Temperature-dependent genetics of thermotolerance between yeast species

Abstract

Many traits of industrial and basic biological interest arose long ago, and manifest now as fixed differences between a focal species and its reproductively isolated relatives. In these systems, extant individuals can hold clues to the mechanisms by which phenotypes evolved in their ancestors. We harnessed yeast thermotolerance as a test case for such molecular-genetic inferences. In viability experiments, we showed that extant Saccharomyces cerevisiae survived at temperatures where cultures of its sister species S. paradoxus died out. Then, focusing on loci that contribute to this difference, we found that the genetic mechanisms of high-temperature growth changed with temperature. We also uncovered an enrichment of low-frequency variants at thermotolerance loci in S. cerevisiae population sequences, suggestive of a history of non-neutral selective forces acting at these genes. We interpret these results in light of models of the evolutionary mechanisms by which the thermotolerance trait arose in the S. cerevisiae lineage. Together, our results and interpretation underscore the power of genetic approaches to explore how an ancient trait came to be.

Keywords: Adaptation; Ancient; Cline; Evolution; Saccharomyces; Thermotolerance.

Figure 1.. S. paradoxus alleles of thermotolerance genes confer temperature-dependent defects in the S. cerevisiae background.

Each line reports the results of growth experiments of a strain of S. cerevisiae DBVPG1373 harboring the indicated gene from S. paradoxus Z1, or the respective wild-type parent strains (WT), across a temperature gradient. The y-axis reports growth efficiency, the optical density reached by the culture after 24 hours at the temperature indicated on the x-axis, as a difference from the analogous quantity at time zero. Each point reports results from one biological replicate (n = 3), and the line represents the average growth efficiency of the indicated strain across the temperature gradient. *, P < 10⁻³ for the strain by temperature interaction term of a two-factor ANOVA, in a comparison between the indicated strain and wild-type S. cerevisiae (Table S2).