-
PDF
- Split View
-
Views
-
Cite
Cite
Koichiro Kako, Jun-Dal Kim, Akiyoshi Fukamizu, Emerging impacts of biological methylation on genetic information, The Journal of Biochemistry, Volume 165, Issue 1, January 2019, Pages 9–18, https://doi.org/10.1093/jb/mvy075
Close -
Share
Abstract
The central dogma of molecular biology explains the fundamental flow of genetic information for life. Although genome sequence (DNA) itself is a static chemical signature, it includes multiple layers of information composed of mRNA, tRNA, rRNA and small RNAs, all of which are involved in protein synthesis and is passing from parents to offspring via DNA. Methylation is a biologically important modification, because DNA, RNAs and proteins, components of the central dogma, are methylated by a set of methyltransferases. Recent works focused on understanding a variety of biological methylation have shed light on new regulation of cellular functions. In this review, we briefly discuss some of those recent findings of methylation, including DNA, RNAs and proteins.
Seventy years ago, the term ‘biological methylation’ was first used by Bach (1) and Challenger (2), who were pioneers of a rising research interest in methylation reactions. Multiple lines of evidence have demonstrated that methylation of biopolymers with large molecular weights associated in the central dogma, DNA, RNAs and proteins (in particular histones), plays a key role in genetic and epigenetic regulations. In addition, recent works revealed the connection between nutrients and cellular functions, indicating that metabolism influences homeostasis. In this article, we will review the importance of dynamic biological methylation.
Methylation of biomolecules is mediated by catalytic enzymes, so called as methyltransferases, which use S-adenosyl-L-methionine (SAM) as a methyl donor, synthesized from L-methionine and ATP. In almost organism, an estimated 1% of genes in the genome encodes methyltransferases (3) that act on DNA, mRNA, noncoding RNA (ncRNA) including rRNA, tRNA, miRNA and long noncoding RNA (lncRNA) and snRNA, and proteins (Fig. 1). Methyltransferase-mediated reactions are substrate specific, and different enzymes exhibit different intracellular localization patterns. For example, methyltransferases that methylate histones and other nuclear proteins are localized to the nucleus, whereas those that act on transmembrane signal transduction proteins are typically abundant in the cytoplasm. In addition, some methyltransferases shuttle between the nucleus and cytoplasm (4, 5). These differences in substrate specificity and intracellular localization allow methylation to occur in a wide variety of biological processes.
Biological methylations of various organisms. Methylations have various impacts on each step of the genetic information flow (DNA, RNA and protein) in almost all living organisms.
Biological methylations of various organisms. Methylations have various impacts on each step of the genetic information flow (DNA, RNA and protein) in almost all living organisms.
DNA Methylation
DNA methylation has been extensively studied in many eukaryotes including fungi, plants, invertebrates and vertebrates (Fig. 2). In particular, 5-methylcytosine (5mC) in cytosine-guanine (CpG) dinucleotides is an important modification of DNA in mammals, because it is widely distributed in the genome, where 60–80% of CpG dinucleotides are speculated to be methylated (6). Alteration of methylation patterns of CpG dinucleotides plays a crucial role for X-chromosome inactivation (7), parent-of-origin-specific genomic imprinting (8), transcriptional silencing (9), epigenetic memory competence (10) and gene expression during differentiation in embryonic stem cells (11). Therefore, aberrant changes in DNA methylation levels and patterns are frequently perturbed in human pathology such as cancer (12) and metabolism-related (13), cardiovascular (14) and neurodegenerative diseases (15).
Profiles of the biological methylations on DNA. Although DNA is methylated at the 5-position of cytosine base in mammals, N6-position methylation of adenine base is observed in invertebrate, such as nematode.
Profiles of the biological methylations on DNA. Although DNA is methylated at the 5-position of cytosine base in mammals, N6-position methylation of adenine base is observed in invertebrate, such as nematode.
In a contrasting situation to mammals, in bacteria, N6-methylation of adenine (6 mA) is a dominant modification in comparison with 5mC found in prokaryotic genomes (16). Bacterial 6 mA was initially identified as a part of defense mechanisms, called as the restriction-modification system, to eliminate invader DNAs including phages and plasmids with no 6 mA. When invader DNAs lacking 6 mA enter bacterial cells, the hosts can distinguish and digest them by DNA methylation-sensitive restriction enzymes. Independently of the defense function, 6 mA catalyzed by DNA methytransferases plays a role as an epigenetic signal in certain groups of bacterial strains for DNA replication and repair, and transcription (17).
