination is made, the plants are subject to USDA oversight for importation, interstate movement and environmental release (for an outline, see ref. 2).
The EPA's oversight focuses on the pesticidal substance produced (such as Bt protein or δ-endotoxin) and the genetic material necessary for its production in the plant (such as cry genes). The EPA calls this unique class of biotechnology-based pesticides 'plant-incorporated protectants' (PIPs) and describes procedures specific for PIPs in "Procedures and Requirements for Plant-Incorporated Protectants" 3 . The EPA grants experimental use permits for field testing and registrations that permit the sale and use of pesticides in commerce under the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA) 4 . The EPA also issues tolerances or tolerance exemptions that permit pesticide residues in food and/or feed under the Federal Food, Drug, and Cosmetic Act (FFDCA) 5 .

The reassessment
The Bt Crops Reassessment was designed to ensure that the decisions about the renewal of these registrations were based on the most current health and ecological data. As such, the EPA incorporated recommendations made by the agency's FIFRA Scientific Advisory Panel (SAP), a US National Academy of Sciences (NAS) report on Genetically Modified Pest-Protected Plants issued in 2000 (ref. 6), and the findings of the 2000 administration-wide biotechnology review led jointly by the Council on Environmental Quality (CEQ; Washington, DC, USA) and the Office of Science and Technology Policy (OSTP; Washington, DC, USA).
During the reassessment, the EPA became aware of unexpected results from scientific studies and other information related to potential adverse effects on monarch butterfly populations and to the presence of an unapproved PIP in the US food supply. The agency reviewed data and consulted with experts regarding monarch butterfly safety and also worked with other US federal partners to respond to the reports of StarLink corn (Aventis' Cry9C corn) in the US food supply 7 .
The registration for StarLink corn was voluntarily cancelled and the registrations for Event 176 corn, the PIP variety most closely associated with effects on monarchs in the scientific literature, were allowed to expire while the reassessment efforts were proceeding. On the basis of the lessons learned from the StarLink episode, the EPA anticipates that the type of split pesticide registration that allowed StarLink to be used in animal feed, but not in human food, will no longer be considered a regulatory option.
Nine Bt crop PIPs had been registered by the EPA under FIFRA as of October 15, 2001; of these, the four still registered but expiring Bt crops were reassessed. Although the Bt Cry3A potato registration was not reassessed because its registration is nonexpiring, summary results were presented in the Agency's Fall 2001 reassessment. Data requirements for Cry1Ac cotton, two Cry1Ab corns and Cry1F corn are shown in Box 1.
In each case, a detailed scientific assessment of the Bt crop was undertaken to characterize each product ( Table 2). Corn products registered at the EPA were transformed by protoplast electroporation to introduce the desired DNA or by methods involving bombardment of particles coated with DNA encoding the intended insert. Agrobacterium tumefaciens-mediated transformation was used for both cotton and potato products.

Human health assessment
Bt plant-incorporated protectants are proteins. Commonly found in the diet, proteins present little risk, except for a few welldescribed cases (such as food allergens, acute toxins and antinutrients). In addition, for the majority of Bt proteins currently registered, the source bacterium has been a registered microbial pesticide previously approved for use on food crops without specific restrictions. Because of their use as microbial pesticides, a long history of safe use is associated with many proteins found in these Bt products.
The EPA requires several types of data for the Bt plant-incorporated protectants to provide a reasonable certainty that no harm will result from the aggregate exposure to these proteins. The information is intended to show that the Bt protein behaves as would be expected of a dietary protein, is not structurally related to any known food allergen or protein toxin and does not show any oral toxicity when administered at high doses. These data consist of an in vitro digestion assay, amino acid sequence homology comparisons and an acute oral For all Bt crops: • Analytical methods for detecting Bt residues in commerce.
• Protein expression level data in various plant organs, (expressed in terms of dry weight for consistency among different PIPs).
• Protein levels in soil.
• Field data regarding possible impacts on nontarget insects.

For Bt corn:
• Monarch butterfly studies evaluating fitness and reproductive costs from subchronic exposure to Bt corn.
• Insect resistance management data regarding (1) potential for north-to-south movement of Helicopvera zea (a polyphagous pest known as the corn earworm when a pest of corn and as the cotton bollworm when a pest of cotton), as movement of H. zea exposed to Bt from the corn belt and their overwintering in cotton regions could affect resistance; (2) impact of conventional chemical insecticide use on the effectiveness of a refuge producing susceptible insects; and (3) development of discriminating concentration bioassay for Cry1f corn to help in monitoring for resistance in European corn borer, corn earworm and southwestern corn borer.
For Bt cotton: • Insect resistance management data regarding (1) potential for north-to-south movement of cotton bollworm; (2) alternative plant hosts, to demonstrate whether they serve as an effective refuge in generating Bt susceptible insects; and (3) insect resistance management (IRM) value of sprays with different chemical insecticides used in conventional and Bt cotton.

