A molecular modification called m6Am has been found to regulate the stability of messenger RNAs in mammalian cells. The mechanism casts fresh light on how reversibly modified RNA bases control the fate of mRNA. See Article p.371
In the cells of eukaryotic organisms, which include plants, animals and fungi, the 5′ end of every messenger RNA is modified to carry a cap. This modification incorporates a methyl group and acts as a key regulator of mRNA translation and decay. Additional methyl modifications can also be found next to the 5′ cap, including one known as m6Am. On page 371, Mauer et al.1 identify a role for m6Am in controlling mRNA stability. They also reveal that this modification is the preferred substrate of the methyl-removing enzyme known as fat mass and obesity-associated protein (FTO), instead of a related modification known as m6A that occurs in the body of mRNAs. A fascinating implication of these findings is that the diverse effects previously associated with FTO gene variants might be caused by dysregulation of m6Am — rather than m6A — homeostasis.
The 5′ cap of eukaryotic mRNAs is called m7Gppp, and consists of a methylated guanosine molecule connected to mRNA by an unusual triphosphate linkage (Fig. 1). In many eukaryotes, the first nucleotide of an mRNA is also methylated on its sugar and, when that nucleotide is adenosine monophosphate, it can be further methylated at the nitrogen-6 position of its base (adenine) to form N6,2′-O-dimethyladenosine (m6Am) monophosphate.
N6-methylation of adenine was first identified as a common mRNA modification in the 1970s2,3,4,5. Chemical analyses revealed that the methylated base could be found next to the cap as m6Am or in the mRNA body as N6-methyladenosine (m6A). These modifications received little attention until the discovery6 that m6A marks could be removed by FTO. The reversibility of m6A suggested that this methyl mark has a role in the dynamic regulation of the transcriptome (the complete set of RNA transcripts in the cell), leading to the idea of an 'epitranscriptome code' comprised of reversibly modified bases that expand the functionality of mRNA. Roles for m6A have now been described for essentially every step of the mRNA life cycle, including splicing, translation and decay7. But most studies of N6-methylation of adenine have focused on internal m6A modifications, leaving the role of cap-proximal m6Am a mystery.
Motivated by the relatively subtle effects8 that FTO disruption has on the transcriptome-wide distribution of m6A, Mauer et al. hypothesized that m6Am rather than m6A might be FTO's preferred substrate. Remarkably, the authors' detailed biochemical analyses revealed that the catalytic efficiency of FTO was more than 100-fold greater for m6Am than for m6A. The preference of FTO for m6Am was due not only to the presence of the methylated sugar, but also to the modification's proximity to the m7Gppp cap. The authors' in vivo studies of the impact of FTO disruption and overexpression satisfyingly confirmed that m6Am is the physiological substrate of FTO in both mouse and human cells.
What are the functional consequences of the m6Am modification? Because removal of the m7Gppp cap, by the decapping enzyme DCP2, is a key step in mRNA decay, Mauer et al. focused on how the cap-adjacent m6Am mark affects RNA stability. They found that the mark was associated with a long mRNA half-life, and was sufficient to stabilize a synthetic mRNA substantially. Moreover, the half-lives of naturally occurring m6Am-containing mRNAs could be modulated by altering the cytoplasmic levels of FTO, raising the possibility that dynamic regulation of the transcriptome could be achieved in a similar manner under certain physiological conditions. The authors further showed that the increased stability provided by m6Am, compared with analogous mRNAs in which the cap-adjacent adenosine monophosphate is methylated only on the sugar, can be explained, at least in part, by the cap's reduced susceptibility to removal by DCP2. Thus, m6Am can be added to the growing list of factors that contribute to the wide range of decay rates observed for cellular mRNAs.
Mauer and colleagues' in vitro and in vivo analyses clearly pinpoint DCP2 as a 'reader' of m6Am and FTO as an 'eraser' of the methyl mark. However, there may be other readers and erasers that act on m6Am. The cap-binding protein eIF4E, which is involved in translation initiation, is a candidate reader that could explain the enhancing effect of m6Am on translational efficiency reported by the authors. Another possible reader is the nuclear cap-binding complex CBC20/CBC80, which is crucial for splicing — a process in which FTO has been implicated9 and in which pre-mRNA molecules are converted into mature mRNAs. Members of the YTH family of RNA-binding proteins recognize m6A (ref. 7) and might also read m6Am. Moreover, the identity of the crucial 'writer' of m6Am remains unknown — a fact that prevented the authors from eliminating m6Am in vivo altogether and determining the consequences. Discovering the N6-methyltransferase enzyme that installs m6Am marks is an important next step for the field.
Another key area for future work will be determining the molecular basis of the selectivity of FTO for, and of DCP2 against, m7Gpppm6Am-capped mRNAs. The structure of FTO has so far been characterized only in complex with DNA that contains the methylated base 3-methylthymine10. The knowledge that m7Gpppm6Am is the preferred substrate of FTO will facilitate structural analyses of FTO in complex with RNA, and potentially reveal the structural basis of the enzyme's selectivity.
In contrast to FTO, DCP2 has been the subject of many structural studies11. On the basis of these structures, it seems possible that, when m6A marks the first nucleotide of an mRNA, the bulk of the methyl group clashes with DCP2. Crucially, however, Mauer et al. find that m6Am-modified mRNA can still be decapped in vitro, albeit less efficiently than mRNA that lacks N6-methylation. This observation suggests that m7Gpppm6Am-capped mRNAs can be decapped directly in vivo without first being demethylated by FTO. It should also be noted that the active decapping complex contains other proteins in addition to DCP2, and might therefore have a different activity towards m7Gpppm6Am from the isolated DCP2 used by the authors.
A central question raised by this work is when and how m6Am marks are modulated in vivo. Like m6A, global m6Am levels will be influenced by the abundances and subcellular localizations of N6-methyltransferases and demethylase enzymes — any of which could rapidly change in response to environmental cues. Furthermore, the strict requirement for transcription to be initiated by adenosine triphosphate to form m6Am raises the intriguing possibility that transcript-specific methylation could be regulated by the process in which transcription start sites (TSSs) in DNA are selected — even a single nucleotide shift in a TSS that changes the 5′ end to or from adenosine triphosphate will affect whether an mRNA can contain m6Am.
Mauer and colleagues' discovery also means that previous studies of N6-methylation of adenine that involve FTO should be re-examined through the lens of cap-adjacent m6Am rather than through that of internal m6A sites. For example, FTO is associated with obesity12,13 and neurological defects8; dysregulation of m7Gpppm6Am-capped mRNAs now emerges as a possible underlying cause. Finally, the fact that removal of the N6-methyl group from m6Am by FTO proceeds through an intermediate suggests that m7Gpppm6Am might be just one of several forms of the eukaryotic mRNA cap, each of which could have distinct activities and regulation.