SCOPE 44 - Introduction of Genetically Modified Organisms into the Environment

Within the next few years, a number of new varieties of plants will be tested. Outwardly, these plants will appear identical to the agricultural varieties currently in use. However, whereas the older kinds of plants are sensitive to herbicides, or retarded by virus infection, or ravaged by insect plagues, the new forms will grow more or less unaffected. These new varieties will be protected by one or two genes derived from prokaryotes or viruses and introduced by means of modern techniques of plant molecular biology. This review will (1) summarize how these techniques are applied, (2) describe several examples of the kind of genes being introduced into crop plants, and (3) evaluate the impact the resulting plants might have on agricultural and non-agricultural plant communities.

7.1 GENE TRANSFER BETWEEN BACTERIA AND PLANTS

There are a number of neoplastic growths of plants that are induced by plant-associated bacteria. In 1943, Braun showed that one such type of tumor, called crown galls, continued to grow in vitro even when bacteria were no longer present (Braun, 1943). The explanation proved to be that the disease-causing organism, Agrobacterium tumefaciens, had somehow transferred to the infected plant cells a new set of genes that promoted dedifferentiation and proliferation (Chilton et al., 1977; Schell et al., 1979). A genetic analysis carried out over the last seven years has revealed the sequence of events leading to this transformation process.

One of the first steps leading to the formation of crown gall tumors is carried out by products of a number of genes lying in the bacterial chromosome. At least three of these cistrons, chvA, chvB, and pcsA, are involved in attaching the bacterium to plant cells (Douglas et al., 1985; Thomashow et al., 1987). During this time, a second set of genes encoding at least 10 proteins comes into play (Stachel and Nester, 1986; Engström et al., 1987). These virulence genes lie within a 35-kilobase region of a very large, tumor-inducing (Ti) plasmid. Only two of them, virA and virG, are expressed initially. However, the other virulence proteins are induced when the bacterium senses the presence of phenolic compounds, such as acetosyringone (Stachel et al., 1985), that are excreted from wounded plant cells during the course of cell wall repair. The precise functions of all of these proteins remains unresolved, but it is clear that they are responsible for producing a single-stranded DNA copy of the actual tumor-inducing genes and for transferring this copy from bacteria into plant cells (Stachel et al., 1986; Wang et al., 1987) in a process similar to some forms of bacterial conjugation (Buchanan-Wollaston et al., 1987).

The tumor-inducing genes of Agrobacterium lie within the T-DNA region of the Ti plasmid. Two of the genes cooperate to produce the plant growth factor, auxin (Inzé et al., 1984; Schröder et al., 1984; Thomashow et al., 1984), and one additional gene is responsible for the synthesis of a second factor, cytokinin (Akiyoshi et al., 1984; Barry et al., 1984). It is the unregulated production of these two hormones that promotes the proliferation of the infected plant cell so that it forms an undifferentiated crown gall tumor. Plant cells, particularly from dicotyledonous plants, have enormous developmental potential. Thus, if either of the auxin-producing genes is deleted, then the crown gall tissue contains an excess of cytokinin relative to auxin with the result that the tumor produces numerous shoots and leaves (Garfinkel et al., 1981; Joos et al., 1983). On the other hand, if the single cytokinin-synthesizing gene is removed, the tumors produce an excess of auxin and some of the tissues differentiate into roots. If all of the T-DNA genes are removed, leaving only the DNA sequences at the very ends of the T-DNA, then no tumors arise (Leemans et al., 1982; Joos et al., 1983; Zambryski et al., 1983). Nevertheless, the remaining parts of the T-DNA (and any new sequence inserted between the ends) are copied, transferred to the plant, and integrated in the nuclear genome in a colinear unit, identical to the copy originally in the bacterium.

