Cell.2014.08.05-Emerging Roles of RNA

Leading Edge

Review

Emerging Roles of RNA

Modi?cation:m6A and U-Tail

Mihye Lee,1,2Boseon Kim,1,2and V.Narry Kim1,2,*

1IBS Center for RNA Research,Institute for Basic Science,Seoul151-742,Korea

2School of Biological Sciences,Seoul National University,Seoul151-742,Korea

*Correspondence:narrykim@snu.ac.kr

https://www.360docs.net/doc/181211120.html,/10.1016/j.cell.2014.08.005

Although more than100types of RNA modi?cation have been described thus far,most of them were thought to be rare in mRNAs and in regulatory noncoding RNAs.Recent developments have un-veiled that at least some of the modi?cations are considerably abundant and widely conserved. This Minireview summarizes the molecular machineries and biological functions of methylation (N6-methyladenosine,m6A)and uridylation(U-tail).

RNAs undergo chemical modi?cations that can affect their activ-ity,localization,and stability(Machnicka et al.,2013),which is conceptually analogous to the modi?cations of DNA and protein. Abundant noncoding RNAs such as rRNAs and tRNAs are exten-sively modi?ed,whereas mRNA modi?cations are thought to be relatively low in frequency apart from the common terminal mod-i?cations,m7G cap and poly(A)tail.

The most abundant internal modi?cation on mRNA is N6-meth-yladenosine(m6A),discovered in the1970s.Early studies esti-mated m6A to be present at a level of approximately three to ?ve sites per mRNA(0.1%–0.4%of adenosines)in mammalian cells(Bokar,2005;Dubin and Taylor,1975;Perry et al.,1975; Wei et al.,1975).But the initial interest was dampened by con-cerns that m6A may have come from contamination from rRNAs and snRNAs and also due to the fact that mutation of speci?c m6A sites did not affect mRNA abundance and processing. Another concern was that,although viral mRNAs are frequently modi?ed,m6A was rarely detected in cellular mRNAs examined, which led to the idea that adenosine methylation may occur in a limited subset of viral and cellular mRNAs(Bokar,2005).After four decades of latency,however,the interest in m6A has been re-newed recently by developments of new sequencing techniques and discoveries from genetic and biochemical studies. Another‘‘old’’modi?cation that recently regained attention is tailing(30terminal untemplated nucleotidyl addition),which was initially noted in the1970s(Norbury,2013).Apart from canonical polyadenylation,it is now clear that RNAs undergo multiple types of tailing,including noncanonical adenylation,ur-idylation,and guanylation(Chang et al.,2014;Norbury,2013). Tailing plays important roles in the small RNA pathways(Ameres and Zamore,2013;Ha and Kim,2014).Accumulating evidence suggests that mRNAs are also subject to widespread regulatory tailing such as uridylation(Chang et al.,2014;Morozov et al., 2010;Mullen and Marzluff,2008;Rissland and Norbury,2009; Sement et al.,2013).The recent development of a high-throughput method now offers an opportunity to examine the 30terminal modi?cation of mRNA at the transcriptome level (Chang et al.,2014).

This Minireview will focus on two emerging modi?cations,m6A and U-tail,and will summarize recent technical developments, as well as the enzymology and biological signi?cance of methyl-ation and uridylation.

N6-Methyladenosine in mRNA

m6A modi?cation is found in rRNA,snRNA,and mRNA of viruses, yeast,plants,and animals and is particularly prevalent in higher eukaryotes(Bokar,2005).The consensus sequence of m6A was initially predicted from several m6A-containing mRNA se-quences and was de?ned as‘‘RRACH’’([G/A][G>A]m6AC[U> A>C])through in vitro assays(Harper et al.,1990).However, the in vivo methylation state of the predicted m6A sites remained largely unknown due to technical limitations.This changed when two groups independently developed similar methods,MeRIP-seq(methylated RNA immunoprecipitation followed by se-quencing)and m6A-seq(Dominissini et al.,2012;Meyer et al., 2012).To brie?y explain the methods,RNA is fragmented into $100nt long segments and is immunoprecipitated with anti-m6A antibody,which results in selective enrichment of methyl-ated RNA fragments.Eluted RNAs and input control samples are deep sequenced,and the reads are mapped to the genome. Using peak-calling algorithms,regions enriched in the immuno-precipitate relative to input samples are identi?ed as‘‘m6A peaks.’’These methods allowed genome-wide mapping of m6A modi?cation with a resolution of$200nt,detected over 12,000m6A peaks in transcripts of>7,000genes in human cells and mouse tissues.Many of the m6A peaks are conserved be-tween mouse and human,and they are enriched near the stop codon.These studies also revealed that the methylation status of some m6A sites dynamically changes in stress conditions, implicating a potential role of m6A in stress responses.The genome-wide m6A mapping in two yeast species(S.cerevisiae and S.mikatae)also demonstrated a conserved and dynamically regulated methylation during meiosis(Schwartz et al.,2013). Other various technical approaches have been taken by other groups to detect m6A.Methylation of a speci?c site can be quantitated by a digestion-based method called

SCARLET

980Cell158,August28,2014a2014Elsevier Inc.

(site-speci?c cleavage and radioactive-labeling followed by ligation-assisted extraction and thin-layer chromatography),although it cannot be adopted to transcriptomic analyses because it is based on thin-layer chromatography (Liu et al.,2013).When applied to several individual mRNAs and long ncRNAs,SCARLET assays estimated the percentages of m 6A at speci?c sites to range from 6%to 80%.Additional methods such as m 6A-sensitive ligase reaction and reverse transcription have also been proposed and need to be tested further (Dai et al.,2007;Harcourt et al.,2013;Vilfan et al.,2013).

