Journal or Publishing Institution: Final Report of DEFRA Project RG0216: An experimental and mathematical study of the local and regional scale movement of an oilseed rape transgene.
Author(s): Ramsay, G., Thompson, C. and Squire, G.
Article Type: Peer Reviewed Study
Record ID: 2027
Abstract: All the methods required for this study were adapted and put into place. These consisted of protocols for the use of male-sterile plants to estimate potential cross- pollination, recording airborne pollen using suction traps and impaction on static surfaces, recovering pollen from bees, microscopic identification of pollen types, DNA extraction and fingerprinting methods for pollen and vegetative plant material, PCR detection of transgenes, and screens for marker genes giving resistance to antibiotics and sulphonylurea herbicides. Male-sterile plants were used to generate large data sets of maximum landscape-scale gene flow. This method over-estimates gene flow found in more normal situations due to the lack of competing self-pollen. In one season, direct comparisons were made of frequencies of fertilisation of male-sterile recipients and the movement of a marker gene into the seeds produced by male-fertile recipients. The data from male- sterile plants over-estimated the flow of a marker gene into a normal male-fertile population by about one order of magnitude. Gene flow into larger populations of male-fertile oilseed rape is less than that into smaller populations. In one experiment to investigate this effect, increasing recipient populations from 10 plants to areas of approximately 0.1 ha reduced the inflow of genes about 4-fold. This reduction was greater in internal parts of these larger blocks. Rates of cross-pollination, measured either by pollination of male-steriles or by tracking a marker gene in the progeny of male-fertile plants, dropped rapidly over the first few tens of metres from the edge of a field, but beyond that the decline with distance was slight over long distances. The exact shape of the decline varied between seasons and did not follow one mathematical description. However, declines with increasing distance are best described by functions giving a sharp initial decline then a long tail. Cross-pollination was highly variable from site to site and dependent on additional factors than simply distance to the nearest source. This applied whether gene flow was measured by fertilisation of male-sterile plants, or the presence of marker genes in seeds from male-fertile plants. Gene flow was detected over long distances. To facilitate a mathematical description of the decline in fertilisation and to characterise better the long tail in the distribution, sites with male-sterile plants were set out at 5 and 26 km from the nearest known source with little expectation of finding pollination. Low levels of fertilisation occurred at both sites. Although the origin of such events is hard to ascribe with absolute certainty, it appears that they were due to normal natural pollination via vectors which operate over long distances. A critical appraisal of this long-distance gene flow is presented in an appendix. Gene flow over the landscape involving a specific gene reflects the proportion of competing donor fields carrying that gene within a radius of a few kilometers. Pollen movement and fertilisation in oilseed rape have been assumed by some researchers to be due only or mainly to wind transport. Our data indicated that insects were predominantly responsible for pollination in the areas and years examined, even close to a source field. In one experiment, gene flow took place several km upwind of the source as effectively as downwind. Evidence was gathered which is suggestive of bee-to-bee contact in the hive being a major means of effective pollen dispersal through the foraging area of the colony. However, pollination of male-sterile plants took place over longer distances than those flown by worker honeybees. Of the many dispersive insects visiting OSR flowers, the pollen beetle was found on flowers at the most distant site and is already known to move over long distances. Patterns of pollination in these experiments were relatively insensitive to airborne pollen deposition and therefore should not be modelled on that basis. Factors such as the numbers of fields being worked simultaneously by single bee colonies, the degree of mixing achieved, other insect activity, and the relative areas of sources and sinks will partially govern cross-pollination rates. The data generated in this project permitted estimates to be made of cross-pollination into different types of recipient population over widely different distances. These are presented along with a note of the assumptions made and uncertainties requiring consideration. Increasing separation distance is an inefficient means of maintaining stringent crop purity, and complete freedom from impurity is unlikely to be guaranteed by geographical separation. However, even relatively small separation distances reduce impurity through cross-pollination in fields of fully-fertile oilseed rape varieties to a low level, around 0.1% or below.
Keywords: Gene Flow, Genetic Exchange, Populations, Evolution, Biological Integrity, Pollen, Propagules, Seeds, Crop Purity, GM, Crop Production, Oilseed Rape, Insects, Wind, Male Fertility
Citation: Ramsay, G., Thompson, C. and Squire, G., 2003. Quantifying landscape-scale gene flow in oilseed rape. Final Report of DEFRA Project RG0216: An experimental and mathematical study of the local and regional scale movement of an oilseed rape transgene.
