Genomic analysis reveals major determinants of cis-regulatory variation in Capsella grandiflora

Abstract

Understanding the causes of cis-regulatory variation is a long-standing aim in evolutionary biology. Although cis-regulatory variation has long been considered important for adaptation, we still have a limited understanding of the selective importance and genomic determinants of standing cis-regulatory variation. To address these questions, we studied the prevalence, genomic determinants and selective forces shaping cis-regulatory variation in the outcrossing plant Capsella grandiflora. We first identified a set of 1,010 genes with common cis-regulatory variation using analyses of allele-specific expression (ASE). Population genomic analyses of whole-genome sequences from 32 individuals showed that genes with common cis-regulatory variation are 1) under weaker purifying selection and 2) undergo less frequent positive selection than other genes. We further identified genomic determinants of cis-regulatory variation. Gene-body methylation (gbM) was a major factor constraining cis-regulatory variation, whereas presence of nearby TEs and tissue specificity of expression increased the odds of ASE. Our results suggest that most common cis-regulatory variation in C. grandiflora is under weak purifying selection, and that gene-specific functional constraints are more important for the maintenance of cis-regulatory variation than genome-scale variation in the intensity of selection. Our results agree with previous findings that suggest TE silencing affects nearby gene expression, and provide novel evidence for a link between gbM and cis-regulatory constraint, possibly reflecting greater dosage-sensitivity of body-methylated genes. Given the extensive conservation of gene-body methylation in flowering plants, this suggests that gene-body methylation could be an important predictor of cis-regulatory variation in a wide range of plant species.