ASSESSING THE RISKS OF TRANSGENE ESCAPE: A CASE STUDY IN SUNFLOWERS
September 5, 2003
Viginia Tech University (ISB)
John M. Burke
http://www.isb.vt.edu/news/2003/news03.sep.html#sep0302
In recent years, the potential impact of transgenic crops on the environment
has been a topic of intense international debate.
Arguments in favor of genetic engineering point to the possible
environmental benefits of genetically modified (GM) crops. These include a
reduction in the amount of chemicals applied to agricultural systems, a
transition to less toxic chemical treatments, and the facilitation of
zero-till agriculture. Environmental objections to GM crops, on the other
hand, are largely based on factors such as the possible negative effects of
transgenes on non-target organisms and the potential for transgene escape
via crop-wild hybridization to facilitate the evolution of increasingly
weedy or invasive plants.
Although it may take years for the true environmental effects of transgene
escape to be known, predictions regarding the particular crops or traits
that are likely to pose the greatest environmental risks can be made. For
example, crops that hybridize readily with wild relatives represent greater
risks than those that do not. Likewise, transgenes that are advantageous in
wild or weedy forms of a plant are most likely to pose a significant risk,
whereas those that are neutral or disadvantageous will do little to disrupt
the evolutionary dynamics of the recipient population(s). Current concern
stems from the fact that many of the traits that are the target of genetic
manipulation abiotic stresses In many cases, the conditions necessary for
hybridization between crop
plants and their wild relatives are met, and hybridization appears to be
frequent. For example, there is evidence that twelve of the world's thirteen
most important food crops hybridize with at least one wild relative in at
least part of their range of cultivation (1). Thus, the question of whether
or not genes will ultimately escape from cultivation has been largely
answered; in most cases, they will. Research on the risks associated with
transgene escape should, therefore, focus on the fitness consequences of the
gene(s) in question, rather than on rates of gene flow. Until recently,
however, virtually nothing was known about the fitness effects of pest or
pathogen resistance transgenes in wild plant populations.
In a recent study (2), we examined the fitness effects of a transgene that
confers resistance to white mold (Sclerotinia sclerotiorum) following its
`escape' from cultivated into common sunflower (Helianthus annuus). Of the
more than three dozen pathogens that afflict sunflower, white mold is one of
the most common and widespread, having been reported from all sunflower
growing regions throughout the world. White mold infection, which typically
begins at the base of the stem, results in the rapid wilting and death of
cultivated sunflower plants, greatly reducing seed output. Infection rates
as high as 100% have been reported in North American sunflower fields, and
white mold has been known to reduce yield by as much as 70%. Attempts to
develop resistant cultivars via traditional plant breeding techniques have
met with little success in sunflower, and chemical control methods are often
costly and ineffective. Attention has turned, therefore, to genetic
modification. Because oxalic acid plays a key role in the pathogenicity of
white mold, it has been hypothesized that the insertion of an oxalate
oxidase (OxOx) transgene would provide otherwise susceptible plants with a
mechanism of resistance (3). This approach has now been used by Pioneer
Hi-Bred, Intl. to successfully enhance white mold resistance in cultivated
sunflower.
Unfortunately, the potential for transgene escape is especially high in
sunflower. Nearly all of the cultivated sunflower acreage in the United
States is contained within the geographic range of common sunflower, and
range-wide surveys of the potential for reproductive contact have revealed
that approximately two-thirds of all cultivated sunflower fields in the
United States occur in close proximity to, and flower coincidentally with,
common sunflower populations (4). Moreover, the results of previous research
indicate that, where they come into contact, cultivated and wild sunflower
often hybridise (5). Thus, crop-wild gene flow is a virtual certainty
throughout the range of sunflower cultivation in the United States.