For Cry1Ab corn and Cry1Ac cotton:
• Comparison of amino acid sequence to known toxins and allergens via stepwise 8-amino-acid analysis.
For MON810 Cry1Ab corn • Processing and/or heat stability data.

Box 1 Data required from EPA reassessment of Bt crops
For the reassessment, the EPA required companies/applicants to provide the following data: toxicity test. The acute oral toxicity test is done at a maximum-hazard dose using purified protein of the plant-incorporated protectant as a test substance. Because of limitations in obtaining sufficient quantities of pure protein test substance from the plant itself, an alternative production source of the protein is often used, such as the B. thuringiensis source organism or an industrial fermentation microbe.
The EPA believes that protein instability in digestive fluids and the lack of adverse effects using the maximum-hazard dose approach eliminate, in general, the need for longer-term testing of Bt protein plantincorporated protectants. Dosing of animals with the maximum-hazard dose, along with the product characterization data, should identify potential toxins and allergens and provide an effective means to determine the safety of these proteins.
In vitro digestibility assay. The in vitro digestibility test confirms that the protein is unstable in the presence of digestive fluids and that it is not unusually persistent in the digestive system. The digestibility test is not intended to provide information on the toxicity of the protein or imply that similar breakdown will happen in all human digestive systems. The assay may also provide information about the potential of a protein to be a food allergen. A limitation of the test is that it usually only tracks protein breakdown to fragments still recognized by the immunological reagents employed.
Although only gastric fluid is typically tested, because Cry protein is known to be stable in intestinal fluid, in the initial Bt products registered, gastric and intestinal fluids were examined separately. To track the breakdown of the product, the proteins are added to a solution of the digestive fluids and a sample is either removed or quenched at given time points (usually at time 0, one to several minutes later and one hour later). The samples are then either subjected to electrophoresis in a sodium dodecyl sulfate-polyacrylamide gel (SDS-PAGE) and further analyzed by western (immunologic) blotting, or tested in a bioassay using the target pest. Each of the currently registered Bt proteins were tested and all were degraded in gastric fluid in 0-7 minutes.
Heat stability and amino acid homology. Two additional characteristics that may indicate possible relation to a food allergen are a protein's ability to withstand heat or food processing conditions, and its amino acid sequence as compared to those of known food allergens. For a few protein plantincorporated protectants registered to date, information is available on the heat or processing stability of the δ-endotoxins, as indicated by bioactivity or immunological recognition after typical food processing. The Cry1Ab protein in one corn product and the Cry1Ac protein were demonstrated to be inactive in processed corn. A full-length amino acid sequence homology comparison for one Cry1Ab product against the database of known proteins (allergens and gliadins) has been formally reviewed by the EPA.
Acute oral toxicity. Acute toxicity testing relies on the fact that toxic proteins generally express toxicity at low doses. Therefore, when the protein plant-incorporated protectants have no apparent effects in the acute oral toxicity test, even at relatively high doses, the proteins are considered nontoxic. The acute oral toxicity test is performed in mice with a pure preparation of the plant-incorporated protectant protein at doses from 3,280 to 5,000 mg per kilogram body weight. None of the tests performed to date have shown any significant treatmentrelated effects on the test animals.
Health conclusions. The mammalian toxicity data gathered by the EPA currently are sufficient to support the Bt plant-incorporated protectant registrations. None of the products registered at this time, all of which have tolerance exemptions for food use, show any characteristics of toxins or food allergens.

