*For offprint requests: tel: (204) 474-6074, fax: (204) 261-5732, email: McVetty@bldgagric.lan1.umanitoba.ca

SUMMARY

Ten metre wide border areas of non-transgenic Brassica napus strains which flower synchronously with confined transgenic B. napus strains are frequently used, instead of large isolation zones, to limit the spread of transgenic B. napus pollen for summer rape field trials conducted in Canada. Since there is limited information on the effectiveness of these border areas under Canadian conditions, a study was undertaken to determine the effectiveness of border areas in confining the spread of transgenic B. napus pollen. Border areas 15 m to 30 m wide were planted around a 30 m x 30 m central plot of bromoxynil (a Rhône-Poulenc product) herbicide resistant transgenic B. napus strains for field trials conducted at Carman and Winnipeg, Manitoba in 1994. Seed samples were taken from the border areas at distances of 0, 2.5, 5, 10, 15 and 30 m in all four cardinal compass directions. These seed samples were planted in the field in 1995 and screened for the presence of bromoxynil resistant plants (i.e. the outcross type) in otherwise bromoxynil susceptible populations. Outcross rates were affected by both direction and distance. Outcross rates varied from approximately 0.5% at 0 m to 0.015% at 30 m for data combined over direction and site. An exponential decay model, for outcross rates as a function of distance, y = 0.54e^-0.25x (R² =0.99) was fitted to data combined over direction and site. Border areas were found to be very effective in confining transgenic B. napus pollen.

Key words: transgenic pollen, gene flow, containment, herbicide resistance

INTRODUCTION

While genetically modified (i.e. transgenic) plants are developed and initially assessed indoors in controlled environments, the goal of researchers developing these materials is often unrestricted use in the outdoor environment. The initial release of genetically modified plants into the outdoor environment is frequently done using small plot field trials. These field trials provide the opportunity for gene movement to other plants, including related weed species and related crops.

Even though B. napus is self-fertile, fully capable of self-fertilization, it possesses sticky, entomophilous pollen well suited for transfer by insects (Eisikowitch, 1981). Crosses can occur naturally between B. napus and other cultivated Brassica species, including B. rapa (syn campestris), B. juncea, and B. oleracea (Bing et al., 1991; Kerlan et al., 1991; 1992; 1993; Leckie et al., 1993). B. napus readily outcrosses within its species to varying degrees, as reported from studies using elevated erucic acid levels as a marker, from as little as 5% to as much as 74% (Huhn and Rakow, 1979; Rakow and Woods, 1987; Lewis and Woods, 1991). Huhn and Rakow (1979) reported outcross rates from plot to plot in B. napus ranged from 5 to 15%. In contrast, Lewis and Woods (1991) reported that plant to plant within-plot outcross rates for field grown B. napus ranged from 22 to 74% and Rakow and Woods (1987) reported that plant to plant within-plot outcross rates for field grown B. napus ranged from 12.5 to 38%. Outcross of transgenic B. napus to related Brassica crops, especially B. napus crops is possible, perhaps even probable.

Gene flow in B. napus small plot field trials is affected primarily by 1) isolation zones and 2) border areas or trap crops (Morris et al., 1994). The use of isolation zones and border areas are therefore the two main strategies which are employed in field trials to confine the spread of transgenic B. napus pollen.

Bilsborrow and Evans (1994) studied the effect of isolation distance on gene movement in B. napus, using erucic acid levels as a marker, and reported that a 10 m isolation distance between commercial fields of B. napus grown in England reduced the outcross rate to approximately 2%. Morris et al.(1994) conducted field trials of transgenic kanamycin resistant B. napus in America, and found that isolation distances of 4 m resulted in outcross rates of 0.90 to 1.1%, while isolation distances of 8 m resulted in outcross rates of 1.2 to 1.9%. Scheffler et al. (1995) grew genetically modified glufosinate resistant B. napus in field trials in England with large isolation distances of 200 and 400 m and reported outcrossing rates of 0.0156% at 200 m and 0.0038% at 400 m.

Morris et al. (1994) also investigated the effects of border areas or trap crops on B. napus gene flow in America, using a continuous border area surrounding a central genetically modified plot of kanamycin resistant B. napus. These authors reported that outcross rates were 2.0 to 3.5% at 0 m distance from the transgenic plot, and that these declined rapidly to 0.6 to 0.7% at 5 m from the transgenic plot.

