Biotechnology Project
Center for Science in the Public Interest
  Comments to EPA Opposing the Exemption from Regulation of Several Categories of Genetically Engineered Pesticidal Plants
    The U.S. Environmental Protection Agency (EPA) recently published a new rule (PIP rule) formalizing its role in regulating genetically engineered plants. EPA now proposes possible categorical exemptions from that rule that would weaken protection of the environment. CSPI opposes those categorical exemptions. EPA is the primary agency responsible for assuring food and environmental safety of crops that are genetically engineered (GE) to produce substances that protect plants from pests (plant pest protectants or plant pesticides). Such crops include those engineered to contain Bt toxins, primarily corn and cotton so far, including the infamous Starlink. While the proposed exemptions would not include Bt crops, they would include other pesticidal GE plants, such as many of those that protect against viruses. Several of those GE crops have already proven extremely useful or show considerable promise. Therefore such GE crops may become much more prevalent and important in the future, and concomitantly may provide greater opportunity for environmental harm. The comments below arguing against the proposed exemptions were submitted to EPA as part of its decision making process on whether to adopt the proposed exemptions.

Comments on Plant-Incorporated Protectants (Formerly Plant Pesticides), Supplemental Proposal; Docket Number OPP-300370B

Center for Science in the Public Interest (“CSPI”) submits the following comments in response to EPA solicitation for additional comment on proposed alternative regulatory approaches for certain categories of Plant-Incorporated Protectants (hereafter PIP), originally proposed in 1994.

If found acceptable, some or all of those alternatives proposals might be included in a revised version of the current PIP rule. CSPI argues in these comments that none of the proposed categorical exemptions should be adopted.

Executive Summary

Center for Science in the Public Interest (CSPI) opposes the modification of the current PIP rule by adoption of any of the proposed alternatives. In issuing its PIP rule, EPA has established that transgenic PIP require regulation to prevent unreasonable adverse effects to people or the environment. In so doing, CSPI believes that the agency establishes the need to demonstrate in a convincing manner that the proposed categorical exemptions could be accomplished while still fulfilling the agency’s mandate for preventing unreasonable harm. We argue that EPA has not demonstrated that members of the proposed exempted categories would not cause unreasonable harm. In fact, some PIP from the proposed categories are as likely to cause harm as PIP plants from non-exempted categories, and pose many of the same concerns. Therefore, the proposed categories of PIP do not merit exemption.

Some of the reasons that PIP in the proposed exempt categories may present unreasonable harm to the environment are set forth below.

  • Out-crossing of the transgenic PIP traits may cause a competitive advantage to wild relatives of the crop, making them more aggressive weeds.
  • Some crops, especially those that are relatively undomesticated, may themselves become more weedy.
  • The PIP trait, in the crop or a wild relative, may cause adverse effects to non-target species.
  • The cloning process may introduce unexpected or pleiotropic effects due to tissue culture, insertional mutagenesis, or interactions of the transgene or protein with the new genetic background.

Additionally, EPA’s conclusion that the PIP in the proposed exemption categories would not cause unreasonable risks is based on the agency’s ability to anticipate possible risks caused by those PIP and based on current knowledge. However, in the case of several registered PIP discussed below, EPA has been unable to anticipate possible risks during the registration process. Had those PIP been exempted, EPA would have had fewer regulatory options to address those possible risks. Those examples illustrate both the difficulty of anticipating potential unreasonable risks in PIP, and the importance of maintaining regulatory oversight.

CSPI does not propose that all or even most regulated PIP would cause unreasonable harm. But the potential for such harm clearly exists and sufficient experience to predict such harmful cases in advance, and thereby exempt them from regulation, has not been acquired.

An alternative to exempting the proposed PIP that allows an appropriate level of regulatory oversight would be a modified version of the tiered risk evaluation model familiar to EPA. Under a tiered system, demonstration of minimal risk in the first tier obviates or reduces the need for higher tier testing that is often more time consuming and costly to all involved. That approach would allow adequate regulatory flexibility combined with case-by-case evaluation needed to adequately regulate this new and powerful technology.

