Epi-epigenetics: RNA methylation (updated)

Two papers appeared online on February 10th, to claim the first whole-transcriptome study of specific RNA modification, N1-methylation of adenine (m1A). To both teams’ credits, they cited each other as they learned about “competing” study.  Both papers, Li, Xiong et al. in Nature Chemical Biology and Dominissini, Nachtergaele, Moshitch-Moshkovitz et al. in Nature, overlap quite significantly but also complete each other in several aspects, and give starting insights into the role of RNA methylation in gene regulation.

The N1 modification of adenine is nothing new and has been known for DNA and some non-coding RNAs (tRNA and rRNA) for quite a while. Also, one of the standard methods for mapping RNA secondary structure is actually based on N1-methylation of adenine by dimethyl sulfate. However, nobody so far looked into distribution of endogenous m1A residues in mRNA transcripts. One of the technical difficulties for such a study is that m1A can spontaneously isomerize into m6A under alkaline conditions used for RNA isolation (Dimroth rearrangement). Both teams, however, used the rearrangement for their advantage, to validate the results of immunoprecipitation selective for m1A residue and to precisely locate N1-methylated adenines.

Dimroth rearrangement of N1-methyl adenine

Since both studies overlap quite significantly, it’s really interesting to see where their findings agree and where they don’t.

  Nature Nature Chem Bio
Ratio m1A/A 0.015−0.054% 0.02%
Occurrence in genes 4214 gene transcripts 887 transcripts from 600 genes
Studied cell lines Yeasts: BY4741, Sp1

Mouse: MEF, mESC

Human: HeLa, HepG2, HEK293


Human: HEK293T, U2OS, WPMY-1, DU145, LNCap, A549
Localization  Nat_m1A_distrib  NCB_m1A_distrib
Sequence context of m1A  Nat_consens  NCB_consens
Functional context AUG start codon, 1st splice site, alternative transcription initiation sites AUG start codon
Metabolic regulation of m1A L-Methionine-dependent methylation; ALKBH3-dependent demethylation ALKBH3-dependent demethylation
Influence on translation Methylation increases translation Not studied

As can be seen from the table above, both studies give qualitatively similar results. The quantitative discrepancies could stem from different methodologies and criteria for identification of a ‘signal’. This resulted in a different count of methylated gene transcripts, which could bias either study to unknown extent.

In general, N1-methylated adenines cluster in the 5′ untranslated region (UTR) of mRNA, next to start codon (or more precisely to the 1st splice site). These regions are also more GC-rich and more structured. All this leads to the elevated translation for methylated transcripts compared to non-methylated when normalized for mRNA abundance. As these properties of RNA are quite heavily cross-correlated, the authors of Nature paper performed ANCOVA analysis of factors underlying protein expression. They identified that m1A adds its own share for the increase, suggesting a mechanistic link. Principal component analysis could add some quantitative insight for the contribution, though. It is quite remarkable that for the most methylated transcripts a single m1A residue was enough to cause the observed differences. Looks like a sort of ‘magic methyl effect‘ equivalent in molecular biology.

Demethylation of m1A by DNA-repairing ALKBH3 and altered methylation pattern under number of tested stress conditions support functional significance of this RNA modification. Stark contrast to the distribution of isomeric m6A residues, which cluster in 3’UTR of mRNAs, suggests that N1-methylation has a different mechanism. So the next quest will be to find the enzyme involved in m1A formation.

Correlation of GC-content and minimum free energy density (MFEden) for methylated and non-methylated genes. Also a nice illustration why p-values suck at very big sample sizes. See the papers for better illustration of the point.

All in all these result nicely show that post-transcriptional chemical modification of mRNA itself can alter its translation. This adds one more lever to already complex epigenetic machinery comprised of transcription factors, histone modifications, and DNA methylation. As well as a bunch of other modificaions of RNA, collectively named ‘epitranscriptomics‘.

Normally it’s impossible to estimate the errorbars on the reported values in a whole-genome study because there’s nothing to compare them to.  So simultaneous publication of these two papers on the same subject gives a unique illustration on the fuzziness of such kind of analysis. Therefore, the results will probably need a thorough validation in case of each particular mRNA if one decides to exploit this methylation pathway.

Update. Added some links to previous epitranscriptomics discoveries.


Author: Slava Bernat

I did my PhD in medicinal chemistry/chemical biology of G protein-coupled receptors and then explored some chemical biology of non-coding RNA as a postdoc. Currently I'm working in a small biotech company in San-Francisco Bay area as a research chemist. I'm writing about science, which catches my attention in rss feed reader and some random thoughts or tutorials.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s