*For offprint requests: fax: (910)334-5839, email: nstewart@goodall.uncg.edu

SUMMARY

Canola (Brassica napus) transgenic for a Bacillus thuringiensis (Bt) transgene is insecticidal towards certain caterpillars. To simulate an escape of the transgenics from cultivation, a field experiment was performed in which transgenic and non-transgenic plants were planted in natural vegetation and semi-cultivated plots and subjected to various insect selection pressure in the form of herbivory. Only two plants survived the winter to reproduce in the natural-vegetation plots which were dominated by perennial grasses such as crabgrass. However, in semi-cultivated plots, medium to high levels of defoliation decreased survivorship of non-transgenic plants relative to Bt-transgenic plants, and increased differential reproduction in favor of Bt plants. Thus, where suitable habitat is available, such as roadsides, fallow fields, and field edges, there is a strong likelihood of enhanced ecological risk associated with the release of certain transgene/crop combinations such as insecticidal canola.

Keywords: Brassica napus, Bacillus thuringiensis, crystal delta endotoxin, population replacement, invasiveness

INTRODUCTION

Genetically modified plants are rapidly becoming commercialized and released into the environment (Dale et al., 1993; Sawahel, 1994; Rogers and Parkes, 1995). Several studies have addressed various safety and risk issues regarding food safety and the ecological risks of releasing transgenic plants into the environment (Kareiva et al., 1994; Kok et al., 1994; Regal, 1994; Purrington and Bergelson, 1995). Most often, the risk of transgene spread and introgression have been studied, while the ecological effects of the transgenes themselves have been largely ignored (Crawley et al., 1993; Linder and Schmitt, 1994; Mikkelsen et al., 1996). While neutral transgenes may have minimal ecological effects to their hosts and ecological systems, a transgene conferring increased fitness may have significant effects (Kareiva et al., 1994; Linder and Schmitt, 1995). Fitness-conferring transgenes may have their greatest effects when they are in a host that may persist outside of cultivation, and/or may be spread from an agricultural host to a weedy relative. It is conceivable that a crop plant may become weedy or related weeds could become weedier.

To assess a possible worse-case scenario, we transformed canola (Brassica napus), a crop produced for oilseed, with a synthetic Bacillus thuringiensis crystal endotoxin transgene (Bt CryIAc) (Stewart, Jr. and Via, 1993), and performed an experiment to simulate a transgenic crop escape. We deployed two cultivars, transformed and non-transformed, in a tented field study in which herbivorous insect availability was manipulated. Fitness was compared between transgenics and non-transgenics to assess whether population replacement was likely. We tested the null hypothesis that there would be no difference in fitness between transgenic vs. non-transgenic plants in vegetated and in semi-cultivated plots.

MATERIALS AND METHODS

A split plot design with complete randomization was used to determine the effects of vegetation type and insects on canola fitness (Bt-transgenic vs. non-transgenic). The field experiment was performed on the University of Georgia Horticulture Farm in Watkinsville, GA, USA, which is located in the Piedmont physiographic region. The field, which had been out of cultivation for 10-15 years, had moderately low fertility and over 90% cover of perennial grasses such as Agrostis ssp., Andropogon ssp., Digitaria ssp., and Paspalum ssp. One half of the 72 plots were cultivated and one half were left naturally vegetated. The semi-cultivated plots were roto-tilled using common equipment. After the seeds were planted no further cultivation was performed. The semi-cultivated plots contained numerous annuals by early spring, 1996. The plots (1 m²) contained 100 seeds of each of 4 lines: two cultivars (Oscar and Westar), transformed (OBT and WBT) and non-transformed (O and W). OBT was a moderately high expressing line (0.05% Bt CryIAc), and WBT was a low expressing line (0.005% Bt CryIAc) (Stewart, Jr., Via, 1993). Plant line was randomized among rows within plots. Plots were randomized within blocks. Seeds were planted September 23, 1995. Each plot was subjected to one of six treatments: Tent--1 m² insect cages that enclosed the plants and vegetation; Insecticide--tented treatment in which Malathion (6.25 ml containing 56.8% active ingredient and diluted in 83.3 ml of water) was applied throughout the fall growing season to keep plots free of insects; CEW--tented treatment in which corn earworm neonates were applied to all plants with a "bazooka" (Wiseman et al., 1980) at a rate of 800 larvae per plot row; DBM--tented treatment in which diamondback moth neonates were applied to all plants with a bazooka at a rate of 3150 larvae per plant; CEW+DBM--tented treatment in which corn earworm and diamondback moth were applied together at one half the above rates; Control--a treatment with no tent. Corn earworm eggs were obtained from the USDA-ARS Insect and Population Management Research Laboratory, Tifton, GA. Diamondback moth eggs were obtained from Abbott Laboratories, Chemical and Agricultural Products Division, Chicago, IL. The eggs were hatched in the laboratory and 2-4 hour-old neonate larvae were mixed with corn grits prior to application. The insects were applied 30 days after seeds germinated. Insect survivorship and damage data were taken 15, 40, and 150 days after insect applications.

