The roles of positive and negative selection in the molecular evolution of insect endosymbionts

The evolutionary rate acceleration observed in most endosymbiotic bacteria may be explained by higher mutation rates, changes in selective pressure, and increased fixation of deleterious mutations by genetic drift. Here, we explore the forces influencing molecular evolution in Blochmannia, an obligate endosymbiont of Camponotus and related ant genera. Our goals were to compare rates of sequence evolution in Blochmannia with related bacteria, to explore variation in the strength and efficacy of negative (purifying) selection, and to evaluate the effect of positive selection. For six Blochmannia pairs, plus Buchnera and related enterobacteria, estimates of sequence divergence at four genes confirm faster rates of synonymous evolution in the ant mutualist. This conclusion is based on higher dS between Blochmannia lineages despite their more recent divergence. Likewise, generally higher dN in Blochmannia indicates faster rates of nonsynonymous substitution in this group. One exception is the groEL gene, for which lower dN and dN / dS compared to Buchnera indicate exceptionally strong negative selection in Blochmannia. In addition, we explored evidence for positive selection in Blochmannia using both site-and lineage-based maximum likelihood models. These approaches confirmed heterogeneity of dN / dS among codon sites and revealed significant variation in dN / dS across Blochmannia lineages for three genes. Lineage variation affected genes independently, with no evidence of parallel changes in dN / dS across genes along a given branch. Our data also reveal instances of dN / dS greater than one; however, we do not interpret these large dN / dS ratios as evidence for positive selection. In sum, while drift may contribute to an overall rate acceleration at nonsynonymous sites in Blochmannia, variable selective pressures best explain the apparent gene-specific changes in dN / dS across lineages of this ant mutualist. In the course of this study, we reanalyzed variation at Buchnera groEL and found no evidence of positive selection that was previously reported.