Mutation rate variation in eukaryotes: evolutionary implications of site-specific mechanisms

Baer et al.'s Review “Mutation rate variation in multicellular eukaryotes: causes and consequences”¹ accurately conveys the tenor of most literature on mutation rate evolution. Even while recognizing that mutation rates can vary dramatically from site to site within a genome, this literature mostly ignores or minimizes the special characteristics and consequences of site-specific mutational mechanisms. Therefore, to the authors' list of “five critical questions” for future investigation, we would like to propose a sixth: to what extent can elevated site-specific mutation rates vary (and evolve) independently from low average nucleotide-substitution rates?

Following historical precedent, Baer et al.¹ consistently refer to “the mutation rate” as if a single value could adequately summarize the complex outcome of many diverse mutational mechanisms. However, such a simplistic statistic should no longer be tolerated, at least not without precise qualification such as “average rate of nucleotide substitutions within protein-coding sequences”. Even then, somatic mutability of antibody genes demonstrates that localized hypermutation can be exploited by adaptive evolution². Site-specific sources of germline mutation are also familiar, as evidenced by the prolific variation that is produced by simple sequence repeats (SSRs; microsatellites and minisatellites)³.

Although Baer et al.¹ acknowledge the remarkably high mutation rates of SSRs, they do not discuss the profound implications of such site-specific mutability for understanding mutation rate evolution. Phenotypic effects of SSR mutations include reversible on–off switching and quantitative variation in many aspects of gene function^3,4,5,6,7,8. Such effects need not be predominantly deleterious, especially when adaptation is suboptimal, as in variable environments⁹. But an assumption that “the vast majority of mutations with observable effects are deleterious” has informed most discussion of mutation rate evolution throughout the past century¹. Traditionally, theoretical discussion also assumes that recombination in diploid organisms must eventually separate 'mutator alleles' from resulting mutations, thereby preventing mutators from 'hitch-hiking' on selection for beneficial mutants¹. However, recombination cannot unlink site-specific mutational mechanisms, such as those based on SSRs, from the mutations that they generate. Hence, selection for any beneficial mutation at a mutable site must also, indirectly, favour the site's intrinsic mutability.

The best evidence for positive selection of sites with high mutation rates comes from the SSR-based contingency genes of haploid microorganisms¹⁰. In eukaryotes, the reported abundances, genomic distributions, phylogenetic conservation and patterns of variation for SSRs are also strongly suggestive of positive selection^4,6,7. The utility of SSR mutability for adaptive evolution has already been implicated in several cases, including skeletal evolution of domestic dogs⁵, temperature compensation of the Drosophila melanogaster circadian clock^11,12,13 and social behaviour of voles¹⁴, among others⁷. Whether such mutable sites somehow prevail in spite of their high mutability (as conventional theory requires), or whether indirect selection for their high mutability is the reason for their prevalence^8,9, remains to be established. In either case, the special characteristics of site-specific mutational mechanisms deserve attention in any future consideration of mutation rate evolution.