Investigating the enzymology and molecular function of 5-methylctosine in RNA

Abstract

Internal nucleoside methylation of eukaryotic RNA in the form of N6-methyladenosine (m⁶A) and 5-methylcytosine (m⁵C) have been known to exist for decades, however absence of facile methods to map modified sites have limited the understanding of their role. With the availability of next-generation sequencing, these drawbacks have been overcome, revealing non-random distribution of internal methylation across different transcript biotypes. Recently, we implemented a bisulfite sequencing-based technique for transcriptome-wide detection of m⁵C (bsRNA-seq) and mapped thousands of m⁵C sites in the human transcriptome including in ncRNA and mRNA. Biased distribution of m⁵C within mRNAs was evident, consistent with roles in post-transcriptional gene regulation. Two RNA methyltransferases are known to be active in humans, NSUN2, which is overexpressed in various tumours, and TRDMT1. The aims of this thesis were to determine the RNA targets of NSUN2 and TRDMT1, explore the molecular functions of m⁵C in mRNA and compare RNA m⁵C profiles of normal prostate cells (PrEC), prostate cancer cells (LNCaP) and cervical cancer cells (HeLa) using improved high-throughput methods. RNAi-mediated knockdown of NSUN2 and TRDMT1 in HeLa cells coupled with bsRNA-seq revealed 4,241 candidate m⁵C sites across tRNA, ncRNA and mRNA; 393 of these sites were called as NSUN2 targets and 120 as TRDMT1 targets. NSUN2 mediated the bulk of tRNA methylation, targeting several structural positions and a broad range of isotypes and isoacceptors, while TRDMT1 was specific to one structural position across four isoacceptors, consistent with previously published work. Interestingly, m⁵C was evident in tRNAs prior to end processing and splicing, indicating that methylation occurs early during tRNA biogenesis. Five tRNAs exhibiting m⁵C sites were independently validated by locus-specific bsRNAsequencing. Importantly, five m⁵C sites in ncRNAs and nine in mRNAs were independently validated as NSUN2 targets with NSUN2 exhibiting multisite specificity, indicating its importance as a ncRNA and mRNA modifying enzyme. Enrichment of Ago2 footprints upstream of m⁵C sites and depletion of miRNA target regions at m⁵C sites was evident in the CDS and 3’ UTR, suggestive of regulatory interactions between m⁵C and RISC. A subtle decrease in average steady-state levels of NSUN2-targeted mRNAs was evident in the NSUN2 knockdown sample, suggesting a role for m⁵C in mRNA stability. Investigation of the correlation of m⁵C with translation state of individual mRNAs revealed two examples with opposing trends, suggesting context-dependent effects of m⁵C on translation. Comparison of the RNA m⁵C profiles of PrEC and LNCaP cells revealed 5,653 candidate m⁵C sites in common across tRNA, ncRNA and mRNA, with few discernible differential m⁵C sites. Inspection of differential sites suggested that many might have resulted from technical imperfections of bsRNA-seq, which was independently confirmed. A comparison of prostate cells with HeLa cells showed high concordance of m⁵C patterns across all three cell lines, with independent examples validated, indicating that m⁵C is conserved across multiple cellular contexts. The discovery that m⁵C is abundant in mRNA across different cellular contexts and correlative links between the presence of m⁵C in mRNA and several aspects of posttranscriptional gene regulation highlight a functional importance for m⁵C in mRNA. Show more