m¹A within cytoplasmic mRNAs at single nucleotide resolution: a reconciled transcriptome-wide map

Abstract

Following synthesis, RNA can be modified with over 100 chemically distinct modifications. Recently, two studies—one by our group—developed conceptually similar approaches to globally map N1-methyladenosine (m¹A) at single nucleotide resolution. Surprisingly, the studies diverged quite substantially in their estimates of the abundance, whereabouts, and stoichiometry of m¹A within internal sites in cytosolic mRNAs: One study reported it to be a very rare modification, present at very low stoichiometries, and invariably catalyzed by TRMT6/61A. The other found it to be present at >470 sites, often at high levels, and suggested that the vast majority were highly unlikely to be TRMT6/61A substrates. Here we reanalyze the data from the latter study, and demonstrate that the vast majority of the detected sites originate from duplications, misannotations, mismapping, SNPs, sequencing errors, and a set of sites from the very first transcribed base that appear to originate from nontemplated incorporations by reverse transcriptase. Only 53 of the sites detected in the latter study likely reflect bona-fide internal modifications of cytoplasmically encoded mRNA molecules, nearly all of which are likely TRMT6/TRMT61A substrates and typically modified at low to undetectable levels. The experimental data sets from both studies thus consistently demonstrate that within cytosolic mRNAs, m¹A is a rare internal modification where it is typically catalyzed at very low stoichiometries via a single complex. Our findings offer a clear and consistent view on the abundance and whereabouts of m¹A, and lay out directions for future studies.