Introduction

What's gene flow?

Is it possible to quantify  the level of gene flow from a GM crop?

Does the dispersal of pollen/seed imply the successful transfer of a GM trait?

What will happen if a GM crop crosses with a weed?

Are 'superweeds' unique to GM cropping?

Introduction

The potential impact on the environment of growing GM crops is continuously cited as a source of societal concern. It is important to note that prior to the commercial release of a GM crop in Europe, the GM crop must first undergo a comprehensive environmental risk assessment, which must be passed by the competent agency  of the member state (e.g. EPA in Ireland) in which the crop is to be first grown.

An Environmental Risk Assessment examines the potential direct and/or indirect impact the GM trait (e.g. insect resistance) might have on wildlife. The environmental risk assessment will also examine whether there is potential for the GM trait to spread (via gene flow) into those wild plant species that are related to the GM crop species.

What's gene flow?

For crops, gene flow occurs through the spread of pollen (i.e. 'pollen-mediated gene flow') and seed (i.e. 'seed-mediated gene flow'). This is a natural phenomenon as all crops spread pollen and shed seed, irrespective of whether they are GM or not.

What is critical for the environmental risk assessment of a GM crop however is not whether gene flow occurs, but what is the outcome of a gene flow event. So when pollen is blown from a GM crop (e.g. GM oilseed rape) and lands on the flower of a wild relative (e.g. wild turnip), what will be the outcome of this 'gene flow' event and will it impose a negative or neutral impact on the environment?

Successful pollen-mediated gene flow is limited to when the crop is flowering and is dependent upon:

• the amount of pollen produced
• the durability of the pollen
• the mechanism by which the pollen is dispersed
• an overlap in flowering time with the related plant species
• the distribution of the related plants
• the  fertility of the receiving plants

While pollen is only dispersed via wind or insect during the flowering period of the crop, seed dispersal may occur during any one of several phases in the cropping cycle including;

• sowing,
• harvesting
• post-harvest transport
• post-harvest storage

As stated earlier, the mechanisms of gene flow in GM crops is identical to that of non-GM crops because the modification made in the GM plant has no impact on pollen and/or seed production. Therefore the dispersal pattern of pollen from a particular crop is the same whether it is GM or non-GM.

In this manner, non-GM and GM crops are equal and this is why studies that examine the dispersal distance of pollen in a certain non-GM crop (e.g. oilseed rape) can be used to model the dispersal patterns of pollen from an equivalent GM crop (e.g. GM herbicide tolerant oilseed rape).

Is it possible to quantify the level of gene flow from a GM crop?

Yes. Previously the level of gene flow from a crop (be it GM or not) was ranked as 'high', 'medium' or 'low'.  However, this qualitative scale is both confusing and uninformative. For a farmer/consumer the introduction of a distinguishable scale or Gene Flow Index would greatly assist them in understanding the potential impact of a GM crop.

The creation of such an index would have to be flexible enought to cover all cutlivated crops and address the presence/absence of related weeds. It would also have to address the ability of any hybrid seed to germinate and establish as a fit GM crop-weed hybrid. Such an index has been completed by the risk assessment research group at Teagasc.

This gene flow index (GFI), examines the tendencies of conventional crops to undergo gene flow and based on these results we are able to predict how gene flow from a particular GM crop is likely to occur.  In designing the GFI model, we collected data from a broad literature base (peer-reviewed scientific journals and reports) and considered in conjunction with field data from Teagasc Oak Park information from the Teagasc Farm Advisory Service.

Importantly, only information pertaining to systems comparable to the Irish agricultural environment was included. Our GFI model has been used to evaluate the principal crops cultivated in Ireland (see graph below): wheat, barley, maize, ryegrass, oilseed rape and potato and the results of the modelling work can be accessed here.

Briefly, oilseed rape and ryegrass attained large GFI values (19/27 for oilseed rape and 25/27 for ryegrass) which means they have a higher tendency for gene flow. In contrast, wheat, barley and maize recorded lower GFI values and thus have a lower tendency to spread their genes through gene flow.

Does the dispersal of pollen/seed imply the successful transfer of a GM trait?

No. For successful pollen-mediated gene flow, the following events must occur:

1. the pollen has to find and fertilise a related weed or crop
2. this plant must then produce ‘hybrid' seed
3. this seed must be able to over winter and survive foraging animals and the weather
4. surviving seed must germinate in to a hybrid plant
5. this hybrid plant must then mature and produce its own seed

It is only at this stage can it be concluded that the GM trait has established itself in the receiving wild plant species. Although hybrids may develop spontaneously, it is not automatically implied that they will be able to establish, survive and reproduce in the wild (Hauser et al. , 2003).

The same is true for seed-mediated gene flow: it is only a reality if the germinating seed gives rise to a viable plant that has the ability to reproduce and produce an equivalent quantity of seed as its neighbouring plants.

What will happen if a GM crop crosses with a weed?

