SCOPE 44 - Introduction of Genetically Modified Organisms into the Environment

This paper had its genesis both in the Bellagio Meeting that led to the present volume and in the challenge of a joint special issue of Trends in Ecology and Evolution and Trends in Biotechnology, both published by Elsevier. The basic points to be made on both occasions were the same, but it seemed important to have the views expressed represented in both contexts. Hence, this chapter is, with only minor modification, a reprint of the paper `Safety considerations for the environmental release of genetically engineered organisms', which appeared in Trends in Biotechnology/ Trends in Ecology and Evolution. A similar version also appeared in Chemical Technology and Biotechnology published by Elsevier.

The number of recent meetings on the topic of the release of genetically engineered organisms is impressive. Unfortunately, the large number of meetings and opportunities to reprint this paper does not represent a rapidly expanding body of knowledge on potential risks and benefits. Most of what needs to be said of a general nature has been said long ago. What is needed now are approaches that abandon the generalities that have been prevalent and that deal with the specifics of individual cases and categories. A number of recent efforts have shown progress in that direction, and this volume is intended to be the first of a series that explore the potential for genetic engineering to assist in developing improved approaches to environmental management and ecological and evolutionary research.

¹This article is reprinted from 'Planned release of genetically engineered organisms', in: Hodgson, J. and Sugden, A.M. (Eds.) Trends in Biotechnology/ Trends in Ecology and Evolution Special Publication, pp. S47-S49, April 1988, with permission from Elsevier Publications, Cambridge.
Introduction of Genetically Modified Organisms into the Environment Edited by Harold A. Mooney and Giorgio Bernardi Published 1990 by John Wiley & Sons Ltd.

The considerable debate that has been associated with proposals to introduce genetically engineered organisms into the environment obscures the fact that among most scientists there is strong basis for consensus concerning risks. Where disagreements do exist, they are concerned less with what the historical record says than with which part of it is most relevant to introductions of organisms developed with the new genetic techniques (Regal, 1988; Colwell, 1988). It is generally acknowledged that the overwhelming majority of introductions are likely to be benign, but that some will be problematical; nonetheless, the question can still be posed as to whether extrapolations should be based upon the average case, which will be low in risk, or upon the extreme high-risk cases, which will be low in probability.

Fortunately, such a simplistic question need not be answered, since it presupposes an artificial dichotomy. Most informed discussion recognizes that biotechnology holds tremendous potential for improving the human condition, including environmental applications that offer significant opportunities to reduce the pollutant load associated with the use of chemicals. There is general agreement that most introductions will pose minimal environmental risk and that thus there is no basis for generic arguments against the introduction of genetically modified organisms. On the other hand, there is no question that certain types of possible introductions pose a non-negligible degree of risk; thus, generic safety arguments must equally be rejected. Certainly, new genetic technologies will allow the creation of genotypes that never before have appeared and the introduction of genotypes into environments that are new for them. Thus, each potential introduction must be evaluated on its own merits. The inescapable conclusion is that some consideration is needed of the risks associated with the deliberate introduction of organisms, and furthermore that evaluation of risks must be specific to the particular application. This conclusion is universal in applicability. It is independent of whether or not one feels that existing legislation and procedures are adequate, an issue that requires different considerations for different nations.

13.1 WHY IS THERE DEBATE?

Given such strong elements of agreement regarding the basic issues, why has consensus been so long in developing? One reason, already suggested above, has been the inclination on all sides to generalize from special cases and the fact that the special cases of choice are not the same for everyone. Some refer to the long history of safe introductions associated with breeding in agriculture; others point to the problems that have been associated with the deliberate and accidental introductions of exotics. The relevance of these particular referents will be discussed below, but the differences of opinion and perspective point again to the necessity to get down to the specifics of individual cases and classes of introductions and to avoid broad generalizations.

A second and related reason for the debate reflects a difference in points of view. Some argue that the risks associated with the products of biotechnology are no different in kind than those associated with conventional procedures and that therefore no special attention is merited. Many who argue to the contrary accept the basic premise that there are no qualitative differences in risk, but emphasize the fact that there are risks in any introduction. They argue either for stricter controls on all introductions or that lax procedures regarding other types of introductions do not justify extending this laxity to an even broader class of introductions (Thompson, 1987).