Wilkinson et al. (1995) studied gene flow from commercial fields of summer B. napus to adjacent commercial fields of winter B. napus in Scotland, using vernalization requirement as a marker and Randomly Amplified Polymorphic DNA (RAPD) analysis to confirm the identity of putative outcross hybrids. In this experiment, the winter crop was the functional equivalent of a border area or trap crop. Wilkinson et al. (1995) reported that hybrid frequency decreased rapidly,(frequency at 0 m was not reported), with distance from the pollen source but stabilized by 32 m at rates of 0.035 to 0.05%.

In Canada, small plot field trials of genetically modified B. napus strains frequently use 10 m wide non-transgenic B. napus border areas to contain the spread of pollen. The purpose of this study was to determine the effectiveness of non-transgenic B. napus border areas in minimizing pollen flow from transgenic B. napus plots. While most previous studies have used erucic acid levels to study gene movement, a trait that is difficult to assess for large seed samples, the use of a single, Mendelian dominant gene trait, such as bromoxynil herbicide resistance allows for easier and more accurate assessment of outcross rates and effectiveness of border areas in confining transgenic B. napus pollen spread in the field in Canada.

MATERIALS AND METHODS

Non-transgenic B. napus border areas 15 to 30 m wide were planted around central plots of genetically modified bromoxynil (Rhône-Poulenc product) herbicide resistant B. napus (Freyssinet et al., 1995) at Carman and Winnipeg, Manitoba in 1994. Samples were taken in two strips on all four sides (the four cardinal compass directions) of the border at distances of 0, 2.5, 5, 10, 15, and 30 m. Seedlings from the samples were screened in the field in 1995. Approximate seed numbers screened at each distance sampled were: 0, 2.5 and 5 m: 130,000; 10 m: 260,000; 15 m: 390,000; 30 m: 420,000. Screening involved planting the seed samples in the field in 1995 in 6 row plots 1.2 m wide by 30 m long. Total number of emerged seedlings per plot were determined prior to application of bromoxynil herbicide at 700 g ai/ha (approximately 2.5 times the field dosage normally used in cereal crops) when the B. napus plants were in the 3 to 4 leaf stage. After the bromoxynil herbicide activity was obvious, about three days after application, the number of plants surviving in each plot, i.e. the apparently bromoxynil herbicide resistant individuals, were counted. Bromoxynil herbicide was then re-applied to confirm that surviving plants were truly resistant to the herbicide and not escapes. The outcrossing rate for each plot was determined as the ratio of the number of surviving seedlings to the total number of seedlings originally present in the plot. Data was statistically analysed using SAS V.5 to determine the effects of direction, distance and site on outcross rate.

An exponential decay model (y=ae^-cx) was fitted to the data, where y = percent outcrossing, x = distance from pollen source (m), c = nonlinear coefficient determining slope of the curve, a = intercept at 0 m distance, and e = base of the natural logarithm. Parameter standard errors were determined and R² values were calculated to determine the fit of the model to the data.

RESULTS

The exponential decay model parameter estimates, for outcross rates as a function of distance for each direction and site, for combined directions and for combined directions and sites are presented in Table 1. The outcross rates at 0 m, ("a" values) ranged from 0.13 to 1.14%, while the slope estimates, ("c" values) ranged from -0.12 to -1.23 (Table 1). All parameter estimates were significantly different from zero, while R² values ranged from 0.77 to 0.99 (Table 1).

The outcross rates in the east direction at both Winnipeg and Carman at 0 m and in the south direction at Carman at 0 m were significantly higher than for all other directions and sites (Table 1). Additionally, the outcross rates for the east direction at Winnipeg were significantly higher than for the other directions, for the 0, 2.5, 5 and 10 m distances from the transgenic plot (Figure 1), while the Carman east direction outcross rates were significantly higher only for the 0 m and 2.5 m distances (Figure 1). The Carman south direction had a significantly higher outcross rate only at the 0 m distance (Figure 1). All other directions at both sites had much lower outcross rates at all distances from the transgenic plot (Figure 1). The outcross rates combined over directions for Carman and Winnipeg were approximately 0.6% at 0 m and these rapidly declined to approximately 0.015% at 30 m (Table 2). The decline in outcross rate as a function of distance was very rapid for all directions at both sites.

DISCUSSION

The exponential decay model for outcross rate as a function of distance fitted the data very well, as indicated by the high R² values for all directions at both sites (Table 1). Clearly, distance from the transgenic pollen source had a major effect on outcrossing rate in this study. The exponential model provided a much better fit to the data than the Weibull model (statistics not shown), another commonly used probability distribution.

