ACRE Publications

BACKGROUND PAPER: GENE FLOW FROM GENETICALLY MODIFIED CROPS

Background Paper by the Advisory Committee on Releases to the Environment (ACRE)

Introduction

1. Gene flow from cultivated crops to sexually compatible neighbouring crops, volunteer plants or wild relatives is a process as old as agriculture itself. There are many examples of the introgression of genes from cultivated crops into the gene pools of wild relatives growing around agriculture fields ¹. Equally, there is also opportunity for gene flow from wild plants to nearby crops; with implications for the quality of the harvest. However, never has the subject of gene flow in agriculture generated as much debate as when it involves genetically modified (GM) crops.
2. Uncertainty about the likelihood and the consequences of gene flow from GM crops was a prominent issue for the Royal Commission on Environmental Pollution (RCEP) in 1989. The RCEP's report was influential in shaping the then emerging regulatory framework, which today controls the release, and marketing of GMOs in the European Community ². The assessment of gene flow and the possible risks posed to human health and the environment is therefore one of the main issues for ACRE when reviewing applications for consent to release or market GM crops ³.
3. The aim of this paper is to describe briefly the background science and knowledge base that informs ACRE in the assessment of gene flow and its possible consequences. The paper (like the regulatory framework) deals only with the science and safety of gene flow. Socio-economic considerations, identity preservation and judgements about the 'acceptability' of gene flow are beyond ACRE's remit and the scope of the regulations ⁴.

Definition of gene flow from GM crops

4. Gene flow may be defined as the transfer of genes by pollination. In the case of GM crops the focus of interest is on the movement of transgenes ⁵. Gene flow may occur between a GM crop and a sexually compatible neighbouring crop or wild relative. Gene flow is not the same as pollen dispersal. ACRE acknowledges that pollen can travel many miles, but only very rarely will this result in a pollination event. Gene flow also includes horizontal gene transfer by asexual means. For the purpose of this paper, this is considered in terms of the possible movement of genes from a GM crop to unrelated organisms such as bacteria in soil or the gastrointestinal tract of farm animals.

Gene flow from GM crops to neighbouring crops

5. When ACRE reviews an application for consent to release a particular GM crop the Committee makes a judgement about the likely extent of gene flow to neighbouring crops and whether that poses a risk to human health and the environment. This judgement is made case by case and depends on the crop and the particular genetic modification.
6. The extent of gene flow between sexually compatible neighbouring crops (whether they are GM or not) depends largely on the separation distance between them, the breeding system of the particular plants (whether they are self-pollinating or out-crossing) and the mode of pollination (by wind or by insect vectors). In practice, other factors also play a part in dictating gene flow. The most important of these are coincidence of flowering, the period of pollen viability, the variation in weather conditions, the size of the source and recipient populations and the nature of the intervening geography or vegetation.
7. Gene flow declines rapidly with distance away from the pollen source. This is illustrated by work performed by Champolivier et al. ⁶. They investigated three sites in France in which herbicide-tolerant oilseed rape varieties were grown in adjacent fields; doubly tolerant genotypes ⁷ were detected in plants grown from seed collected at various distances from the edge of the crop and in volunteers emerging after harvest. The results indicated average hybridisation rates of about 2% at one metre, 0.2% at 20m, and less than 0.01% at 65m ⁸. The best mathematical description of such a decline varies from species to species but is very characteristic with a rapid fall from near-neighbour pollinations within a metre or two, according to some exponential power function, and a very long tail with gene flow occurring at extremely rare frequencies, sometimes over considerable distances.
8. Although it is difficult to make accurate predictions about exactly how much gene flow will occur at a given distance, especially those involving the rare long distance events, we do have robust separation distances for most crop species. These have been determined as part of the process of producing conventional seed stocks of required purity and were recently reviewed ⁹. Seed production experience from around the world has led to internationally-accepted separation distances for various levels of seed purity. In the UK these are governed by a range of legislation (e.g. for oilseed rape The Oil and Fibre Plant Seeds Regulation 1993) as amended)). They are based on practical experience and extensive seed testing over many years and, as would be expected, the separation distances required vary from crop to crop. Thus, to produce Pre-basic and Basic standard seed (with not less than 99.9% varietal purity), oilseed rape varieties must be separated by at least 400m. To produce Certified seed, the distance in this species is 200m (99.7% purity), and at 50m a level of 99.5% purity is achieved. By contrast, for wheat, barley and oats, a compulsory separation distance of only 2m is required, although a distance of 50m between different varieties is recommended. For maize, the highest standard of varietal purity requires 200m isolation.
9. The separation distance required to achieve a certain level of purity can also vary among different varieties of the same crop species. For conventional varieties of oilseed rape a separation distance of 100m will result in less than 0.1% cross-pollination in the great majority of field situations. However, for partially restored hybrid varieties of rape and varietal associations, which contain a proportion of male sterile plants ¹⁰, the separation distance needs to be greater to achieve the same purity. This is because the male sterile plants produce little or no pollen and are therefore unable to self-pollinate, which makes them particularly susceptible to pollination from outside sources of pollen.
10. Thus, international seed certification standards provide an empirical guide to the physical separation of GM and non-GM crops needed to minimise gene flow. Because they are based on practical field experience they take due account of year to year variations in prevailing weather conditions and the activities of bees and other pollinating insects. Evidence shows that such separation distances really work in a variety of field situations. For example, of 647 oilseed rape checks in the last five years only two have failed to meet the varietal purity standard of 99.7% and these failures were thought to be for reasons other than cross-pollination ¹¹.
11. International seed certification experience coupled with many supporting scientific studies on gene flow, provide the scientific and empirical knowledge base used by ACRE to judge the likely extent of gene flow between GM crops and their neighbours. Separation distances cannot (and have never been meant to) provide complete genetic isolation of the GM plants released, but they do undoubtedly contribute to a practical and precautionary reduction in the extent of gene flow.

