Control of gene output by intron RNA structure

Abstract

During mRNA biogenesis, RNA folding can promote or antagonize transcript processing. The effects of intron structure on mRNA and protein levels remain largely unexplored, although introns account for the bulk of nascent RNA. Here we systematically probe the effect of intron structure using massively parallel reporter assays. We show that base pairing modulates gene expression across orders of magnitude through inhibitory RNA structures at splice sites and at newly identified regions. Conversely, poor splicing of human {beta}-globin pre-mRNA could be improved by sequence alterations that reduce base pairing. For large libraries of RNA structures differing in stability, machine learning models could nearly fully explain observed gene output. Structure destabilizing mutations emerge rapidly under selective pressure. Thus, formation of RNA structures as dictated by intron sequence provides a simple, powerful means to adjust gene expression.