The following “General Considerations” apply to all of the proposed exempt categories, while further comments apply to the specific sections of the Federal Register notice.

General Considerations

There are well founded scientific reasons to suppose that PIP within the alternative categories could produce unreasonable risks to humans or the environment as defined by FIFRA and FFDCA. Many of those reasons were addressed in the National Research Council/National Academy of Science 2000 report, “Genetically Modified Pest-Protected Plants” (hereafter NAS report) (7). CSPI agrees with that NAS report that EPA has not provided a sufficient scientific basis to categorically limit the regulation of PIP by adopting the proposed alternatives. Our review of the scientific literature upon which the NAS report is based as well as subsequent scientific literature does not suggest that the NAS assessment should be altered.

Additionally, our knowledge of the complex interactions between transgenic PIP and the environment is currently too incomplete to anticipate the risk scenarios that may arise in individual cases, arguing against categorical exemptions. The likelihood that PIP of the proposed exemption categories could cause unreasonable risk cannot be directly addressed because few such PIP have been commercialized, and those are only grown on limited acreage. However, the principles upon which these categories of risk are based are sufficiently well established that they should be accepted as reasonable.

Until EPA can address the concerns raised in these comments, it should proceed cautiously, and regulate all transgenic PIP using the conditional registration process, rather than by blanket exemption that would preclude most effective regulatory action.

General Considerations: Risks Due to Out-Crossing

In all three proposed exemption categories, out-crossing of the PIP trait or traits to wild relatives of crop plants might cause a selective advantage, causing or enhancing the weediness of the related species. This may cause unreasonable risk directly to other wild or crop plants by preventing or reducing their growth, or indirectly to non-target organisms that are not plants (see “Risks due to adverse effects to non-target organisms other than plants” section below). Many domestic crops such as sunflower, wheat, squashes, rice, sorghum, etc. have wild relatives in the U.S., and out-crossing of conventional genes has been verified (3). Some of those weedy relatives, such as Johnson’s grass, are major non-indigenous invasive weeds. That example illustrates the additional concern that new wild relatives may continue to be introduced to the U.S., providing additional opportunities for out-crossing in the future. Since the introduction of non-indigenous plants is commonplace, we cannot assume in all cases that wild relatives not currently present will remain outside the U.S.

It is widely agreed that in cases of gene transfer, environmental impact depends, in part, on whether the transferred trait confers a selective advantage to the wild relative. Scientists are not confident that they can always predict which traits would confer a selective advantage to a wild relative (9,14). But PIP traits, possibly increasing the fitness of wild relatives by increasing growth or reproduction through control of pest damage, generally have a reasonable possibility of conferring a selective advantage.

Most plants have a number of pests that likely contribute to keeping their populations in check. However, some individual pests may have a significant role reducing the fitness of wild species. Unfortunately, there are far too few data to allow predictions of the importance of any pest in reducing the weediness of a PIP without testing or examining that PIP individually. Therefore, blanket exemptions as proposed in several of the alternatives are not supported by the available data.

General Considerations: Risks Due to Increased Weediness of Crop Plants

The potential for increased invasiveness or weediness does not apply only to wild relatives of crops. Some crops that are not domesticated may be competitive in the non-agricultural environment, and PIP may enhance that competitiveness by the same means as were discussed above for out-crossing. While most crop species have been selected for properties such as non-shattering seed heads that make their survival outside of the highly modified agricultural environment highly unlikely, some crops, such as many timber or wood pulp species, do not differ substantially from wild varieties.

