The Distribution of Fitness Effects of Synonymous Mutations in a Gene under Selection

Abstract

Models of adaptive evolution consider the path that a population takes to a fitness optimum through a series of beneficial single-nucleotide changes, although synonymous mutations, which do not alter the amino acid sequence, have been conventionally seen as ‘silent’ changes. However, recent studies have reported the existence of spontaneous adaptive synonymous substitutions, raising the possibility that this type of mutation is an overlooked source of adaptive potential. Using site-directed mutagenesis, we constructed and quantified the Darwinian fitness of 40 synonymous and 71 nonsynonymous individual mutants of a gene essential for glucose uptake in Pseudomonas fluorescens. The gene in question, gtsB, is the target of natural selection in an environment where glucose is the growth-limiting factor. Of the 56 beneficial mutations, half were synonymous and covered as wide a range of beneficial fitness effects as the nonsynonymous sample. Drawing on extreme value theory, we show that the fitness effects of synonymous mutations correspond to a distribution with a right-truncated tail belonging to the Weibull domain of the Generalised Pareto Distribution. Moreover, the distribution of beneficial fitness effects of synonymous mutations is indistinguishable from that of nonsynonymous mutations, revealing their previously unrecognised adaptive potential.