RNA structure landscape of S. cerevisiae introns

Pre-mRNA secondary structures are hypothesized to play widespread roles in regulating RNA processing pathways, but these structures have been difficult to visualize in vivo. Here, we characterize S. cerevisiae pre-mRNA structures through transcriptome-wide dimethyl sulfate (DMS) probing, enriching for low-abundance pre-mRNA through splicing inhibition. These data enable evaluation of structures from phylogenetic and mutational studies as well as identification of new structures across 161 introns. We find widespread formation of “zipper stems” between the 5' splice site and branch point, “downstream stems” between the branch point and the 3' splice site, and previously uncharacterized long stems that distinguish pre-mRNA from spliced mRNA. Multi-dimensional chemical mapping reveals that intron structures can form in vitro without the presence of binding partners, and structure ensemble prediction suggests that these structures appear in introns across the Saccharomyces genus. We develop the functional assay VARS-seq to characterize variants of RNA structure in high-throughput and we apply the method on 135 sets of stems across 7 introns, finding that some structured elements can increase spliced mRNA levels despite being distal from canonical splice sites. Unexpectedly, other structures including zipper stems can increase retained intron levels. This transcriptome-wide inference of intron RNA structures suggests new ideas and model systems for understanding how pre-mRNA folding influences gene expression. Overall design: 1. RNA structure probing after splicing inhibition in yeast: RNA was treated with DMS and reverse transcription with TGIRT (DMS-MaPseq) for samples and controls. The samples include: two replicates of DMS treatment after splicing inhibition (one replicate with 3 lanes of sequencing, and another with one lane), one sample for DMS treatment without splicing inhibition, and one sample for a no-modification control with splicing inhibition. 2. Splicing measurements for intron variant library: Intron structure variant libraries were integrated into the S. cerevisiae genome in their native gene context. Splicing levels were measured with RNA-sequencing, and gDNA sequencing was used to identify intron variants by pairing randomized barcodes to intron sequences.