The hybrid zones studied by Nolte et al. (2009) have developed from very recent secondary contact (in the early 1990s) between the Rhine stream sculpin (Cottus rhenanus) and an invasive sculpin species that is expanding its range upriver. In fact, the invasive sculpin itself appears to have originated via hybridization between C. rhenanus and a distinct stream sculpin, Cottus perifretum, an event that presumably occurred less than 200 years ago after an artificial canal connected these formerly allopatric species (Nolte et al. 2005). Currently, steep clines form where C. rhenanus and the invasive sculpin species overlap, suggesting there is strong natural selection against immigrant genotypes and/or associations with narrow ecotones at the confluence of small streams and larger rivers (Nolte et al. 2006). It is this ecological context that makes Cottus (Fig. 1) an excellent system for evaluating the relative contributions of intrinsic hybrid fitness problems (endogenous selection) and adaptation to novel environments (exogenous selection) in the process of speciation.
The genetic architecture of reproductive isolation can dictate the potential for introgression between sympatric species (Wu 2001). Depending on their number, distribution, and phenotypic effects, isolation loci might prevent tightly linked genomic regions from moving freely between species. So, to what degree does reproductive isolation restrict the potential for gene flow between hybridizing sculpins? In this study, Nolte et al. (2009) use a new method to test whether introgression for a particular locus deviates from what is expected based on the pattern of genome-wide admixture (Gompert & Buerkle 2009). A major advantage of this approach is that it can detect both the limitation and promotion of introgression across species boundaries. In fact, roughly half of the 168 microsatellites tested by Nolte et al. (2009) exhibit a significant deviation from the neutral expectation (generated via parametric simulations that approximate a model of neutral diffusion). Of these ‘deviant’ loci, a majority shows a pattern of genetic variation consistent with underdominance (reduced fitness of heterozygotes). One possible explanation for this pattern is trivial; Nolte et al. (2009) might have underestimated the proportion of microsatellite markers segregating null alleles. But, assuming this explanation cannot entirely account for the deficit of heterozygotes, it is possible that chromosomal inversions and/or assortative mating contribute to reproductive isolation between sculpin species. Surprisingly, only a single marker showed evidence for epistasis, an indication that few hybrid incompatibilities have evolved. But, what about the potential for adaptive gene flow? Given the recent hybrid origin of the invasive sculpin, the authors have good reason to suspect that not all genomic regions cause reproductive isolation. Indeed, Nolte et al. (2009) discover evidence for adaptive introgression of invasive sculpin alleles at a substantial proportion of markers (17% averaged between the two sites).
Because the invasive sculpin species has encroached upon the range of C. rhenanus within multiple, independent tributary streams, the authors can also address the important question of whether there has been parallel evolution of particular loci. Interestingly, Nolte et al. (2009) find little correspondence between two hybrid zones in patterns of variation for isolation loci. The proportion of underdominant loci observed in both hybrid zones is no greater than what is expected by chance. On the other hand, there is some overlap between the two study sites in loci that show adaptive introgression of invasive sculpin alleles.
Of course, the next step is to understand why there might be variation in reproductive isolation among replicate hybrid zones. Because of the young age of the invasive sculpin, Nolte et al. (2009) argue that variation between hybrid zones is either due to differentiation among populations of C. rhenanus, or to variable selection pressures at the two sites. The first possibility — divergence within C. rhenanus— seems reasonable; populations used in this study show both genetic (Nolte et al. 2006) and phenotypic variation (Nolte & Sheets 2005). In line with these results, classical crossing experiments in plants and animals have often shown variability for barriers to interspecific reproduction, and several recent studies have discovered variation in hybrid inviability or sterility loci in natural populations (Bomblies et al. 2007; Sweigart et al. 2007; Good et al. 2008; Seidel et al. 2008).
This study of hybrid zones by Nolte et al. (2009) is a promising first step towards an understanding of the ecological and genetic basis of sculpin divergence. Future research efforts should focus on genetic mapping to confirm the phenotypic effects of candidate regions for reproductive isolation. The authors are also eager to look for potential site-specific ecological determinants of sculpin fitness, which can then be included as additional variables in their genomic clines analysis (Gompert & Buerkle 2009). New technologies like high throughput sequencing promise to open the door to many more genetic studies of reproductive isolation in natural populations. The challenge ahead will be to connect the dots between the initial genetic changes that accompany the early stages of speciation and the genetic differences that eventually mark well-defined species.