General Considerations: Risks Due to Adverse Effects to Non-Target Organisms Other than Plants

The ability of PIP to resist pests may pose a direct unreasonable risk to the environment by harming non-target organisms. For traditional chemical pesticides and many genetically engineered PIP, pest resistance is due to toxicity to the target pest and, unintentionally, to other organisms as well. In those cases, it is clear that there may be unreasonable risks to the environment. However, even in cases where the specificity of the PIP is extremely narrow, such as for viral coat protein resistance, indirect adverse effects may occur to non-target organisms. For example, increased weediness due to the viral coat protein PIP could reduce the population of a plant species necessary or important to the survival of a non-target herbivore.

While PIP that act primarily by affecting the plant generally do not directly harm non-target species (but see section on “PIP that Act Primarily by Affecting the Plant” below) by a toxic mechanism, those PIP may harm non-target species by eliminating or reducing an important food source by preventing feeding on weedy relatives. These possible adverse effects are discussed further in the appropriate sections below.

General Considerations: Unexpected or Pleiotropic Effects May Cause Unreasonable Harm

Unexpected or pleiotropic effects caused by the transformation process or transgene could also cause unreasonable harm The cloning process may introduce variation in the expression of indigenous genes or novel products due to the tissue culture process, insertional mutation by random genomic insertion of the cloned gene into the recipient genome, or unpredictable interactions of the new gene product with existing genes or gene products (1,10). As has been pointed out elsewhere, traditional breeding also has the potential to cause unanticipated changes in the recipient plant genome as well. A recent NRC/NAS report (7) notes several important examples of unexpected changes in toxicants in potato and celery caused by “traditional” breeding methods. However, we are aware of no attempts to determine whether such changes occur more or less frequently in transgenic plants compared to traditionally bred plants.

It has been suggested that the potential for adverse effects to occur during the breeding of crops not currently regulated by EPA argues that genetically engineered PIP, such as those in the proposed exemption categories, also should not be regulated. However, the overriding issue for determining the regulation of genetically engineered PIP should be the unreasonable harm standard, not whether other plants are currently unregulated.

Registrants and others have argued that during the process of screening the PIP plant to assess its agronomic and nutritional traits, potential adverse pleiotropic or unexpected effects are likely to be discovered, and those transformation events would be eliminated prior to commercialization. However, there is no assurance that the expression of those few screened traits are typically linked with the expression of harmful traits, and therefore the screened traits cannot act as indicators for harmful traits. While several plant toxicants and antinutrients are among the traits typically tested for during this process, those traits are not an exhaustive list of possible adverse traits. This is especially true for traits causing possible environmental harms, since the toxicants that are currently evaluated are based on human toxicity, which is not always synonymous with environmental risks.

General Considerations: EPA’s Record of Regulating Genetically Engineered PIP Illustrates how Difficult it is to Anticipate Environmental Harms From PIP

EPA’s history of regulating genetically engineered PIP shows that it is difficult to anticipate unreasonable risks for plants in complex ecosystems and suggests that important regulatory decisions should be made cautiously without categorical exemptions. Individual cases of unanticipated potential risks that were discovered after registration illustrate the difficulty of predicting all potential unreasonable risks that could arise from genetically engineered PIP of any type. In the case Bt corn and monarch butterflies, EPA failed to recognize that larval food plants (milkweed) were apparently present in many corn fields, presenting a possible Bt exposure scenario not previously considered. Even if other factors preclude harm to monarchs, such as low levels of Cry protein in pollen, or if Bt pollen is shed when monarch larva are not present in corn fields, that should not obscure the inability of the agency to anticipate the possible risks that have come to light since registration. If event 176 Bt corn had been more widely planted, its much higher level of expression of Cry1Ab in pollen may have caused unreasonable harm, at least in some regions of the corn belt.

EPA also did not adequately anticipate possible exposures to soil non-target organisms in initial risk assessments of currently registered products. In particular, EPA had initially determined that single soil incorporations of plant material would adequately represent possible exposure of soil organisms, without adequately considering possible continuous exposure through the roots (5,13). And EPA did not require data on the level of expression of Bt in the roots of all Bt crops even though this is a likely exposure route for soil organisms (5). Again, the important issue is not whether the new exposure considerations result in an unreasonable risk, but the inability to anticipate relevant exposure parameters prior to registration. If those Bt crops had been exempted from registration, the current evaluation processes may not have occurred.

