5-methylcytosine mediates nuclear export of mRNA

5-methylcytosine was shown before to be an epitranscriptomic mark. Yang et al. now explored the unique topology of this mRNA modification, identified its writer and demonstrated its involvement in nuclear-cytoplasmic shuttling mediated by a specific reader.

More than one hundred modified nucleotides, in addition to the four canonical A, C, G and U letters, are found in RNA molecules, mainly in highly abundant RNA species such as rRNA and tRNA. The last decade witnessed the mapping and characterization of a growing number of mRNA-decorating modified nucleotides collectively constituting the nascent field known as epitranscriptomics¹. Starting with the deciphering of inosine (I) and N⁶-methyladenosine (m⁶A) and followed by the identification of 5-methylcytosine (m⁵C), 5-hydroxymethylcytosine (hm⁵C), pseudouridine (ψ), N¹-methyadenosine (m¹A), 2′-O-methylnucleotides (Nm) and N⁶-2′-O-dimethyladenosine (m⁶Am), the epitranscriptome alphabet is rapidly expanding¹.

m⁵C is the time-honored prototypic DNA modification, a major player in gene expression control and epigenetic regulation. It turned out that RNA is also decorated by m⁵C. rRNA marking by m⁵C was demonstrated in bacteria, archaea and eukaryotes whereas tRNA marking was found only in the latter two². Archaeal and viral mRNA m⁵C decoration was clearly demonstrated, and some indications for eukaryotic mRNA marking also surfaced². In recent years, based on the advance of bisulfite deep RNA sequencing and additional high throughput methodologies, more than 10 000 m⁵C sites were identified in the human transcriptome, and were reported to be enriched in the untranslated regions (UTRs) of mRNA transcripts³. A prominently augmented m⁵C marking was reported in the vicinity of Argonaute protein binding sites, suggesting a role in gene expression regulation. m⁵C decoration of mRNA was found to be dynamic, with some indications for a role in protein translation. Yet, much remained to be studied regarding cellular machineries involved in regulation of m⁵C RNA decoration and in mediation of m⁵C-regulated activities.

The extent of information available regarding the biological significance of the various epitranscriptomic marks differs widely. Some criteria enable us to evaluate, based on current knowledge, the relevance of each type of modification to cellular and organismal functions in the context of RNA epigenetics. These criteria include characteristic topology, evolutionary conservation, dynamic and reversible marking, identification of specific writers, readers and erasers, and documented links to biochemical and cellular outcomes. m⁶A marking currently provides the gold standard for a key player in the epitranscriptomic network. This modification is non-randomly distributed along mRNA landmarks, with enhanced decoration of stop codon and the proximal 3′ UTR regions as well as of long exons⁴. m⁶A is dynamic in response to environmental signals and is highly conserved⁴. Both writers and erasers of m⁶A are known, further alluding to the dynamic nature of this new type of epigenetic regulation^1,4. A major key to deciphering cellular functions mediated by m⁶A was the identification of specific readers, members of the YTH and the HNRNP protein families, that are recruited to the modified nucleotide embedded in a typical motif and mediate its downstream activities⁴. Much like DNA methyl-CpG-binding proteins at the time, it was the discovery of m⁶A-binding proteins that proved instrumental in uncovering its functions and mechanisms of action. After revealing how cells read m⁶A, the first molecular mechanisms soon followed. m⁶A recognized by specific readers was shown to be involved in the regulation of RNA splicing, mRNA recruitment to P bodies and degradation, translation, 3′ UTR processing, microRNA biogenesis and activity, and X-chromosome inactivation¹. m⁶A was shown to be essential for cell fate decisions in early stages of embryogenesis⁵, was linked to circadian control⁶, and was found to be relevant to diseases such as cancer, neurodegeneration, infertility and obesity¹.

