Over 30 species of grasses are utilized for turf (Huff 1998), while others are important in agriculture as forage crops. The commercial value of this group of plants makes them attractive for improvement through modern genetic engineering techniques (Johnson and Riordan, in press). Because of the diversity of species and the consequent differences in biology among them, broad generalizations regarding the ecological effects of pest resistance genes introduced into these crops cannot be made. Rather, questions regarding the potential for pest resistance genes must be directed toward specific cases in which the species and the particular introduced gene are known. In keeping with this approach, particular attention at this meeting was paid to the turfgrasses—in particular creeping bentgrass (Agrostis palustris Huds.) and Kentucky bluegrass (Poa pratensis L.)—since these are the two grass species that have had transgenic lines tested in the field, and therefore are the species that present the greatest likelihood of being commercialized in the near future.

Specialized cultivation practices are also employed (P. Johnson, pers. comm.). Greens, tees, and fairways are "core aerified." During this procedure, cores of soil measuring 2-4" long, 1/4-3/4" in diameter, and spaced 3/4-1" apart, are pulled from the turf surface. The resulting holes are filled with sand. Many of the new cultivars of creeping bentgrass are frequently mowed vertically; blades that are held perpendicularly to the soil on a rapidly spinning shaft are used to cut stolons and reduce thatch buildup in the turf. This is done regularly, varying from every day on some greens to twice a year on some fairways.

The working group considered evidence that introduction of a pest resistance trait could increase the ability of the crop to become established, persist, or spread. Diseases are clearly a factor in the distribution of turfgrasses as crops. Thus, disease resistances have enabled more extensive planting of turfgrasses within the range in which they are used. Examples of traits that have allowed more extensive planting are resistance to summer patch, Phythium blight (Pythium spp.), brown patch, and snow mold. It is important to note that turfgrass spread is due to human planting rather than any inherent ability conferred on these crops to spread on their own as a consequence of disease resistance. In the same way, tolerance to heat, salt, and other environmental stress tolerance have expanded the range in which these crops can be grown.

To evaluate potential ecological effects of introduced pest resistance genes, the first step is to examine evidence that pests have a significant effect on populations of plant species that are sexually compatible with the crop. The working group was unaware of any evidence that introduction of a disease or pest resistance trait had resulted in the release of turfgrasses or their sexually compatible relatives from any control exerted by those diseases or pests. However, the potential utility of studying endophytes was discussed. Turfgrasses are known to obtain significant benefits from endophytic fungi, which confer greater overall vigor on the plants and therefore might provide greater tolerance to pests or disease. In this respect, evidence obtained from the comparative study of plants with and without endophytes may provide insight into effects of genes that confer resistance to pests, pathogens, or environmental stress. The effect of endophytes might also be useful as a model of the effects of broad pest resistance genes on the fitness and potential weediness of the crop species.

As with the lack of evidence concerning the effect of pest resistance genes on turfgrasses, there was also a lack of knowledge regarding the similarity of pests and pathogens attacking crops and their sexually compatible relatives. It was assumed that the pests affecting the crops also affected sexually compatible species. However, this lack of information was seen as an important gap that needs to be filled.

The working group began its consideration of this issue with a discussion on the potential effects of herbicide tolerance. It was concluded that herbicide tolerance would probably not make these crops more invasive. Creeping bentgrass and Kentucky bluegrass possess several traits that render these species ill-adapted for unmanaged situations; hence, the single trait of herbicide resistance was judged to be insufficient to cause these species to become weedy. However, the effect that engineering glyphosate resistance in bentgrass could have on annual bluegrass was raised as a concern. In this case, the effect would not be caused by the transfer of the herbicide resistance trait from bentgrass to annual bluegrass; rather, use of glyphosate on the bentgrass would raise the selection pressure exerted upon annual bluegrass. The emergence of resistance within this species would be accelerated, leading to the consequent loss of glyphosate as a weed management tool in golf course greens. The net result would be reversion to the present situation in which control of annual bluegrass in bentgrass by spraying with glyphosate is not possible. Despite this concern, the potential benefits of glyphosate resistance (reduced use of herbicide, a more environmentally friendly herbicide) was judged to counteract the concern presented.

After further discussion of the consequences of pest resistance, the group concluded that an assessment of the potential effects of pest resistance traits required a case-by-case evaluation. Important considerations in this assessment included the type and mechanism of resistance. In particular, broad spectrum pest resistance was viewed as being of special concern. Therefore, a hypothetical example was considered in which broad spectrum pest resistance was engineered into creeping bentgrass and subsequently transmitted to the sexually compatible species, redtop (Agrostis gigantea). For this specific example, it was judged unlikely that pests or pathogens limited populations of bentgrass or redtop. Therefore, even the addition of broad spectrum pest resistance would be unlikely to convert either species to a weed. Additionally, bentgrasses are perennials that do not display many traits seen for typical weeds. Based on these reasons, the addition of pest resistance was seen to be of little concern in the cases of bentgrass and redtop. On the other hand, it was recognized that the conclusion could be different for a species such as buffalograss, which is more likely to have populations controlled by pests or pathogens.

The group concluded that we are currently lacking important information for evaluating the effect of pest resistance genes on the establishment, persistence, and spread of the crop or its sexually compatible relatives. Basic information on the natural history and biology of turfgrasses, as well as weeds in general, was seen as important in evaluating the effect on weediness of an introduced resistance gene in the crop or sexually compatible relatives. Specific areas where information should be obtained or compiled are:

• The life history and invasiveness of the various turfgrass species.
• The geographic range of related species, as well as the cross-compatibility of those species with crop species. Some information on crossing relationships is already known (see for example, Johnson and Riordan, in press). However, local variations in genotype and ploidy will result in different rates of transmission of a transgene to sexually compatible relatives.
• The range of pests and pathogens that attack the sexually compatible relatives.
• The factors (including pests and pathogens) that limit populations of sexually compatible relatives.
• The rate of increase of populations of sexually compatible relatives, and the factors that control them.
• A greater understanding of the characteristics of weeds in general. A more thorough study of the characteristics that predispose plants to becoming weeds is needed.

The information listed above can be obtained from a number of sources or through experimentation. The committee discussed the following sources and general approaches:

Extrapolation from small to large scale was not seen to present as great a problem in turfgrasses as it may in other crops. In the case of creeping bentgrass, releases will be on a relatively small scale, since golf courses are typically only 100-200 acres, and bentgrasses make up even a smaller proportion of that area (about three acres). Consequently, any information that might be obtained in small scale risk assessment studies could be readily extrapolated to commercial- scale release. The fact that management practices are relatively uniform throughout the range of commercial releases also increases the applicability of data obtained from small scale tests to wide-scale releases. However, certain factors could affect the ability to extrapolate from small scale to large scale:

The use of small scale trials to predict performance on a large scale is a standard tool of plant variety development. In plant breeding, there is a long history of extrapolating performance based on small plot trials. In the case of turfgrasses, knowledge of performance is gained through National Turfgrass Evaluation Trials. Although the information obtained from these trials does not usually address the issue of wild relatives becoming weeds, considerable observational data on the weediness potential of the crops themselves could be gathered from these types of trials.

References: