Risk assessment of outcrossing of transgenic brassica, with focus on B. rapa and B. napus


R.K. Downey


AAFC Saskatoon Research Centre, 107 Science Pl., Saskatoon, SK  S7N 0X2, Canada, e-mail:  DowneyK@em.agr.ca




The rapid Canadian and Australian adoption of herbicide tolerant Brassica napus, resulting from both, mutation and gene insertion, indicates a wide acceptance of this technology by producers and consumers in North and Cental America as well as Asia.  Concern remains, however, as to the potential impact this new technology may have for evolutionary change in agriculture and natural habitats.  Studies have shown that B. napus and B. rapa plants exhibiting herbicide tolerance are no more invasive of cultivated or natural habitats than herbicide susceptible plants of the same species, unless the relevant herbicide is applied to remove competing vegetation.  The possible escape of herbicide tolerant genes from B. napus and B. rapa into related weedy species such as Sinapis arvensis (wild mustard), Raphanus raphanistrum (wild radish), Hirschfeldia incana (Hoary mustard), B. nigra (black mustard) and Erucastrum gallicum (dog mustard) is still under investigation.  Data to date indicate that natural barriers to introgression into the genomes of S. arvensis, R. raphanistrum and H. incana are strong.  Further studies with E. gallicum are underway.  Pollen outflow from fields of B. napus and B. rapa by wind and insects can be substantial and extend over long distances.  Such pollen movement is an important factor in assessing the risks of both inter and intraspecific crossing.  Outcrossing among fields of the same species can endanger the purity of seed stocks, raise concerns of organic growers and result in gene stacking among volunteer canola plants.  The Canadian experience as to the extent of natural outcrossing among different sized pollen donor and recipient populations and the effectiveness of distance isolation will be examined.  A case of gene stacking in volunteer B. napus will be reviewed and the need for more intensive agronomic management and grower education outlined.

KEYWORDS:  Sinapis arvensis, Raphanus raphanistrum, Hirschfeldia incana,
Erucastrum gallicum.




The use of herbicide tolerant canola, resulting from both induced mutation and gene transfer, has grown rapidly in Canada (Table 1).  Also in Australia cultivars tolerant to the S-triazine family of herbicide account for nearly 50% of canola plantings (Rieger et al., 1999).  However, herbicide tolerance is only the first of many possible agronomic and product modifications that are expected to be genetically engineered into B. napus and B. rapa in the not to distant future.  The outcrossing nature of the canola species has raised concerns of possible gene flow among cultivars as well as closely related species and genera.


The risks associated with B. napus and B. rapa outcrossing will vary with the characteristic of the introduced novel trait, as well as the region and intensity of production in which the genetically altered crop is grown.  Modifications to fatty acid or amino acid composition, for example, constitute a much lower agronomic or environmental risk than herbicide or stress tolerance.  The crucifer weed spectrium can also vary widely between production regions.  For instance, in Western Canada, dog mustard (Erucastrum gallicum) is on the increase but the region is essentially free of wild radish (Raphanus raphanistrum), hoary mustard (Hirschfeldia incana) and black mustard (Brassica nigra) that are present in many parts of Europe, South America and Australia. The

Table 1.    Estimated area sown (‘000 ha) to herbicide tolerant canola (B. napus) cultivars by herbicide type and total canola production in western Canada,1995-1999.










Total canola

area sown

% Herbicide
































* Significant usage of non-certified seed included

** Projected production and usage


intensity of production, in terms of length of crop rotation and distance between canola fields, can also impact on the opportunity for interspecific crossing and the combining of two or more novel traits in the same plant (gene stacking).


Since herbicide tolerance was the first transgenic trait to be incorporated and commercialized in B. napus, the main risk assessments have dealt with the questions, will herbicide tolerant plants become weedy or invasive of natural habitats or will the herbicide tolerant gene(s) escape into weedy relatives making them harder to control or more invasive of natural ecosystems?


Several studies both in Canada and Europe have demonstrated that the presence of a herbicide tolerant trait does not confer a competitive advantage to B. napus or B. rapa plants, unless the specific herbicide is applied and natural competition is eliminated (Belyk and MacDonald, 1994; 1995a; 1995b; MacDonald 1994; Crawley et al., 1993; Fredshavn et al., 1995).




The possibility of gene escape to close relatives has been under investigation since the late 1980’s.  It has long been known that natural interspecific crossing can and does occur among the oilseed Brassica species, B. napus, B. rapa and B. juncea (Bing et al., 1991; Jorgensen and Anderson, 1994; Frello et al., 1995) and in some countries a weedy form of B. rapa occurs.  Not unexpectedly Jorgenson and Anderson (1994), under conditions highly favorable to interspecific crossing and backcrossing between B. napus and wild B. rapa, were successful in introgressing a herbicide tolerant gene from B. napus into B. rapa with no adverse effect on the agronomic fitness of the recipient  plants (Snow et al. 1999).  However, it should be realized that B. rapa is also a major oilseed crop occupying over two million ha in Canada, India, China, Sweden and Finland and that glufosinate and glyphosate tolerant B. rapa varieties are already being grown commercially in Canada.  Regardlesss of whether a herbicide tolerant gene is present in  tame or wild B. rapa plants, such plants are no more difficult to control than the original populations with any one of three different herbicide groups and/or cultivation.


Of perhaps greater concern for rapeseed/canola producing regions is the possibility of gene transfer to non-oilseed cruciferous species such as the wide spread and persistent weed, wild mustard (Sinapis arvensis).  Fortunately, there is general agreement that although a very low level of the interspecific cross, B. napus x S. arvensis, may occur, the chance of an inserted gene being integrated into the S. arvensis genome is extremely remote (Bing et al., 1991; Eder et al., 1994; ChPvre et al., 1996; Lefol et al., 1996).  However, it has been speculated that if a herbicide tolerant gene was inserted into B. juncea it might migrate to S. arvensis via the bridging species B. nigra.  Fortunately B. nigra rarely occurs in Western Canada and large scale B. juncea production is largely confined to Western Canada and the Indian subcontinent.


The ease with which natural interspecific crossing can occur in both directions between B. napus and wild radish (Raphanus raphanistrum), as documented in France by ChPvre et al. (1999) and in Australia by Rieger et al. (1999), has been a continuing concern.  However, ChPvre et al. (1999) reports that despite four generations of backcrossing to R. raphanistrum and selecting herbicide tolerance in each generation, all herbicide tolerant plants contain one or more extra chromosomes.  Thus, despite intensive selection pressure, the herbicide tolerant gene as well as marker genes from B. napus, were not integrated into the wild radish genome (ChPvre et al., 1999; ChPvre personal communication).


Natural intergeneric crossing between B. napus and hoary mustard (Hirschfeldia incana) has also been documented (Lefol et al., 1995).  However, Darmency (personal communication) has recently reported that on backcrossing the intergeneric hybrids to H. incana, and subjecting the progeny to herbicide tolerance selection, fewer seeds were set with each successive backcross.  By the third backcross only a single seed was set and that developed into a herbicide susceptible plant.

These studies indicate, that with the weed populations and B. napus cultivars used, there are strong natural barriers to the integration of B. napus genes into these three weedy species.  Hand-crosses of B. napus and B. rapa with dog mustard (E. gallicum), produced F1 intergeneric hybrids (Lefol et al., 1997).  The B. napus x E. gallicum hybrid was weak and would not likely survive in nature.  Further, when the F1 was pollinated by E. gallicum the backcross progeny were found to be poor competitors and all progeny appeared to have completely reverted to E. gallicum, suggesting the loss of B. napus chromosomes.  The data to date indicate the possibility of gene transfer from B. napus to E. gallicum is very low.  On the other hand, the B. rapa x E. gallicum hybrid was vigorous and fertile, producing 67 selfed seeds as well as 690 seeds from the backcross B. rapa x the F1 and 14 seeds from the reciprocal backcross.  No seed was obtained when E. gallicum was the female parent.  The selfed and backcross seed is now under evaluation.  Field crossing blocks in 1998 unexpectedly produced an F1 E. gallicum x B. rapa hybrid (Seguin-Swartz, personal communication).  Data to date suggest that, gene transfer from B. rapa to E. gallicum is a real possibility.  However, considerably more information on the survival of the progeny as well as the ease of gene integration into the E. gallicum genome is required before firm conclusions on risk can be drawn.


What would be the consequences of herbicide tolerance gene(s) escaping to E. gallicum?  From what is known about the biology of this weed it is more of a nuisance than as ecological or agronomic threat.  The species may continue to build up populations in infested Canadian canola fields, as it does now, but it is readily controlled in cereal and other crops with  synthetic auxin and photosystem II inhibitor herbicides.