Until recently, very little progress had been made in understanding biological functions of 6 mA in eukaryotes, because of its extremely low contents in the genome. In the nematode Caenorhabditis elegans (C. elegans), an excellent model for aging, it has been known that the worm genome is thought to be not methylated at the CpG sites due to the lack of an enzyme that encodes a conserved DNA methyltranferase to catalyze methylation of the dinucleotides found in mammals (18, 19). In 2015, however, Greer et al. (20) found that adenine N6-methylation is an important epigenetic mark in C. elegans that is responsible for transgenerational progressive loss of fertility in spr-5 (suppressor of presenilin defect-5, a nematode homologue of histone demethylase LSD1) mutant. Interestingly, spr-5 mutant shows longevity and this phenotype partially depends on 6 mA DNA methylation (21). Fu et al. (22) and Zhang et al. (23) also discovered that 6 mA is an epigenetic mark for active transcription start sites in Chlamydomonas reinhardtii and for the regulation of transposon expression in Drosophila melanogaster, respectively. These three works have inspired further researches not only on potential 6 mA functions as epigenetic signs (24, 25) but also on other DNA methylation than 5mC and 6 mA in the future.
RNA Methylation
More than 100 chemical modifications have been identified in coding mRNA and ncRNA such as lncRNA and snRNA in the context of tRNA, rRNA and small RNAs, which are key elements of the translation machinery (Fig. 3). For example, snRNA is not only hypermethylated to trimethylguanosine at the 5′-cap site (26) but also is 2′-O- or base-methylated in the internal region (27). The chemistry of RNA methylation involves diverse sites on nucleosides, so called as 2′-O-methylation (ribose methylation) and base methylation, and related structural findings have been stored in the two comprehensive databases for RNA modifications, Modomics (a databases of RNA modification pathways: International Institute of Molecular and Cell Biology in Warsaw, Poland) (28) and The RNA Modification Database (The RNA Institute, University at Albany, State University of New York, USA) (29). A large number of RNA methyltransferases have been identified in unicellular organisms such as Escherichia coli and Saccharomyces cerevisiae. In multicellular organisms, however, much remains unknown regarding the mechanisms that trigger RNA methylation, and researches upon their biological significance have just begun.
Profiles of the biological methylations on RNA. In RNA, methyl groups are found not only in the base (both Watson–Crick face and non-Watson–Crick faces) but also at 2′-oxygen in the ribose.
Profiles of the biological methylations on RNA. In RNA, methyl groups are found not only in the base (both Watson–Crick face and non-Watson–Crick faces) but also at 2′-oxygen in the ribose.
rRNA methylation
Methylation of rRNAs is required to construct both large and small subunits structures. For example, as the rRNA methylation sites that are universally conserved among all living organisms, N6-dimethylations of A1850 and A1851 (Homo sapiens 18S rRNA numbering) of helix 45 in the small subunit contributes to the formation of ‘contact hub’ for the P-site of tRNA, mRNA and the ribosomal large subunit. Similarly, the 2′-O- methylations of G4196 and U4498 (H. sapiens 28S rRNA numbering) located in helix 80 (P-loop) and helix 92 (A-loop) of large subunit, respectively, play a role to maintain active conformation for base pairing in P- and A-sites of the tRNA necessary for aminoacyl-tRNA docking into the large subunit (30).
A recent study using C. elegans demonstrated that NSUN-5 (Nop2/Sun-like domain containing protein 5), a homolog of a yeast protein that methylates a specific position at 5-methylcytosine (m5C) in the conserved region of 25s rRNA, modulates worm lifespan (31). Notably, reduced levels of NSUN-5 are associated with elevated stress resistance. The exact molecular mechanism underlying the relationship between NSUN-5-mediated RNA methylation and longevity remains to be elucidated. Because rRNA is an important element of the translation machinery, methylation of this molecule is likely to have a major impact on translation efficiency.