The T-DNA of Agrobacterium is unique as a gene transfer vector. The integrated form does not undergo successive cycles of replication or integration like a virus or transposon. Moreover, the proteins required for intercellular transfer remain (as far as is known) outside of the plant cell (Otten et al., 1984). As a result, integrated T-DNAs do not increase in number or change position even when cells are reinfected by Agrobacterium. In fact, integrated T-DNAs lack intact copies of the DNA sequences that are required by the bacterial virulence genes to process the T-DNA for the DNA transfer operation (Zambryski et al., 1980; Wang et al., 1984). Consequently, once a T-DNA is inserted in the plant genome, it is as stable as other plant sequences (Müller et al., 1987). These properties have contributed to the widespread use of the T-DNA in much of the genetic engineering currently being done with plants.

7.2 THE BASIC PRINCIPLES OF GENETIC ENGINEERING

The genetic improvement of an organism depends on the breeder's ability to manipulate the genetic diversity within the species. Every cross that introduces a desired trait potentially introduces undesirable ones. These latter genes must be removed by subsequent crosses, sometimes over the course of many years.

The alternative approach is to use techniques developed for DNA cloning to first isolate a gene from one plant and then to introduce it into another. In fact, there is no need for the gene to come from the same species. It is possible to take any gene, prokaryotic or eukaryotic, provide it with suitable transcription and translation signals, and introduce it into an organism of another kingdom. In the case of plants, this entails inserting the isolated and/or modified gene between the terminal sequences of a T-DNA and then bringing this into specially modified agrobacteria that no longer contain any other T-DNA sequences (Deblaere et al., 1985; Koncz and Schell, 1986). Using appropriate culture conditions, these bacteria can infect and transfer the artificial T-DNA either to pieces of leaf tissue (Horsch et al., 1985) or to leaf protoplasts (De Block et al., 1984). If the tissues or cells are given suitable amounts of auxin and cytokinin, then they will grow in vitro as callus or shoot tissues, and later become fully developed plants when the hormones are removed (Nagata and Takebe, 1970). Three examples of the kinds of genes that can be introduced are presented in the following pages. 

7.2.1 CASE 1. PROTECTING PLANTS AGAINST LEPIDOPTERA LARVAE

Since the latter half of the 1950s, several companies have been marketing a biological insecticide based on the crystal protein of Bacillus thuringiensis. This protein seems to have no effect on vertebrates or most insects, but kills Lepidoptera larvae within 3-5 days of ingestion (Carlton and González, 1986). Despite its specificity and effectiveness, it is not extensively used. This is in part due to the fact that it is perceived to have a low degree of field persistence: in order to be fully effective, it is necessary to apply it uniformly, and sometimes repeatedly, over the insect-infected field.

One possible solution to this problem has recently been found. The crystal protein gene was cloned and subsequently modified so that it could be expressed efficiently in plant cells. This modified gene was then introduced, using the technique described earlier, into cells of tobacco (Vaeck et al., 1987) and tomato (Fischhoff et al., 1987), and transgenic plants were regenerated. In order to determine whether the new gene conferred the expected trait, pieces of these plants were fed to tobacco hornworm (Manduca sexta), tobacco budworm (Heliothus virescens), or corn earworm (Heliothus zea) larvae. Most of the tobacco hornworm larvae ceased feeding within one day and died within six days. Fewer of the other two insects died, although the growth of the survivors of all three species was severely retarded.

This sort of approach to biological control has several clear advantages over the use of broadcast insecticides. First, the product can be distributed uniformly through the plant by using transcriptional controls that are equally active (`constitutive') in all tissues. Alternatively, it is possible to produce the protein only in tissues most frequently attacked by caterpillars by linking the gene to tissue-specific regulatory signals. Second, the pesticide is contained within the tissue itself and so there is no risk of inadvertant contamination of uncultivated fields. Third, the protein appears to be highly specific and consequently unlikely to affect directly any insect other than the one for which it is intended.

7.2.2 CASE 2. PROTECTING PLANTS AGAINST VIRUS

Liberal application of insecticides can generally halt, and very often eradicate, insect plagues. The spread of virus infection is less readily controlled. For this reason, considerable effort has gone into isolating crop genotypes that either decrease the chances that the plant will be attacked by the insect or fungus agent that transmits the virus or in some way alter the physiological properties of the plants so as to minimize the spread of the virus after infection (Keen, 1986). Unfortunately, the number of genes known to confer some degree of disease resistance is still lower than the number of diseases that afflict the crops. At present, it is considerably easier to isolate and characterize the disease-causing organism than to search the thousands of varieties of a particular host for the loci conferring immunity.

