Slip into something more functional: selection maintains ancient frameshifts in homopolymeric sequences

Mutational hotspots offer significant sources of genetic variability upon which selection can act. However, with a few notable exceptions, we know little about the dynamics and fitness consequences of mutations in these regions. Here, we explore evolutionary forces shaping homopolymeric tracts that are especially vulnerable to slippage errors during replication and transcription. Such tracts are typically eliminated by selection from most bacterial sequences, yet persist in genomes of endosymbionts with small effective population sizes (N(e)) and biased base compositions. Focusing on Blochmannia, a bacterial endosymbiont of ants, we track the divergence of genes that contain frameshift mutations within long (9-11 bp) polyA or polyT tracts. Earlier experimental work documented that transcriptional slippage restores the reading frame in a fraction of messenger RNA molecules and thereby rescues the function of frameshifted genes. In this study, we demonstrate a surprising persistence of these frameshifts and associated tracts for millions of years. Across the genome of this ant mutualist, rates of indel mutation within homopolymeric tracts far exceed the synonymous mutation rate, indicating that long-term conservation of frameshifts within these tracts is inconsistent with neutrality. In addition, the homopolymeric tracts themselves are more conserved than expected by chance, given extensive neutral substitutions that occur elsewhere in the genes sampled. These data suggest an unexpected role for slippage-prone DNA tracts and highlight a new mechanism for their persistence. That is, when such tracts contain a frameshift, transcriptional slippage plays a critical role in rescuing gene function. In such cases, selection will purge nucleotide changes interrupting the slippery tract so that otherwise volatile sequences become frozen in evolutionary time. Although the advantage of the frameshift itself is less clear, it may offer a mechanism to lower effective gene expression by reducing but not eliminating transcripts that encode full-length proteins.