Additional breakthroughs came from the identi?cation of fac-tors involved in m 6A regulation.The pathway is composed of three classes of protein factors:‘‘writers’’(that methylate aden-osine at N6position),‘‘erasers’’(that demethylate m 6A for reversible regulation),and ‘‘readers’’(the effectors that control the modi?ed mRNA’s fate)(Figures 1A and 1B).

The writer of m 6A is a multicomponent complex composed of METTL3(also known as MT-A70),METTL14,and WTAP in mam-mals (Bokar et al.,1997;Liu et al.,2014;Narayan and Rottman,1988;Ping et al.,2014;Wang et al.,2014b ).Knockdown of METTL3or METTL14reduced m 6A/A ratios in human cell lines and mouse embryonic stem cells (ESCs)(Liu et al.,2014;Ping et al.,2014;Wang et al.,2014b ).Each of METTL3and METTL14has methylation activity.But the heterodimer of the two methyl-transferases has a strongly enhanced activity in vitro,and they stabilize each other in vivo.METTL3and METTL14colocalize in nuclear speckles together with the third component,WTAP.WTAP was initially known as a splicing factor,but its biochemical role remained unknown.Knockdown of WTAP lowered the m 6A levels and decreased the amount of RNA bound to METTL3,sug-gesting that WTAP may recruit substrate RNAs to the enzymatic complex (Liu et al.,2014;Ping et al.,2014).MeRIP-seq and PAR-CLIP (photoactivatable ribonucleoside-enhanced crosslinking and immunoprecipitation)independently performed in different cell lines revealed that the targets of METTL3and METTL14largely overlap (Liu et al.,2014;Wang et al.,2014b ).Moreover,WTAP targets detected by PAR-CLIP also share the binding motif with that of METTL3and METTL14,which is similar to the m 6A consensus sequence (Liu et al.,2014;Ping et al.,2014).These results indicate that the three components function as a heterotrimeric methyltransferase in vivo (Figure 1B).

Illustrating their physiological relevance,mutations of the METTL3or WTAP homologs led to diverse defects in eukary-otes.In S.cerevisiae ,the MIS complex containing

Ime4

Figure 1.The m 6A Pathway and Its Cellular Functions

(A)N 6-methyladenosine methylation and demethylation reactions.Methyltransferase complex containing METTL3,METTL14,and WTAP catalyzes m 6A methylation,whereas FTO and ALKBH5,the demethylases,catalyze oxidative demethylation of m 6A.

(B)Domain structures of writer,eraser,and reader/effector proteins in human.METTL3,METTL14,and WTAP are components of the methyltransferase complex.METTL3and METTL4have a SAM-binding domain required for m 6A methylation,whereas WTAP contains no characteristic domain.Eraser proteins,FTO and ALKBH5demethylases,have an AlkB domain in https://www.360docs.net/doc/181211120.html,pared to ALKBH5,FTO has an additional C-terminal domain.YTHDF1,2,3containing a YTH RNA-binding domain are effector proteins.The P/Q/N-rich domain is known to be important for the localization of YTHDF2to P body (Fu et al.,2014).

(C)Proposed model of the cellular function of m 6A on mRNA.Reversible methylation/demethylation is thought to occur in nuclear speckles where the enzymes are concentrated.Methylation may affect the export and splicing of mRNAs in the nucleus.Exported methylated mRNAs are recognized by YTHDF2in the cytoplasm and then localize to P bodies,where mRNA decay factors are enriched.

Cell 158,August 28,2014a2014Elsevier Inc.981

(a METTL3homolog),Mum2(a WTAP homolog),and Slz1is responsible for mRNA methylation during meiosis,and muta-tions in the MIS components display meiotic defect(Agarwala et al.,2012;Clancy et al.,2002).In Arabidopsis,disruption of MTA(a METTL3homolog)causes early developmental arrest and cell division defects(Zhong et al.,2008).Its depletion at a later stage results in altered growth patterns and?oral defects (Bodi et al.,2012).Drosophila Ime4(a METTL3homolog)is critical for viability and is associated with Notch signaling during oogenesis(Hongay and Orr-Weaver,2011).In zebra?sh, knockdown of METTL3or WTAP led to multiple defects in early development and increased apoptosis(Ping et al.,2014). Furthermore,RNAi of Mettl3or Mettl14in mouse embryonic stem cells downregulated pluripotency markers and proliferation rates and resulted in morphological changes in ESCs(Wang et al.,2014b).These studies collectively suggested that RNA methylation is physiologically relevant,though it is yet to be proven that the observed phenotypes are due to methylation of mRNAs rather than that of other RNA classes.

Although the studies of writer made important contributions,a critical breakthrough that revived the m6A biology actually came from the discovery of an eraser,fat mass and obesity-associated (FTO).The?nding of FTO as an m6A demethylase implicated the dynamic and reversible nature of m6A modi?cation and estab-lished a link to human health.Genome-wide association studies found a strong correlation between a single-nucleotide polymor-phism in the?rst intron of FTO and increased body mass index and obesity risk in multiple human populations.In addition,ho-mozygous loss-of-function mutation(Arg316Gln)is associated with postnatal growth retardation and multiple malformations in humans(Fawcett and Barroso,2010).Consistent with the hu-man data,deletion of Fto in mice resulted in altered body weight, increased lethality,and growth retardation(Fawcett and Bar-roso,2010).Fto is most abundantly expressed in the brain, particularly within hypothalamus,which is important for meta-bolism.Another demethylase,ALKBH5,was recently discovered (Zheng et al.,2013a).ALKBH5has the highest expression in testes,and the ALKBH5-de?cient mice show defective sper-matogenesis and impaired fertility,though the mice are viable and develop to adult normally.The homologs of m6A demethy-lases are yet to be investigated in other species.ALKBH5 appears to be conserved only in vertebrates,whereas FTO ho-mologs have been detected in vertebrates as well as in algae (Zheng et al.,2013b)(Figure1B).