The outcross rate to border areas was also affected by direction in this study (the highest outcross rate was in the east direction, especially at short distances from the central transgenic plot) (Figure 1). There is no obvious explanation for this significant effect of direction. There are no references in the literature citing significant effects of direction on outcross rate in insect pollinated crops. Further studies are in progress to confirm the reproducibility of this significant direction effect on B. napus outcross rate.

Outcross rates were low at 0 m for all directions and sites (less than 1.2%), and declined rapidly to very low levels at 30 m (less than 0.025%)(Table 2). For Carman, Winnipeg and for the two sites combined, over 93% of the outcross plants were present in the first 10 m of the border (Table 2). Conversely, less than 7% of the outcross plants detected in this study occurred in the next 20 m of border (the 10 to 30 m distance) (Table 2). Combined over directions and sites, average outcross rates decreased from approximately 0.5% at 0 m to less than 0.1% at 10 m, and further decreased to approximately 0.015% at 30 m (Table 2). Expressed in terms of frequency of outcross plants per total plants sampled, 1 plant in 200 was a result of an outcross event at 0 m, while 1 plant in 1000 and 1 plant in 5000 were the result of outcross events at 10 m and 30 m, respectively.

The effectiveness of border areas, as observed in this study, versus the effectiveness of isolation distances as reported in the literature can be compared, albeit tenuously, since insect pollinator activity levels may affect, possibly differentially, observed outcross rates for border areas versus isolation zones. The B. napus outcross rate into the border area at 10 m distance in this study is an order of magnitude lower than for 8 to 10 m isolation zones as reported by Bilsborrow and Evans (1994) or Morris et al. (1994). In addition, outcross rates into the border area at 30 m observed in this study are comparable to the outcross rates observed by Scheffler et al.(1995) for a 200 m isolation zone. These results suggest, but can not confirm, because of several confounding factors, that border areas are quite effective at minimizing B. napus pollen flow compared to isolation zones.

The outcross rates observed in this study, for a central plot of genetically modified B. napus strains to surrounding border areas, are approximately one order of magnitude lower than those reported by Rakow and Woods (1987) and Lewis and Woods (1991) for plant to plant within-plot outcross rates in B. napus. Plant to plant within-plot outcrosses in B. napus can occur through physical contact of the plants, something which could not occur in this study. The outcross rates observed in this study were also nearly an order of magnitude lower than those observed by Huhn and Rakow (1979) for winter B. napus plot to plot outcross rates in Germany. It would appear that insect pollinator activity in Germany is much higher than in Manitoba.

The average outcross rates to border areas observed in this study are substantially lower (0.54% at 0 m declining to 0.14% at 5 m) than the average outcross rates to border areas reported by Morris et al.(1994), (2.75% at 0 m declining to 0.65% at 5 m). Insect pollinator activity was apparently much greater in Morris et al.(1994) American location field trials than in this study.

The average outcross rates to border areas at 30 m observed in this study are also lower (0.015% at 30 m) than those reported by Wilkinson et al.(1995)(0.04% at 32 m) for B. napus field trials in Scotland. Again, insect pollinator activity was apparently much greater in Wilkinson et al.(1995) Scotland location field trials than in this study.

The preliminary results presented in this paper are presently being confirmed through the assessment of samples taken from three additional B. napus outcross study sites grown in Manitoba in 1995.

In conclusion, the results of this study confirm that border areas are very effective in confining the spread of transgenic B. napus pollen in the field in Canada.

ACKNOWLEDGMENTS

This research was supported in part by a contract with the Agriculture and Agri-Food Canada branch of the Federal Government of Canada, and in part by a contract with Rhône-Poulenc. Rhône-Poulenc also kindly granted permission to use bromoxynil resistant B. napus strains in this study and to publish the results of this study.

REFERENCES

Bilsborrow PE, Evans EJ (1994) Pollen transfer between high and low erucic acid oilseed rape crops. Asp-Appl-Biol 40: 143-149.

Bing DJ, Downey RK, Rakow GFW (1991) Potential of gene transfer among oilseed Brassica and their weedy relatives. GCIRC Eighth International Rapeseed Congress: 1022-1027.

Eisikowitch D (1981) Some aspects of pollination of oil seed rape (Brassica napus L.). J. Agric. Sci. Camb. 96:321-326.

