Shifting fitness landscapes in response to altered environments

One of the most controversial questions in evolutionary biology is the role of adaptation in molecular evolution. After decades of debate between selectionists and neutralists, new high-throughput methods are beginning to illuminate the full distribution of fitness effects of new mutations. Here, we shed light on the adaptive potential in Saccharomyces cerevisiae by presenting systematic high-throughput fitness measurements for 560 point mutations in a region of Hsp90 under six environmental conditions. Under elevated salinity, we observe numerous beneficial mutations, all of which are observed to be associated with high costs of adaptation. We thus demonstrate that an essential protein can harbor adaptive potential upon an environmental challenge, and report a remarkable fit of Fisher's geometric model. In addition, we compare the differences in the DFEs resulting from mutations covering 1, 2 and 3 nucleotide steps from the wild type - showing that multiple-step mutations harbor more potential for adaptation in challenging environments, but also tend to be more deleterious in the standard environment. We utilize a Bayesian MCMC modeling framework to evaluate the statistical significance of the results - showing a remarkable accuracy of the experimental approach that allows us, e.g., to identify a deleterious synonymous mutation under standard conditions.