Newly arising mutations create the phenotypic variation upon which selection can act. But mutations do not necessarily act independently or on a single phenotype, making their effects on evolution difficult to predict. Mutational interactions, or epistasis, can constrain the path of natural selection. Additionally, pleiotropic mutations, which impact more than one phenotype, have fitness consequences that represent the totality of their phenotypic effects and not simply their impact on any individual character.

I present three studies that examine the impact of epistasis and pleiotropy on protein evolution. First, my colleagues and I identified a single nucleotide insertion in a vineyard isolate of Saccharomyces cerevisiae that has cascading effects through the gene-expression network. Using isogenic laboratory strains, we confirm that this allele causes dramatic differences in gene-expression levels of key genes involved in amino acid biosynthesis. We conclude that this allele's relatively high mutation rate, combined with its mild phenotypic effects, account for its persistence in natural populations.

Second, in order to understand the importance of regulatory and structural mutations in multistep evolutionary pathways, we carried out experiments in which the expression of β-lactamase in Escherichia coli was under the control of a tunable arabinose promoter. We find that the fitness effect of an increase in gene expression is highly dependent on the catalytic activity of the coding sequence. The mapping of enzyme activity to fitness strongly influences the temporal incorporation and importance of regulatory mutants on evolutionary pathways.

Finally, utilizing a Saccharomyces cerevisiae model of dihydrofolate reductase (DHFR) evolution in Plasmodium falciparum, we examine the robustness of growth rate to mutations that confer drug resistance. Assays of all 48 combinations of 6 naturally occurring resistance mutations reveal that growth and resistance phenotypes freely associate and do not demonstrate a strong negative tradeoff. The three evolutionary pathways that dominate DHFR evolution show that subsequent resistance-increasing mutations can compensate for initial declines in growth rate. Our results suggest that growth rate in P. falciparum is robust to drug resistance mutations at the DHFR locus.