Of most relevance to this issue is the presence/absence of weeds that are related to the GM crop. For example, in Ireland, farmers cultivate several native and non-native (alien) crops, which may or may not have related weeds growing near them.

So, whereas wheat, barley, potatoes, and maize are 'alien' species without fertile wild relatives; ryegrass and oilseed rape are native and have wild relatives here in Ireland. Therefore, if varieties of GM maize, wheat, barley or potato were commercialised and sown in Ireland there is zero risk of the GM trait entering a wild plant species.

On the other hand, if varieties of GM ryegrass or GM oilseed rape were commercialised and sown in Ireland it is possible, and indeed likely, that they could cross with their wild relatives already growing in Ireland.

However, as discussed above the critical issue is not whether pollen flows but what are the consequences of such pollen flow?

So while it is important to gauge the occurrence of gene flow, the crucial factor is in determining the result of the pollen flow.  This can only be achieved by focussing on the GM trait that has been transferred and not the fact that gene flow has occurred.

Lets assume pollen flow has occurred between a GM oilseed rape crop (e.g. resistant to the herbicide glyphosate) and neighbouring wild rape located in the hedgerows of the GM field; what will be the consequence?

While recent research (Warwick et al. 2007) has clearly outlined the potential for a herbicide tolerant trait to persist for up to 6 years in a wild weed population, this will only present an environmental risk if the glyphosate resistant trait provides the weed with an increased level of ‘fitness'. Critically, this will only occur if the weed is sprayed with the glyphosate herbicide: that is the weed will only possess increased fitness in the presence of glyphosate.

Therefore the herbicide tolerant trait could remain in the weed population for 20 years, but it will no environmental impact if the plants are not exposed to glyphosate. In the scenario, where the weed (often labelled a 'superweed') is exposed to glyphosate, it could then have a competitive advantage over its neighbours who are likely to be susceptible to glyphosate. 

Hence, they will die off and the glyphosate resistant weed will have a 'free patch' for its seed to grow in. However, if the farmer wishes to control the wild rape, this can be easily achieved through applications of other herbicides (e.g. glufosinate, paraquat or imidizolnone-based herbicides) or mechanical measures. 

Therefore, in the absence of glyphosate, the weed population that has the GM trait will have no advantage over its neighbours. Hence the spread of pollen from the GM crop imposes a neutral impact on the environment for this scenario.

Are 'superweeds' unique to GM cropping?

No. The emergence of herbicide resistant weeds has been an issue for farmers since they first ever starting using herbicides to clear fields between rotating crops. Quite simply, the more a single herbicide is used the greater the chance weeds will evolve resistance to the herbicide. 

By definition, 'herbicide resistance' refers to the ability of a plant to survive and reproduce under a normally lethal dose of herbicide. Over the last 20 years, more than 300 types of weeds have evolved resistance to one of the main class of agricultural herbicides (for more information visit http://www.weedscience.org).

Indeed, a recent review has highlighted the prevalence of herbicide resistant weeds in over 56 different countries, many of which do not cultivate GM crops. The paper has also catalogued the most prevalent agricultural weeds in global agriculture that have evolved resistance to several non-GM herbicides as outlined in the below table.

In recent years there has been a phenomenal increase in the use of herbicides containing the active ingredient glyphosate (e.g. Roundup™). This herbicide is popular because quite simply it is very good at doing its job and because it is safer to use than other herbicides such as paraquat.

Indeed, the usage of glyphosate is not restricted to agricultural lands with many homeowners using it to clear paths and drives of moss and weeds. However where glyphosate is used intensively, this increased usage has led to the evolution of glyphosate resistant weeds. For example, horseweed (Conyza canadensis) and morning glory (ipomea purpurea) in the USA and rigid ryegrass (Lolium rigidum) in Australia.

Clearly therefore the use of herbicide resistant crops has the potential to lead to increased herbicide resistance in weed populations and this phenomenon is not solely a GM issue, as it was apparent before GM ever appeared on the scene. But how can it be solved?

As stated earlier, the more a chemical (e.g. herbicide or fungicide) is applied the greater the chance of resistance to that chemicial emerging. So to lessen the chance of this occurring, it is necessary to reduce the frequency with which the farmer uses the chemical. The practical way to do this is to divesify and use a range of different herbicides and also to adopt mechanical means where appropriate.

In addition, the development of novel herbicide strategies (e.g. dicamba resistance, Behrens et al. (2007)) will also augement weed management strategies and minimise the potenital for 'superweeds' to develop for all crop systems, be they GM or non-GM.

References:

Behrens et al. (2007). Dicamba resistance: enlargin and preserving biotechnology-based weed management strategies. Science, Vol. 316, p.1185-1117.

Warwick et al. (2007). Do excaped transgenes persist in nature? The case of an herbicide resistance transgene in a weedy Brassica rapa populations. Molecular Ecology (http://www.blackwell-synergy.com/doi/abs/10.1111/j.1365-294X.2007.03567.x?prevSearch=allfield%3A%28Brassica+rapa%29).