Insect resistance management
The unrestricted use of Cry1Ab and/or Cry1F in corn is likely to lead to the emergence of resistance in target insect pests unless measures are used to delay or halt its development. As some pests attack more than one crop, not only would the emergence of resistance affect the benefits of the Bt crop, but it also could affect the efficacy of Bt microbial formulations. The loss of Bt as an effective pest management tool could have adverse consequences for the environment to the extent that growers might shift to the use of more toxic pesticides and a valuable tool for organic farmers might be lost. The emergence of resistance could also have significant economic consequences for growers of Bt crops. Therefore, the EPA continues to require the registrants to implement an insect resistance management (IRM) program to mitigate the possibility that pest resistance will occur.
Certain measures are required to delay or halt resistance from developing for Bt corn and Bt cotton. These include planting of a non-Bt refuge in conjunction with the planting of any acreage of Bt field corn or cotton ( Table 3); agreements with growers which impose binding contractual obligations on the grower to comply with the refuge requirements; grower education; compliance assurance programs; monitoring for changes in target insect susceptibility to Bt Cry proteins; remedial action plans regarding measures the companies would take in the event that any insect resistance was detected; and annual reports on sales, IRM grower agreements results, compliance and educational programs. The companies registering the PIPs are responsible for seeing that that these measures are taken; failure of a farmer to follow the required IRM plan (measures) could result in the farmer losing the right to buy Bt seeds.

Environmental assessment
The EPA has conducted an environmental reassessment of the registered Bt plantincorporated protectants. The general topics covered include gene flow and the potential for weeds to develop if pollen from Bt crops plants were to fertilize other plants; horizontal gene transfer; expression of Bt Cry proteins in plant tissues; ecological effects, especially considering the available data on monarch butterflies; and fate of Bt Cry proteins in the environment Gene flow and weediness. Under FIFRA, the EPA has reviewed the potential for gene capture and expression of the Bt endotoxins by wild or weedy relatives of corn, cotton and potatoes in the United States, its possessions or territories. Bt plant-incorporated protectants that have been registered to date have been expressed in agronomic plant species that, for the most part, do not have a reasonable possibility of passing their traits to wild native plants. Feral species related to these crops, as found within the United States, cannot be pollinated by the crops According to the registrant submission, before transformation, the purified tryptic core proteins from both plant the plasmid was digested with the endonuclease NotI to remove amp r . and microbe were similar in molecular weight Although no data were submitted to confirm removal of the amp r gene (by SDS-PAGE), western blot, ELISA, partial from the transforming DNA, subsequent analysis by the applicant showed amino acid sequence analysis, lack of that amp r was not present in Bt11 corn genome. The cry1Ab gene was glycosylation, and bioactivity against also altered to increase its GC ratio for expression in corn and to increase European corn borer and corn earworm. This its GC ratio for expression in corn and to truncate the original protein analysis justified use of microbially produced (corn, potato and cotton) because of differences in chromosome number, phenology (that is, periodicity or timing of events within an organism's life cycle as related to climate, e.g., flowering time) and habitat. The only exception is the possibility of gene transfer from Bt cotton to wild or feral cotton relatives in Hawaii, Florida, Puerto Rico and the US Virgin Islands. The EPA has restricted the sale or distribution of Bt cotton in these areas to prevent the movement of the registered Bt endotoxin from Bt cotton to wild or feral cotton relatives. Horizontal gene transfer. The EPA has evaluated the potential for horizontal gene transfer from Bt crops to soil microorganisms and has considered possible risk implications if this occurred. Several experiments published in the scientific literature have been conducted to assess the likelihood of horizontal gene transfer and have not detected gene transfer under typical conditions. Horizontal gene transfer has only been detected under conditions designed to favor transfer. In addition, the genes that have been engineered into the Bt crops are mostly found in, or have their origin in, soil inhabiting bacteria. Therefore, the EPA concluded that horizontal gene transfer is at most an extremely rare event and that the traits engineered into the Bt crops are already present in soil bacteria or are unlikely to have selective value for soil microorganisms.