General Considerations: Regulating the Proposed Exempt Categories can be Done in a Flexible and Cost Effective Manner

EPA desires the elimination of regulatory burden that would unnecessarily hamper the development of agricultural biotechnology. Of course, these considerations should be secondary to adequately protecting human and environmental safety. But EPA’s tiered risk evaluation process has the ability to incorporate reasonable flexibility into the regulatory process without imposing undue burden, while retaining needed regulatory oversight of the proposed exempted PIP. EPA can limit testing of a PIP or other biopesticides to minimal lower tier requirements if those tests find it unlikely that the pesticidal substance will cause unreasonable harm. Higher tier testing which can consist of more extensive, expensive, and longer term tests may then be foregone. EPA can design additional lower tier tests specific to the proposed exempt categories of PIP.

General Considerations: Stacked PIP May Cause Unreasonable Risk Where the Individual PIP Would Not

Perhaps more importantly, EPA has apparently not considered the additional potential selective advantage conferred by several stacked genes in the alternative PIP categories. Even if single genes in the proposed categories do not individually cause unreasonable harm, combinations of several such genes, each providing resistance to different pests, may substantially increase the likelihood of harm. Several commercial examples of multiply viral resistant plants (squash and potatoes) already exist. The technology for multiple trait transfers already has advanced compared to earlier transgenic crops (2), making stacked traits more likely in the future.

In general, those traits would often be integrated together at a single genetic locus to facilitate transfer to new crop varieties. Unfortunately that would also facilitate transfer of the entire locus of multiple traits by out-crossing to wild relatives. It is easily conceivable that multiple traits from proposed exempt categories, resisting several taxa of plant pests, could be included in a single plant. Such linkage of multiple different pest resistant traits is not usually found in nature. The reduced regulatory cost under the exemption proposals would provide incentive to make such crop plants.

General Considerations: Reliance on Adverse Incident Reports is Inadequate

CSPI believes that adverse incident reporting is not a sufficient means to assure adequate regulatory oversight for the proposed exemption categories of PIP. Such a passive approach will often not discover problems until they are well advanced, are more difficult to remedy, and possibly have caused considerable harm. In addition, reliance on adverse incident reporting alone would forego the current proactive and thorough reevaluation process afforded by the use of conditional registration of PIP.

Comments on III. Proposed Alternative Regulatory Approaches to Plant Incorporated Protectants Derived Through Genetic Engineering from Sexually Compatible Plants

Proposal 1. Exemption of all plant-incorporated protectants derived from plants sexually compatible with the recipient plant.

This proposed alternative exemption would largely remove PIP genes from EPA registration requirements under FIFRA if they originate in sexually compatible plants and where the “specificity or function of the pesticidal substance” is not altered. This proposal does not adequately consider possible unreasonable environmental harm due to higher expression of the PIP or expression in different parts of the plant or at different times compared to the same gene in the sexually compatible plant.

The exemption would apply even if the PIP was expressed at much higher levels than in the donor plant since the regulatory regions that determine how and when the transgenic proteins are expressed are explicitly excluded from this requirement of “no functional alteration". The PIP gene may be expressed at much higher or lower levels, at different stages of crop growth, and in different tissues than was the case in the donor crop or through conventional breeding. It is a widely accepted tenant of toxicology that “the dose makes the poison", which is in fact the general scientific basis that allows the setting of "safe” tolerance levels in conventional pesticides. The proposed exemption appears to ignore this fundamental principal of toxicology, which applies to non-target organisms as well as to humans, and would thereby exempt PIP that could cause unreasonable environmental harm.

