5-Formylcytosine is not a prevalent RNA modification in mammalian cells

Abstract

The RNA modification 5-formylcytidine (f5C) is poorly explored in mammals. Low f5C levels reported in mRNA may reflect spurious 5-methylcytidine (m5C) oxidation or targeted demethylation by TET or ALKBH1 dioxygenases. We analyzed f5C in RNA of mouse embryonic stem cells (mESCs) using LC-MS/MS and chemical-assisted sequencing. We reveal that the previously reported pyridine-borane-sequencing misidentifies N4-acetylcytidine (ac4C) and unmodified, hyper-reactive cytidines in a CUMC context as f5C. To overcome these limitations, we developed FIBo-seq with enhanced specificity and sensitivity for f5C-sequencing. We find no evidence for a role of TET enzymes in generating f5C, unlike for ALKBH1. Moreover, no f5C sites are detectable in mRNA. Instead, the bulk of mammalian f5C resides in the well-established mitochondrial tRNA Methionine (mt-tRNAMet) and is mediated by ALKBH1. The results argue against an instructive function for f5C outside tRNA in mammals.