Environmental exposure
For each of the four Bt crops, the nominal protein expression levels as determined by field and/or greenhouse conditions are described in Table 4. The Bt protein values reported by each company may vary as a result of differences in the antibody-based reagents used for quantifying the Bt protein.
There are also differences caused by reporting Bt protein values based on tissue fresh weight. Although these differences may make it difficult to compare directly the tissue expression levels reported by different companies, the reported levels provide enough information for risk assessment purposes, especially when considered along with the reported tissue bioactivity values.
Soil. Soil organisms may be exposed to Cry proteins from current transgenic crops by exposure to roots, incorporation of above-ground plant tissues into soil after harvest or pollen deposition on the soil. Root exposure may occur by feeding on living or dead roots-or, theoretically, by ingestion or absorption after secretion of Cry proteins into the soil. In addition, evidence suggests that some soil components, such as clays and humic acids, bind Cry proteins in a manner that makes them recalcitrant to degradation by soil microorganisms, but without eliminating their insect toxicity. Therefore, exposure to Cry proteins bound to soil particles may also be a route of exposure for some soil organisms.
The possible accumulation of Cry proteins has been examined by determining degradation rates of Cry proteins, either in isolation or as expressed in the plant tissue and incorporated into the soil at a single point in time. Estimates of total Cry protein incorporated into the soil have been based on the biomass of total plant tissue, although it is not clear whether root biomass has been included in these calculations.
Most of the Cry protein deposited into soil by Bt crops is degraded within a few days, although a residue may persist in biologically active form for a much longer period of time ( Table 5). It is also reported that the same amount of Bt Cry protein persists in soils that have been exposed to repeat Bt spray applications when compared to soil exposed to Bt crops. Although field tests of Cry protein degradation in soil under a range of conditions typical of Bt crop cultivation are needed to provide relevant data on persistence and natural variation, the limited data available do not indicate that Cry proteins have any measurable effect on microbial populations in the soil. Current studies of Bt in soil show no effect on bacteria, actinomyces, fungi, protozoa, algae, nematodes, springtails or earthworms. In addition, new plants planted in Bt Cry protein-containing soil do not take up the Bt protein.
Effect of Cry1Ab and Cry 1F corn on nontarget wildlife. In light of concerns that commercialization of Bt crops will effect the environment, the EPA reviewed new and existing data regarding nontarget wildlife effects for Bt corn with a special emphasis on Lepidoptera and monarch butterflies, and re-evaluated the data to support continued registration of Bt crops. The weight of evidence from the data reviewed indicated that there is no hazard to nontarget wildlife from the continued registration of Bt corn ( Table 6).
The toxicity of Bt to butterflies is a well known and widely published phenomenon. For the purpose of its original risk assessment of Bt plant products, the EPA accepted Pink bollworm cotton Refuge embedded in Bt cotton field 6-10 rows Bt/1 row non-Bt In corn, the 20% refuge is required in areas outside cotton-growing regions, the 50% refuge in cotton-growing regions. This is because of the polyphagous pest Helicopvera zea, known as the corn earworm when a pest of corn and as the cotton bollworm when a pest of cotton. that Bt proteins could be toxic to Lepidoptera and relied exclusively on data on lepidopteran exposure to Bt Cry protein. Because exposure to butterflies and moths from the agricultural uses of Bt was not expected to be as high as that from the forest spraying of Bt for pests such as the gypsy moth (where no widespread and recurring or irreversible harm to lepidopteran insects was observed), Bt crops likewise were not expected to cause widespread or irreversible harm to nontarget lepidopteran insect populations. The weight of evidence of currently published research data reviewed indicates that milkweeds in the corn fields and within 1 meter of cornfields are unlikely to be dusted with toxic levels of Bt pollen from the currently registered Bt corn varieties, MON810, Bt11 and TC1507. In addition, a variety of factors-the distribution of corn pollen within and outside corn fields, the distribution of milkweeds within corn habitat and other types of habitat, monarch oviposition and feeding behavior, limited temporal overlap between monarch larvae and pollen shed (and similar issues) in much of the corn growing regions of the United Statesindicate a low probability of adverse effects of Bt corn pollen on monarch larvae.
Data available to date indicate no difference in the number of total insects or the numbers of insects of specific orders between the transgenic crop plots and either the isogenic or the wild-type control crops. No shift in the taxonomic distribution of insects was seen, except in cases where the predators are dependent on the pest insect as prey as their major food source.
Toxicity data show that the only endangered species of any potential concern are in the Lepidoptera. The majority of endangered species in this order have very restricted habitat ranges, and do not feed on Bt crops or approach the planting areas closely enough to be exposed to toxic amounts of Bt pollen. Potential concern regarding range overlap with corn production was restricted to the Karner blue butterfly. However, the Karner blue host plant, the wild lupine, does not occur in corn fields and it appears highly unlikely that significant numbers of lupine would occur within a few (2) meters of corn field edges, where the toxic levels of corn pollen may be present. Moreover, there is only limited overlap between the time of the year when corn pollen is shed and the times when Karner blue larvae are likely to be present.
Effect of Cry1Ac cotton on nontarget organisms. The EPA determined that the nontarget organisms most likely to be exposed to the protein in transgenic cotton fields were beneficial insects feeding on cotton pollen and nectar and upland birds feeding on cotton seed. Thus, tests were required using representatives of those organisms ( Table 6). Waterfowl, fish and aquatic invertebrate tests were waived because of probable lack of exposure. Studies on the effects of earthworms were not required. It was originally thought that because long-term exposure of soil organisms such as earthworms is possible when crop residues are incorporated or left upon the soil surface, the EPA would require studies evaluating effects upon earthworms. Data submitted indicate that Cry protein production ceases at senescence, allowing some time for protein degradation before harvest. Additionally, as the environmental fate data indicate that only 1.44 g of Cry1Ac protein per acre would enter the soil as a result of post-harvest incorporation of Bt cotton, and such proteins degrade rapidly, the potential for effects to nontarget soil organisms is not anticipated. Thus, an observable deleterious effect on earthworms is not expected to result from the growing of Cry1Ac-containing cotton plants.
Data available to date indicate that the transgenic cotton lines had no significant effect on populations of beneficial predator insects. However, the impact of chemical spray drift clearly affected the abundance of beneficial insects.
Cotton is an insect-pollinated crop, and only very small amounts of pollen containing the Cry1Ac protein can drift out of fields. Pollen containing Cry1Ac protein, at relatively very high dosages, was not toxic to the test species representative of organisms likely to be exposed to such pollen (e.g., lady beetles, green lacewings, honeybees). The habitats of the larvae of endangered Lepidoptera species in cotton-growing counties (Quino Checkerspot butterfly, Saint Francis' Satyr butterfly and Kern Primrose Sphinx moth) do not overlap with cotton fields. Hence, none of these larvae feed on cotton and thus they will not be exposed to Cry protein in pollen. The amount of pollen that would drift from these cotton plants onto plants fed upon by endangered or threatened species would be very small (if measurable) compared to the levels fed to the test species (Table 6). Therefore, the EPA does not expect that any endangered or threatened species will be affected by pollen containing the Cry1Ac protein.
In addition, because the EPA is imposing conditions for geographic areas that have sexually compatible wild or weedy relatives of cotton, the Cry1Ac protein gene cannot escape into related wild plants that could serve as a source of Bt pollen for plants on which endangered or threatened species may feed on in these areas. Because the EPA expects that no listed endangered species of Lepidoptera will be exposed to the Bt Cry protein expressed in cotton plants, and because the most probable exposure scenario does not appear to affect listed species, the agency believes that Cry1Ac Cotton will have no effect on listed species.