It is argued by EPA that, for plants like forest plantation crops that are not as domesticated as many food and fiber crops, or in the case of wild relatives used for breeding new traits into established crops, the levels of expression of particular traits varies considerably, suggesting less concern for possible changes in expression levels due to genetic engineering of the trait. However, there is much greater variation in expression levels available from the thousands of plant or other promoters that may be used in genetic engineering than the variation in the strength of the promoter of a single gene, even if from non-domesticated plants. Further, some newer transformation processes, such as chloroplast transformation, can result in extremely high expression levels due to the high copy number of these organelles (2). Since most plant genes are carried on the chromosomes, not the plastids, such chromosomal genes would not typically have similar levels of expression.

EPA seems to recognize, in proposed tolerance exemption 174.479(d), that increased level of gene expression, as would frequently occur by the use of different promoters, needs to be considered in the case of human exposure through food. Higher expression can similarly increase toxicity to non-target organisms, and it is unclear why EPA does not acknowledge this fact.

However, level of expression is not the only possible unreasonable environmental harm due to cloning of genes from sexually related plants. Changes in the timing or tissue specificity of expression can also cause a profound difference in exposure of non-target organisms, which may be present at different times in the growing season, or are exposed to particular plant tissues of a crop or wild relative. Current consideration of the impact of pollen expressing Bt Cry proteins on monarch butterflies should serve as an important illustration of this issue, even if it turns out that Bt pollen does not have an adverse impact on monarch populations. It could be asked, for example, what the impact of protease or alpha-amylase inhibitors would be if transferred into a sexually compatible plant by cloning, and expressed in the pollen at high levels but where it was not so expressed in the donor plant? This would undoubtedly provide an entirely new risk scenario to non-target organisms than would have been the case through sexual transfer.

Comments on Proposal 2. Case-by case review of eligibility for exemption through notification process [of PIP from sexually compatible plants].

As an alternative to the blanket exemption discussed above, EPA proposes a notification process to determine whether the transgene is “substantially equivalent” in the recipient plant compared to the donor. This proposal does not adequately consider possible unreasonable risk from pleiotropic or background effects discussed in the “General Considerations” section above. The SAP would be used to develop substantial equivalence criteria, but presumably not determine the feasibility of the concept of substantial equivalence itself. It therefore puts the cart before the horse in proposing the notification process prior to vetting with the Scientific Advisory Panel (SAP).

Examination of several examples of unanticipated potential risks that were discovered subsequent to registration of current transgenic PIP products, also discussed in the “General Considerations” section, suggests that the evaluation of risks from GE plants is not advanced enough to predict all possible unreasonable environmental harms for any transgenic PIP, including those from the proposed exempt categories.

Comments on Proposal 3. Variant of notification process for broader group of plant-incorporated protectants.

For all of the considerations discussed above, this third variant proposal is not adequately supported by the currently available science. In addition to the above considerations, issues of potency differences and differences in substrate specificity for otherwise similar proteins must be considered. The science of protein structure-function relationships is not sufficiently advanced to allow predictions of functional differences based on structural, sequence based differences in the proteins, as would be required for this alternative proposal to be capable of giving adequate regulatory oversight. While EPA lists relevant criteria for establishing levels of similarity, it is unclear what a sufficient amount of testing would be. How many potential substrates should be tested, for example, to establish adequate similarity? Perhaps more importantly, since sequence identity between the donor plant gene and the recipient of sexually compatible species is presumably unlikely, how much dissimilarity would be acceptable and how would this be decided without field testing for environmental impacts or for toxicity? There are far too many fundamental questions unresolved to consider this proposal to be a viable alternative.

For all of the above reasons, CSPI believes that it is premature for EPA to either grant a categorical exemption of PIP derived from sexually compatible plants, or exemptions based on the proposed notification processes

Comments on Question 4. Does use of antibiotic or herbicide resistance or other selectable markers represent risk?