A third reason that consensus has not been immediate is a specious philosophical dichotomy concerning who bears the burden of proof. Environmentalists point to the problems that often attend new technologies and argue that when new technologies are proposed the burden of proof regarding safety should rest upon the proposer (Levin and Harwell, 1986; Levin, 1986). On the other hand, proponents of biotechnology argue that the discussion of problems with deliberate introductions is hypothetical, dealing with the possible rather than the probable. They argue that there have been no examples of environmental problems associated with any introductions of recombinant-DNA organisms; indeed, only a handful of introductions have been carried out. Therefore, they conclude, the democratic process requires an assumption of innocence rather than guilt, and the burden of proof should rest upon those who construct unlikely scenarios for disaster.

Both of these lines of argument mislead. The fact of the matter is that we are not dealing with a new technology when we are dealing with the introduction of modified organisms into the environment. Humans have modified organisms genetically for centuries by a variety of techniques, and the repertoire of available tools has evolved and grown continually. Recombinant-DNA methods are the latest in the continuum of techniques, but do not represent anything inherently different from other advances that have occurred, except an ever-increased ability to carry out modifications efficiently and precisely. To the extent that increased efficiency leads ultimately to a change in the scale or frequency of introductions, that is worth consideration in terms of cumulative risks; indeed, it is this possible quantum increase in patterns of use that has been the major concern of many who have advocated caution. However, for any individual introduction, it is the characteristics of the engineered organism and its recipient environment that should be the basis for risk assessment, not the method by which the organism was modified (National Academy of Sciences, 1987). Thus, the appropriate line of argument, and one that leads much more rapidly to consensus, is as follows: We have sufficient understanding of the roles of genetics and environment in controlling the expressed phenotypic properties of an organism to allow us to state that it is not the method of modification that matters; what is relevant is the product and how it will be used. Therefore, as the recent US National Academy of Sciences report concludes, `The risks associated with the introduction of recombinant-DNA-engineered organisms are the same in kind as those associated with the introduction into the environment of unmodified organisms and organisms modified by other genetic techniques' (National Academy of Sciences, 1987). This does not argue that those risks should be ignored, but simply that the considerations are no different in kind than those that apply for other kinds of introductions. Thus, we are not operating in an information vacuum, and have a large body of experience to draw on in assessing risks.

Although there is a growing recognition that it is more appropriate to focus on the product and patterns of use rather than on the method of production and that this should be the most important principle underlying regulatory procedures, current practice in the United States and other countries has not completely adjusted. Recently, for example, the National Institutes of Health concluded that a researcher at Montana State University, Gary Strobel, had not violated their guidelines when he introduced a genetically modified bacterium into the environment as part of an effort to combat Dutch elm disease. The explanation was that the introduced organism, a non-pathogenic Pseudomonas syringae, was the product of mating a second organism, which was recombinant-DNA modified, with a Pseudomonas bacterium that was not. The end product, it was adjudged, technically was not recombinant. This legal legerdemain is the inescapable consequence of having decisions made by a committee, the Recombinant Advisory Committee (RAC), whose very title is based on method rather than product. The irony of the whole situation is that Strobel's introduction of a known pathogen, the Dutch elm fungus, as part of the experiment, attracted virtually no attention because the fungus was not genetically engineered. Although Strobel took precautions to prevent the spread of the fungus, its introduction undoubtedly posed a greater hazard than did the introduction of the bacterium that was designed to control it, and a rational regulatory system should recognize that.

13.2 WHAT ARE THE CONCERNS AND THEIR RELEVANCE TO RISK ASSESSMENT? 

13.2.1 INTRODUCED ALIEN SPECIES

One of the most compelling concerns for ecologists has emerged from their awareness of the problems that have sometimes attended the deliberate or accidental introduction of species taken from other environments. Charles Elton (1958) in his classic text on biological invasions, was the first to bring focus to this problem. Increasingly, in a growing number of environments, invasions are being recognized as serious threats to the preservation of what we choose (by our choice of time scale) to regard as native fauna and flora (Mooney and Drake, 1986). Although the great majority of accidental introductions undoubtedly fail to become established (the exact statistics are impossible to determine, since we generally do not know about failed introductions), a substantial number do become established, and some of these become serious pests. Examples are legion: the European rabbit in Australia, the kudzu vine in the United States, the muskrat in Europe, the Africanized bee in South America, and forest and agricultural pests, as well as human pathogens and disease vectors, world-wide. Yet the relevance of these examples to most cases of introduction of genetically modified organisms has been severely challenged, because the transfer of a species from one environment to another, lacking its co-evolved biotic controls, involves fundamentally different considerations than does the reintroduction of a slightly modified species into the environment from which it was taken.