FTO and ALKBH5belong to the AlkB family of the Fe(II)and a-ketoglutarate-dependent dioxygenases.This family also in-cludes demethylases of DNA and histones(Kurowski et al., 2003;Zheng et al.,2013b).FTO effectively carries out oxidative demethylation of m6A on RNA in vitro(Jia et al.,2011),and ALKBH5has comparable m6A demethylation activity with a pref-erence for the m6A consensus sequence(RRACH)(Zheng et al., 2013a).Silencing or overexpression of these factors resulted in an increase or decrease of m6A on mRNA,respectively.FTO forms nuclear foci that partially colocalize with nuclear speckles, whereas ALKBH5is tightly localized in nuclear speckles that contain splicing factors(Jia et al.,2011;Zheng et al.,2013a). Knockdown of ALKBH5,but not that of FTO,affects the spliceo-some assembly.ALKBH5-depleted cells also show cytoplasmic accumulation of poly(A)+RNA,suggesting that methylation may be necessary for nuclear export of mRNA(Zheng et al.,2013a).It remains to be determined what the molecular consequences of FTO-mediated demethylation are and how they differ from those of ALKBH5-mediated demethylation.

The m6A readers and their effector functions begin to be un-covered.From a pull-down assay using a synthetic m6A RNA bait,three YTH domain proteins(YTHDF1–3)have been identi-?ed in mammalian cells(Dominissini et al.,2012;Wang et al., 2014a).The YTH domain family is widespread in eukaryotes and is known to bind to ssRNA through the YTH domain,but their functions were not known(Zhang et al.,2010).Though all YTHDF1–3show selective binding to m6A embedded in con-sensus sequences,YTHDF2has the highest af?nity(Wang et al.,2014a).YTHDF2target sites identi?ed by PAR-CLIP largely overlap with m6A peaks mapped by MeRIP-seq,and they pri-marily localize near the stop codon.YTHDF2knockdown led to an increase of abundance and half-lives of target mRNAs,indi-cating that YTHDF2is involved in RNA decay.Indeed,mRNA half-lives were increased in accordance with the number of bind-ing sites present in target mRNAs.Via its P/Q/N-rich N-terminal domain,YTHDF2localizes to processing bodies(P bodies)in the cytoplasm where mRNA turnover factors are concentrated. Thus,YTHDF2may select m6A-modi?ed mRNA through the YTH domain and localize them to the P bodies using N-terminal domain so as to facilitate mRNA decay(Wang et al.,2014a) (Figures1B and1C).

Apart from mRNA decay(Wang et al.,2014a,2014b),several functions of m6A have been proposed.Initial studies proposed splicing as the main regulatory target of m6A,based on the ob-servations that methylation occurs in the nucleus and that methylation sites are detected in introns of pre-mRNAs(Carroll et al.,1990;Stoltzfus and Dane,1982).Global discovery of m6A sites showed that some m6A sites are indeed located in in-tronic regions and that alternatively spliced exons and introns are considerably more methylated than constitutively spliced ones (Dominissini et al.,2012).A similar pattern was observed in PAR-CLIP of METTL3and WTAP,and depletion of METTL3or WTAP led to alteration in splicing isoforms(Ping et al.,2014). In addition to splicing,other RNA processing defects have been reported to be related to methylation.METTL3knockdown in mouse or human cell lines caused an apparent delay in the nuclear exit of mature mRNA of circadian genes and showed circadian period elongation phenotype(Fustin et al.,2013). Conversely,depletion of ALKBH5leads to accelerated mRNA export(Zheng et al.,2013a)(Figure1C).It is noteworthy that m6A peaks are detected frequently near stop codons.This observation implies a link to translation and/or translation-coupled decay,though these possibilities have yet to be tested directly.

U-Tail in mRNAs and MicroRNAs

‘‘Writers’’of uridylation belong to a large family of ribonucleotidyl transferases that have a catalytic domain with sequence homol-ogy to DNA polymerase b(Norbury,2013).These proteins were ?rst described as noncanonical poly(A)polymerases(PAPs), based on the?ndings that ribonucleotidyl transferases such as PAPD4(GLD2or TUT2)and mtPAP(TUT1)catalyze adenylation

982Cell158,August28,2014a2014Elsevier Inc.