Cry3A potatoes on nontarget wildlife
Data presented in Table 6 indicates that Bt potato has no adverse effects on nontarget wildlife likely to be exposed to the crop. In addition, the data available to date indicate that beneficial arthropods were substantially more abundant in plots containing genetically modified potato plants and microbial Bt toxin applied to plant foliage than in those treated with conventional chemical insecticides. Aphid control was achieved in the plots containing transgenic potatoes solely through predation by natural enemies, whereas aphid populations rose to high levels in plots where beneficial arthropods were eliminated as a result of the conventional chemical insecticide treatment and no chemical aphid control was applied. The EPA has determined that Cry3A potatoes will not affect any threatened or endangered species. The known host range for the Cry3A protein is restricted to Coleoptera species. The listed coleopteran threatened or endangered species in potatogrowing areas are the American burying beetle, Hungerford's crawling water beetle, Mount Hermon June beetle, Northeastern Beach Tiger beetle, Puritan Tiger beetle and Valley Elderberry Longhorn beetle. These will not be exposed to Cry3A protein because their habitat does not overlap with potato fields and/or their larvae do not feed on potato tissue and will not be exposed to Cry protein in pollen or to toxic Cry3A levels in the soil. The amount of pollen that would drift from the potato plants onto plants fed upon by endangered or threatened species can be expected to be very small compared to the levels fed to the test species. Submitted data confirm that some coleopteran species tested are not affected, including lady beetles. Generally potato plants do not produce large amounts of pollen, which limits exposure. No endangered or threatened avian species feed on potatoes and no aquatic species are known to feed on potato plants.

Conclusions
In the fall of 2001, the EPA completed a comprehensive reassessment of the timelimited registrations for all existing Bt corn and cotton PIPs. As part of this reassessment, the agency decided to extend the registrations with additional terms and conditions, including requiring confirmatory data to ensure protection of nontarget organisms and lack of accumulation of Bt proteins in soils, measures to limit gene flow from Bt cotton to wild (or weedy) relatives, and a strengthened IRM program, especially in regard to compliance.
The Bt cotton registration is now set to automatically expire on September 30, 2006 except for the external, unsprayed refuge option, which will expire September 30, 2004. The Bt corn registrations are now set to automatically expire on October 15, 2008.
This reassessment was designed to assure that the decisions on the renewal of these registrations were based on the most current health and ecological data, and that the process was conducted in an open and transparent public process that incorporated sound and current science and substantial public involvement.