Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
5 The Release of Genetically Engineered Organisms into the Environment A LITTLE MORE THAN A DECADE after the basic techniques of genetic engineering were pioneered, biotechnology is about to enter an entirely new domain. Over the next few years the first organisms genetically engineered for use in the environment will be field tested and put to work. The first of these organisms to reach the marketplace will probably have agricultural applications. Genetically engineered microorganisms will be used to control insect pests, to fix nitrogen, and to reduce frost damage. Bioengineered crops may be hardier, resistant to different kinds of pesticides, and more productive. (Chapter 3 discusses potential agricultural applications in detail.) Genetically engineered organisms could also find nonagricultural uses in wastewater treatment facilities, in mining operations, and in oil wells. (Some of these possible applications are described in Chapter 2.) In addition to offering dramatic new capabilities, the use of geneti- cally engineered organisms in the environment will raise a host of new concerns. Most natural ecosystems are exceedingly complex assem- blages of many different organisms and abiotic influences, and many of the relationships among an ecosystem's components are still poorly understood. A genetically engineered organism introduced into an This chapter includes material from the presentations by Martin Alexander and Daniel Nathans at the symposium. 54
THE RELEASE OF GENETICALLY ENGINEERED ORGANISMS 55 ecosystem will therefore have the potential of affecting it in unantici- pated, and possibly detrimental, ways. Moreover, unlike air or water pollution, which tends to dissipate over time, organisms have the capacity to reproduce and spread, magnifying any problem that does arise. The history of conventional plant and animal breeding suggests that the likelihood of a problem is very low. Plant and animal breeders have been creating new crop varieties and livestock breeds for many centu- ries without posing untoward risks to the environment. On the other hand, plants, animals, and microorganisms introduced into new locales from other parts of the world have caused major and lasting environmental damage. A fungus introduced into America from Asia killed almost all of North America's majestic chestnut trees. Another fungus has eliminated most Dutch elm trees from the eastern United States. A virus introduced into Australia almost completely annihilated that continent's rabbit population. Over half the insect pests in the United States today come from abroad. Kudzu and hydrilla are two examples of weeds introduced into the United States that have caused monumental problems. Similarly, starlings, house sparrows, and gypsy moths are all introduced animals that America would almost surely be better off without. By the same token, most of America's major crops, including soy- beans, wheat, and rice, are not indigenous to this continent. Most livestock breeds and many poultry breeds have their roots in Asia, Africa, or Europe. Farmers and gardeners even apply Rhizobium bacteria from other parts of the world to their fields and plots to boost yields. Such arguments by analogy are valuable reminders of biotech- nology's potential to do great good or great harm, but they leave much unexplored when it comes to examining the release of genetically engineered organisms into the environment. Genetic engineering of- fers capabilities that range far beyond those of conventional breeding programs. Understanding the issue therefore requires a much more detailed examination of the specific factors involved when a new organism is introduced into the environment. The Components of Environmental Risk According to Martin Alexander of Cornell University, five indepen- dent factors come into play in determining what effect a genetically engineered organism will have on other organisms. These are (1) whether the organism is released into the environment, (2) whether it
56 BIOTECHNOLOGY Common soil bacteria that have been genetically engineered to contain the gene for a pesticidal toxin colonize the roots of a corn plant. Such genetically engineered organisms must be thoroughly tested to ensure that they will not have unintended detrimental effects on ecosystems into which they are introduced. survives, (3) whether it multiplies, (4) whether it moves to an area where it can have an effect, and (5) what that effect actually is. In addition, DNA may be transferred between organisms in the environ- ment, either sexually or asexually, and this must also be taken into account in calculating the risk posed by a genetically engineered organism. Release Obviously, an organism must first enter the environment to cause harm. Most of the organisms that have been genetically engineered to date have been designed for laboratory research or fermentation processes, and so long as they stay within the fermentor they pose no risk to the environment. Moreover, these organisms have generally been genetically crippled to make it very difficult for them to survive outside the flask or fermentor.