of cytoplasmic mRNA and mitochondrial mRNA,respectively (Tomecki et al.,2004;Wang et al.,2002).Some ribonucleotidyl transferases have uridylation activity;hence,noncanonical PAPs are also called terminal uridylyl transferases (TUTs)or poly(U)polymerases (PUPs)(Norbury,2013).Humans have seven proteins with potential TUT activity.Figure 2A shows ?ve human TUTs that have been proposed to uridylate mRNAs or mi-croRNAs.Although the catalytic domains show sequence ho-mology among TUTs,their overall domain structures have little similarity between species,making it dif?cult to predict ortholo-gous relationships.Budding yeast S.cerevisiae is the only eu-karyotic model organism that lacks an apparent TUT homolog,whereas ?ssion yeast S.pombe possesses at least one TUT.The ?rst glimpse of uridylation in cytoplasmic RNA came from the description of nontemplated U residues on the 50cleavage products from miRNA-mediated slicing (Shen and Goodman,2004).Soon after,miRNA uridylation was found in Arabidopsis HEN1mutant that is defective in 20-O-methyl modi?cation at the terminal ribose (Li et al.,2005).Uridylation plays regulatory roles in the miRNA pathway,with the precursor of let-7(pre-let-7)being the most extensively studied example (Heo et al.,2008;Newman et al.,2011)(Figure 2B).In differentiated cells,precursors of most let-7members are mono-uridylated by ZCCHC11(TUT4),ZCCHC6(TUT7),or PAPD4(GLD2or TUT2)(Heo et al.,2012).Mono-U tail extends the 30overhang of pre-let-7and thereby facilitates Dicer processing and miRNA maturation.But in ESCs and in certain cancer cells in which an oncofetal protein LIN28or LIN28B is expressed,pre-let-7is oligo-uridylated,and the long U-tail blocks Dicer processing and inhibits miRNA maturation (Hagan et al.,2009;Heo et al.,2008,2009).Thus,there is a functional duality in uridylation de-pending on the length of the U-tail.What LIN28does is to stabi-lize the interaction between the substrate and TUTs,thereby enhancing processivity of reaction (Yeom et al.,2011).In animals and plants,there have been many reports of uridylation and ad-enylation of mature miRNAs,which affect stability and activity

of

Figure 2.Uridylation in the miRNA and mRNA Pathways

(A)Domain organization of ?ve human terminal uridylyl transferases (writers).Red,nucleotidyl transferase domain;yellow,PAP-associated domain;dark red,inactive nucleotidyl transferase domain;blue,C2H2-type zinc ?nger domain;bluish green,CCHC-type zinc ?nger domain.

(B)Pre-miRNA uridylation.Microprocessor composed of DROSHA and DGCR8crops pri-miRNA to release pre-miRNA.Pre-miRNA is exported to the cytoplasm by EXPORTIN5(EXP5).In embryonic stem cells and in certain cancer cells in which LIN28or Lin28B is expressed,pre-let-7is oligo-uridylated by TUT4or TUT7and the U-tail blocks Dicer processing and recruits exonuclease DIS3L2.However,in differentiated cells in which LIN28is absent,precursors of most let-7members are mono-uridylated by TUT2,TUT4,or TUT7,which facilitates DICER processing.

(C)Model for mRNA degradation.Most mRNAs undergo deadenylation at the beginning of decay.Poly(A)tail is removed by deadenylases (shown here as Ccr4-Caf1complex as an example).Following deadenylation,mRNAs are degraded bidirectionally by Xrn1following decapping by Dcp1-Dcp2(50-to-30decay pathway)and by exosome (30-to-50decay pathway).Uridylation is thought to act redundantly with deadenylation in S.pombe to stimulate decapping.TUTase uridylates poly(A)+mRNAs (Cid1in S.pombe ,CutA in A.nidulans ,or URT1in A.thaliana ).Lsm1-7binds to the U-tail and enhances decapping,which triggers 50-to-30decay by Xrn1.

(D)Degradation pathway for replication-dependent histone mRNAs.Degradation is initiated by oligo-uridylation.Then LSm1-7binds to the U-tail and stimulates decapping and 50-to-30decay by XRN1.Some histone mRNAs are also degraded in the 30-to-50direction by 30hExo and PM/Scl-100(one of exosome components).

Cell 158,August 28,2014a2014Elsevier Inc.983

miRNAs.Tailing in the small RNA pathway has been reviewed elsewhere in detail(Ameres and Zamore,2013;Ha and Kim, 2014).

Mounting evidence indicates that mRNAs are generally sub-ject to uridylation,which challenges the way that we think about mRNA tails(Figure2C).Short uridine tails(typically mono-or di-) were initially found on mRNAs of?ssion yeast S.pombe(Riss-land et al.,2007;Rissland and Norbury,2009).Deletion of the TUTase gene cid1resulted in an increase of mRNA stability, though it was tested for a single gene(urg1).Uridylated tran-scripts accumulated when deadenylation factor(Ccr4)or de-capping factors(Dcp1and Lsm1)were mutated.A related phenomenon was observed in?lamentous fungus Aspergillus ni-dulans,where mRNAs frequently contain an extended tail with mixed C and U residues with the consensus CUCU(Morozov et al.,2010,2012).When ribonucleotidyl transferase CutA(or CutB)gene was deleted,mRNA decay rate was decreased. These results implicate that uridylation and deadenylation may act redundantly to stimulate decapping and decay of mRNA.Thus,despite differences in sequences,functions of downstream effectors may be conserved among fungi species. Uridylation has also been found on poly(A)+mRNAs of higher eu-karyotes,but it remains to be determined whether or not the function of uridylation is conserved.In Arabidopsis,short U-tails occur preferentially on deadenylated mRNAs,suggesting a link to mRNA decay(Sement et al.,2013).However,in the mutant of urt1,which is responsible for uridylation,30trimmed decay in-termediates accumulated without signi?cant changes in mRNA half-lives.Accordingly,it was proposed that uridylation may sim-ply block30trimming to establish the50-to-30directionality of degradation rather than to enhance the rate of decay.