Commercial releases of GM crops under Part C of Directive 90/220/EEC

12. The factors that influence gene flow from small-scale trials-plot releases are largely the same as those for larger scale commercial releases. However, gene flow may be different on the landscape scale to that observed in trial-plots. In one study ¹² using non-GM oilseed rape in a patchwork of fields, pollen traps, male sterile bait plants and mathematical modelling were used to demonstrate greater complexity and more gene flow than would have been be predicted from measurements of single source fields. In addition, volunteers and feral populations (both are very common in rape growing regions) provide a means of transferring genes from GM to non-GM crops over time (so-called "green bridges"). Since seed from oilseed rape may persist in the soil for several years, there is the potential over time for transgenes to move around in regions where the crop is grown in high density year after year.
13. If a GM crop is considered for commercial cultivation then it must be assumed that gene flow will occur when the crop enters agricultural practice. Of the GM crops that already have European Part C marketing consent, ACRE has taken the view that gene flow to neighbouring crops does not pose a risk to human health or the environment. However, if there were to be concerns about the desirability (in terms of safety) of gene flow from a GM crop then consent should not be given or risk management provisions, such as separation distances, to limit gene flow should be mandatory. Examples already exist of commercial crops (non-GM) that demand separation distances to protect the integrity of sexually compatible neighbours, e.g. high erucic acid oilseed rape.

Gene flow from GM crops to wild relatives

14. The potential for gene flow from GM crops to wild relatives in the UK has been reviewed previously ¹³. The crops grown most commonly in the UK fall into three groups. The first group contains crops such as maize and potato where the likelihood of gene flow to wild relatives is effectively zero because there are no sexually compatible wild relatives in the UK. The second group of crops, such as oilseed rape and some cereals, do have wild relatives, but their compatibility is poor and, in most cases, hybrids with the crops can only be isolated under laboratory or controlled field conditions. There are however examples of spontaneous hybridisation in the field between oilseed rape and a number of its wild relatives ¹⁴. Of these only wild turnip (Brassica rapa) is likely to be a recipient of transgenes by introgression into natural populations. Recent studies ¹⁵ have shown that rates of hybridisation are very low and introgression will be slow unless the transgene confers a selective advantage.
15. The third group of crops have relatives with which they are more-or-less fully compatible. Among this group is sugar beet (Beta vulgaris ssp. vulgaris) which has been the subject of many GM field trials. Sugar beet can hybridise readily with wild sea beet (Beta vulgaris ssp. maritima). Where gene flow to wild beet is considered undesirable then the removal of bolters (flower spikes) before flowering manages the risk.