Recently, we demonstrated that a nucleolar factor nucleomethylin (NML) catalyzes the N1-methyladenosine (m1A) modification in human 28S rRNA at position 1309 (A1309), and that this is indispensable to the 60S ribosomal subunit formation. When NML was knocked down, RPL11, a ribosomal protein of 60S subunit, was released from ribosome and it formed a complex with MDM2. The formation of RPL11–MDM2 complex suppressed proliferation and cell cycle arrest in NML-depleted cells (32). Moreover, we have clarified the m1A modification of 26S rRNA at position 674 (A674, corresponds to human A1309) in C. elegans, and this transmethylation is catalyzed by rRNA adenine methyltransferase-1, a nematode homologue of NML. These findings suggested that the m1A modification at A674 of 26S rRNA may be essential for the adequate assembly of the ribosomal large subunit in C. elegans (33).
mRNA methylation
Regarding the posttranscriptional regulation of mRNA molecules that possess 5′- and 3′-untranslated regions, many studies have been conducted with respect to N6-methyladenosine but comparatively less on methylation in 5′-untranslated sequences. Very recently, researchers found that heat shock induces selective methylation of the 5′ mRNA untranslated region via the METTL3–METTL14 complex and other methyltransferases. This work clarified the significance of methylation in the stress response (34).
On the other hand, N7-methylguanine (m7G) binds to the opposite direction across three phosphate groups at the 5′-cap structure of the mRNA. This base-methylation plays an important role not only for protecting the mature mRNA from the degradation via 5′-exonuclease but also for triggering the translation. eIF4E, a subunit of translation initiation complex eIF4F, interacts with the positively charged m7G in an aromatic cleft with this cationic base sandwiched between two aromatic residues, Trp56 and Trp102, and then, recruits m7G mRNA to eIF4F (35, 36). Furthermore, N7-methylation of guanosine is usually followed by 2′-O-methylation of the ribose at the first (Cap1, 5′-m7Gppp2′-O-Xm…), and the second (Cap2, 5′-m7Gppp2′-O-Xm2′-O-Ym…) transcribed nucleotides. Given the first nucleoside is adenosine, Am is methylated at N6-position by an unidentified RNA methyltransferase to form N6, 2′-O-dimethyladenosine (m6Am). The stability of m6Am-modified mRNA is increased via preventing their DCP2 (decapping protein 2)-mediated decapping, and this modification is demethylated (m6Am to m6A) by FTO (fat mass and obesity-associated protein) through the m7G modification-dependent manner (37).
Besides 5′-cap structure, it is revealed that m6A has been enriched around stop codon, 3′-noncoding region, and 3′-terminal coding region by transcriptome-wide profiling. In the nucleus, this modification may affect mRNA export, nuclear retention and splicing machinery, possibly through the interactions of m6A-binding proteins. Furthermore, m6A of cytoplasmic mRNA is recognized by YTHDF2 (YTH domain-containing family protein 2) and destabilized in the m6A-dependent manner. This dynamic retention of mRNA via methylation (writing) and its recognition (reading) is speculated to be involved in a variety of biological events such as development, stemness and circadian rhythm (38).
tRNA methylation
A number of chemical modifications have been described in tRNA giving it functional and biological diversities. More than 80% of RNA modifications including methylations are observed in tRNA (29). Some of environmental factors including foods affect these tRNA modifications. Queuine, a nutrient derived from microorganisms alters the translational fidelity through the tRNA anticodon modification from guanosine to queuosine (39). The modification of uridine of the mitochondrial tRNA with taurine, one of sulfur-containing ingredient of foods, is involved in mitochondrial translation through stabilizing codon-anticodon interaction (40). Although most of methylation reactions (base-methylations and the 2′-O-methylations) are catalyzed by Rossmann fold-type methyltransferases, some of them are driven by SPOUT (SpoU-TrmD) family methyltransferases (41).
Not only base methylation is known to protect from cytosolic nucleases (42) but also methylations of the base and of the ribose at 2′-oxygen are involved in the maintenance of cloverleaf structures of tRNA (43). Contrary to these advantages, undesirable methylations on the Watson–Crick faces in the double-strand parts lead the misfolding of tRNA and translational defects. ALKBH family demethylases drive functional recovery of RNA exposed to such methylation damages through the oxidative demethylation reactions (44). Furthermore, the methylation of the 34th and the 37th nucleotides to be located in the anticodon of the tRNA is essential to decoding of the mRNA in the ribosome. The 5-methoxycarboxymethyl modification of U34 at the wobble position participates in read-through of the stop codon (conversion into selenocysteine), and 1-methylguanine (m1G) modification at G37 suppresses frame-shift to begin in cytidine (45).