Some poly(A)àRNAs are also known to be uridylated.The replication-dependent histone mRNAs have a conserved stem-loop structure instead of a poly(A)tail at their30end,and they are rapidly degraded at the end of S phase or when DNA replica-tion is inhibited during S phase(Mullen and Marzluff,2008).The 30ends of histone mRNAs are oligo-uridylated during their degradation(Figure2D).Recent studies analyzed degradation intermediates of histone mRNA using a circular RT-PCR assay or a high-throughput sequencing approach,revealing that uridy-lation facilitates decay in both directions(Hoe?g et al.,2013; Mullen and Marzluff,2008;Slevin et al.,2014).The50-30degrada-tion is carried out by XRN1following decapping.The30-50degra-dation is mediated initially by30hExo(also known as Eri1)and subsequently by PM/Scl-100(a component of exosome;also known as RRP6).When oligo-uridylation is inhibited,the rate of degradation decreased(Mullen and Marzluff,2008;Schmidt et al.,2011;Su et al.,2013).As for uridylating enzymes for the histone mRNA,an initial study proposed mtPAP(TUT1) and PAPD5(TUT3)(Mullen and Marzluff,2008),whereas two following reports instead implicated ZCCHC11(TUT4)(Schmidt et al.,2011;Su et al.,2013).Other poly(A)àmRNAs such as endonucleolytic cleavage products also seem to be subject to tailing and decay.An example is miRNA-directed50cleavage products that are oligo-uridylated in plants and mammalian cells, though the TUTs for this reaction are yet to be identi?ed(Shen and Goodman,2004).Notably,however,a related phenomenon has been studied in green algae Chlamydomonas(Ibrahim et al.,2006).An adenylyl transferase MUT68mediates oligo-adenyla-tion of the siRNA-mediated cleavage product and induces decay of adenylated RNA.

Further studies are needed to understand the commonalities and variations among these seemingly related phenomena from diverse species.Also,because the above studies were per-formed on a few individual genes from each species,it is impor-tant to examine how widespread uridylation is and how generally it contributes to gene regulation.

Recently,global investigation of mRNA tail was made possible by the development of a high-throughput technique called TAIL-seq(Chang et al.,2014).Brie?y,RNAs are ligated to a bio-tinylated30adaptor.Following fragmentation,the30-most frag-ments are puri?ed with streptavidin beads and ligated to the50 adaptor.Paired-end sequencing of the cDNA is used to identify the transcript(51nt from the50end)and determine the30tail se-quences(231nt from the30end).Due to the dif?culties associ-ated with sequencing homopolymeric poly(A)tail,?uorescence signals are used for a machine learning algorithm,which allows the accurate determination of poly(A)tail length.TAIL-seq re-vealed that a surprisingly large proportion of poly(A)+mRNAs are uridylated in mammalian cells(Chang et al.,2014).More than85%of mRNA species are uridylated at a frequency of higher than1%.Most U-tails are short(1–4uridines)at the steady state.Similarly to plants,U-tails are found mostly on mRNAs with shortened A tails(<$25nt).Moreover,uridylation frequency correlates negatively with mRNA half-life at the global level.Thus,uridylation may be involved in mRNA decay in most eukaryotes(with some exceptions,such as S.cerevisiae). Because the current model for mRNA decay is largely based on studies of S.cerevisiae that lacks the uridylation machinery, future studies may introduce substantial changes to our under-standing of general mRNA decay.It will be important to identify the writer(TUT)for mammalian poly(A)+mRNA and to under-stand how the uridylyl transferase recognizes nonfunctional mRNAs selectively.It is also necessary to reveal what the direct molecular consequences of uridylation are and if and how uridylation is regulated during developmental and pathological transitions.

It is currently unclear whether uridylation of mRNA and miRNA is a reversible process—that is,whether deuridylase plays a reg-ulatory role as an‘‘eraser.’’A30-50exonuclease DIS3L2has recently been shown to act preferentially on oligo-U-tails of mRNAs(in?ssion yeast)and on precursor of let-7(in mice and humans)(Chang et al.,2013;Malecki et al.,2013;Ustianenko et al.,2013).The structure of mouse Dis3l2in complex with an oligo-U(U14)has recently been reported,which showed exten-sive uracil base-speci?c interaction(Faehnle et al.,2014).How-ever,as DIS3L2degrades the whole RNA body,it would be more appropriate to consider this enzyme as a putative‘‘reader/ effector.’’It is expected that there are additional effectors that selectively recognize U-tails.For example,the Lsm1-7complex is known to bind to oligo-U-tracts and enhance decapping in vitro(Song and Kiledjian,2007).In yeast,Lsm1-7complex contributes to the decay of uridylated poly(A)+mRNAs(Rissland and Norbury,2009),and in mammals,the Lsm1-7complex par-ticipates in the decay of uridylated histone mRNAs through its interaction with SLBP(stem-loop binding protein)and30hExo

984Cell158,August28,2014a2014Elsevier Inc.

(Hoe?g et al.,2013;Lyons et al.,2014;Mullen and Marzluff, 2008).Thus,the Lsm1-7complex may be another conserved reader/effector that selectively recognizes U-tails.It is currently unclear what the molecular basis of U-tail recognition is and how many downstream effectors exist.

Perspectives

Pioneering studies on the methylome and30-terminome of RNA have revealed interesting new layers of gene expression.How-ever,there is yet no consensus as to the exact function of m6A and U-tail,and many questions are pending without clear an-swers.Are there additional writers,erasers,and readers for m6A?What are the writers(TUTs)and readers for U-tails of mammalian mRNAs?How do these factors achieve speci?city? Is uridylation reversible?How are they regulated and connected to cell-signaling pathways?Which human diseases are associ-ated with such modi?cations?