Brassica oilseed pollen is carried over long distances by both wind and insects.  Timmons et al. (1996) noted that, although the amount of wind born B. napus pollen at 360 m was only 10% of that at the field edge, pollen was still detected 1.5 to 2.5 km from 3 to 10 ha sized fields.  They calculated that some 4 km isolation would be needed to prevent unwanted outcrossing from commercial scale plantings.  Raney and Falk (1998) found pollen outflow, from a small (0.4 ha) block of high erucic acid B. rapa to a surrounding large field (64 ha) of low erucic canola, to result in outcrossing up to 260 m.


In 1998, commercial fields of glyphosate susceptible (RS) B. napus and B. rapa, growing near glyphosate tolerant (RR) fields, were sampled at various distances (0 to 100 m) from the closest edge of the RS fields to their paired RR field.  Random samples of over 900 seeds per sampled site were sown in the greenhouse and the seedlings sprayed twice with a 2x concentration of glyphosate.  The level of outcrossing between large (> 16 ha) commercial fields was low (Table 2).  Outcrossing rates were substantially lower than those found by Stringam and Downey (1978, 1982) where pollen flow from large commercial contaminant fields onto small (46 m2) plots resulted in an average of 0.6% and 3.7% outcrossing for B. napus and B. rapa, respectively, at 366 m.  Additional field monitoring is needed to determine the level of outcrossing that occurs over the total field surface.  Such sampling would allow a better estimate of outcrossing for the whole field, not just the first 100 m as measured in 1998.


Table 2.  Percent outcrossing1 between large fields (> 16 ha) of glyphosate tolerant and susceptible B. rapa and B. napus in western Canada,



Species &


Distance from field edge (m)



Field No.

of fields (m)






B. rapa


-------------% of outcrossing------------------





















(small 4 ha field)





























B. napus



---------------% of outcrossing-------------


















1 Based on herbicide reaction of > 900 seedlings per sampled distance.




Despite the low level of outcrossing recorded in 1998 a substantial number of outcrossed seed could still be produced per ha.  Assuming a 0.2% outcrossing rate in a field yielding 1400 kg/ha with a harvest loss of 5% some 35,000 outcrossed seeds (3.5 seeds/m2) would remain in the recipient field.  Although nearly all the plants originating from such seed would be killed by frost, herbicide treatment and/or cultivation, gene stacking in volunteer B. napus plants has already been observed in a farm field (Downey, 1999).  With the increasing use of cultivars carrying three or four different herbicide tolerance genes, plants carrying multiple resistances will soon become common.  Although plants with multiple resistances can be readily and easily controlled with other selective herbicides, extension workers from government and industry need to ensure that producers are well versed in the do’s and don’ts of growing herbicide tolerant cultivars.


Given the data on pollen movement and the level of outcrossing recorded in both species, complete containment of a novel trait, using distance isolation alone, becomes impractical under commercial production conditions.  Although pollen and gene flow may result in some agronomic and environmental risk, there is also the question as to the impact of gene flow on the promise of future multiple biotech improvements and product diversification in Brassica oilseed crops.  Some means of controlling pollen or gene flow is required if the full promise of biotechnology in Brassica oilseed crops is to be achieved.  The incorporation of transgenes in the chloroplast  genome could be one solution.  Alternatively, the introduction of a suicide trait, such as the “Terminator Technology” (Oliver et al. 1998) now under development could provide the solution.



Thanks are due to United Grain Growers personnel for obtaining the 1998 B. rapa seed samples, Monsanto Canada for the 1998 B. napus data and funding for growouts and to Dr. D.J. Bing for supervising the growing and rating the B. rapa seedlings.




Belyk, M. and MacDonald, R. (1994).  Invasive potential of transgenic B. napus (HCN92) under agronomic conditions.  In: Report No. AC194-06, pp. 20.  AgrEvo Canada Inc.: Regina SK.


Belyk, M. and MacDonald, R. (1995a).  Invasive potential of transgenic pHOE4 B. rapa (HCN28) under agronomic conditions.  In: Report No. AC195-18, pp. 21.  AgrEvo Canada Inc.:

         Regina SK.


Belyk, M. and MacDonald, R. (1995b).  Invasive potential of transgenic B. rapa (pHOE4) under disturbed and undisturbed field conditions.  In: Report No. AC195-17, pp. 18.  AgrEvo Canada Inc.: Regina SK.


Bing, D. J., Downey, R. K., and Rakow, G. F. W. (1991).  Potential of gene transfer among oilseed Brassica and their weedy relatives.  Proceedings 8th International Rapeseed Congress – Saskatoon, pp. 1022-1027.