THE RELEASE OF GENETICALLY ENGINEERED ORGANISMS 57 Genetically engineered organisms undergo initial testing in isolated growth chambers such as the one shown here. But growth chambers cannot re-create the full complexity of a natural ecosystem, and at some point small-scale field testing becomes necessary. A few bioengineered organisms designed for use in the environment, including microbial pesticides and genetically engineered crops, have already been produced and studied within laboratories, growth cham- bers, and greenhouses. Experiments involving these organisms carried out with public funds are required to adhere to the appropriate containment procedures specified in NIH's Guidelines for Research Involving RecombinantDNA Molecules (the history of these guidelines appears in Chapter 6). These procedures are meant to ensure that the likelihood of a potentially hazardous organism escaping into the environment remains very low. For the purposes of calculating a probability, however, it should be remembered that accidental releases of hazardous organisms from research facilities have occurred in the past. As Alexander says, "no tank never leaks." Even stringent efforts at containment are only "reducing the probability of an accident, not converting it to zero probability." Regardless of the developmental research involved, an organism genetically engineered for environmental applications will eventually
58 BIOTECHNOLOGY be ready for small-scale field testing. At that point, the probability of release becomes one. Survival Once an organism has been released into the environment, it must survive to have an effect. As Alexander points out, predicting the survival of an introduced organism is one of the most difficult problems in ecology. "Natural ecosystems have what are known as homeostatic mechanisms," he explains. "There are a variety of interactions among plants, animals, and microorganisms that tend to keep in check the rare species, eliminate alien species, and prevent the dominant species from overexploiting the environment. If we go into a particular area, we see the same plants, and an occasional introduction of a new organism will not result in its establishment in the field. "Homeostasis is effective in eliminating aliens, but it is not always wholly successful. In plant ecology, animal ecology, and microbial ecology, it is known that an introduced organism does occasionally survive." Alexander's own research and the studies of other microbial ecolo- gists have revealed many instances of microorganisms from foreign locations that survive for days, weeks, months, and even years when introduced into new environments. The past history of introduced plants, animals, and microorganisms that have done great environ- mental damage, while admittedly worst-case examples, also indicates that some percentage of introduced organisms will survive. The problem lies in determining which organisms will survive and why. "The general feeling is that homeostasis will eliminate nearly all, but not all, species," says Alexander. "Even given the successful establishment of an organismâplant, animal, or microorganismâwe cannot explain why that one was successful and many others failed." Thus there is considerable uncertainty surrounding this component of environmental release. However, if a genetically engineered organism is to have its intended effect, it must survive for at least some period of time. Multiplication Most genetically engineered organisms must also multiply if they are to have an effect. In general, the number of organisms originally released will be too small to do much harm. "The organism must reach a population density high enough to upset other organismsâeither a
THE RELEASE OF GENETICALLY ENGINEERED ORGANISMS 59 microorganism upsetting host plants or animals, a plant that becomes a weed because it is abundant, or an animal that disturbs its natural environment," says Alexander. The determinants of successful multiplication are in most cases as unknown as those of survival. Says Alexander, "Except for pathogens, we do not know under what conditions almost any microorganism can multiply in nature. We cannot predict the organisms that will multi- ply. This is true of most species of plants as well as of microorganisms. All we have are instances with particular economically important species." One factor that may work against the survival of a genetically engineered organism is that the organism contains extra DNA, which diverts part of its metabolic energy from the pursuit of survival and multiplication to the production of agriculturally important proteins. In this way, the genetically engineered organism is at a competitive disadvantage with organisms that do not bear the burden of extra DNA. However, as Alexander points out, the ecological consequences of extra DNA need not be wholly negative. "If the acquisition of one characteristic results in an ecological advantage, then the organism may be able to overcome one of the environmental barriers to its establishment. Unfortunately, at this time, we can't tell whether additional DNA that is disadvantageous in one way will also be advantageous in another way." Furthermore, one could argue that it is unlikely that a genetically engineered organism would acquire a newfound persistence, as in the case of a weed or rampant pathogen, because many interacting genes are needed to generate such characteristics. Here too, however, counterexamples can be cited in which small genetic alterations lead to major changes in an organism's behavior. Although not related to genetic engineering, slight changes in the antigenicity of an influenza virus can lead to reduced immunity among humans and a greater severity of the disease. Similarly, the genetically straightforward formation of a capsule around some bacteria can make them resistant to normal human and animal defenses. In agriculture, the addition of a gene for resistance to pests or herbicides or the acquisition of genes for more efficient photosynthesis could give plants an edge over their nonengineered competitors. Dispersal An organism usually will not cause harmful effects in the area where it is releasedâa farmer's field, a waste dump, a tailings pond. Instead,
