Abstract
Many traits of interest in biology evolved long ago and are fixed in a particular species, distinguishing it from other sister taxa. Elucidating the mechanisms underlying such ancient traits has been a central goal for evolutionary biologists. The yeast Saccharomyces cerevisiae is unique among its relatives for its ability to thrive at high temperature, via a genetic architecture that remains incompletely understood. We sought to understand the contribution of species variation in mitochondrial DNA to yeast thermotolerance. We used mitochondrial transgenesis to show that S. cerevisiae mitotypes were sufficient for a partial boost to thermotolerance and respiration in the S. paradoxus background. These mitochondrial alleles worked best when the background also harbored a pro-thermotolerance nuclear genotype, attesting to positive epistasis between the two genomes. The benefits of S. cerevisiae alleles in terms of respiration and growth at high temperature came at the cost of worse performance in cooler conditions. Together, our results establish this system as a case in which mitoalleles have fitness benefits and work well in multiple backgrounds; given the high mutation rate of the mitochondrial compartment, they were also likely extra easy for the ancestral population to acquire. We thus propose a broader model in which mitochondrial variants may prove to be especially good candidates in the search for genetic mechanisms of the adaptive process.