RESULTS AND DISCUSSION

Vegetation effects. To test for vegetation effects, one-half of the 72 plots were cultivated and one half were left naturally vegetated. The naturally vegetated plots, dominated by perennial grasses, contained only two (of 3200 seed planted) canola plants (WBT in a control plot) that survived the winter to reproduce. So, although these two plants were transgenic, it seems that the transgene did not confer the ability to allow canola to outcompete the perennial grasses that were already established.

In contrast with the above result, in semi-cultivated plots there were no differences among transgenic and non-transgenic cultivars in establishment after planting but Oscar (transgenic and non-transgenic) had better establishment than Westar, (21.2 vs. 17.5 plants per plot row 30 days after planting) regardless of transgene status. There was no catastrophic survivorship decline after winter (Table 1). Furthermore, there were no apparent morphological differences between transgenic and transgenic plants prior to insect treatments.

Insect effects. To assess the effect of a single defoliation episode of CEW and DBM, we measured defoliation and damage rating of the plants, and also censused insects of semi-cultivated plots. There were means of 2.8 CEW, 19.6 DBM, and 17.2 CEW and DBM in each of respective plot 15 days after application. By any means, the level of infestation represents a small single infestation. There were no appreciable herbivores present in any of the tented treatments other than a small number of aphids. However, during the winter, the Control (untented) treatment canola were apparently eaten by rabbits or deer. For this reason, the Control treatment is excluded from most analyses. There were significant effects in defoliation and damage rating as the result of the DBM and the CEW+DBM treatments (Table 1). In addition, the damage rating for CEW was significant. It is clear that the Brassica specialist, DBM had a prominent effect on the plants at the end of the fall growing season. The effect of a CEW infestation was less clear.

It is not unexpected that there would be profound differences in defoliation between transgenics and non-transgenics as the result of a DBM infestation. To determine whether defoliation differences translated to differences in reproductive success, we monitored plant survivorship and reproductive effort throughout the following winter and fall. There were no differences in overwinter survivorship among plants except for within the DBM treatment, where there was a marginally significant effect noted among transgenics and non-transgenics at P = 0.10 (Table 1). Thus, the large amount of defoliation did not effect plant survivorship except in the extremely-defoliated plots. Nonetheless, there were apparent effects of defoliation on seed production (Figure 1). When only the DBM treatment is considered, the non-transgenics suffered various degrees of defoliation (from 5 to 98%) and never produced more than 80 seeds per row (Figure 1). However, the transgenic plants never suffered more than 5% defoliation and ranged from 0 to over 750 seeds per row (Fig 1). When all tented treatments are considered and ratios of non-transgenic to transgenic plants are computed, the same significant (at P = 0.05) inverse relationship between defoliation and seed production is observed (Figure 2). In this analysis, when the non-transgenic/transgenic seed ratio equals one, then there are equal amounts of seed produced by each type of plant. When this ratio is greater than one, there are more seeds produced by non-transgenics compared with transgenics. The ratio greater than one was observed in only 17% of the plots where insects were applied, but was twice that frequency in plots where insects were not applied. Thus, herbivory by insects was an effective selection agent in favor of transgenic insecticidal plants. Furthermore, effective selection only required low differentials of insect numbers (on non-transgenic vs. transgenic) and a single episode of herbivory.