It is for this reason that a number of strategies have been devised to subvert a virus or a portion of that virus to protecting a plant against infection. One approach has been based on the observation that certain non-pathogenic `satellite' viruses seem to attenuate the effects of pathogenic forms. The satellite virus may encode its own coat protein (Reichmann, 1964) or be packaged with the products of the pathogenic virus (Schneider, 1971). In either case, once the pathogen has acquired a satellite, the two forms are cotransmitted. The satellite virus cannot replicate without the gene products of an autonomous form. As a result, the viruses reach an equilibrium determined by the efficiency with which the satellite sequesters essential products of the autonomous virus, yet still leaves enough of that product for the pathogen to persist.

This balance can be upset to favour the satellite virus. One way to demonstrate this is to establish the satellite in the plant before introducing a pathogen. Since most plant viruses are composed of RNA, this is done by first producing a DNA copy of the satellite and then providing it with the necessary promoters and transcription termination signals. Modified forms of the satellite of cucumber mosaic virus (CMV) and of tobacco ringspot virus (TobRV) were prepared in this way and put into tobacco (Gerlach et al., 1987; Harrison et al., 1987). These viral genomes were phenotypically silent until plants were infected with the appropriate pathogen. The pathogenic virus then provided the replication functions needed to permit the satellite to multiply. The satellite, in turn, prevented the pathogen from replicating, and greatly reduced the severity of the infection.

This approach to protecting plants against virus infection appears quite promising. The system is self'-regulating: the more the pathogen replicates initially, the more the satellite replicates subsequently to halt the spread of the disease. Although every virulent form is not parasitized by a satellite virus, it may be possible to create or select ones in the laboratory by modifying the pathogen itself. Alternatively, some viruses can be controlled by overexpressing negative regulators of viral genes or proteins such as the virus coat protein that can block virus replication (Abel et al., 1986; LoeschFries et al., 1987).

7.2.3 CASE 3. INTRODUCING HERBICIDE RESISTANCE INTO CROP PLANTS

The more similar a particular agricultural pest is to the plant under cultivation, the more difficult it is to eradicate. Monocotyledonous weeds, for example, can be killed by certain herbicides, and dicotyledonous weeds by others (Gressel, 1986). Few herbicides, however, are so selective that they can be used to kill monocotyledonous weeds in fields of corn or rice, or dicotyledonous weeds growing amidst tomatoes or soybeans.

One way to circumvent this problem has been to select for particular crop lines that are resistant to commercially important herbicides. This strategy, unfortunately, has not always proved effective, in part because it has depended on preexisting sources of genetic variation. Recombinant DNA technology can be used to broaden this base by bringing together genes from wholly unrelated organisms. Several different approaches are being evaluated currently.

One means of conferring herbicide resistance has been to introduce into a crop bacterial genes that are less sensitive to the effects of the herbicide than the endogenous plant gene. The herbicide glyphosate, for example, blocks the synthesis of aromatic amino acids (Jaworski, 1972). The gene target for this chemical has been cloned from Escherichia coli and modified so that it is expressed in tomato (Fillatti et al., 1987) and petunia (DellaCioppa et al., 1987). The resulting transgenic plants clearly tolerate levels of glyphosate that kill unmodified lines, although the levels of resistance are not yet high enough to be commercially useful.

An alternative type of resistance can be created by using enzymes that inactivate the herbicide before it reaches toxic levels in the cell. For example, it has been possible to isolate a gene from the prokaryote, Streptomyces hygroscopicus, that detoxifies the broad spectrum herbicide, phosphinothricin (Murakami et al., 1986). When joined to suitable transcriptional signals, the gene produces enough enzyme to inactivate as much as eight times the amount of herbicide commonly used on agricultural land (De Block et al., 1987). With genes such as this in a plant, it would be possible to use the herbicide phosphinothricin to kill virtually any weed, regardless of its relatedness to the crop species.

