Poplars consist of all species of the genus Populus, including cottonwoods, aspens, and the many interspecies hybrids in common use (Dickmann and Stuart 1983). Our working group focused on the fungal pathogens, arthropod herbivores, and weed competitors of Populus in the United States. However, bacterial and viral diseases of Populus are significant in Europe, and genetic engineering approaches toward their control or management are being studied. The key aspects of poplar biology important to understanding the use of pest resistance genes are described below.

Disease is believed to be the most important factor limiting adoption and productivity of poplar plantations (Royle and Ostry 1995). Poplars are susceptible to many pathogens (Newcombe 1996), and intensive culture has triggered changes in pathogen populations. Changes in North America have included the introduction of Eurasian pathogens (Newcombe 1996), the movement of regional pathogens within North America (Newcombe 1998b; Newcombe and Callan 1997), and hybridization between exotic and native species of the leaf rust pathogen Melampsora (Newcombe, unpubl. data). Leaf rust is the most important disease of Populus worldwide. Host resistance has been the only widespread and economical control method for which both pathotype-specific and non-specific types of resistance are known (Newcombe 1996). Exotic species of Populus frequently are resistant to native pathogens (Newcombe 1998a), and resistance is often simply inherited in F₁ interspecific hybrids. Genome analysis methods have allowed mapping of the genes for resistance to races E1, E2, and E3 of Melampsora larici-populina (Cervera et al. 1996) and the Mmd1 gene for resistance to Melampsora medusae (Newcombe et al. 1996). The Mmd1 gene is expected to be physically isolated and transformed into a susceptible genotype in the near future, demonstrating the feasibility of genetically engineering disease resistance using native genes.

Attempts to increase resistance using heterologous genes have so far given poor results, but work has been limited (e.g., Strauss et al. 1988). The prospect of a heterologous resistance gene having broad and durable effectiveness against major pathogens, without negative pleiotropic effects on fitness, appears remote using current and foreseeable technology. These transgenes therefore do not appear to have the potential to significantly impact poplars in wild systems via introgression of transgenes. Moreover, transgenes that might give a useful degree of resistance in a well-tended genetic monoculture such as a clonal plantation are unlikely to be comparably important to pathogen resistance in genetically and environmentally diverse wild populations. Simple alterations in expression of native poplar or pathogen genes, such as by inducing constitutive overexpression or cosuppression, were also considered unlikely to be of significant ecological consequence. The transfer of unmodified resistance genes between Populus species is commonplace in conventional poplar breeding and should bring about similar risks if accomplished via gene isolation and genetic transformation. This should apply equally to leaf rust (Newcombe 1996) and other diseases of Populus.

Insect damage is a major limitation to plantation viability and productivity in many regions (Ostry et al. 1988). Currently, the primary control method uses pesticides rather than resistant genotypes. Genetically based resistance is known but is often either incomplete or would require major alterations of breeding programs to accommodate, such as the use of different species as hybrid parents, with a consequent reduction in genetic improvement of other traits. The cottonwood leaf beetle (CLB) is the major pest of poplars in the United States and is believed to be largely restricted to poplars and other species in the same family (Salicaceae). The cry3a toxin from Bt (Bacillus thuringiensis) is highly toxic to CLB when applied topically or expressed in transgenic poplars (Strauss et al. 1998).

Other chrysomelid leaf beetles are also locally important. Lepidopteran defoliators are episodically significant, many have broad host ranges (e.g., gypsy moth and forest tent caterpillar), and most are sensitive to Cry1A Bt toxins (e.g., Kleiner et al. 1995). Wood borers of several taxa can be important pests in specific areas; because they are hard to reach with topical pesticides, the use of transgenes could be an important control option. Insect damage in wild stands is sporadic in space and time though rarely results in genotypic mortality because of poplar’s resprouting capability. Thus, invasion of established stands by progeny of insect resistant transgenic trees is expected to be very slow. Field trials of transgenic poplars with beetle- and caterpillar-active Bt transgenes are underway in several areas (e.g., Ellis and Raffa 1997; Yingchuan et al. 1993; Strauss et al. 1998), most notably China, France, and the northwestern United States. Other than Bt, work with alternative insect resistance transgenes has been limited. Proteinase inhibitors expressed in poplars have given either modest levels of resistance or none at all (e.g., Leplé et al. 1995; Confalonieri et al. 1988) and thus do not appear to be under consideration for commercial use. Genes with different modes of action, but as effective as Bt against poplar pests, are unknown.