Although the field portion of the study is still in its first year, there is a clear indication that insecticidal canola could pose an ecological risk upon environmental release. Since canola is a minor weed (Williamson, 1992), the ability to strongly resist defoliation may allow it to selectively persist to a greater extent (Hoffman, 1990; Andow, 1994; Dale, 1994; Schmitt and Linder, 1994). As Andow (Andow, 1994) has pointed out using mathematical modeling, a drastic alteration in one single trait could increase a plant's competiveness, thereby allowing it to impact natural plant populations and communities. There are several examples of single trait alterations that have impacted plant competiveness. One of which is chestnut blight (Andow, 1994), in which the lack of resistance to a single pathogen reduced American chestnut (Castanea dentata) from a dominant tree to an incidental, non-flowering shrub throughout its distribution. It is foreseeable that the introduction of a single profound trait could also cause a drastic shift in the weediness potential of canola. Perhaps even more importantly, since canola readily outcrosses with other members of the Brassicaceae, such as wild mustard, a gene conferring insect resistance will likely rapidly become fixed in weedy Brassica's. Insect resistant wild mustard is certainly not a desirable organism in either agricultural or natural ecosystems (Kareiva et al., 1994).

Successive years' data will be collected that should indicate the speed which population replacement will occur. The speed at which Bt CryIAc canola replaces non-transgenic canola will likely depend on not only selection pressure but also other environmental conditions, and therefore may not be readily predictable. Thus, only through long-term monitoring and experiments will the ecological risk of such crop/transgene combinations be known.

ACKNOWLEDGEMENTS

This work was supported by a grant from the USDA Biotechnology Risk Assessment Research Grants Program. We thank Jacque Bailey, Mitch Gilmer, Staci Leffel, Steve Mabon, Derrick Pulliam, Berry Tanner, and Donna Wyatt for help in data collection.

REFERENCES

Andow DA 1994. Community response to transgenic plant release: using mathematical theory to predict effects of transgenic plants. Molecular Ecology 3:65-70.

Crawley MJ, Hails RS, Rees M, Kohn D, Buxton J. 1993. Ecology of transgenic oilseed rape in natural habitats. Nature 363:620-623.

Dale PJ, Irwin JA, Scheffler JA. 1993. The experimental and commercial release of transgenic crop plants. Plant Breed 111:1-22.

Dale PJ. 1994. The impact of hybrids between genetically modified crop plants and their related species: general considerations. Molecular Ecology 3:31-36.

Hoffman CA. 1990. Ecological risks of genetic engineering of crop plants. BioScience 40:434-437.

Kareiva P, Morris W, Jacobi CM. 1994. Studying and managing the risk of cross-fertilization between transgenic crops and wild relatives. Molecular Ecology 3:15-21.

Kok EJ, Noteborn HPJM, Kuiper HA. 1994. Food safety assessment of marker genes in transgenic crops. Trends in Food Science and Technology 5:294-298.

Linder CR, Schmitt J. 1994. Assessing the risks of transgene escape through time and crop-wild hybrid persistence. Molecular Ecology 3:23-30.

Linder CR, Schmitt J. 1995. Potential persistence of escaped transgenes performance of transgenic, oil modified Brassica seeds and seedlings. Ecol Appl 5:1056-1068.

Mikkelsen TR, Andersen B, Jorgensen RB. 1996. The risk of crop transgene spread. Nature 380:31

Purrington CB, Bergelson J. 1995. Assessing weediness of transgenic crops: industry plays plant ecologist. Trends in Ecol and Evol 10:340-342.

Regal PJ. 1994. Scientific principles for ecologically based risk assessment of transgenic organisms. Molecular Ecology 3:5-13.

Rogers HJ, Parkes HC. 1995. Transgenic plants and the environment. J Exp Bot 46:467-488.

Sawahel WA. 1994. Transgenic plants: performance, release and containment. W J Microbiol Biotechnol 10:139-144.

Schmitt J, Linder CR. 1994. Will escaped transgenes lead to ecological release? Molecular Ecology 3:71-74.