Antibiotic markers generally do not pose unreasonable risk in PIP

CSPI does not believe that antibiotic resistance genes incorporated into PIP pose an unreasonable risk. Although there is a possibility of horizontal gene transfer to bacteria in the soil or intestines of humans or livestock, experimental evidence suggests that the rate of transfer is extremely low. Therefore, for PIPs, antibiotic gene transfer should be a rare, if not nonexistent, event (8).

Even if horizontal gene transfer did occur, it is unlikely to cause unreasonable adverse effects. Several studies show that the antibiotic genes commonly used for selection already are prevalent in the environment (12). There are also alternatives available for clinical use of these particular antibiotics if the very unlikely event of resistance did arise.

While the risk from the use of antibiotic resistance markers is low for the commonly used genes, the existence of adequate alternatives suggest that it would be prudent to eliminate the use of antibiotic resistance genes in the future. For example, alternatives such as sucrose synthase are often available for bacterial selection. Alternatively, the PIP gene along with any other desirable DNA can be easily separated from bacterial antibiotic resistance genes, for example by restriction and separation by gel electrophoresis, prior to plant transformation.

On the other hand, while the likelihood of horizontal gene transfer is very low, the resulting risk due to such transfer may be unreasonable for antibiotics that are clinically important, that is for which there are few available or suitable alternatives (for example, having low levels of serious side effects). In addition, it may be less likely that such resistance genes are already widely dispersed in the environment. Use of those antibiotic markers should be avoided as a matter of prudence, especially since alternative selectable markers are available. Since homology with recipient bacteria greatly increases the rate of horizontal transfer, vector DNA of bacterial origin could also be eliminated, but presents limited risk. Similarly, vector DNA from transposable elements should be avoided because of the mobility of those elements.

Herbicide resistance genes may pose unreasonable risks

Herbicide resistance can present several risk issues if transferred to a weedy relative or if the crop itself can be a problem as a “volunteer” in succeeding crops or if otherwise transferred to fields of different crops. If the herbicide in question is important in controlling a weedy relative of the crop plant, this should be considered to be an important risk factor. Similar consideration should be given to the control of the crop itself as a volunteer weed. For example, it was recently reported that volunteer glyphosate resistant canola had become difficult to control in some regions of Canada. While other herbicides can sometimes be used to control volunteer canola, those that are available for particular rotation crops cannot be used under some circumstances. In this case, forced late planting in some seasons prevents the use of some otherwise effective herbicides, leaving no good control options (4). Glyphosate has lower persistence than many other herbicides, and so can be used closer to planting than those other herbicides even when the crop itself is not resistant. Furthermore, the issue of multiple herbicide resistance enhancing those concerns should be considered.

While we encourage EPA to consider herbicide resistance risk, the question must be asked as to whether EPA will regulate herbicide resistance as a primary genetically engineered trait as well. Otherwise, the agency will be faced with regulating herbicide resistance in one instance, and not regulating exactly the same property in another.

Comments on Question 5. Should protoplast fusion be included in the definition of wide cross?

EPA should further explore this issue. There is currently some anecdotal evidence, cited by EPA, as possibly poising unacceptable risk. However, there does not appear to be sufficient data to make a decision on this issue at this time. As above, similar issues have been raised concerning other relatively recent breeding practices such as tissue culture and random mutagenesis.

Comments on IV. Notice of Data Availability and Request for Comment

EPA notes that the recent NAS report referenced in the current Federal Register document appears to contain an internal contradiction. On one hand, the report is cited as suggesting that there are no categorical risks of genetically engineered PIP (GE PIP) that cannot occur in traditionally bred PIP plants. On the other hand, the report suggests that EPA should regulate genetically engineered PIP plants, including those containing genes from sexually related plants, while exempting traditionally bred plants. This dilemma has been noted elsewhere and may be called the product vs. process dilemma. Simple put, some suggest that the product of the crop production process should be regulated, not a process such as genetic engineering.