60 BIOTECHNOLOGY it must move to an area where it encounters organisms susceptible to its effects. Thus the greater the range of a genetically engineered organism, the greater its chances of causing a problem. The dispersal of some organisms has been studied extensively, according to Alexander, but much less is known about other organisms. Yet it is crucial that the range of an organism be determined before its release, since genetically engineered organisms will generally have been designed to survive and multiply in the environment. Effects Finally, the effects of a released organism on other living things in the environmentâmicrobes, plants, animals, and humansâmust be calculated. In some cases these effects, if they occur, will be obvious; in others they will be indirect and subtle. To take just one example from traditional plant breeding, a specific cultivar of potato had to be removed from supermarket shelves because it was found cap- able of producing hazardous levels of toxins under certain conditions of stress. Such unintended effects may be less likely to occur with recombinant DNA techniques than with traditional plant and ani- mal breeding, since the genes and metabolic pathways to be altered are likely to be more fully characterized with recombinant DNA. Nevertheless, the search for effects will be difficult in many cases, because the interrelations among organisms in the environment are often poorly characterized. The Transfer of Genetic Information An additional complication in calculating the risks of environmental release is that organisms in the environment can transfer DNA to other organisms through a variety of means. If a genetically engi- neered organism transfers its new traits to another organism, the string of risk factors, from survival to effects, must be calculated anew. The most common means of genetic transfer among plants and animals is sexual recombination. For instance, pollen from a geneti- cally engineered crop could fertilize the seeds of a similar but nonengineered crop. Much more worrisome is the possibility of crosses between crops and related noncrops or weeds. Just as plant breeders transfer traits from cross-fertile weeds into agriculturally important plants through crossbreeding followed by successive backcrossesâa process known as introgressionâso a gene from a crop could be
THE RELEASE OF GENETICALLY ENGINEERED ORGANISMS 61 transferred to a related weed through natural introgression. In fact, this seems to have occurred several times in the past, such as in the relationship between weedy Johnson grass and sorghum. There are several forces that will probably keep this from being a serious problem in biotechnology. For one thing, only a few species of weeds cross in nature with major crop plants. Nor is it clear just how much advantage any weedy species has ever gained from natural introgression. If problems were anticipated, geographic limitations on the use of genetically engineered crops could be imposed. Finally, natural introgression is a concern with the introduction of any new plant or animal breed, not only with those that will be produced through biotechnology. Plants are not known to transfer DNA between one another through nonsexual means, and such transfers appear to be rare among animals (viruses are possible intermediaries of nonsexual genetic transfer). However, microorganisms exchange genetic material nonsexually in several different ways. Such exchanges of DNA have been known to transfer traits like resistance to antibiotics among microorganisms in laboratory and hospital settings. But it is not known if such transfers among microorganisms occur in natural environments, according to Alexander, and their impact of such transfers on the risk of environ- mental release is likewise unknown. Risks and Uncertainties The probability that a genetically engineered organism will have a detrimental effect on the environment is the product of the five factors discussed above: release, survival, multiplication, dispersal, and ef- fects. (The last four factors also come into play for an organism that acquires foreign DNA from a genetically engineered organism.) Since the probabilities associated with one or more of these factors are likely to be small, the overall probability of a harmful effect is likely to be very small. But a low probability is not a zero probability. And, as Alexander points out, "the consequences of this low-probability event could be very significant." The uncertainties surrounding each of the six components of envi- ronmental risk make it impossible to calculate precisely how small the risk is. Claims of zero risk or great risk are therefore inappropriate, according to Alexander, and merely muddy the debate surrounding the issue. Furthermore, the uncertainties will loom larger as more and more organisms are altered, as the number and kind of introduced genes grow, and as genetically engineered organisms are released into