7.3 EVALUATING THE IMPACT OF TRANSGENIC PLANTS ON AGRICULTURAL AND NON-AGRICULTURAL ECOSYSTEMS

Although each of the three cases that have been discussed involved the use of recombinant DNA, it is possible that alternative solutions could have been found using conventional plant breeding strategies. Genes conferring resistance to plant hoppers and leaf hoppers have been identified in rice (Khush and Virmani, 1985) and to beetles in summer squash (Dhillon and Sharma, 1987). In a similar way, sets of genes have been identified that confer resistance to specific viral pathogens of cotton and barley (Catherall et al., 1970; Siddig, 1970), while herbicide resistance has been extensively investigated in tobacco (Chaleff and Ray, 1984). Recombinant-DNA techniques simply provide a more direct means to add to existing or potential genetic resources.

There are certain aspects of recombinant-DNA technology that set this art apart from traditional genetic breeding practices. Insect or virus resistance genes identified by conventional breeding programmes have, at best, been transferred only between evolutionarily related species. By contrast, once a successful resistance gene has been obtained by means of recombinant DNA, it can be introduced into a number of very different plant genera. Ironically, the ease with which it is possible to repeat a successful gene transfer may help maintain the genetic diversity within a crop species, whereas crossbreeding tends to homogenize species and reduce variation. If, for example, an indigenous variety of potato was being displaced from its traditional region of cultivation because it was more sensitive than a commercial variety to a particular virus, then it might be possible to cure the deficiency and preserve the genotype by introducing a single new gene. This could not be easily done using genetic crosses without concomitantly disrupting and diluting many of the desirable traits of the indigenous variety.

A second distinguishing feature of recombinant-DNA technology is the sheer number of gene combinations that can be tested in a comparatively short time. For example, one of the most frequently mentioned objectives of genetic modification is the improvement of the nutritional quality of plants. Plant breeders cannot select directly for characters such as protein quality. Instead, they must assay the protein content of hundreds of varieties of a particular grain or tuber, and then, over the course of several generations, combine and accentuate the best traits by crosses between different lines. By comparison, the desired sort of genetic variation in the form of the cloned storage proteins of maize, peas, and potatoes can be introduced in one step into a plant to compensate for amino acid deficiencies. Some of these additions will fail if the new proteins are not produced in sufficient quantities to improve the quality of the crop, or if producing the protein reduces the yields, or if the proteins prove unpalatable to the people they are intended to help. Still, so many different combinations can be tested that some will be found to prove successful. In this event, what would be the ecological consequences of producing and releasing these kinds of organisms?

Traits such as improved nutritional quality are unlikely to confer a selective advantage in natural ecosystems. In fact, this type of attribute is more likely to make a plant more attractive to herbivores so that it will be less competitive in non-agricultural plant communities.

Genes conferring herbicide resistance are similarly unlikely to change the fitness of a plant that has entered a natural habitat. These genes are providing proteins that either duplicate steps already present in a metabolic pathway (Della-Cioppa et al., 1987) or modify highly unusual compounds not normally found in plant cells (De Block et al., 1987). In the absence of the herbicide, either type of gene product is likely to be selectively neutral or slightly deleterious according to the metabolic cost of its production.

In comparison to the previous examples, traits such as insect or disease resistance will confer a selective advantage in any natural habitat. This does not automatically imply that the species will flourish. The Bacillus toxin only kills caterpillars, not all insects, while the satellite virus only protects against the co-evolved pathogen (and a few closely related forms), not all viruses. This specificity is qualitatively no different from the insect- and virus-resistance plants selected by conventional procedures and already released. The fact that these plants have not become established in natural ecosystems is an indication that their distribution is not limited solely by one form of disease or predation. The transgenic plants derived from the same stock material are unlikely to behave any differently. Clearly, though, as techniques in genetic engineering improve, it will be possible to modify the utilization of abiotic factors in the environment such as water, light, and minerals. Improvements of this sort may enhance the invasiveness of a species and therefore will have to be applied more cautiously.

