Born to run: competition enhances the spread of genes from crops to wild relatives


(Lorne M. Wolfe: tel +1 912 6810848; fax +1 912 6810845; email AmyC.Blair: email

‘By introgressive hybridization elements of an entirely foreign genetic adaptive system can be carried over into a previously stabilized one, permitting the rapid reshuffling of varying adaptations … Natural selection is presented with segregating blocks of genic material belonging to entirely different adaptive systems.’(Anderson & Stebbins, 1954)

From poodles to great Danes, Granny Smith to Macintosh apples, Yukon Gold to Russet potatoes, and tabby to Siamese cats, humans have been tinkering with animal and plant genomes for thousands of years. In his 1868 book titled, ‘The Variation of Animals and Plants Under Domestication’, Charles Darwin eloquently summarized how selective breeding had resulted in the production of varieties of species with desirable traits (Darwin, 1868). Such artificial selection can be a prolonged process that requires multiple generations to propagate those individuals with favorable phenotypes. Currently, with our ability to genetically modify organisms by moving genes (traits) between species, we have overcome many of the pre-existing constraints and can create crops with desirable traits in a single generation. For example, Golden rice may someday supplement beta-carotene in the diets of people from nations that rely almost exclusively on rice as their staple food; a lack of this vitamin results in blindness and premature death in children. The past few years have witnessed a dramatic growth in the genetic modification of commercial crops, with some crops in the USA now being composed of a majority of transgenic material. Crops that have undergone selection for advantageous traits and/or have been genetically modified have conferred a vast benefit on agriculture, by increasing yields and reducing chemical inputs to control weeds and insect herbivores. As 30–40% of agricultural productivity is reduced worldwide by insect herbivory (Oerke et al., 1994), plants selected or created to resist such herbivory are highly beneficial to farmers.

Despite the financial benefits of growing such plants, there are concerns that the same genes (alleles) which confer a growth advantage to the crop plant could cause ecological problems by escaping and becoming introduced into plant species in the wild (Ellstrand et al., 1999; Haygood et al., 2003; Pilson & Prendeville, 2004; Lu & Snow, 2005; Chapman & Burke, 2006). This is not an empty concern, as Reichman et al. (2006) recently provided the first evidence of the escape of transgenes into native and naturalized plant populations in the USA. Glyophosphate-resistant creeping bent-grass was identified up to 3.8 km from the control area. Movement of crop genes into wild relatives could potentially result in the evolution of a weedier or more invasive plant species. It is already known that 22 out of the 25 most important crop species hybridize with wild relatives, so it seems probable that such a hybridization event could occur in most systems (Ellstrand, 2003).

Such gene flow depends on two processes. First, in order for a gene to move to a wild relative, there must be a hybridization event between the crop and the wild species. Thus, factors such as pollinator behavior and density, and timing of flowering, will directly influence the rate of such gene flow. Wind-pollinated plants will potentially undergo even less-constrained hybridization owing to their independence from pollinators. Second, despite the intial assumption that the rate of gene flow is the primary determinant of successful hybridization, there is growing recognition that the fate of hybrid offspring under natural selection has an even greater influence on gene escape. Theoretical work has demonstrated that even with low rates of allele migration, the success of hybridization depends mainly on the selective advantage provided by the allele (Slatkin, 1976; Morjan & Rieseberg, 2004; Chapman & Burke, 2006). Assuming that the mating event was successful, is the offspring more or less ‘fit’ compared with its parents? How does it fare in the natural environment, where pathogens, herbivores and competitors conspire to make life difficult? Surprisingly, we know very little about these key questions. Owing to the fact that a phenotype is the product of an interaction between the genotype and its environment, it is perhaps only a matter of time until a certain combination of genes is in the appropriate habitat, thus allowing the hybrid to establish outside the crop setting.

Research on the spread of crop genes to wild relatives is the study of rare events coupled with difficult-to-predict outcomes. It is an inherently complicated field. An article by Campbell & Snow, in this issue of New Phytologist (pp. 648–660), highlights this complexity and makes a wonderful contribution to our understanding of this process. Their study is a multigenerational experiment that examines how realistic competitive interactions impact the performance (growth, reproductive output, etc.) of hybrids. A particular strength of this article is that the study uses late-generation hybrids, rather than the direct products of the primary hybridization event. The increased or decreased fitness, resulting from heterosis or outbreeding depression seen in early generation hybrids (i.e. F1 hybrids), is often exaggerated relative to advanced-generation hybrids. Campbell & Snow have provided a more insightful assessment of the impact of crop-to-wild plant hybridization by investigating the performance of advanced-hybrid genotypes under realistic field conditions. Specifically, they used a well-studied system, consisting of a weedy radish species (Raphanus raphanistrum) and third-generation hybrids between R. raphanistrum and Rsativus. Interestingly, it appears in nature that the crop–wild hybrids have replaced the original populations of R. raphanistrum throughout California (Hedge et al., 2006). Campbell & Snow initiated this experiment in 2002 by planting three F1 hybrid populations and three wild populations of R. raphanistrum in Michigan. The populations experienced simulated agricultural management and natural environmental conditions through time. Once the F3 generation was produced, the parent species and hybrids were grown in a seminatural agricultural garden, under varying plant densities, to examine the effect of competition on life history traits and adult fecundity. Plants were grown (1) alone, (2) with intrabiotype competition (i.e. R. raphanistrum vs R. raphanistrum) or (3) with interbiotype competition (i.e. R. raphanistrum vs F3 hybrids). Using an elegant path analytic approach, the key finding of this large experiment was that whereas wild plants, when grown alone, generally outperformed the hybrids, overall fitness measures of hybrids were enhanced under competitive conditions. Thus, plant–plant competition may actually serve to increase the evolutionary impact of hybridization by promoting the movement (introgression) of crop alleles into wild populations. As noted by Campbell & Snow, ‘the persistence of crop genes within weed populations also depends on the competitive ability of advanced-generation hybrids when growing near its wild relatives, as well as other weed species’.


One of the fundamental underpinnings of evolutionary biology is that natural selection operates on variable phenotypic traits to effect changes in allele frequencies at the population level. The factors that result in the individual phenotype will have direct impact on the speed and intensity of a response to natural selection. These phenotypic factors include both the organism's genetic blueprint (‘nature’) and its environment (‘nurture’), such that phenotypic variation will reflect how genes interact with different environments. Such genetic × environment interactions are notoriously difficult to predict. This rather simple point highlights the importance of the current study by Campbell & Snow. The introgression of genes from crop to wild radishes impacted not only the hybrid competitive ability, but also had cascading effects on life history characters. Traits, such as the time to flower, are probably controlled by multiple genes. To then take these genes, place them into a real-world situation (such as a community of weeds) and predict their behavior, is virtually impossible. However, an awareness of the potential outcomes is critical.

Clearly, the world in which we live would be a vastly different place without the selective breeding for desirable crop traits that has occurred across several millennia, and more recently, the sophisticated ability to move single genes among species. If we were to revert back to growing ‘natural’, preselected fruits and vegetables, the world could support only a fraction of the people currently on the planet owing to greatly decreased crop yields. Yet, there are ecological risks with such tinkering of genomes. Campbell & Snow demonstrate the potential for crop × wild hybrids to be successful in realistic environments, and remind us that gene flow between such organisms may prove to one day be more common than previously anticipated.