Alternative splicing contributes to plasticity and regulatory divergence in locally adapted house mice from the Americas

Abstract

Alternative splicing is a major driver of transcriptome and proteome variation, but the role of alternative splicing in regulatory evolution remains understudied. Alternative splicing can also contribute to phenotypic plasticity, which may be critical when taxa colonize new environments. Here, we investigate variation in alternative splicing among new wild-derived strains of mice from different climates in the Americas on both a standard and high-fat diet. We show that alternative splicing is widespread and highly context-dependent. Comparisons between strains on different diets revealed abundant gene-by-environment interactions affecting alternative splicing, with most genes showing strain- and sex-specific diet responses. More often than not, genes that were differentially spliced between strains were not differentially expressed, adding to evidence that the two regulatory mechanisms often act independently. Moreover, differentially spliced genes were more widely expressed across tissues but also less central to biological networks than differentially expressed genes, suggesting differences in pleiotropic constraint. Importantly, divergence in alternative splicing was found to be predominantly driven by cis-regulatory changes. However, trans changes affecting splicing make be central to plasticity as they were impacted more by environmental variation. Finally, we performed scans for selection and found that, while genes with splicing divergence more often co-localized with genomic outliers associated with metabolic traits, they were not enriched for genomic outliers. Overall, our results provide evidence that alternative splicing plays an important role in gene regulation in house mice, contributing to adaptation and plasticity.

Keywords: Mus; adaptation; diet; gene regulation; gene-by-environment interactions.

Figure 1.

Alternative splicing variation in wild-derived house mouse strains from the Americas. A. Locations from which each strain was derived: EDME: Edmonton, Alberta, Canada; GAIB: Gainesville, Florida, USA; MANB: Manaus, Amazonas, Brazil; SARA, SARB: Saratoga Springs, New York, USA. B. Males and females from each strain were reared on either a high-fat (HF) or standard diet (STD). Liver tissue was collected from each mouse for RNA-seq. C. Five alternative splicing event types were surveyed across strains. D. Principal component analysis of splicing variation showed that strains cluster by locality on PC 1 and PC2.

Figure 2.

Alternative splicing differences associated with diet. A. UpSet plot showing the overlap in differentially spliced genes across strains between a high-fat and standard diet. Alternative splicing was strain-specific for the majority of genes. B. Event types for differential spliced genes across diets for each strain. The “shared” bar represents genes that showed differential splicing in more than one strain.

Figure 3.

Cis and trans changes contribute to differences in splicing between strains from different climates. Bar charts below show the number of genes attributed to either cis (left bars) or trans (right bars) for each splicing event (A3SS: alternative 3’ splice site; A5SS: alternative 5’ splice site; RI: retained intron; SE: skipped exon).

Figure 4.

Evidence for diet-specific allele-specific alternative splicing (ASAS). Plots show comparisons of percent spliced in (PSI) proportions between alleles on the high-fat versus standard diet. Cases where ASAS is specific to one diet (purple and blue) highlight cis-by-diet interactions affecting alternative splicing.