7.4 THE DISPERSION OF NEW TRAITS

A commonly voiced concern about releasing genetically modified species is that the new traits will be transmitted from the domesticated organism to ones better able to compete in nature, with unpredictable consequences. Virus and bacterially mediated transfer of genes between unrelated eukaryotes is very rare outside the laboratory, so for the moment it must be assumed that the horizontal dispersal of a gene is governed by the frequency of interspecific hybridization. Consequently, each release must be considered individually and take into account the probability a crop will be grown near reproductively compatible wild species. If it is judged that the chance for hybridization is high or that the reproductive advantage of the new trait will allow it to spread beyond the hybrid zone, then it may be necessary to restrict the fertility of the introduced species. This could be done by introducing the recombinant genes only into male-sterile or reproductively constrained lines. Alternatively, it may be possible to genetically engineer genes for male sterility. These could be linked to the agronomically important traits, even before the latter are introduced, in order to reduce undesirable outcrossing.

7.5 PERSPECTIVES

Few, if any, of the anticipated changes wrought in the next few years by plant genetic engineering will be truly novel. By contrast, the genes introduced into the plants in order to accomplish this will be. To this extent, we should anticipate some changes in the rate of evolution of some species. The species most modified, however, will be those that have already been the most studied and changed during their domestication. Few of these past modifications have made these crops more invasive. It is unlikely that many of the new genes will change this, but the fitness of each modified organism should be assessed in a variety of small-scale tests prior to more general release. At the same time, if the engineered traits are successful, then it should be possible to discontinue the use of certain persistant herbicides or non-selective insecticides, and thereby allow some currently disturbed ecosystems to recover. It can be hoped that population and molecular biologists will cooperate to record and interpret both aspects of the release of a modified organism in order to refine our ability to predict the course of evolution of the organism and the other members of its ecosystem.

REFERENCES

Akiyoshi, D.E., Klee, H., Amasino, R.M., Nester, E.W. and Gordon, M.P. (1984) T-DNA of Agrobacterium tumefaciens encodes an enzyme of cytokinin biosynthesis. Proc. Natl Acad. Sci. USA 81, 5994-8.

Braun, A.C. (1943) Studies on tumor inception in crown gall disease. Am. J. Bot. 30, 674-7.

Buchanan-Wollaston, V., Passiatore, J.E. and Cannon, F. (1987) The mob and oriT mobilization functions of a bacterial plasmid promote its transfer to plants. Nature (Lond.) 328, 172-5.

Bacillus

Catherall, P.J., Jones, A.T. and Hayes, J.D. (1970) Inheritance and effectiveness of genes in barley that condition tolerance to barley yellow dwarf virus. Ann. Appl. Biol. 65, 147.

Chilton, M.-D., Drummond, M.H., Merlo, D.J., Sciaky, D., Montoya, A.L., Gordon, M.P. and Nester, E.W. (1977) Stable incorporation of plasmid DNA into higher plant cells: the molecular basis of crown gall tumorigenesis. Cell 11, 263-71. 

Deblaere, R., Bytebier, B., De Greve, H., Deboeck, F., Schell, J., Van Montagu, M. and Leemans, J. (1985) Efficient octopine Ti plasmid-derived vectors for Agrobacterium-mediated gene transfer to plants. Nucl. Acids Res. 13, 4777-88.

De Block, M., Botterman, J., Vandewiele, M., Dockx, J., Thoen, C., Gosselé, V., Movva, R., Thompson, C., Van Montagu, M. and Leemans, J. (1987) Engineering herbicide resistance in plants by expression of a detoxifying enzyme. EMBO J. 6, 2513-18.

Dhillon, N.P.S. and Sharma, B.R. (1987) Genetics of resistance of red pumpkin beetle (Aulacophora foveicollis) in summer squash (Cucurbita pepo L.). Theor. Appl. Genet 5, 711-15.

Engström, P., Zambryski, P., Van Montagu, M. and Stachel, S. (1987) Characterization of Agrobacterium tumefaciens virulence proteins induced by the plant factor acetosyringone. J. Molec. Biol. 197, 635-45.

