Natural variation in abiotic stress responsive gene expression and local adaptation to climate in Arabidopsis thaliana

Abstract

Gene expression varies widely in natural populations, yet the proximate and ultimate causes of this variation are poorly known. Understanding how variation in gene expression affects abiotic stress tolerance, fitness, and adaptation is central to the field of evolutionary genetics. We tested the hypothesis that genes with natural genetic variation in their expression responses to abiotic stress are likely to be involved in local adaptation to climate in Arabidopsis thaliana. Specifically, we compared genes with consistent expression responses to environmental stress (expression stress responsive, "eSR") to genes with genetically variable responses to abiotic stress (expression genotype-by-environment interaction, "eGEI"). We found that on average genes that exhibited eGEI in response to drought or cold had greater polymorphism in promoter regions and stronger associations with climate than those of eSR genes or genomic controls. We also found that transcription factor binding sites known to respond to environmental stressors, especially abscisic acid responsive elements, showed significantly higher polymorphism in drought eGEI genes in comparison to eSR genes. By contrast, eSR genes tended to exhibit relatively greater pairwise haplotype sharing, lower promoter diversity, and fewer nonsynonymous polymorphisms, suggesting purifying selection or selective sweeps. Our results indicate that cis-regulatory evolution and genetic variation in stress responsive gene expression may be important mechanisms of local adaptation to climatic selective gradients.

Keywords: abiotic stress; landscape genomics; phenotypic plasticity; regulatory evolution; transcriptome.

Fig. 1.

Representative eSR (A) and eGEI (B) expression responses to drought (Des Marais et al. 2012). (A) AT1G15290, a tetratricopeptide repeat-like superfamily protein, has a significant eSR (but not eGEI) effect in response to drought. The AT1G15290 promoter is characterized by an uncommonly high and invariant number of ABREs across 80 accessions. Additionally, AT1G15290 contains a relatively high number of SNPs with high PHS. (B) AT1G33760, a member of the DREB subfamily A-4 of the ERF/AP2 transcription factor family, has a significant eGEI effect in response to drought and is near (<1 kb distant) SNPs with outlier associations with survival in the United Kingdom and growing season precipitation. Additionally, ABRE counts in the AT1G33760 promoter showed relatively high variation across 80 accessions. The ten early-flowering accessions from the original experiment (Des Marais et al. 2012) are shown.

Fig. 2.

Enrichment of gene sets with associations to cold-related climate variables for early- and late-flowering accessions. Observed enrichments are calculated as a z score using the distribution from null permutations. A high z score indicates a gene set has greater outlier climate-SNP associations compared with the genome-wide expectation, whereas a low z indicates few outlier climate-SNP associations. Enrichments are shown with large circles with eSR genes in the top rows and eGEI genes in the bottom rows. Null permutations are shown as small gray dots (°P < 0.1, *P < 0.05, **P < 0.01, ***P < 0.005).

Fig. 3.

Enrichment of gene sets with associations to drought-related climate variables for early- and late-flowering accessions. Observed enrichments are calculated as a z score using the distribution from null permutations. A high z score indicates a gene set has greater outlier climate-SNP associations compared with the genome-wide expectation, whereas a low z indicates few outlier climate-SNP associations. Enrichments are shown with large circles with eSR genes in the top rows and eGEI genes in the bottom rows. Null permutations are shown as small gray dots (°P < 0.1, *P < 0.05, **P < 0.01, ***P < 0.005).

Fig. 4.

The enrichment of candidate gene sets with promoter diversity or ratio of nonsynonymous/synonymous polymorphisms in 80 resequenced accessions. Observed enrichments are calculated as a z score using the distribution from null permutations of gene sets. Enrichments are shown with large circles with eSR genes in the top rows and eGEI genes in the bottom rows. Null permutations are shown as small gray points (*P < 0.05, **P < 0.01, ***P < 0.005).

Fig. 5.

The enrichment of candidate gene sets with mean motif frequency or variance in motif frequency among 80 resequenced accessions. Observed enrichments are calculated as a z score using the distribution from null permutations. Enrichments are shown with large circles with eSR genes in the top rows and eGEI genes in the bottom rows. Null permutations are shown as small gray dots (°P < 0.1, *P < 0.05, **P < 0.01, ***P < 0.005).