The Population Genomics of Trans-Specific Inversion Polymorphisms in Anopheles gambiae

In the malaria mosquito Anopheles gambiae polymorphic chromosomal inversions may play an important role in adaptation to environmental variation. Recently, we used microarray-based divergence mapping combined with targeted resequencing to map nucleotide differentiation between alternative arrangements of the 2La inversion. Here, we applied the same technique to four different polymorphic inversions on the 2R chromosome of An. gambiae. Surprisingly, divergence was much lower between alternative arrangements for all 2R inversions when compared to the 2La inversion. For one of the rearrangements, 2Ru, we successfully mapped a very small region (∼100 kb) of elevated divergence. For the other three rearrangements, we did not identify any regions of significantly high divergence, despite ample independent evidence from natural populations of geographic clines and seasonal cycling, and stable heterotic polymorphisms in laboratory populations. If these inversions are the targets of selection as hypothesized, we suggest that divergence between rearrangements may have escaped detection due to retained ancestral polymorphism in the case of the youngest 2R rearrangements and to extensive gene flux in the older 2R inversion systems that segregate in both An. gambiae and its sibling species An. arabiensis.MORE than 70 years ago Dobzhansky and Sturtevant (1938) first discovered polymorphic inversion arrangements carried by various Drosophila pseudoobscura populations. After observing correlations between environmental conditions and inversion frequencies, Dobzhansky proposed that inversions are under strong selection due to their role in promoting local adaptation to the heterogeneous conditions a species encounters both spatially and temporally (Dobzhansky 1944, 1948; Powell 1997). More recent studies have implicated chromosomal inversions in the adaptation of a diversity of eukaryotes including humans (Coluzzi et al. 1979; Feder et al. 2003; Hoffmann et al. 2004; Stefansson et al. 2005). Long known to be common in dipteran insects, more recent HapMap data suggest that polymorphic inversions may be numerous in human populations and by extension other mammals (Bansal et al. 2007). Given their potential importance in facilitating adaptation, surprisingly little is known about the mechanism(s) or the genes responsible for maintaining inversion polymorphisms in natural populations.Gene exchange between inverted and standard arrangements, although reduced, can still occur through gene flux: the action of gene conversion and multiple crossovers in inversion heterozygotes (heterokaryotypes) (Chovnick 1973; Navarro et al. 1997; Schaeffer and Anderson 2005). Over time allelic variation unrelated to ecological adaptation should become homogenized between arrangements, while alleles which are under divergent selection pressures should remain in linkage disequilibrium with each other and with the inversion itself, leading to heightened differentiation between standard and inverted arrangements at and near the target loci. In principle, this process allows the identification of specific loci involved in adaptive divergence (Schaeffer et al. 2003; Schaeffer and Anderson 2005; Storz 2005). Consistent with this model, previous low-resolution studies of Drosophila inversions revealed heterogeneous patterns of nucleotide diversity relative to divergence, as well as the interspersion of regions of high and low genetic association potentially due to the interaction of selection and gene flux (Schaeffer et al. 2003; Kennington et al. 2006; but see Munte et al. 2005). The application of high-resolution tools flowing from completely sequenced genomes will facilitate the mapping of genes that are the targets of divergent natural selection within gene arrangements.Although Drosophila has been the favored model, the African malaria vector Anopheles gambiae sensu stricto also provides an excellent system for studying the maintenance of inversion polymorphisms, not only within a species but across speciation events of different ages in the An. gambiae sibling species complex. The nominal species An. gambiae s.s. (hereafter, An. gambiae) is synanthropic: almost exclusively biting humans, resting indoors, and exploiting anthropogenic larval habitats (Coluzzi 1999). This close association with humans, vital to making An. gambiae one of the most proficient vectors of malaria, is likely to have been facilitated by chromosomal inversions thought to confer adaptive benefits in heterogeneous climatic and