We have given a number of reasons in these comments about why GE PIP should be regulated. We find the product vs. process dilemma to present a false dichotomy. To suggest that product and process exist independently of each other is not realistic. The nature of products depend on the particular process used to make them. In the case of genetic engineering, while the categories of possible risk are similar to those from traditional breeding, the GE process greatly extends the range of possible products that have never been in crop plants, and in combinations that nature has not been exposed or adapted to. Thereby, the possibility of risks is extended compared to traditional breeding. Furthermore, a significant new technological process and its products necessarily present uncertainties not faced with older technologies that have a long history of use. The products of GE should therefore face a higher level of regulatory scrutiny than the products of traditional breeding. Further experience may reveal that GE presents no higher (or possibly lower) levels of risk than traditional breeding, but we have not achieved that experience at this time.

More importantly, the standard for determining whether to exempt the proposed categories of transgenic PIP should be whether such exemptions could cause unreasonable harm. Whether other currently exempt PIP categories may also cause unreasonable harm should be decided separately based on the characteristics of those PIP.

In addition, while breeders have substantial experience with identifying traits in conventionally bred plants that may be deleterious, such as alkaloids in solanaceous species, they do not have experience with the many new traits that can be introduced by genetic engineering.

Comments on V. Proposed Alternative Regulatory Approaches to Plant-Incorporated Protectants Based on Viral Coat Proteins (vcp)

Comments on 1994 Proposal Option 1, Exemption of vcp PIP from all FIFRA requirements except adverse incident reporting

An important environmental risk issue for viral coat protein (vcp) resistance PIP is that out-crossing may confer a selective advantage to wild relatives of the crop, and in so doing significantly increase "weediness". This issue is addressed in the “General Considerations” section. Weediness should be defined not only as increased economic impact on agricultural yield or quality, but also invasiveness of natural environments. The latter property can be described as the ability of the plant to increase its range at the expense of the populations of other species.

Option 1, which would exempt all vcp PIP from FIFRA regulation, would not provide adequate regulatory oversight since it would exempt PIP that could increase the weediness of wild relatives. However, the issue of potential increased invasiveness does not necessarily pertain only to wild relatives of crops. As discussed in the “General Considerations” section, some crops that are not domesticated may be competitive in the environment, and vcp PIP may enhance that competitiveness or weediness.

Comments on Option 2, More Limited Exemption

The issue of possible increased weediness of some crops also pertains to this option. In addition, as a practical matter, adequate fulfillment of the conditions of option 2 can be difficult. For example, the recent NAS report (7) criticized the environmental assessment of virus resistant squash by USDA/APHIS despite that agency’s attempt to determine whether virus resistance conferred a selective advantage to wild relatives. In particular, the APHIS assessment was criticized for not looking at sufficient populations of wild relatives for a long enough period of time or looking hard enough for viral infection. That PIP was not registered by EPA because it was exempted under the requirements of the provisional 1994 precursor to the PIP rule. Therefore, viral protected squash cannot be reassessed to determine whether, in light of recent concerns, it presents an unreasonable risk. By maintaining regulatory oversight, that is by not granting an exemption from most FIFRA requirements, EPA would maintain the ability to take mitigating action should unaccounted for risk arise after registration.

Therefore, while option 2 is a considerable improvement over option 1, it is not comprehensive enough to provide adequate environmental protection.

Comments on VI. Proposal on Plant-Incorporated Protectants that Act Primarily by Affecting the Plant

CSPI believes that there are possible unreasonable risks in this category resulting from out-crossing to wild relatives of crop plants. We are also concerned, as with vcp resistance, with situations where the crop can normally compete in a non-agricultural setting, as with many forest crops. The mechanism by which the plant resists the pest(s) is not the only concern when considering its potential for adverse environmental impact in these cases. Rather, the deciding factor is often whether the trait(s) confer a selective advantage to the plant. Potential adverse effects on herbivores that may depend upon wild relatives is also of concern. Some non-target insects have very restricted host ranges, and could be especially vulnerable to a new mechanism of resistance in the plant.

