Reversion is most likely under high mutation supply, when compensatory mutations don't fully restore fitness costs

Abstract

Adaptive mutations are often associated with a fitness cost. These costs can be compensated for through the acquisition of additional mutations, or the adaptations can be lost through reversion in settings where they are no longer favored. While the dynamics of adaptation, reversion and compensation have been central features in several studies of microbial evolution, few studies have attempted to resolve the population genetics underlying how and when either compensation or reversion occur. Specifically, questions remain regarding how certain actors--the evolution of mutators and whether compensatory mutations alleviate costs fully or partially-- may influence evolutionary dynamics of compensation and reversion. In this study, we attempt to explain findings from an experimental evolution study by utilizing computational and theoretical approaches towards a more refined understanding of how mutation rate and the fitness effects of compensatory mutations influence evolutionary dynamics. We find that high mutation rates increase the probability of reversion towards the wild type when compensation is only partial. However, the existence of even a single fully compensatory mutation is associated with a dramatically decreased probability of reversion to the wild type. These findings help to explain specific findings from experimental evolution, where compensation was observed in non-mutator strains, but reversion (sometimes with compensation) was observed in mutator strains, indicating that real-world compensatory mutations are often unable to fully alleviate the costs associated with resistance. Our findings emphasize the potential role of the supply and quality of mutations in crafting the evolution of antibiotic resistance, and more generally highlight the importance of population genetic context for explaining findings from experimental evolution.