The most important consideration when using Bt transgenes is the significant potential for development of Bt-resistant insect biotypes if the extensive transgenic poplar plantations are established without accompanying resistance management considerations (Raffa et al. 1997). High levels of CLB resistance have readily been bred in laboratory colonies under Cry3A selection (L. Bauer, pers. comm.). For most poplar plantations, wild stands are expected to provide large refugia that can slow resistance development and may obviate the need for planted refugia. However, the role of natural stands in the dispersal and mating behavior of target pests in areas where transgenic trees are being deployed should be studied. The working group considered that the potential for resistant biotype development from plantation use was a far greater concern than the risk of Bt transgenes providing a significant fitness advantage in wild trees after introgression. Sterility or other strategies for stringent gene containment were therefore not viewed as essential for use of pest resistance transgenes.

High levels of weed control for the first one to three years are essential for obtaining high rates of survival and tree growth in poplar plantations (Tuskan 1998). Plantation managers in many parts of the US believe that herbicide resistance (HR), particularly to glyphosate, can significantly reduce weed management costs and increase tree growth by providing more effective weed control and increasing moisture availability to trees (W. Schuette and J. Finley, pers. comm.). Because of the common use of poplars as windbreaks between agricultural fields and the future likelihood of their increased use for biofiltration plantings near streams in agricultural areas, HR poplars resistant to spray drift may be important components of agroecosystems dominated by glyphosate tolerant crops. Transgenic poplars with high levels of field resistance to glyphosate and phosphinothricin herbicides have been demonstrated in field trials (e.g., Strauss et al. 1998).

If transgenes are allowed to spread via seed, sprouting of HR poplars could complicate their control (Strauss et al. 1997). In some systems, poplars are considered "mild" weeds; examples include perennial crops (e.g., conifers), rights of way, and drainage ditches. Spread of HR trees would remove certain herbicides as control options, which could be an important loss in systems that must rely on one or few herbicides for control. HR trees also may complicate plantation management in significant ways, such as making the "volunteers" from seed produced in flowering stands more difficult to control in regenerating stands and requiring use of other chemicals for killing resprouts from stumps after harvest.

Poplars and other trees present substantial difficulties for extrapolating from small trials to large-scale effects for several reasons:

The working group identified seven research areas important to regulators and needed to improve overall scientific risk assessment (see Table 1). None of the knowledge gaps were considered so large that they should preclude commercial uses (i.e., none of the areas were rated as "urgent" under the regulatory decisions category in the Table), however, this conclusion presumes that reasonable research and monitoring are done as part of commercialization.

The most important research area identified was the continued acquisition of highly effective pest resistance genes. The transgenes presently available either do not provide sufficiently strong resistance to disease or, in the case of insect resistance, may not provide the functional diversity needed to adequately deter evolution of resistant pest biotypes. Use of multiple transgenes with different mechanisms for toxicity is highly desirable. Although engineered sterility systems (Strauss et al. 1995) are not considered essential to prevent ecological impacts, the group recognized that such systems would simplify scientific and regulatory assessments and avoid some important agronomic factors that would impact plantation management. Therefore, research on sterility mechanisms was considered as important as obtaining new resistance genes.

Intermediate priority was given to learning more about the following: reproductive biology, particularly rates of gene flow through pollen, seed, and vegetative spread; the degree of hybrid fertility; and factors limiting the spread of hybrid-derived genes into wild populations. The group felt that too little is known about the impacts of herbivores and diseases in wild populations, the nature and genetic variation of resistance mechanisms, and the way in which changes in resistance genetics might directly and indirectly affect species interactions and ecosystem processes. Interdisciplinary, long-term studies are required for advances in knowledge in these areas.

The group also concluded that too little socioeconomic and environmental impact data are available to assess the value of pest resistance transgenes on a broad basis. Studies are needed to identify the kinds of land uses and landowners that might be economically impacted by the spread of pest or herbicide resistant transgenes and quantify the extent of that impact. This information is also important for assessing the need for sterility systems. Finally, information is needed on the broader impacts of transgenic poplars to help society and government assess their socioeconomic and environmental importance. For example, what degree of economic and environmental values are expected in the medium term on farm, landscape, and regional levels for trees with multiple functional transgenes (e.g., herbicide resistance, insect resistance, sterility, and disease resistance)? If these transgenic technologies are used wisely, the economies they provide farmers and industries may significantly increase the representation of poplars in agroecosystems in place of annual crops, with multiple environmental benefits. This economic aspect needs to be understood on a regional and national level to guide research, policy, and regulatory decisions.

Table 1. Priority research concerning pest resistance genes in poplars needed to inform scientific analysis and regulatory decision making.

The group considered several broad hypotheses frequently encountered when considering the risks of transgenic poplars and other plants. For each hypothesis we accepted, refuted, or qualified the stated hypotheses.