PerspectiveMolecular Biology

Internal mRNA Methylation Finally Finds Functions

See allHide authors and affiliations

Science  14 Mar 2014:
Vol. 343, Issue 6176, pp. 1207-1208
DOI: 10.1126/science.1249340

Modification of internal adenosines in messenger RNA (mRNA) by methylation of the N6 position (m6A) was first observed almost four decades ago. Early work demonstrated that the m6A modification was quite common, occurring at an estimated frequency of three to five residues per mRNA. Nevertheless, the function, if any, of this modification remained enigmatic. The cloning in 1997 of a subunit of the RNA methylase complex (1) was followed by a period of fitful inactivity before the field reawakened in 2011 with the discovery of an enzyme, FTO (fat mass and obesity-associated protein), that was shown to catalyze the demethylation of internal m6A residues (2). The existence of a demethylase and the demonstration that m6A levels were raised when the enzyme was knocked down strongly suggested that at least some m6A modifications were reversible and might be subject to dynamic regulation. Since then, a series of papers have appeared in rapid succession, together providing a wealth of unequivocal evidence for m6A function. But these findings still have not led to a coherent picture of the number and variety of functions of the m6A modification.

Two recent RNA “methylome” studies (3, 4) provide a transcriptome-wide map of m6A-modified mRNAs. Both report the presence of m6A in about 7000 mRNAs, indicating that the modification is fairly promiscuous with regard to mRNA targets. However, methylation is highly selective when considering which specific sites are modified. Although both studies showed that methylation sites are largely confined to the consensus sequence Pu[G>A]m6AC[A/C/U], only about 10% of sites conforming to this consensus are actually modified. Neither study was able to determine the stoichiometry of any specific modified site, nor could clusters of modified sites be identified. Both studies showed enrichment of modification sites near stop codons; however, one study saw enrichment in long internal exons while the other saw enrichment in 3′ untranslated regions. Additionally, one study found elevated levels of m6A in the brain and the other study did not. The reasons for these discrepancies are not obvious. Nevertheless, both studies found that modification sites were well conserved between human and mouse transcriptomes—a finding strongly suggestive of biological function. Indeed, knockdown of the m6A methylase (3) resulted in large-scale alterations in splicing patterns, consistent with a striking enrichment of m6A modifications within alternatively spliced exons (see the figure), with constitutive exons being correspondingly underenriched (3). How exactly the depletion of m6A could alter splicing patterns is currently unknown.

Methylating RNA regulates its function.

A schematic diagram of the RNA m6A methylation cycle. METTL3, methyltransferase-like 3.

A provocative study recently reported that inhibition of methylation by either a drug or knockdown of the m6A methylase perturbed the circadian clock (5). When methylation was reduced by either treatment, the period of the clock was lengthened. By contrast, overexpression of the methylase caused a shorter period. Reduction of internal methylation resulted in a pronounced delay in export of mRNAs from the nucleus and, consequently, marked nuclear retention of those mRNAs. A second, perhaps related, phenomenon linked to inhibition of methylation was the widespread stabilization of mRNAs, in particular mRNAs encoding proteins known to be involved in the circadian clock. This observation may explain how the clock is perturbed when methylation is inhibited, but how specificity is achieved remains unknown. It is also unclear why overexpression of the methylase speeds up the clock; the methylation status of specific mRNAs was not examined.

Whereas these studies focused on the effects of inhibiting RNA methylation, another report examined the effect of a genetic null in an RNA demethylase. FTO, the first demethylase discovered, belongs to a family of enzymes that catalyze a wide range of oxidation reactions (2). Systematic study of these enzymes revealed that another family member, ALKBH5, was also a m6A demethylase (6). Both FTO and ALKBH5 are localized to nuclei, and both colocalize with nuclear speckles, sites of concentrations of splicing factors. It is tempting to speculate that this localization is tied to the effects of methylation on splicing. Indeed, more than 3000 altered mRNA isoforms were observed when ALKBH5 was knocked down in tissue culture cells, indicating that splicing was affected. Despite this, mice null for ALKBH5 were anatomically normal and grew to adulthood. However, male mice lacking the demethylase were sterile, with gene expression and splicing massively altered in their testes. The sterility phenotype was manifested as a metaphase arrest during spermatogenesis. The tissue specificity of the phenomenon might be explained by the fact that ALKBH5 is more highly expressed in testes relative to other tissues, but perhaps the most puzzling observation is that deletion of the demethylase resulted in only a very modest increase in overall levels of m6A in the testis.

The existence of a methylase and at least two demethylases coupled with phenotypes observed when either activity was altered strongly suggested that one or more factors recognize m6A. A first step in identifying such factors has come from the demonstration that an RNA binding protein, YTHDF2, specifically recognizes m6A-modified RNAs (3, 7). In vivo cross-linking and immunoprecipitation showed that YTHDF2-bound RNAs were highly enriched for mRNAs known to be methylated (7). Moreover, a variety of approaches, including ribosome profiling and immunofluorescence microscopy, provided evidence that the binding of YTHDF2 to mRNA resulted in relocalization of the mRNA from translating ribosomes to cytoplasmic foci (P bodies) known to be enriched in RNA degradative activities. As a consequence, targeted mRNAs were destabilized, and the magnitude of destabilization correlated with the number of methylation sites in the mRNA. Knockdown of YTHDF2 led to an increase in mRNA stability for targeted transcripts.

Although these studies have provided new insight into the distribution and functional importance of the m6A modification, several fundamental questions remain. Prime among these is how specificity of modification is achieved. Clearly, the sequence that constitutes the consensus site is not sufficient on its own, nor does secondary structure appear to play a role. If methylation is cotranscriptional, it may be possible that chromatin status could play a role in site selection. The function(s) of m6A in nuclear RNA metabolism are also unclear. Although it is possible that nuclear factors recognize the modification, it seems equally plausible that modifications could function by preventing or altering the binding of some proteins. The importance or function of modifications in the vicinity of stop codons remains to be established, as does the importance of the FTO demethylase. Perhaps the most challenging question—and most difficult to answer—is how such a widespread modification has apparently quite specific effects. Are all modified sites equally important, or is only a small subset of them important? Hopefully, answers to at least some of these questions will emerge soon in this now fast-moving field.


Navigate This Article