ChPvre, A-M., Eber, F., Renard, M. and Darmency, H. (1999).  Gene flow from oilseed rape to weeds.  Proceedings of Gene Flow and Agriculture Symposium No. 72 – University of Keele, UK.  British Crop Protection Council, pp. 125-130.


ChPvre, A-M., Eber, F., Baranger, A., Kerlan, M. C., Barret, P., VallPe, P., and Renard, M. (1996).  Interspecific gene flow as a component of risk assessment for transgenic Brassicas.  Ninth Crucifer genetic workshop.  ISHS. Ed. J. S. Dias, I. Crute, A. A. Monteiro, Acta Horticulturae 407, 169-179.


Crawley, M. J., Halls, R. S., Rees, M., Kohn, D., and  Buxton, J. (1993).  Ecology of transgenic oilseed rape in natural habitats.  Nature 363, 620-623.


Downey, R. K. (1999).  Gene flow and rape- the Canadian experience.  Proceedings of Gene Flow and Agriculture Symposium No. 72 – University of Keele, UK.  British Crop Protection Council , pp. 109-116.


Eber, F., ChPvre, A-M., Baranger, A., VallJe, P., Tanguy, X., and  Renard, M. (1994).  Spontaneous hybridization between a male sterile oilseed rape and two weeds.  Theoretical and Applied Genetics 88, 362-268.


Fredshavn, J. R., Poulsen, G. S., Huybrechts, I., and Ruselsheim, P. (1995).  Competitiveness of transgenic oilseed rape.  Transgenic Research 4, 142-148.

Frello, S., Hansen, K. R., Jensen, J., and Jkrgensen, R. B. (1995).  Inheritance of rapeseed (Brassica napus) specific RAPD markers and a transgene in the cross B. juncea x (B. juncea x B. napus).  Theoretically Applied Genetics 91, 236-241.


Jkrgensen, R. B., and Andersen, B. (1994).  Spontaneous hybridization between oilseed rape (Brassica napus) and weedy B. campestris (Brassicaceae): a risk of growing genetically modified oilseed rape.  American Journal of Botany 81, 1620-1626.


Lefol, E., Danielou, V., Darmency, H., Boucher, F., Maillet, J., and  Renard, M. (1995).  Gene dispersal from transgenic crops: I. Growth of interspecific hybrids between oilseed rape and the wild hoary mustard.  Journal of Applied Ecology 32, 803-808.


Lefol, E., Danielou, V., and Darmency, H. (1996).  Predicting hybridization between transgenic oilseed rape and wild mustard.  Field Crops Research 45, 153-161.


Lefol, E., SJguin-Swartz, G., and Downey, R. K. (1997).  Sexual hybridization in crosses of cultivated Brassica species with Erucastrum gallicum and Raphanus raphanistrum: potential for gene introgression.  Euphytica 95, 127-139.


MacDonald, R. (1994).  Invasive potential of transgenic B. napus (HCN92) under disturbed and undisturbed field conditions.  In: Report No. AC194-05, pp. 14.  AgrEvo Canada Inc.: Regina SK.


Oliver, M. J., Quisenberry, J. E., Trolinder, N. L. G., and  Keim, D. L. (1998).  Control of plant gene expression. U.S. Patent No. 5,723,765.


Raney, J. P. and Falk, K. C. (1998).  Outcrossing in turnip rape (Brasica rapa L.).  11th International Crucifer Genetics Workshop, Montreal PQBook of Abstracts, p. 36.


Rieger, M. A., Preston, C., Potter, T., and Powles, S. B. (1999).  Gene flow from transgenic canola to wild radish – a model system to determine the risks.  Proceedings of Gene Flow and Agriculture Symposium No. 72 – University of Keele, UK.  British Crop Protection Council , pp. 131-136.


Snow, A. A., Andersen, B., and Jkrgensen, R. B. (1999).  Costs of herbicide resistance introgressed from Brassica napus into weedy Brassica rapaMolecular Ecology, (in press).


Stringam, G. R., and Downey, R. K. (1978).  Effectiveness of isolation distance in turnip rape.  Can. J. Plant Sci. 58, 427-434.


Stringam, G. R., and Downey, R. K. (1982).  Effectiveness of isolation distance in seed production of rapeseed (Brassica napus).  Agronomy Abstracts, 136-137.


Timmons, A. M., Charters, Y. M., Crawford, J. W., Burn, D., Scott, S. E., Dubbels, S. J., Wilson, N. J., Robertson, A., O’Brien, E. T., Squire, G. R., and Wilkinson, M. J. (1996).  Risks from transgenic crops.  Nature 380, 487.