Most introductions currently under consideration fall into the category of reintroductions, but certainly there will be introductions proposed that involve alien species or that involve engineering to expand the range of some desirable species, such as popular sport fish. To address this, Levin and Harwell (1986) introduced four categories of introductions, and argued that separate considerations must apply to each. Some, and probably most, introductions will be reintroductions of native species; for such introductions, the alien species model is not a very good one, and conventional experience in agriculture provides a much more appropriate basis for extrapolation. The other two major categories involve nonnative species that are not found in the target environment because either (1) they cannot survive there without continual supplement or (2) they can survive there, but never before have been introduced. It is the latter category that encompasses the destructive examples mentioned earlier and presents the greatest cause for concern. The final category mentioned by Levin and Harwell is a catchall for the remaining cases, namely those that would be sufficiently novel to have no close analogues in any environment. For these cases, no good basis for extrapolation exists.

13.2.2 PATHOGENIC MICROORGANISMS

It is well recognized that small genetic changes in pathogenic species can convert avirulent types into virulent ones, or can permit expansions of the host range. Such small genetic changes underlie the outbreaks of new strains of human diseases such as influenza or of fungal pathogens such as the black stem rust of wheat. Thus, engineering of pathogenic organisms must be carried out with the utmost precaution, and with proper attention to safety. Similarly, if one were engineering nonpathogenic organisms with the intention of altering properties related to pathogenicity, similar care would be mandated. However, if nonpathogenic organisms are being altered with respect to properties unrelated to pathogenicity or invasibility, the likelihood of accidentally producing a pathogenic organism is negligible; pathogenicity involves not one but a complicated suite of characters (National Academy of Sciences, 1987). Much of the confusion on this point centers around the definition of a pathogen. To a pathologist, an avirulent strain of a pathogen is still a pathogen and should be handled with appropriate care. However, a single bacterial species, such as Escherichia coli or Pseudomonas syringae, may contain both pathogenic and nonpathogenic strains that, despite their common taxonomic identification, are genetically very different, and not interconvertible. The complexities of this issue and the subtleties and inadequacies of bacterial systematics have made it difficult to clarify the regulatory issues, and current practice is not yet totally satisfactory.

13.2.3 WEEDINESS

Introduced weedy species, such as Melaleuca or Hydrilla in the United States, are a familiar problem. However, weediness is highly unlikely to arise by accident when plants are being engineered for unrelated characteristics. A more substantial concern involves the possible exchange of genetic material between the domesticated targets of genetic engineering and their wild and weedy relatives. To evaluate this risk, one requires a detailed understanding of the ecological relationships in the particular environment. In the United States, for example, the potential for such exchange would be negligible for most row crops, although sorghum provides a striking counter example. However, for forage grasses in the United States, the potential for introgression between wild and domesticated species varies widely (Mack, 1988). In South America the situation would be far more problematical for row crops: there is considerable exchange and introgression between domesticated crop species and wild and weedy relatives.

Properties such as herbicide resistance, if introduced into crop species, could well find their way into weedy populations if they impart a fitness advantage to the bearer. Such transfer could exacerbate chemical pollution problems if it led to the need to apply new herbicides to the weedy populations. It should be remarked that this is not a problem that has anything to do with the method of genetic engineering per se; once again, it is the properties of the introduced organism that are important, not how those properties were achieved. For any introduction of plants, whether of genetically modified or other organisms, what is needed is better information on the potential for exchange with wild and weedy relatives.

Sexual exchange of material between domesticated plants and weedy relatives is but one of a variety of mechanisms of exchange known to occur among species. Bacterial populations, which usually do not exchange chromosomal DNA (Selander et al., 1987), can exchange mobile genetic elements freely among species, and this has been important in the evolution of pathogenic traits (National Academy of Sciences, 1987). The most familiar example of such exchange involves the wide proliferation of antibiotic resistance among distinct bacterial species (Levy, 1982). Such transfer and spread is facilitated when there is selective pressure favoring it, for example when the traits borne on plasmids impart a selective advantage in polluted environments; thus the possibility for such spread must be given serious consideration when the environmental conditions are likely to favor the spread of the engineered trait. Although the frequency of horizontal exchange in nature is not established, such exchange is unlikely to be an important consideration in the absence of selective advantage favoring spread.