Finding the factors involved in the pathway will provide crucial information on the mechanism and function of methylation and uridylation.It is expected that many factors will emerge from af?nity puri?cation/immunoprecipitation followed by proteomic analyses.Candidate approaches based on homology to known factors will also be fruitful when combined with RNAi.The following step will be to investigate the factors in detail to gain mechanistic understanding.Structural and biochemical studies of METTL3-METTL14-WTAP and YTHDF2are much needed to understand how the substrates are selected.Structures of TUT and the higher eukaryotic Lsm1-7in complex with RNA are also necessary to reveal how U-tail is generated and recog-nized.Furthermore,genetic studies will teach us what develop-mental roles the modi?cation factors play,though a caveat is that it would be dif?cult to segregate their effects on different RNA classes.Genome editing tools are likely to accelerate the func-tional studies of the modi?cation factors.

There are intriguing commonalities between m6A modi?cation and U-tailing.Both modi?cations are conserved widely in eu-karyotes.Both m6A and U-tail serve as molecular marks that pro-vide binding sites for their effectors.m6A is selectively bound to the YTH domain proteins,whereas U-tail is recognized by DIS3L2or the Lsm1-7complex.Intriguingly,the most plausible function of m6A and U-tail is to facilitate mRNA decay.It will be interesting to ask whether there is any interplay between these two modi?cations.

The rapid progress in the biology of m6A and tailing was possible,in part,due to the development of new technologies that allowed genome-wide detection of the modi?cations. Although these methods will continue to serve as potent tools, further technical improvement is necessary.For instance, MeRIP-seq and m6A-seq suffer from a low resolution,so it is currently impossible to identify the modi?cation sites at a sin-gle-nucleotide resolution.In addition,MeRIP-seq and m6A-seq can only detect the relative abundance of modi?cation and cannot quantitate the absolute stoichiometry of m6A on each site.Additional approaches may become available by combining crosslinking or chemical modi?cation to increase resolution.As for the detection of U-tails,the current method,TAIL-seq,suc-cessfully offers genome-wide and quantitative pro?les of tail se-quences.However,it needs further improvement to effectively remove abundant ncRNA contaminants so as to increase sequencing depth and lower the cost.TAIL-seq is expected to be used not only to investigate uridylation,but also to study other features of30terminome,such as deadenylation,endonucleo-lytic cleavage,and guanylation.

In this Minireview,we only dealt with m6A and U-tail,but some of the main messages from this area of research are that RNA is not different from DNA or protein in its capacity to be regulated by postsynthetic modi?cations and that there may be additional underestimated modi?cations that contribute to gene regulation. Other types of such‘‘epitranscriptomic’’modi?cations may deserve attention in the future,and the potential candidates include pseudouridylation,5-methyl cytosine,and20-O-methyl-ation.Effective tools that detect such modi?cations may open new windows in RNA biology and beyond. ACKNOWLEDGMENTS

The authors are grateful to the members of their laboratory,particularly Jae-chul Lim and Minju Ha for helpful discussions and comments.This work was supported by the Research Center Program(EM1402)of the Institute for Basic Science(IBS)from the Ministry of Science,Information,and Communication Technology(ICT)and Future Planning of Korea(M.L.,B.K.,and V.N.K.)and by BK21Research Fellowships from the Ministry of Education of Korea(B.K.).

REFERENCES

Agarwala,S.D.,Blitzblau,H.G.,Hochwagen,A.,and Fink,G.R.(2012).RNA methylation by the MIS complex regulates a cell fate decision in yeast.PLoS Genet.8,e1002732.

Ameres,S.L.,and Zamore,P.D.(2013).Diversifying microRNA sequence and function.Nat.Rev.Mol.Cell Biol.14,475–488.

Bodi,Z.,Zhong,S.,Mehra,S.,Song,J.,Graham,N.,Li,H.,May,S.,and Fray, R.G.(2012).Adenosine Methylation in Arabidopsis mRNA is Associated with the30End and Reduced Levels Cause Developmental Defects.Front Plant Sci3,48.

Bokar,J.A.(2005).The biosynthesis and functional roles of methylated nucle-osides in eukaryotic mRNA.In Fine-Tuning of RNA Functions by Modi?cation and Editing,H.Grosjean,ed.(Berlin:Springer).

Bokar,J.A.,Shambaugh,M.E.,Polayes,D.,Matera,A.G.,and Rottman,F.M. (1997).Puri?cation and cDNA cloning of the AdoMet-binding subunit of the human mRNA(N6-adenosine)-methyltransferase.RNA3,1233–1247.

Carroll,S.M.,Narayan,P.,and Rottman,F.M.(1990).N6-methyladenosine residues in an intron-speci?c region of prolactin pre-mRNA.Mol.Cell.Biol. 10,4456–4465.

Chang,H.M.,Triboulet,R.,Thornton,J.E.,and Gregory,R.I.(2013).A role for the Perlman syndrome exonuclease Dis3l2in the Lin28-let-7pathway.Nature 497,244–248.

Chang,H.,Lim,J.,Ha,M.,and Kim,V.N.(2014).TAIL-seq:genome-wide determination of poly(A)tail length and30end modi?cations.Mol.Cell53, 1044–1052.

Clancy,M.J.,Shambaugh,M.E.,Timpte,C.S.,and Bokar,J.A.(2002).Induc-tion of sporulation in Saccharomyces cerevisiae leads to the formation of N6-methyladenosine in mRNA:a potential mechanism for the activity of the IME4 gene.Nucleic Acids Res.30,4509–4518.

Dai,Q.,Fong,R.,Saikia,M.,Stephenson,D.,Yu,Y.T.,Pan,T.,and Piccirilli, J.A.(2007).Identi?cation of recognition residues for ligation-based detection and quantitation of pseudouridine and N6-methyladenosine.Nucleic Acids Res.35,6322–6329.

