Genomic islands of speciation in Anopheles gambiae

Abstract

The African malaria mosquito, Anopheles gambiae sensu stricto (A. gambiae), provides a unique opportunity to study the evolution of reproductive isolation because it is divided into two sympatric, partially isolated subtaxa known as M form and S form. With the annotated genome of this species now available, high-throughput techniques can be applied to locate and characterize the genomic regions contributing to reproductive isolation. In order to quantify patterns of differentiation within A. gambiae, we hybridized population samples of genomic DNA from each form to Affymetrix GeneChip microarrays. We found that three regions, together encompassing less than 2.8 Mb, are the only locations where the M and S forms are significantly differentiated. Two of these regions are adjacent to centromeres, on Chromosomes 2L and X, and contain 50 and 12 predicted genes, respectively. Sequenced loci in these regions contain fixed differences between forms and no shared polymorphisms, while no fixed differences were found at nearby control loci. The third region, on Chromosome 2R, contains only five predicted genes; fixed differences in this region were also verified by direct sequencing. These "speciation islands" remain differentiated despite considerable gene flow, and are therefore expected to contain the genes responsible for reproductive isolation. Much effort has recently been applied to locating the genes and genetic changes responsible for reproductive isolation between species. Though much can be inferred about speciation by studying taxa that have diverged for millions of years, studying differentiation between taxa that are in the early stages of isolation will lead to a clearer view of the number and size of regions involved in the genetics of speciation. Despite appreciable levels of gene flow between the M and S forms of A. gambiae, we were able to isolate three small regions of differentiation where genes responsible for ecological and behavioral isolation are likely to be located. We expect reproductive isolation to be due to changes at a small number of loci, as these regions together contain only 67 predicted genes. Concentrating future mapping experiments on these regions should reveal the genes responsible for reproductive isolation between forms.

Figure 1. Differentiation between Forms

The significance threshold shown is p = 0.05, Bonferroni-corrected for the number of windows tested per chromosome. The centromeres of Chromosomes 2 and 3 are located between the right and left arms, i.e., between each pair of graphs; the centromere of Chromosome X is located at the right end of the graph. Grey areas are divergent regions identified by our HMM. The grey region at the tip of Chromosome X appears to lie outside of the final window because the chromosomal position given for each window is the location of the central probe in that window; the final window on Chromosome X spans a large region because of low gene density. Sequenced loci are shown with red triangles; overlapping triangles on Chromosomes 3R and 2R obscure multiple sequenced loci (see text for details). 3L, left arm of Chromosome 3; 3R, right arm of Chromosome 3.