Protein Methylation
Methylation of basic amino acids
Although protein phosphorylation and acetylation are anhydrous condensing reactions between amino acid side chain and phosphate/acetyl group, protein methylation is a substitution reaction between proton on functional group and methyl group. Methyl group is chemically inert and is resistant to acid hydrolysis, whereas other covalent modification groups, phosphorylation and acetylation, are easily eliminated through the hydrolytic process (Fig. 4) (46). This chemical feature of protein methylation is beneficial for the quantification of methylated amino acids through the acid hydrolysis followed by the amino acid composition analysis (Supplementary Figs S1 and S2). Protein methylations usually occur on nitrogen atom of basic amino acid side chains, except in the case of α-amino group methylation of N-terminus (Fig. 5). This reaction requires the nucleophilic attack of activated nitrogen toward SAM to deprive the methyl group from the sulfonium part (SN2 reaction) (47).
Representative mass chromatograms of hydrolysates of modified basic amino acids. (A) Nɛ-acetyllysine (Ac-Lys), (B) Nɛ-monomethyllysine (MML) and (C) Nω-monomethylarginine (MMA). Each modified amino acid (100 pmol) is hydrolyzed with 6 N HCl at 110°C for 24 h in glass vials. One tenth of hydrolysates was injected into the LC-MS/MS apparatus. Blue and red arrows indicate elution positions of modified and authentic amino acids, respectively. Elution times of Ac-lys, L-Lys, MML, L-Arg, and MMA are revealed as 8.02, 10.69, 10.51, 10.70 and 10.36 min, respectively.
Representative mass chromatograms of hydrolysates of modified basic amino acids. (A) Nɛ-acetyllysine (Ac-Lys), (B) Nɛ-monomethyllysine (MML) and (C) Nω-monomethylarginine (MMA). Each modified amino acid (100 pmol) is hydrolyzed with 6 N HCl at 110°C for 24 h in glass vials. One tenth of hydrolysates was injected into the LC-MS/MS apparatus. Blue and red arrows indicate elution positions of modified and authentic amino acids, respectively. Elution times of Ac-lys, L-Lys, MML, L-Arg, and MMA are revealed as 8.02, 10.69, 10.51, 10.70 and 10.36 min, respectively.
Profiles of the biological methylations on protein. Protein methylation is mainly taken place on nitrogen atoms of side chains of lysine or arginine residues.
Profiles of the biological methylations on protein. Protein methylation is mainly taken place on nitrogen atoms of side chains of lysine or arginine residues.
Lysine methylation
In the case of lysine methylation, ɛ-amino group is able to receive the progressive mono-, di- or trimethylations. Unlike the acetylation, the positive charge of ɛ-amino group under the physiological pH is still maintained by mono- or dimethylation. Furthermore, trimethylation of ɛ-amino group converts this group into a potent cation (quaternary ammonium ion) independent of the solvent pH (46). The lysine methylation is catalyzed not only by Rossmann fold-type methyltransferases but also by SET (Su(var)3–9 enhancer-of-zeste and trithorax) family methyltransferases whose secondary structure is quite different from Rossmann fold-type methyltransferases (48). The biological significance of lysine methylation of histones had been extensively investigated because this posttranslational modification is indispensable for epigenetic regulation of transcription and chromatin in eukaryotes (49).
On the other hand, lysine methylations also take place on various non-histone proteins, including transcription- and chromatin-regulating proteins. Lysine methylation are involved in the regulation of protein–protein interaction (through the recognition by the methyllysine binder, such as chromodomains and Tudor domains), subcellular localization (e.g. nuclear localization of lysine-methylated Hsp70), competitive inhibition against polyubiquitination to stabilizing substrate (e.g. protection of lysine-methylated p53 from MDM2 ubiquitin ligase) and promoter binding affinity of transcription factors to alter the transcription levels of target genes (e.g. NF-κB activation by lysine methylation). For more details on the lysine methylations of histone and/or non-histone proteins, refer to the other excellent reviews (50, 51).