ecological settings in Africa. Seven common polymorphic inversions exist on the second chromosome. Six of these are located on the right arm (2R): j, b, c, u, d, and k, while 2La is the only inversion on the left arm (Coluzzi et al. 2002). Facilitated by the sequenced reference genome (Holt et al. 2002), some of the breakpoints for these polymorphic inversions have been localized to small genomic regions (Sharakhov et al. 2006; Coulibaly et al. 2007; Sangare 2007). Most of these inversions appear to be the targets of strong selection. Five of the inversions (2La and 2Rb, -c, -d, and -u) are nonrandomly associated with degree of aridity; each cycles seasonally with rainfall, and all except 2Ru form stable geographic clines in frequency from mesic forest to xeric regions bordering the Sahara (Coluzzi et al. 1979; Toure et al. 1994, 1998; Powell et al. 1999). Inversion 2Rj is not clinal, but its distribution in Mali is consistent with adaptation to novel rockpool niches (Coluzzi et al. 1985; Manoukis et al. 2008).In the An. gambiae species complex, inversion polymorphisms can be maintained across the boundaries of emerging and even full species. An. gambiae and its sibling An. arabiensis, strictly sympatric throughout most of their extensive ranges in sub-Saharan Africa, differ by multiple fixed chromosomal rearrangements on the X but share three chromosome 2 inversions: 2La, fixed in An. arabiensis and polymorphic in An. gambiae; and 2Rb and -c, polymorphic in both species (Coluzzi et al. 1979, 2002). Moreover, these same inversions and all other common An. gambiae inversions with the exception of 2Rj are shared and polymorphic in two lineages apparently undergoing ecological speciation within An. gambiae—the assortatively mating M and S molecular forms (della Torre et al. 2002, 2005). Inversion frequencies are correlated with climatic and ecological conditions in parallel in both lineages (Costantini et al. 2009; Simard et al. 2009). Unlike the full species, the M and S incipient species are not distinguished by any fixed inversion differences. Indeed, genomewide divergence mapping between the M and S forms revealed that significant differentiation was confined to two small low-recombination regions adjacent to the centromeres of 2L and X which are distant from any inversions (Turner et al. 2005). Thus, in distinction to models of speciation invoking inversions as facilitating the persistence of hybridizing species (Noor et al. 2001; Rieseberg 2001; Ortiz-Barrientos et al. 2002; Navarro and Barton 2003), the An. gambiae data suggest that chromosome 2 inversions are not directly responsible for reproductive isolation. Instead, the same chromosome 2 inversion polymorphisms appear to confer similar ecological benefits, within and across species boundaries. A long-term research goal is to identify the mechanisms and the genes controlling these processes.Previously we conducted the first high-density genomic scan of divergence across a chromosomal inversion (2La) in An. gambiae (White et al. 2007). By hybridizing genomic DNA from S form mosquitoes homokaryotypic for alternate gene arrangements on chromosome 2L (2La or 2L+^a) to oligonucleotide microarrays we were able to measure divergence across the 22-Mb inversion at nearly 14,000 markers. Differentiation in the rearranged region was significantly higher than in collinear portions of chromosome 2L. Between breakpoints the pattern of differentiation was heterogeneous: two genomic clusters of significantly higher divergence were identified near but not adjacent to the breakpoints. Directed resequencing within the S form confirmed these results and suggested that both clusters contained genes targeted by selection. Observed levels of linkage disequilibrium between the 2La breakpoints and markers in the clusters are highly unlikely under a neutral scenario, in light of known recombination rates and plausible estimates of the age of the inversion.The present study characterizes the patterns of genetic variation in polymorphic rearrangements on the opposite (right) arm of chromosome 2: 2Rj, -b, -c, and -u. With the goal of identifying candidate genes maintaining these inversions in natural populations, we applied microarray-based divergence mapping to measure differentiation between alternative 2R arrangements. Because three of four inversions have taxonomic distributions that span incipient and/or completed speciation events, we validated the microarray findings by targeted sequencing in multiple taxa: sympatric Malian populations of An. gambiae M and S forms, and the sibling species An. arabiensis.