Stewart CN Jr, Adang MJ, All JN, Raymer PL,, Ramachandran S, Parrott WA. 1996. Insect control and dosage effects in transgenic canola, Brassica napus L. (Brassicaceae), containing a synthetic Bacillus thuringiensis cryIAc gene. Plant Physiol. In Press.

Williamson M. 1992. Environmental risks from the release of genetically modified organisms (GMOs)-- the need for molecular ecology. Molecular Ecology 1:3-8.

Wiseman BR, Davis FM, Campbell JE. 1980. Mechanical infestation device used if fall armyworm plant resistance programs. Fl Entomol 63:425-432.

Table 1. Responses of insecticidal and nontransgenic oilseed rape to various treatments (described in the Materials and Methods ). Damage rating of plants and defoliation were assessed at 40 days after treatments. The number of plants per row were assessed the next spring, 6 mo after treatments. The types of the plants were as follows: O--Oscar, OBT--Oscar transgenic for BT CryIAc, W--Westar, WBT--Westar transgenic for Bt CryIAc. Different letters denote significant differences at P=0.05 (Tukeys HSD within columns.

Treatment	Plant type	Plants (number)	Defoliation (%)	Damage Rating*
CEW	O	8.5 ± 4.7	6.3 ± 5.4 b	3.5 ± 2.4 b
	OBT	10.3 ± 5.6	0.50 ± 1.2 b	0 ± 0 c
	W	7.7 ± 4.4	7.3 ± 6.8 b	3.7 ± 2.3 b
	WBT	8.5 ± 3.5	2.3 ± 1.6 b	0.33 ± 0.52 c
DBM	O	4.4 ± 2.9**	54.0 ± 38.9 a	6.0 ± 1.4 a
	OBT	9.0 ± 4.9	0.60 ± 0.55 b	0.20 ± 0.45 c
	W	2.6 ± 1.9	61.0 ± 35.8 a	6.4 ± 1.3 a
	WBT	9.8 ± 6.5**	4.6 ± 1.7 b	1.4 ± 0.89 c
CEW + DBM	O	9.4 ± 4.7	41.0 ± 18.8 a	6.2 ± 0.45 a
	OBT	12.2 ± 6.5	1.6 ± 1.1 b	0.20 ± 0.45 c
	W	8.0 ± 3.1	52.0 ± 13.0 a	6.8 ± 0.45 a
	WBT	8.4 ± 4.5	5.4 ± 8.2 b	0.80 ± 0.84 c
Tent	O	11.5 ± 5.6	0 ± 0 b	0 ± 0 c
	OBT	12.0 ± 5.3	0.83 ± 2.0 b	0.10 ± 0.05 c
	W	9.2 ± 2.4	3.3 ± 5.2 b	0.21 ± 0.12 c
	WBT	11.0 ± 2.5	0.33 ± 0.82 b	0 ± 0 c
Insecticide	O	11.3 ± 3.8	0 ± 0 b	0 ± 0 c
	OBT	13.7 ± 6.9	0 ± 0 b	0 ± 0 c
	W	11.8 ± 8.5	1.17 ± 2.0 b	0.17 ± 0.41 c
	WBT	9.8 ± 4.1	0.67 ± 1.6 b	0 ± 0 c

*0-7 plant damage rating of a row. 0, virtually no damage apparent on plants; 1, <5% plants with some damage; 2, 5-10% of the plants with damage; 3, 11-20% of the plants damaged; 4, 21-30% of the plants damaged; 5, 31-60% of the plants damaged; 6, 61-80% of the plants damaged; 7, >81% of the plants damaged.

**denotes significant differences within the column (P = 0.10).

Figure 1. The relationship of % defoliation recorded in November, 1995 and the resultant seed production of canola in May, 1996. Only data from the diamondback moth treatment are displayed.

Figure 2. Ratios of non-transgenics to Bt CryIAc transgenics for % defoliation and seed production. Each data point represents a plot row of a cultivar non-transgenic/transgenic (e.g., Oscar/Bt CryIAc Oscar). Defoliation were assessed at 40 days after treatments. The number of seeds per row were assessed the next spring, 6 months after treatments. The treatments are described in the Materials and Methods.