13.3 CODA

The use of genetically engineered organisms holds considerable promise for environmental management and other purposes, provided appropriate safety standards are established and observed. The problem with much of the debate concerning deliberate releases has been the difficulty in getting down to specifics. Examples can be advanced that give cause for concern or that demonstrate that introductions can be carried out safely, but none of these has the generality to apply to all cases. Generic arguments for and against the safety of introductions must be rejected and replaced by consideration of the properties of individual introductions. It is the properties of the introduced organism in relation to the environment that must receive attention, not the method by which the genetic modification was achieved. One must go beyond discussions that lump all possible applications together and develop criteria that associate individual cases with the risk categories most appropriate to them.

ACKNOWLEDGEMENTS

This publication is ERC-159 of the Ecosystems Research Center, and was supported in part by the US Environmental Protection Agency, Cooperative Agreement CR812685 with Cornell University. Additional support was provided by National Science Foundation grant DMS-8406472 to the author. The work and conclusions published herein represent the views of the author and do not necessarily represent the opinions, policies, or recommendations of the funding agencies.

Thanks are extended to Clifford Gabriel for valuable comments on an earlier version of this paper and Arthur Kelman for always illuminating discussions and for calling my attention to the Thompson paper. 

REFERENCES

Colwell, R.K. (1988) Ecology and biotechnology: expectations and outliers. In: Fiskel, J. and Covelo, V.T. (Eds.) Risk Analysis Approaches for Environmental Releases of Genetically Engineered Organisms, NATO Advanced Science Institute Series Vol. F, Springer-Verlag, Berlin, Heidelberg, pp. 163-80.

Elton, C.S. (1958) The Ecology of Invasions by Animals and Plants, Methuen, London.

Levy, S.A. (1982) Microbial resistance to antibiotics: an evolving and persistent problem. Lancet 1, 83-8.

Levin, S.A. (1986) Risk assessment, risk management and biotechnology. In: Russell, M.J. (Ed.) Proc. 1986 Washington International Conference on Biotechnology, Center for Energy and Environmental Management, Fairfax, Virginia, pp. 231-44.

 Levin, S.A. and Harwell, M.A. (1986) Environmental risks and genetically engineered organisms. In: Panem, S. (Ed.) Biotechnology: Implications for Public Policy, Brookings Institution, Washington, DC, pp. 56-72.

Mack, R.N. (1988) (and references therein). Temperate grasslands vulnerable to plant invasions: characterstics and consequences. In: Drake, J., diCastri, F., Groves, R., Kruger, F., Mooney, H.A., Rejamanek, M. and Williamson, M. (Eds.) Biological Invasions: A Global Perspective, SCOPE 37, John Wiley, Chichester.

Mooney, H.A. and Drake, J.A. (Eds.) (1986) Ecology of Biological Invasions of North America and Hawaii, Springer-Verlag, New York.

National Academy of Sciences (1987) Introduction of Recombinant DNA-Engineered Organisms into the Environment: Key Issues, Washington, DC, 24 pp.

Regal, P. (1988) The adaptive potential of genetically engineered organisms in nature. In: Special Combined Issue of Trends in Ecology and Evolution 3(4) and Trends in Biotechnology 6(4), Elsevier, Cambridge, England, pp. 536-538.

Selander, R.K., Caugant, D.A. and Whittam, T.S. (1987) Genetic structure and variation in natural populations in Escherichia coli. In: Neidhardt, F.C., Ingraham, J.L., Low, K.B., Magasanik, B., Schaechter, M. and Umbarger, H.E. (Eds.) Escherichia coli and Salmonella typhimurium: Cellular and Molecular Biology, American Society for Microbiology, Washington, DC, pp. 1625-48.

Thompson, P.B. (1987) Agricultural biotechnology and the rhetoric risk: some conceptual issues. National Center for Food and Agricultural Policy Discussion Paper FAP87-01. Resources for the Future, Washington, DC.