Consequences of gene flow

16. In many respects the concern about whether gene flow will occur is secondary to the crucial issue of what will be the consequences of any gene flow. The possible consequences of gene flow are best considered case by case and depend on the particular transgenes, the recipient species and whether the transferred transgenes confer a novel trait(s) to the recipient. There is no simple formula for predicting the consequences of gene flow, but the risk assessment process will highlight areas of uncertainty where, if necessary, risk management procedures should be employed. Risk management procedures might include the precautionary removal or bagging of flower heads to prevent pollen production, separation distances, borders of non-GM plants and/or post release monitoring.

Consequences of gene flow to neighbouring crops

17. Risk assessment of gene flow from GM crops to their neighbours focuses on whether new toxins or allergens will result in the nearby crop. This will clearly depend on the particular transgenes and their level of expression. It is difficult to see how gene flow from any of the GM crops currently approved (or those under consideration) for cultivation in Europe could pose such a hazard ¹⁶. This situation may not be as clear however in the future when a greater variety of genetic modifications come forward for regulatory approval. In particular, the production of pharmaceutical products or industrial raw materials in plants is an area where ACRE will require robust data for lack of toxic or allergenic potential.

Consequences of gene flow to wild relatives

18. Gene flow to wild relatives presents a different set of considerations. Prominent among these is whether the transferred traits will confer a selective advantage (increased fitness) to the recipient wild plants and thus disturb the normal ecological checks and balances.

Herbicide tolerance

19. There is no indication from the scientific literature that the transfer of herbicide tolerance to wild plants confers any selective advantage in the absence of the particular herbicide. In ecosystems outside the agricultural field, the herbicide tolerant plants will be no more invasive or persistent than non-tolerant ones. Nevertheless, if farmers are forced to change the type of herbicide or the pattern of application because of troublesome herbicide tolerant wild relatives or volunteers ¹⁷ then this has the potential for an indirect effect on farmland biodiversity and is thus an important consideration in the risk assessment process.

Insect resistance

20. The consequences of the transfer of insect resistance traits to wild relatives in the UK are less clear. The only insect resistant GM crops approved for cultivation in Europe are two types of GM maize, both modified to express Bt toxins. Maize has no wild relatives in Europe so gene flow has not been a major consideration. However, if more insect resistant crops are developed then this issue will require careful evaluation depending on the GM crop, the nature of its wild relatives, the nature of the trait conferred and its likely insect specificity.
21. There are few field studies which provide clear evidence of how herbivory impacts on plant fitness, although there is some consensus that mammalian herbivores have a more dramatic impact on plant performance than insect herbivores do. However, even intensive long-term mammalian herbivory does not necessarily lead to the invasion of plant communities by unpalatable species because other factors limit their competitive abilities ¹⁸. Thus it is difficult to predict how, for example, the sudden appearance of an effective insect resistance gene in a wild population of plants is likely to effect the fitness of that population.
22. Fitness is likely to depend on the role played by insect herbivory in limiting the persistence or establishment of the wild plant species: if it is a significant one, then gene flow from insect resistant GM crops may result in improved performance, although other environmental factors may still prevent the species gaining a competitive advantage. For example, Bt- oilseed rape exposed to insect herbivory in a recent field experiment showed slightly increased survival and reproduction compared to non-GM rape, but despite this advantage, only two out of the 6000 GM plants survived longer than a few months because they were out-competed by the natural vegetation on the experimental plots ¹⁹. More research is needed into the affect of invertebrates on the population dynamics and fitness of the wild plants (and feral populations of crops) that are able to hybridise with GM crops.
23. It should be noted that potential hazards of gene flow from insect resistant crops are not unique to GM plants. The introgression of any major insect resistance trait into wild relatives, even from conventionally bred crops, could have the same impact. However, there is more opportunity to avoid a possible problem with GM crops because they are subject to risk assessment and regulation whereas conventionally bred crops are not.
24. Another important consequence of gene flow is the possible impact on non-target invertebrates in field margins if their host plants were to suddenly (in evolutionary terms) acquire a major resistance trait. Currently, there is very little direct empirical evidence that plant resistance to insects can affect the population dynamics of insect herbivores in the field, although this is predicted theoretically for at least some plant-herbivore systems ²⁰. Again, more research is needed into both the nature of any possible hazards and likelihood of each occurring, which will inform the risk assessment process. But this uncertainty does not mean that releases of GM insect resistant crops should not be done. On the contrary, carefully controlled, well managed releases, which can be conducted safely with appropriate risk management measures, are needed as part of research initiatives into possible environmental effects and to gather data to build confidence in future marketing applications. One final consideration with insect resistance traits, is that some of them may also confer resistance to diseases (see paragraph 25 below) - this is unlikely to be the case with Bt-resistance but a modification which increased some types of secondary metabolites may have this affect.

