Conserved long-range base pairings associated with pre-mRNA processing of human protein-coding genes

The ability of nucleic acids to form double-stranded structures is essential for all living systems on Earth. While DNA employs it for genome replication, RNA molecules fold into complicated secondary and tertiary structures. Current knowledge on functional RNA structures in human protein-coding genes is focused on locally-occurring base pairs. However, chemical crosslinking and proximity ligation experiments have demonstrated that long-range RNA structures are highly abundant. Here, we present the most complete to-date catalog of conserved long-range RNA structures in the human transcriptome, which consists of 916,360 pairs of conserved complementary regions (PCCRs). PCCRs tend to occur within introns proximally to splice sites, suppress intervening exons, circumscribe circular RNAs, and exert an obstructive effect on cryptic and inactive splice sites. The double-stranded structure of PCCRs is supported by a significant decrease of icSHAPE nucleotide accessibility, high abundance of A-to-I RNA editing sites, and frequent occurrence of forked eCLIP peaks nearby. Introns with PCCRs show a distinct splicing pattern in response to RNA Pol II slowdown suggesting that splicing is widely affected by co-transcriptional RNA folding. Additionally, transcript starts and ends are strongly enriched in regions between complementary parts of PCCRs, leading to an intriguing hypothesis that RNA folding coupled with splicing could mediate co-transcriptional suppression of premature cleavage and polyadenylation events. PCCR detection procedure is highly sensitive with respect to bona fide validated RNA structures at the expense of having a high false positive rate, which cannot be reduced without loss of sensitivity. The catalog of PCCRs is visualized through a UCSC Genome Browser track hub.