Fischhoff, D.A., Bowdish, K.S., Perlak, F.J., Manone, P.G., McCormick, S.M., Niedermeyer, J.G., Fean, D.A., Kusano-Kretzmer, K., Mayer, E.J., Rochester, D.E., Rogers, S.G. and Fraley, R.T. (1987) Insect tolerant transgenic tomato plants. Bio/technology 5, 807-13.

Gerlach, W.L., Llewellyn, D. and Haseloff, J. (1987) Construction of a plant disease resistance gene from the satellite RNA of tobacco ringspot virus. Nature (Lond.) 328, 802-5.

Harrison, B.D., Mayo, M.A. and Baulcombe, D.C. (1987) Virus resistance in transgenic plants that express cucumber mosaic virus satellite RNA. Nature (Loud.) 328, 799-802.

Inzé, D., Follin, A., Van Lijsebettens, M., Simoens, C., Genetello, C., Van Montagu, M. and Schell, J. (1984) Genetic analysis of the individual T-DNA genes of Agrobacterium tumefaciens; further evidence that two genes are involved in indole-3-acetic acid synthesis. Mol. Gen. Genet. 194, 265-74.

Joos, H., Inzé, D., Caplan, A., Sormann, M., Van Montagu, M. and Schell, J. (1983) Genetic analysis of T-DNA transcripts in nopaline crown galls. Cell 32, 1057-67.

Khush, G.S. and Virmani, S.S. (1985) Breeding rice for disease resistance. In: Russell, G.E. (Ed.) Plant Breeding, Vol. 1, Cambridge University Press, Cambridge, pp. 239-80.

Leemans, J., Deblaere, R., Willmitzer, L., De Greve, H., Hernalsteens, J-P., Van Montagu, M. and Schell, J. (1982) Genetic identification of functions of TL-DNA transcripts in octopine crown galls. EMBO J. l, 147-52.

Müller, A.J., Mendel, R.R., Schiemann, J., Simoens, C. and Inzé, D. (1987) High meiotic stability of a foreign gene introduced into tobacco by Agrobacterium-mediated transformation. Mol. Gen. Genet. 207, 171-5.

Nagata, T. and Takebe, I. (1970) Cell wall regeneration and cell division in isolated tobacco mesophyll protoplasts. Planta 92, 301-8.

Reichmann, M.E. (1964) The satellite necrosis virus: a single protein and its genetic code. Proc. Natl Acad. Sci. USA 52, 1009-17.

Schneider, I.R. (1971) Characteristics of a satellite-like virus of tobacco ringspot virus. Virology 45, 108-22.

Siddig, M.A. (1970) Breeding for leaf curl resistance in Sakel cotton. In: Siddig, M.A. and Hughes, L.C. (Eds.) Cotton Growth in the Geziza Environment, Agricultural Research Corporation, Sudan, p. 153.

Stachel, S.E., Messens, E., Van Montagu, M. and Zambryski, P. (1985) Identification of the signal molecules produced by wounded plant cells that activate T-DNA transfer in Agrobacterium tumefaciens. Nature (Lond.) 318, 624-9.

Thomashow, L. S., Reeves, S. and Thomashow, M. F. (1984) Crown gall oncogenesis: evidence that a T-DNA gene from the Agrobacterium Ti plasmid pTiA6 encodes an enzyme that catalyzes synthesis of indoleacetic acid. Proc. Natl Acad. Sci. USA 81, 5071-5.

Vaeck, M., Reynaerts, A., Höfte, H., Jansens, S., De Beuckeleer, M., Dean, C., Zabeau, M., Van Montagu, M. and Leemans, J. (1987) Insect resistance in transgenic plants expressing modified Bacillus thuringiensis toxin genes. Nature (Lond.) 328, 33-7.

Wang, K., Stachel, S., Timmerman, B., Van Montagu, M. and Zambryski, P. (1987) Site-specific nick in the T-DNA border sequence following vir gene expression in Agrobacterium. Science 235, 587-91.

Zambryski, P., Joos, H., Genetello, C., Leemans, J., Van Montagu, M. and Schell, J. (1983) Ti plasmid vector for the introduction of DNA into plant cells without alteration of their normal regeneration capacity. EMBO J. 2, 2143-50.