Other genetically modified traits

25. In addition to insect resistance and herbicide tolerance, a number of other GM traits have come forward for release consent (small-scale R&D trials) in the UK. For example ACRE has advised on the release of GM plants with traits such as modified starch, reduced pod shatter, viral resistance, enhanced fungal resistance and altered fatty acid profiles. All of these raise issues of gene flow to wild relatives (if present) and any possible selective advantage. This is especially the case with disease resistance traits where it may be unknown what ecologically suppressive effect particular diseases have on populations of wild plants. Some work is being done in this area, e.g. looking at the importance of viral diseases in wild brassicas ²¹, but more research is needed, and release consent for widespread cultivation of a GM crop would not be issued where there was legitimate concern about an increase in fitness and its impact.

Horizontal Gene Transfer

26. Horizontal gene transfer (HGT), in the context of GM crops, is regarded as the asexual movement of transgenes from a GM crop into an unrelated organism. Concern is often expressed that HGT will result in the movement of antibiotic resistance genes from GM feedingstuffs into bacteria in the gastrointestinal tracts of farm animals. There is also concern that soil micro-organisms might pick up transgenes from GM crops grown in the field.
27. ACRE has considered HGT carefully and it is always a factor (explicit or implicit) in the risk assessment of any releases of GMOs. Most of the scientific studies to investigate HGT have focussed on the design of laboratory systems to encourage HGT ²². Although HGT could be demonstrated to occur it remained a very rare event even under favourable laboratory conditions. ACRE is not aware of any work that has been able to show HGT from GM plants to bacteria in the field.
28. Despite the lack of evidence for HGT in the field, ACRE takes the precautionary view that HGT will occur and thus concentrates on assessing the possible consequences. As with gene flow via more accepted routes, the consequences of HGT will depend on what DNA sequences are transferred and whether they are active in the recipient bacteria. Most of the promoter sequences used to direct transgene expression in plants show little or no activity in bacteria. Gene sequences are also often optimised for codon usage in plants and express poorly in bacteria. Further, even if the transgenes were to be active in bacteria then without selection pressure they will disappear or remain at a very low frequency in the population.
29. ACRE has advised previously that the transfer of antibiotic resistance genes from GM crops to bacteria would be insignificant compared to the levels of antibiotic resistance that already exist in nature, particularly due to the over use of antibiotics in medicine and as prophylactics/growth enhancers in intensive animal husbandry. However, ACRE has cautioned against the marketing of GM plants and products containing genes that confer resistance to antibiotics of clinical importance. The very small risk that the use of such crops could compromise the clinical value of important antibiotics is a risk not worth taking.
30. ACRE has also considered whether the use of the Cauliflower mosaic virus 35S promoter (CaMV 35S) to express genes in GM crops increases the potential for HGT as suggested by some observers ²³. ACRE concluded that based on the extensive body of information available on this ubiquitous virus and fifteen years experience of using the 35S promoter in nearly all GM plant constructs then there was nothing to support the hypothesis that it increases the likelihood of HGT.

Conclusions and future developments

31. Evaluating the likelihood and, more importantly, the consequences of gene flow from GM crops is a major component of the risk assessment process. There is comparatively more knowledge and practical experience to inform judgements about the likelihood of gene flow than there is to inform judgements about its consequences. Where there are gaps in scientific understanding, particularly as more diverse genetic modifications come forward, then risk management will permit safe testing in a step by step process at the R&D stage. This will inform risk assessment for applications for commercial cultivation, but the onus is on the notifiers to provide robust data in support of their applications.
32. Looking to the future, there are a number of existing and emerging technologies that will help to reduce or prevent gene flow, such as modifying plants so that they do not produce pollen, do not flower or reproduce asexually. ACRE has recently produced guidance on these technologies ²⁴ and highlighted their importance in simplifying risk assessment. It is hoped that those responsible for the production of GM crops will embrace these developments wherever possible.

1 Ellstrand N.C., Prentice H.C. & Hancock J.F. (1999) Gene flow and introgression from domesticated plants into their wild relatives. Annual Review of Ecology & Systematics 30 pp 539 - 563

2 Directive 90/220/EEC on the Deliberate Release of Genetically Modified Organisms to the Environment.

3 The assessment of gene flow and its consequences is important for the release of any genetically modified organisms, not just GM plants.