Arginine methylation
Arginine methylation can be taken place on nitrogen atoms of the guanidium group of this residue, and this reaction is driven by protein arginine methyltransferases (PRMTs) in vivo. Methylation of one of the two ω-positioned nitrogen atoms of the guanidium group (product: Nω-monomethylarginine (MMA)), addition of a second methyl group to the same ω-nitrogen of MMA (product: Nω, Nω-dimethylarginine, or asymmetric dimethylarginine (ADMA)), and the second methylation on the other ω-nitrogen of MMA (product: Nω, N′ω-dimethylarginine, or symmetric dimethylarginine (SDMA)), are catalyzed by Type I, Type II and Type III PRMTs, Type I and Type II PRMTs, respectively. At this moment, nine members of PRMT family are known in mammals and they have mutually structural similarity as Rossmann fold-type methyltransferases. PRMT1, 2, 3, 4, 6 and 8 were cloned as Type I PRMTs, and PRMT5 and PRMT9 were identified as Type II PRMTs. PRMT7 is supposed to be a candidate for Type III (52, 53). In yeast system, monomethylation of δ-position nitrogen atom is reported and its respective enzyme (Rmt2p) is defined as Type IV PRMT (54). Recently, a novel small arginine methyltransferase Mettl23, which has significant homology with the conserved domains of arginine methylation activity in PRMTs, has been reported as a new member of the family (55).
Many lines of evidence have shown that asymmetric and symmetric arginine dimethylations play significant roles in many aspects of cellular processes (52, 53). Especially, certain arginine methylation can inhibit nearby modifications to regulate protein functions as one of their molecular mechanisms. We found that PRMT1-mediated methylation of RxRxxS/T sequences (R, arginine; S, serine; T, threonine; X, any amino acid) in FOXO1 (forkhead box protein O1) (56) and BAD (BCL-2 antagonist of cell death) (57) blocked AKT phosphorylation at serine or threonine residues. In another case, glycolytic enzyme PFKFB3 is saved from proteolysis via the PRMT1-mediated methylation of arginine close to the ubiquitin-targeted lysine residue in cancer cells in which glycolytic pathway is activated (58).
In spite of accumulating evidence regarding cellular roles for protein arginine methylation in vitro, its whole-body functions have remained unclear, because knockout mice of PRMT1 (59) and PRMT5 (5), representative Type I and Type II enzymes, respectively, exhibit early embryonic lethality. However, recent studies revealed severe hypomyelination (60), defects in muscle regeneration and muscle stem cells (MSCs) fate (61), substantial impairment of B-cell activation and differentiation (62) and embryonic death at mid-stage caused by angiodysplasia (63), in central nervous system-specific, MSC-specific, B-cell-specific, and vascular endothelial cell-specific PRMT1-conditional knockout (cKO) mice, respectively. Oligodendrocyte-specific cKO of PRMT5 results in dysmyelination and early postnatal death (64, 65).
In another example as multicellular organisms than mammals, we observed that PRMT-1 (a nematode orthologue of human PRMT1) catalyzes almost all asymmetric arginine dimethylations and regulates lifespan through the blockade of the AKT phosphorylation at serine or threonine residues in the RxRxxS/T sequence via arginine methylation of Daf-16 (a nematode orthologue of human FOXO1) (66). Moreover, deletion of prmt-5 completely abolishes the formation of SDMA, indicating that PRMT-5 is the predominant type II PRMT in nematode (67). Interestingly, both prmt-1 and prmt-5 null mutants show no apparent phenotypes since they are viable and fertile despite the lack of enzymatic activities. Whereas the prmt-1; prmt-5 double mutant worms show that the levels of ADMA and SDMA were abolished as expected, they are viable, and exhibit small brood size and short body length compared to wild type (N2) and each of the single mutant worms. Although prmt-5 did not contribute to the lifespan regulation under the inactivation of prmt-1, both prmt-1 and prmt-5 genes were indispensable for resistance to oxidative and heat stresses (68).