62 BIOTECHNOLOGY a wider range of environments. "The degree of uncertainty is too large for me, as an ecologist, to feel particularly comfortable with," concludes Alexander. Research and Regulation The way to reduce the level of uncertainty now associated with environmental release is through research into the interactions of organisms with their surroundings. Indeed, it is highly desirable that this research be done before the range of biotechnology's applications in the environment begins to expand, but very little of this research is now being supported by federal regulatory or research agencies, Alexander says. Researchers should concentrate on several key points, according to Alexander. Most important, the specific factors that contribute to the probabilities associated with each of the six components of environ- mental risk should be identified. This would help the industry choose the organisms that it should use and rule out those that it should avoid. It could also help in fashioning debilitated organisms that would not survive, multiply, or disperse once their intended purpose was com- plete. The identification of these traits would facilitate the testing needed to evaluate environmental risk. It would also help regulators decide which organisms need extensive testing before approval and which need little or no testing. Several important technical and methodological issues should also be addressed. For instance, ways need to be developed to label a genetically engineered organism so that its fate can be monitored in the field. This would greatly simplify studies of an organism's escape, survival, multiplication, and dispersal and may even help in tracking the movement of DNA among organisms in the environment. Several immunologic and genetic techniques have been adapted for labeling purposes, but they require further development. Research along these lines would reduce many of the uncertainties surrounding environmental release, but it cannot eliminate them. Ecology is not such an exact science as to lend itself to infallible predictions. As Daniel Nathans of the Johns Hopkins University School of Medicine says, "It is difficult to envision how one will get the knowledge to tell you in a concrete way whether transfer of a particular gene into a particular organism is 20 years from now going to cause an ecological disaster. We will never do the experiment if you require that question to be answered in a scientifically acceptable way. . . . We are left with reasoned, conservative judgments of people in the field, and
THE RELEASE OF GENETICALLY ENGINEERED ORGANISMS 63 we are also left with very carefully controlled, step-by-step experi- ments, in which appropriate measurements are made." The difficult task of balancing the remaining uncertainties against the undeniable benefits of biotechnology falls most immediately to the federal regulatory agencies that oversee genetic engineering and its application. As discussed in Chapter 6, industry representatives and government officials agree that the regulations established by these agencies will be a critical and often indispensable factor in the industry's development. According to Alexander, these regulations will reduce the possibility of an ecological upset. They will also ease the public's fears about the new technology. They will help the industry to get liability insurance at reasonable rates. And they will reduce the backlash when a problem does occur or when a problem that arises is mistakenly attributed to industry. The pursuit of the environmental applications of genetic engineering therefore involves three overlapping fronts: the development of the organisms, research on the interactions of the organisms with the environment, and regulation of the organisms' development and appli- cation. By moving forward on these three fronts simultaneously, it should be possible to reap the benefits of biotechnology while holding the risk to the environment at a minimum. "If one has a good base of scientific information and a reasonable testing system, then I think that much of the residual degree of uncertainty can easily be answered by a very modest regulatory program," says Alexander. "But we should have a regulatory system in place, a regulatory system that will reduce the likelihood of a problem arising, and a significant amount of research to find out where the issues are." Additional Readings Martin Alexander. 1985. "Ecological Consequences: Reducing the Uncertainties." Issues in Science and Technology 1(Spring):57-68. Winston J. Brill. 1985. "Safety Concerns and Genetic Engineering in Agriculture." Science 227(January 25):381-384. [See also the responses to this article in Science 229(July 12, 1985):111-118.] Bernard Dixon. 1985. Engineered Organisms in the Environment: Scientific Issues. Washington, D.C.: American Society for Microbiology. Ecosystems Research Center. 1985. Potential Impacts of Environmental Release of Biotechnology Products: Assessment, Regulation, and Research Needs. Ithaca, N.Y.: Ecosystems Research Center, Cornell University. Holly Hauptli, Nanette Newell, and Robert M. Goodman. 1985. "Genetically Engineered Plants: Environmental Issues." Bio/Technology 3(May):437-442. Albert H. Teich, Morris A. Levin, and Jill H. Pace, eds. 1985. Biotechnology and the Environment: Risk and Regulation. Washington, D.C.: American Association for the Advancement of Science.