The extensive study by Yang et al. in May issue of Cell Research significantly upgrades the standing of m⁵C in the epitranscriptome field⁷. The detailed unique m⁵C topology, as explored in this study, differs from that reported before. Enhanced m⁵C deposition was found not only in the neighborhood of Argonaute protein binding sites, but also in regions located immediately after translation initiation sites. No specific conserved methylation-directing motif was identified, yet the m⁵C-decorated sequences are characterized by high GC content. The location of m⁵C in GC-rich sequences in the vicinity of translation start sites resembles the recently unraveled topology of m¹A⁸, suggesting that m⁵C marking, similar to m¹A marking, may affect translation efficiency. Another major finding of the study is that NSUN2 is the main RNA methyltransferase that mediates m⁵C installation on mRNA. The discovery of this m⁵C writer opens the way to experiments exploring the role of this modification by enhancing or silencing its activity. An additional important finding is the identification of the mRNA export adaptor ALYREF as a bona fide mRNA m⁵C reader. The authors convincingly show that ALYREF-mediated nuclear-cytoplasmic shuttling of mRNA is dependent on m⁵C marking by NSUN2 and precisely map K171 as the lysine residue that is essential for recruitment of the methylated transcripts. There are some indications that m⁶A also affects mRNA nuclear-cytoplasmic shuttling⁶, however no clear mechanistic insight is available that connects a specific endogenous cellular component of the export machinery in this case. Interestingly, the HIV-related Rev protein was shown to mediate viral RNA nuclear export, which is needed for viral replication, in an m⁶A-dependent manner⁹. The clear implication of ALYREF in m⁵C-regulated nuclear export provides an excellent mechanistic illustration for modification-dependent mRNA localization mediated by an endogenous export protein (Figure 1). Having in mind the location of both m⁵C and m¹A in GC-rich sequences in the vicinity of translation start sites, it is tempting to explore the possibility that m¹A decoration may also affect transcript export.

Figure 1

Epitranscriptomic marks regulate nuclear export. (A) Decoration of the HIV RNA REV Response Element (RRE) by m⁶A recruits the viral REV reader protein, which accelerates nuclear export of viral transcripts⁹. m⁶A writing is mediated by METTL3/14 and its erasure is mediated by ALKBH5. (B) Decoration with m⁵C of GC-rich regions of mRNA molecules by the NSUN2 writer enables recruitment of the ALYREF m⁵C reader, which mediates nuclear export.

The precise localization of m⁵C marks to a unique region brings to mind the preferred targeting of m¹A and m⁶A around the start and the end of coding regions, respectively. Recent studies indicate that m¹A is located just upstream of the first exonic junction⁸ whereas m⁶A is found immediately downstream of the last exonic junction¹⁰, suggesting that components of the exonic junction complex (EJC) are involved in the exact positioning of these modifications. It can be speculated that the EJC is also involved in the precise deposition of m⁵C in the vicinity of the translation initiation site. The assembly of nuclear export-competent messenger ribonucleoprotein complexes comprises the recruitment of export-facilitating factors to the mature mRNA. It was shown lately that a core set of export proteins containing UAP56, DDX and importantly, ALYREF, nucleate and associate with spliced transcripts in an EJC- and cap-dependent manner¹¹. Further studies are needed to explore the hypothesis that EJC and ALYREF play a major role in both m⁵C deposition and nuclear-cytoplasmic shuttling. Additional immediate questions that are raised by this study include the role of m⁵C in gene expression control in general and translation in particular, a possible interaction or overlap with other modified nucleotides such as m¹A and m⁶A in a kind of “epitranscriptomic code”, whether there is a specific eraser that can demethylate m⁵C (TET protein?) and what the consequences of removal of this mark are.

The data obtained by Yang et al. and previous researchers concerning the unique telltale topology of m⁵C, its evolutionary conservation, dynamic nature and tissue specificity as well as the identification of specific writer and reader proteins and the unequivocal mechanistic role in nuclear export, now turn m⁵C into a very respected citizen of the epitranscriptome world, similar to the well-established position of m⁶A. The findings of this paper will enable to study the role of m⁵C in physiological and pathological states. Similar to the translation of DNA and histone epigenetics into established therapies, the clarification of the players and mechanisms involved in m⁵C decoration and reading may open the way for novel future therapeutic interventions.