Population and comparative genetics of thermotolerance divergence between yeast species

Abstract

Many familiar traits in the natural world-from lions' manes to the longevity of bristlecone pine trees-arose in the distant past, and have long since fixed in their respective species. A key challenge in evolutionary genetics is to figure out how and why species-defining traits have come to be. We used the thermotolerance growth advantage of the yeast Saccharomyces cerevisiae over its sister species Saccharomyces paradoxus as a model for addressing these questions. Analyzing loci at which the S. cerevisiae allele promotes thermotolerance, we detected robust evidence for positive selection, including amino acid divergence between the species and conservation within S. cerevisiae populations. Because such signatures were particularly strong at the chromosome segregation gene ESP1, we used this locus as a case study for focused mechanistic follow-up. Experiments revealed that, in culture at high temperature, the S. paradoxus ESP1 allele conferred a qualitative defect in biomass accumulation and cell division relative to the S. cerevisiae allele. Only genetic divergence in the ESP1 coding region mattered phenotypically, with no functional impact detectable from the promoter. Our data support a model in which an ancient ancestor of S. cerevisiae, under selection to boost viability at high temperature, acquired amino acid variants at ESP1 and many other loci, which have been constrained since then. Complex adaptations of this type hold promise as a paradigm for interspecies genetics, especially in deeply diverged traits that may have taken millions of years to evolve.

Keywords: Saccharomyces; adaptation; ancient; evolution; genetics; thermotolerance.

Figure 1

A peak of high-allele frequency in S. cerevisiae populations at the 5’ end of ESP1. Each panel shows results of analysis of allele frequency at the thermotolerance gene ESP1 in a population of S. cerevisiae from Peter et al. (2018). In each panel, the y-axis reports genotype homozygosity, G1, in a 1200-SNP window around the position shown on the x. The ESP1 open reading frame is demarcated with a dark black arrow (direction of transcription is right to left). (A) Wine/European population. (B) Mosaic Region 3 population. (C) Brazilian Bioethanol population. (D) Sake population. (E) Mixed Origin population.

Figure 2

The S. cerevisiae ESP1 coding region, but not the promoter, is required for thermotolerance. Each column represents results from biomass accumulation assays of a wild-type or transgenic yeast strain cultured at high temperature. The y-axis reports the optical density of a culture of the indicated strain after 24 h at 39°C, normalized to the analogous quantity from wild-type S. cerevisiae (dashed line). Each point reports results from one biological replicate, and each bar height reports the average across replicates (n = 6–18). The first two columns report results from wild-type (WT) strains of S. paradoxus Z1 (Sp) and S. cerevisiae DBVPG17373 (Sc). The last three columns report strains with the indicated region of ESP1 from S. paradoxus swapped into S. cerevisiae at the endogenous location; ESP1 full swap denotes transgenesis of both the promoter and the coding region. *Wilcoxon test P < 0.004 in a comparison against wild-type S. cerevisiae. Culture data at 28°C are given in Supplementary Figure S2.

Figure 3

Growth function of S. paradoxus ESP1 declines sharply with temperature. Each trace reports results from biomass accumulation assays of a wild-type or transgenic yeast strain across temperatures. Strain labels are as in Figure 2. The y-axis reports the optical density of a culture of the indicated strain after 24 h at the temperature on the x, normalized to the optical density of that day’s wild-type S. cerevisiae at 37°C. *P < 10⁻¹² for the strain by temperature interaction term of a two-factor ANOVA, in a comparison between the indicated strain and wild-type S. cerevisiae.

Figure 4

The S. paradoxus allele of ESP1 compromises cell division at high temperature. (A) Each panel reports a representative image of a wild-type or transgenic yeast strain after incubation for 24 h at the indicated temperature. Strain labels are as in Figure 2. (B) Each bar reports quantification of replicated imaging data of the indicated strain cultured at the indicated temperature, as in (A). For each bar, the y-axis shows the fraction of dyads in the indicated size category. *P < 0.015 for the strain by temperature interaction term of a two-factor ANOVA, in a comparison between the indicated strain and wild-type S. cerevisiae. Experiment details are given in Supplementary Table S7.