The efficacy of the OxOx transgene in cultivated sunflower, combined with
the high likelihood of escape, raises the specter of transgene escape,
leading to the evolution of a more weedy and invasive common sunflower. We
simulated the early stages of transgene escape by crossing the OxOx
transgene into common sunflower and growing the resulting plants at field
sites located in California, Indiana, and North Dakota. The final result
revealed a set of populations consisting of wild-like plants that were
segregating for the OxOx transgene. By inoculating a subset of these plants
at each location with white mold and keeping the remainder as controls, we
were able to examine the fitness benefits afforded by the OxOx transgene in
the face of a pathogen challenge, as well as any possible fitness costs
associated with it in the absence of white mold.
Overall, our results indicated that there was no "cost of resistance"
associated with the OxOx transgene in the absence of a pathogen challenge.
This gene did appear, however, to protect its carriers from white mold
infection. Although the effect varied across locations, the frequency of
infection was generally lower in plants carrying the OxOx transgene than in
those that lacked it. In terms of seed output, however, the story was
somewhat different. Following inoculation, there was no detectable
difference in the productivity of transgenic and non-transgenic individuals.
Although the underlying mechanisms remain unknown, this seemingly
paradoxical result has a relatively straightforward explanation: The rate
and severity of infection were effectively decoupled in this experiment. In
California, where the OxOx transgene provided the greatest degree of
protection against infection, onset of the disease had virtually no effect
on fitness. In contrast, while white mold infection had a major (and
negative) impact on fitness in Indiana, infection rates at this location
were unaffected by the OxOx transgene.
Taken together, our results suggest that the OxOx transgene will do little
more than diffuse neutrally following its escape, and therefore, will have
little effect on the evolutionary dynamics of wild sunflower populations. In
other words, it appears that, by giving the OxOx transgene to wild
sunflower, we effectively gave it something that it already had degree of
white mold resistance. This conclusion must be tempered, of
course, with the realization that our work was performed within a single
growing season and on a single genetic background. It is therefore possible
that our results are not generalizable over time or across common sunflower
populations. Longer-term studies replicated across various wild genetic
backgrounds will be necessary to shed light on these issues. Even long-term
studies, however, have their limits; strong but episodic selection can have
a major influence on the evolutionary trajectory of populations, yet may be
rare enough to avoid detection.
In the broader context, our results illustrate the importance of quantifying
transgene fitness more directly than through the use of a presumptive
correlate such as disease incidence. Indeed, if we had relied solely upon
infection rates, rather than looking directly at reproductive output (albeit
only through female function) our conclusions would have been quite
different. This work also represents an important counterpoint to a recently
published report (6) in which a Bt transgene was shown to decrease herbivore
damage and increase fecundity in common sunflower grown under field
conditions.
Our work, combined with the Bt findings, indicates a clear need to assess
the relative risks and benefits of genetic modification on a case-by-case
basis. Although increases in reproductive output do not necessarily
translate into an increase in weediness or invasiveness, the fitness of an
allele remains the best predictor of the likelihood and rate of its spread.
Thus, the best means currently available for assessing the environmental
risks associated with transgene escape are fitness-related measures. The
time has come for us to move beyond hand wringing about the likelihood of
transgene escape and to ask the more important question: What will happen if
and when these genes gets out?
References
1.Ellstrand et al. (1999) Gene flow and introgression from domesticated
plants into their wild relatives. Annual Review of Ecology and Systematics
30: 539-63. 2. Burke & Rieseberg. (2003) Fitness effects of transgenic
disease resistance in sunflowers. Science 300: 1250. 3. Thompson et al.
(1995) Degradation of oxalic acid by transgenic oilseed rape plants
expressing oxalate oxidase. Euphytica 85: 169-72. 4. Burke et al. (2002) The
potential for gene flow between cultivated and wild sunflower (Helianthus
annuus) in the United States. American Journal of Botany 89: 1550-2. 5.
Arias & Rieseberg. (1994) Gene flow between cultivated and wild sunflowers.
Theoretical and Applied Genetics 89: 655-60. 6. Snow et al. (2003) A Bt
transgene reduces herbivory and enhances fecundity in wild sunflowers.
Ecological Applications 13: 279-86. John M. Burke Department of Biological
Sciences Vanderbilt University, Nashville, TN