Dominissini,D.,Moshitch-Moshkovitz,S.,Schwartz,S.,Salmon-Divon,M., Ungar,L.,Osenberg,S.,Cesarkas,K.,Jacob-Hirsch,J.,Amariglio,N.,Kupiec,

Cell158,August28,2014a2014Elsevier Inc.985

M.,et al.(2012).Topology of the human and mouse m6A RNA methylomes revealed by m6A-seq.Nature485,201–206.

Dubin,D.T.,and Taylor,R.H.(1975).The methylation state of poly A-containing messenger RNA from cultured hamster cells.Nucleic Acids Res.2,1653–1668.

Faehnle,C.R.,Walleshauser,J.,and Joshua-Tor,L.(2014).Nature.Published online August3,2014.https://www.360docs.net/doc/181211120.html,/10.1038/nature13553.

Fawcett,K.A.,and Barroso,I.(2010).The genetics of obesity:FTO leads the way.Trends Genet.26,266–274.

Fu,Y.,Dominissini,D.,Rechavi,G.,and He,C.(2014).Gene expression regulation mediated through reversible m6A RNA methylation.Nat.Rev. Genet.15,293–306.

Fustin,J.M.,Doi,M.,Yamaguchi,Y.,Hida,H.,Nishimura,S.,Yoshida,M., Isagawa,T.,Morioka,M.S.,Kakeya,H.,Manabe,I.,and Okamura,H.(2013). RNA-methylation-dependent RNA processing controls the speed of the circa-dian clock.Cell155,793–806.

Ha,M.,and Kim,V.N.(2014).Regulation of microRNA biogenesis.Nat.Rev. Mol.Cell Biol.15,509–524.

Hagan,J.P.,Piskounova,E.,and Gregory,R.I.(2009).Lin28recruits the TUTase Zcchc11to inhibit let-7maturation in mouse embryonic stem cells. Nat.Struct.Mol.Biol.16,1021–1025.

Harcourt,E.M.,Ehrenschwender,T.,Batista,P.J.,Chang,H.Y.,and Kool,E.T. (2013).Identi?cation of a selective polymerase enables detection of N(6)-methyladenosine in RNA.J.Am.Chem.Soc.135,19079–19082.

Harper,J.E.,Miceli,S.M.,Roberts,R.J.,and Manley,J.L.(1990).Sequence speci?city of the human mRNA N6-adenosine methylase in vitro.Nucleic Acids Res.18,5735–5741.

Heo,I.,Joo,C.,Cho,J.,Ha,M.,Han,J.,and Kim,V.N.(2008).Lin28mediates the terminal uridylation of let-7precursor MicroRNA.Mol.Cell32,276–284.

Heo,I.,Joo,C.,Kim,Y.K.,Ha,M.,Yoon,M.J.,Cho,J.,Yeom,K.H.,Han,J., and Kim,V.N.(2009).TUT4in concert with Lin28suppresses microRNA biogenesis through pre-microRNA uridylation.Cell138,696–708.

Heo,I.,Ha,M.,Lim,J.,Yoon,M.J.,Park,J.E.,Kwon,S.C.,Chang,H.,and Kim, V.N.(2012).Mono-uridylation of pre-microRNA as a key step in the biogenesis of group II let-7microRNAs.Cell151,521–532.

Hoe?g,K.P.,Rath,N.,Heinz,G.A.,Wolf,C.,Dameris,J.,Schepers,A.,Krem-mer,E.,Ansel,K.M.,and Heissmeyer,V.(2013).Eri1degrades the stem-loop of oligouridylated histone mRNAs to induce replication-dependent decay.Nat. Struct.Mol.Biol.20,73–81.

Hongay,C.F.,and Orr-Weaver,T.L.(2011).Drosophila Inducer of MEiosis4 (IME4)is required for Notch signaling during oogenesis.Proc.Natl.Acad. https://www.360docs.net/doc/181211120.html,A108,14855–14860.

Ibrahim,F.,Rohr,J.,Jeong,W.J.,Hesson,J.,and Cerutti,H.(2006).Untem-plated oligoadenylation promotes degradation of RISC-cleaved transcripts. Science314,1893.

Jia,G.,Fu,Y.,Zhao,X.,Dai,Q.,Zheng,G.,Yang,Y.,Yi,C.,Lindahl,T.,Pan,T., Yang,Y.G.,and He,C.(2011).N6-methyladenosine in nuclear RNA is a major substrate of the obesity-associated FTO.Nat.Chem.Biol.7,885–887.

Kurowski,M.A.,Bhagwat,A.S.,Papaj,G.,and Bujnicki,J.M.(2003).Phyloge-nomic identi?cation of?ve new human homologs of the DNA repair enzyme AlkB.BMC Genomics4,48.

Li,J.,Yang,Z.,Yu,B.,Liu,J.,and Chen,X.(2005).Methylation protects miRNAs and siRNAs from a30-end uridylation activity in Arabidopsis.Curr. Biol.15,1501–1507.

Liu,N.,Parisien,M.,Dai,Q.,Zheng,G.,He,C.,and Pan,T.(2013).Probing N6-methyladenosine RNA modi?cation status at single nucleotide resolution in mRNA and long noncoding RNA.RNA19,1848–1856.

Liu,J.,Yue,Y.,Han,D.,Wang,X.,Fu,Y.,Zhang,L.,Jia,G.,Yu,M.,Lu,Z., Deng,X.,et al.(2014).A METTL3-METTL14complex mediates mammalian nu-clear RNA N6-adenosine methylation.Nat.Chem.Biol.10,93–95.