In addition, the word “primarily” is ill defined and therefore troubling. A number of mechanisms that would fall under the category of structural barriers are often part of metabolic cascades, such as those termed systemic acquired resistance (SAR) (6) that also include chemical and enzymatic defenses. In general, research into plant pest defense mechanisms is revealing that many individual defense mechanisms, be they structural, chemical or enzymatic, are typically induced together with other types of defense when the plant is attacked by pests. In other words, it appears likely that many of these defense mechanisms are coordinately regulated. Therefore, even if a pesticidal mechanism could be identified as strictly structural, there is concern that such a transgene could affect the regulation of other, non-structural defense mechanisms. It is unclear how EPA would address these complex defense mechanisms under this proposed exemption.

Lignin synthesis, for example, occurs in part from biosynthetic pathways that branch from some secondary metabolite synthesis pathways, where the latter is involved in chemically based pest defense. Trichomes, especially glandular trichomes, which are essentially barriers to many insects, can also be associated with secretory cells which produce pest defense chemicals (11). Will EPA, accept production of these other metabolites under the proposed exemption due to its interpretation of "primarily"?

In particular, the hypersensitive response is also associated with induced systemic resistance that involve toxicants and other mechanisms, not all of which are well characterized. Some plant hormones, such as ethylene and jasmonic acid, are known to be involved in the induction of multiple plant pest defense mechanisms, and should also not be exempted.

For the preceding reasons, CSPI does not believe that the categorical exemption for PIP primarily affecting the plant should be adopted.


We have shown that, while many individual PIP of the proposed exempted categories may be benign or beneficial, we do not have enough experience with these PIP to be able to predict the absence of unreasonable risk in any individual case, let alone as categories. In fact, we have given examples of PIP where new unanticipated potential risks have been discovered after registration, confirming the difficulty of predicting risk for transgenic PIP.

Categorical exemptions would remove much of EPAs ability to perform adequate regulatory oversight if unanticipated risks arise. Therefore, EPA should not adopt any of the proposed exemptions and should register all transgenic PIP on a conditional basis.

Doug Gurian-Sherman, Ph.D.
Co-Director, Biotechnology Project
Center for Science in the Public Interest
1875 Connecticut Ave., N.W.
Suite 300
Washington, D.C. 20009
202-332-9110 ext. 377


1. De Block, M. (1993) Euphytica 71:1-14

2. De Cosa, B et al. (2001) Nature Biotech. 19: 71-75

3. Ellstrand, N.C et al. (1999) Ann. Rev. Ecology Syst. P. 539-563

4. Entz, M. U. Manitoba, Personal telephone communication on June 21, 2001

5. EPA, Background Paper for Bt Crop Reassessment, Environmental Effects, Oct. 2000,

6) Garza R. et al.(2001) J. Econom. Entomol. 94 (4): 958-962 [abstract only cited]

7. National Research Council, “Genetically Modified Pest-Protected Plants: Science and Regulation”, 2000, p.261, National Academy Press, Washington, D.C.

8. Nielsen, K.M. et al. (1998) FEMS Microbiol. Rev. 2(2): 79-103

9. Perrins, J. et al. (1992) Acta-Oecol. International J. Ecol. 13: 517

10. Phillips, R.L. et al. (1994) Proc. Natl. Acad. Sci USA 91: 5222-5226

11. Ranger C.M. and A.A. Hower (2001) J. Econom. Entomol. 94 (4): 950-957 [abstract only cited]

12. Smalla, K et al (1993) FEMS Microbiol. Ecol. 13: 47-58

13. Saxena D. and Stotzky G. (2000) FEMS Microbiol. Ecol. 33: 35-39

14. Williamson, M. (1994) Molec. Ecol. 3: 75-79
Frequently-Asked Questions Articles and Speeches Regulatory Actions Reports and Reviews Take Action!