Demethylations of Methyllysines and Methylarginines
Most characterized lysine demethylation reaction is α-ketoglutarate (α-KG) and iron ion-dependent oxidative reaction. First, an Fe(IV) oxo species is generated through the conversion reaction of molecular oxygen and α-KG into succinate and CO2. The reactive Fe(IV) oxo species oxidizes the methyl group to produce the hemiaminal intermediates. Further oxidation results in the fragmentation of hemiaminal and the release of formaldehyde. The responsible enzyme of this reaction is a group of oxygenases known as Jumonji-C domain (JmjC) subfamily (69). Besides of JmjC demethylases, lysine-specific demethylase 1 (LSD1) subfamily demethylases catalyze demethylation of methyllysine residues by using a flavin adenine dinucleotide as a coenzyme (70). Members of JmjC and LSD1 subfamilies are defined as lysine demethylases (KDMs).
Although the demethylation of methylarginine is supposed to be driven by same mechanism as that of methyllysine, there is still a dispute about enzymes catalyzing elimination of methyl groups from methylarginine residues in proteins (71). Very recently, arginine demethylation activity of some KDMs has been reported by using in vitro demethylation assay (72). Further investigations using another approach including in vivo analyses are required for examining whether in vitro activity of these KDMs accounts for the arginine demethylation or not.
Concluding Remarks
In this review, although we do not discuss regarding glutamine methylation of histone H2A, which is recently identified in yeast and involved in RNA polymerase I-mediated rDNA transcription (73), this methylation is catalyzed by Nop1 that is originally found as a methyltransferase of rRNA (74). Intriguingly, Nop1 is the orthologue of fibrillarin in human cells, where this enzyme catalyzes 2′-O-methylation of rRNA, the most abundant modification of eukaryotic rRNA. What does this substrate specificity of the methyltransferase for rRNA and protein tell us? Is this the species-specific function? or is it conserved not only in unicellular but also in multicellular animals? Does the substrate switch of methyltransferases depend on cofactors or methyl donors? Is it simply dependent on subcellular localization of methyltransferases or substrates or both?
In addition to methyltransferases (writer) and demethylases (eraser) of proteins described in this review, the binding proteins (reader) to methylated amino acids are the third regulator to express the functions of protein methylation. Although it is known that methyllysine are recognized by at least eight domains (Chromo, MBT, Tudor, PHD, PWWP, Ank and BAH domains), the Tudor family is a methylarginine binder possessing the domain known to bind methylarginine. These readers bind the methyllysine/arginine motif with ‘aromatic cage’ comprised the aromatic side chains in their specific domain. The structural studies have shown that the widths for the aromatic cages of the methylarginine binders are narrower than those of methyllysine binders. Since the shape of methyl-guanidino side chain of arginine is flat, methylarginine tightly fits with methylarginine binders but not with methyllysine binders. The Tudor methylarginine binders are supposed to be involved in the regulation of RNA splicing, gene expression, and in a gonad-specific small RNA silencing pathway (75).
In order to uncover increased complexity of biological methylation on genetic information, the specificity relationship between methytransferases and substrates would be one of exciting future prospects and be of great importance to advance understanding of cryptic layers of modifiable network on cellular functions.
Funding
This work was supported by Grant-in-Aid for Scientific Research (A) (17H01519 to A.F.), Grant-in-Aid for Scientific Research (C) (17K01942 to K.K.), and Grant-in-Aid for Scientific Research (C) (18K054298 to J.-D.K.) from the Japan Society for the Promotion of Science.
Conflict of Interest
None declared.
References
Abbreviations
- α-KG
α-ketoglutarate
- 5mC
5-methylcytosine
- ADMA
asymmetric dimethylarginine
- BAD
BCL-2 antagonist of cell death
- CpG
cytosine-guanine
- DCP2
decapping protein 2
- FOXO1
forkhead box protein O1
- FTO
fat mass and obesity-associated protein
- JmjC
Jumonji-C domain
- LSD1
lysine-specific demethylase 1
- MMA
Nω-monomethylarginine
- MSCs
muscle stem cells
- NML
nucleomethylin
- NSUN-5
Nop2/Sun-like domain containing protein 5
- PRMTs
protein arginine methyltransferases
- SAM
S-adenosyl-L-methionine
- SDMA
symmetric dimethylarginine
- YTHDF2
YTH domain-containing family protein 2
Author notes
Koichiro Kako and Jun-Dal Kim two authors equally contributed to this work.