4 This paper should be considered in association with the companion papers from the ACRE Secretariat describing the European regulatory framework and the risk assessment process for the release and marketing of GMOs in the European Union (http://www.defra.gov.uk/environment/acre/index.htm)

5 The term 'transgenes', as used here, means the genetic elements - coding sequences and promoters - inserted into the GM crop by genetic transformation

6 Champolivier J., Gasquez J., Messian A. & Richard-Molard M. (1999). Management of transgenic crops within the cropping system. In British Crop Protection Council Symposium Proceedings no 72. Gene flow and Agriculture - Relevance for Transgenic Crops pp 233 - 240

7 The detection of double tolerant plants indicates that gene flow from one herbicide tolerant line of oilseed rape to another neighbouring line has occurred.

8 Most oilseed rape is self-compatible and between 40 - 80% of pollination events are self-pollinations

9 Ingram J., (2000). Report on the separation distances required to ensure cross-pollination is below specified limits in non-seed crops of sugar beet, maize and oilseed rape. National Institute of Agricultural Botany

10 varietal associations may contain up to 80% male sterile plants, and partially restored hybrids contain about 50% male sterile plants

11 Ingram J., (2000). Report on the separation distances required to ensure cross-pollination is below specified limits in non-seed crops of sugar beet, maize and oilseed rape. National Institute of Agricultural Botany.

12 Squire G.R., Crawford J.W. Ramsay, G. & Thomson C. (1999) Gene flow at the landscape level. In British Crop Protection Council Symposium Proceedings no 72. Gene flow and Agriculture - Relevance for Transgenic Crops pp 57 - 64

13 Raybould A. F & Gray A. J. (1994). Genetically modified crops and their wild relatives - a UK perspective. Research report 1, Published in DETR Blue Series

14 Gray A. J., & Raybould A. F. (1998). Environmental risks of herbicide-tolerant oilseed rape. A review of the PGS hybrid oilseed rape. Research report 15, Published in DETR Blue Series

15 Wilkinson M. J., Davenport I. J., Charters Y. M., Jones A. E., Allainguillaume J., Butler H. T., Mason D. C. & Raybould A. F. (2000) A direct regional scale estimate of transgene movement from genetically modified rape to its wild progenitors. Molecular Ecology 9 pp983 - 991

16 The major traits approved under Part C of Directive 90/220 are herbicide tolerance, using the pat, bar and epsps genes or insect resistance using Bt toxins. In addition to data presented in the application dossiers that none of the gene products poses a risk to human health, they are also consumed, apparently safely, in large quantities outside Europe.

17 Particularly wild relatives or volunteers with 'stacked' herbicide tolerant traits. Gene 'stacking' is where a plant has two or more transgenes each of which confers tolerance to a different herbicide.

18 Augustine D. J. & MacNaughton S. J. (1998). Ungulate effects on the functional species composition of plant communities: herbivore selectivity and plant tolerance. Journal of Wildlife Management 62 pp 1165-1183

19 Stewart C. N., All J. N., Raymer P. L. & Ramachandran S. (1997). Increased fitness of transgenic insecticidal rapeseed under insect selection pressure. Molecular Ecology 6 pp 773-779

20 Underwood N. C. (1999). The influence of plant and herbivore characteristics on the interactions between induced resistance and herbivore population dynamics. American Naturalist, 153 pp 282-294.

21 Raybould A. F., Jones A. E., Alexander M., Pallett D., Thurston M. I., Cooper J. I., Wilkinson M. J. & Gray A. J. (2000) The potential for ecological release following introgression of virus-resistance transgenes into natural populations of wild Brassica species. In: Proceeding of the 6th International Symposium on The Biosafety of Genetically Modified Organisms. Saskatoon, Canada. Ed Fairbairn C., Scoles G. & McHughen A.

22 Such as the selection of transgene sequences, vectors and recipient bacteria to maximise the chance of observing HGT

23 Ho M. W., Ryan A. & Cummins J. (1999) Cauliflower mosaic viral promoter - a recipe for disaster. Microbial Ecology in Health and Disease 11 (4)

24 Guidance on Best Practice in the Design of Genetically Modified Crops: http://www.defra.gov.uk/environment/acre/bestprac/index.htm