Detection of pseudouridine modifications and type I/II hypermodifications in human mRNAs using direct, long-read sequencing

Abstract

AbstractEnzyme-mediated chemical modifications to mRNAs have the potential to fine-tune gene expression in response to environmental stimuli. Notably, pseudouridine-modified mRNAs are more resistant to RNase-mediated degradation, more responsive to cellular stress, and have the potential to modulate immunogenicity and enhance translation in vivo. However, the precise biological functions of pseudouridine modification on mRNAs remain unclear due to the lack of sensitive and accurate tools for mapping. We developed a semi-quantitative method for mapping pseudouridylated sites with high confidence directly on mammalian mRNA transcripts via direct RNA, long-read nanopore sequencing. By analysis of a modification-free transcriptome, we demonstrate that the depth of coverage and intrinsic errors associated with specific k-mer sequences are critical parameters for accurate base-calling. We adjust these parameters for high-confidence U-to-C base-calling errors that occur at pseudouridylated sites, which are benchmarked against sites that were identified previously by biochemical methods. We also uncovered new pseudouridylated sites, many of which fall on genes that encode RNA binding proteins and on uridine-rich k-mers. Sites identified by U- to-C base calling error were verified using 1000mer synthetic RNA controls bearing a single pseudouridine in the center position, demonstrating that 1. the U-to-C base-calling error occurs at the site of pseudouridylation, and 2. the basecalling error is systematically under-calling the pseudouridylated sites. High-occupancy sites with >40% U-to-C basecalling error are classified as sites of hyper modification type I, whereas genes with more than one site of pseudouridylation are classified as having type II hyper modification which is confirmed by single-molecule analysis. We report the discovery of mRNAs with up to 7 unique sites of pseudouridine modification. Here we establish an innovative pipeline for direct identification, quantification, and detection of pseudouridine modifications and type I/II hypermodifications on native RNA molecules using long-read sequencing without resorting to RNA amplification, chemical reactions on RNA, enzyme-based replication, or DNA sequencing steps.