Freyssinet M, Creange P, Renard M, McVetty PBE, Derose R, Freyssinet G (1995) Development of transgenic oilseed rape resistant to oxynil herbicides. GCIRC Ninth International Rapeseed Congress: 974-976.

Huhn M, Rakow G (1979) Einige experimentelle Ergebnisse zur Freindbefruchtungsrate bei Winterraps (Brassica napus oleifera) in Abhangigkeit von Sorte und Abstand. Z. Pflanzenzuchtung 83: 289-307, [In German, English abstract]

Kerlan MC, Chevre AM, Eber F, Baranger A, Renard M (1992) Risk assessment of outcrossing of transgenic rapeseed to related species: I. Interspecific hybrid production under optimal conditions with emphasis on pollination and fertilization. Euphytica 62: 145-153.

Kerlan MC, Chevre AM, Eber F, Bottermand J, DeGreef W (1991) Risk Assessment of gene transfer from transgenic rapeseed to wild species in optimal conditions. GCIRC Eighth International Rapeseed Congress: 1028-1033.

Kerlan MC, Chevre AM, Eber F (1993) Interspecific hybrids between a transgenic rapeseed (Brassica napus) and related species: cytogenetical characterization and detection of the transgene. Genome 36: 1099-1106.

Leckie D, Smithson A, Crute IR (1993) Gene movement from oilseed rape to weedy populations - a component of risk assessment for transgenic cultivars. Asp-Appl-Biol 35: 61-66.

Lewis LJ, Woods DL (1991) Outcrossing in summer oilseed rape (Brassica napusError! Bookmark not defined. L.) under Peace River, Alberta conditions. GCIRC Eighth international rapeseed congress: 1528-1531.

Morris WF, Kareiva PM, Raymer PL (1994) Do barren zones and pollen traps reduce gene escape from transgenic crops? Ecol-Appl 4: 157-165.

Rakow G, Woods DL (1987) Outcrossing in rape and mustard under Saskatchewan prairie conditions. Can. J. Plant Sci. 67: 147-151.

Scheffler JA, Parkinson R, Dale PJ (1995) Evaluating the effectiveness of isolation distances for field plots of oilseed rape (Brassica napus) using a herbicide-resistance transgene as a selectable marker. Plant Breeding 114:317-321.

Wilkinson MJ, Timmons AM, Charters Y, Dubbels S, Robertson A, Wilson N, Scott S, O'Brien E, Lawson HM (1995) Problems of risk assessment with genetically modified oilseed rape. Brighton Crop Protection Conference - Weeds - 1995: 1035-1044.

Table 1. Exponential decay model y=ae^-cx parameter estimates for percent outcrossing of transgenic Brassica napus to border areas in 1994 by direction and by site (values in parentheses are S.E.)

Direction	Site	a	c	R2
EAST	Carman	0.72 (0.05)	-0.27 (0.04)	0.98
	Winnipeg	1.14 (0.18)	-0.12 (0.04)	0.85
WEST	Carman	0.43 (0.05)	-0.42 (0.12)	0.94
	Winnipeg	0.13 (0.03)	-0.44 (0.21)	0.77
SOUTH	Carman	0.95 (0.04)	-1.23 (0.33)	0.99
	Winnipeg	0.25 (0.02)	-0.26 (0.04)	0.96
NORTH	Carman	0.34 (0.02)	-0.31 (0.05)	0.97
	Winnipeg	0.54 (0.04)	-0.33 (0.06)	0.96
EWSN	Carman	0.61 (0.03)	-0.45 (0.06)	0.98
	Winnipeg	0.49 (0.02)	-0.16 (0.02)	0.99
EWSN	Combined	0.54 (0.02)	-0.25 (0.02)	0.99

Table 2. Outcross rates of transgenic Brassica napus to border areas in 1994 by distance combined over direction, for Carman, Winnipeg and for combined sites.

Distance (m)

Site

0

2.5

5.0

10.0

15.0

30.0

Outcross rate in %

Carman	0.61	0.19	0.07	0.06	0.04	0.02
Winnipeg	0.48	0.36	0.22	0.07	0.06	0.01
Combined	0.54	0.28	0.14	0.06	0.05	0.015

Figure 1. Relationship between percent resistant individuals (i.e. percent outcrossing) and distance from pollen source in transgenic Brassica napus using bromoxynil herbicide resistance as a marker. Mean values for outcrossing rate at each distance and an exponential decay function are plotted. Refer to Materials and Methods for a description of the regression model fitted. See Table 1 for equations and parameter estimates.