5-methylcytosine (m5C) is an abundant RNA modification decorating diverse species of RNA. The re-discovery of m5C as an mRNA modification came to light by coupling bisulfite conversion of RNA with transcriptome-wide RNA sequencing (bisRNA-seq). Subsequent studies revealed prominence of deposition of m5C modification in regulating gene expression.
Presently, bisRNA-seq experiments carried out in different cell types have generated several lists for m5C-decorated mRNAs that are often different from each other. This disparity could be partially explained by the variation in the expression levels of the very methyltransferase enzymes (NSUN family of proteins) responsible for the deposition of m5C modification (‘m5C writers’).
In an attempt to make this coherent, we generated a comprehensive, transcriptome-wide human dataset of m5C sites by re-analysing existing bisRNA-seq datasets from 10 different human cell lines and 7 tissues, with a stringent pipeline for m5C site calling. This allowed us to compile a union list of 8,325 high-confidence m5C sites. This union list was then employed to investigate the relationship between m5C modification and RNA-binding proteins (RBPs) binding to mRNA. Therefore, we overlapped the exonic m5C sites with publicly available eCLIP data for >100 RBPs, identifying several proteins whose footprint significantly overlaps with m5C modification. We further generated a knockout HeLa cell line for NSUN2, the main m5C writer protein, to investigate the effect of mRNA methylation level and its influence on DDX3X and UPF1 RBPs function.
CLIP experiments revealed a reduced binding capacity of DDX3X for m5C-decorated mRNAs in the absence of NSUN2 enzyme. Furthermore, knockdown experiments in WT vs NSUN2 knockout HeLa cells showed that the lack of NSUN2 enzyme reduces the effect of UPF1 in regulating the steady state of a subgroup of its methylated mRNA targets. Overall, our data suggests a novel role for DDX3X and UPF1 as m5C reader proteins.