Lyons,S.M.,Ricciardi,A.S.,Guo,A.Y.,Kambach,C.,and Marzluff,W.F. (2014).The C-terminal extension of Lsm4interacts directly with the30end of the histone mRNP and is required for ef?cient histone mRNA degradation. RNA20,88–102.

Machnicka,M.A.,Milanowska,K.,Osman Oglou,O.,Purta,E.,Kurkowska,M., Olchowik,A.,Januszewski,W.,Kalinowski,S.,Dunin-Horkawicz,S.,Rother, K.M.,et al.(2013).MODOMICS:a database of RNA modi?cation path-ways—2013update.Nucleic Acids Res.41(Database issue),D262–D267.

Malecki,M.,Viegas,S.C.,Carneiro,T.,Golik,P.,Dressaire,C.,Ferreira,M.G., and Arraiano,C.M.(2013).The exoribonuclease Dis3L2de?nes a novel eu-karyotic RNA degradation pathway.EMBO J.32,1842–1854.

Meyer,K.D.,Saletore,Y.,Zumbo,P.,Elemento,O.,Mason,C.E.,and Jaffrey, S.R.(2012).Comprehensive analysis of mRNA methylation reveals enrichment in30UTRs and near stop codons.Cell149,1635–1646.

Morozov,I.Y.,Jones,M.G.,Razak,A.A.,Rigden,D.J.,and Caddick,M.X. (2010).CUCU modi?cation of mRNA promotes decapping and transcript degradation in Aspergillus nidulans.Mol.Cell.Biol.30,460–469.

Morozov,I.Y.,Jones,M.G.,Gould,P.D.,Crome,V.,Wilson,J.B.,Hall,A.J., Rigden,D.J.,and Caddick,M.X.(2012).mRNA30tagging is induced by nonsense-mediated decay and promotes ribosome dissociation.Mol.Cell. Biol.32,2585–2595.

Mullen,T.E.,and Marzluff,W.F.(2008).Degradation of histone mRNA requires oligouridylation followed by decapping and simultaneous degradation of the mRNA both50to30and30to50.Genes Dev.22,50–65.

Narayan,P.,and Rottman,F.M.(1988).An in vitro system for accurate methyl-ation of internal adenosine residues in messenger RNA.Science242,1159–1162.

Newman,M.A.,Mani,V.,and Hammond,S.M.(2011).Deep sequencing of mi-croRNA precursors reveals extensive30end modi?cation.RNA17,1795–1803.

Norbury,C.J.(2013).Cytoplasmic RNA:a case of the tail wagging the dog. Nat.Rev.Mol.Cell Biol.14,643–653.

Perry,R.P.,Kelley,D.E.,Friderici,K.,and Rottman,F.(1975).The methylated constituents of L cell messenger RNA:evidence for an unusual cluster at the50 terminus.Cell4,387–394.

Ping,X.L.,Sun,B.F.,Wang,L.,Xiao,W.,Yang,X.,Wang,W.J.,Adhikari,S., Shi,Y.,Lv,Y.,Chen,Y.S.,et al.(2014).Mammalian WTAP is a regulatory subunit of the RNA N6-methyladenosine methyltransferase.Cell Res.24, 177–189.

Rissland,O.S.,and Norbury,C.J.(2009).Decapping is preceded by30uridy-lation in a novel pathway of bulk mRNA turnover.Nat.Struct.Mol.Biol.16, 616–623.

Rissland,O.S.,Mikulasova,A.,and Norbury,C.J.(2007).Ef?cient RNA polyur-idylation by noncanonical poly(A)polymerases.Mol.Cell.Biol.27,3612–3624.

Schmidt,M.J.,West,S.,and Norbury,C.J.(2011).The human cytoplasmic RNA terminal U-transferase ZCCHC11targets histone mRNAs for degrada-tion.RNA17,39–44.

Schwartz,S.,Agarwala,S.D.,Mumbach,M.R.,Jovanovic,M.,Mertins,P., Shishkin,A.,Tabach,Y.,Mikkelsen,T.S.,Satija,R.,Ruvkun,G.,et al.(2013). High-resolution mapping reveals a conserved,widespread,dynamic mRNA methylation program in yeast meiosis.Cell155,1409–1421.

Sement,F.M.,Ferrier,E.,Zuber,H.,Merret,R.,Alioua,M.,Deragon,J.M., Bousquet-Antonelli,C.,Lange,H.,and Gagliardi,D.(2013).Uridylation pre-vents30trimming of oligoadenylated mRNAs.Nucleic Acids Res.41,7115–7127.

Shen, B.,and Goodman,H.M.(2004).Uridine addition after microRNA-directed cleavage.Science306,997.

Slevin,M.K.,Meaux,S.,Welch,J.D.,Bigler,R.,Miliani de Marval,P.L.,Su,W., Rhoads,R.E.,Prins,J.F.,and Marzluff,W.F.(2014).Deep sequencing shows multiple oligouridylations are required for30to50degradation of histone mRNAs on polyribosomes.Mol.Cell53,1020–1030.

Song,M.G.,and Kiledjian,M.(2007).30Terminal oligo U-tract-mediated stim-ulation of decapping.RNA13,2356–2365.

986Cell158,August28,2014a2014Elsevier Inc.

Stoltzfus,C.M.,and Dane,R.W.(1982).Accumulation of spliced avian retro-virus mRNA is inhibited in S-adenosylmethionine-depleted chicken embryo ?broblasts.J.Virol.42,918–931.

Su,W.,Slepenkov,S.V.,Slevin,M.K.,Lyons,S.M.,Ziemniak,M.,Kowalska,J., Darzynkiewicz,E.,Jemielity,J.,Marzluff,W.F.,and Rhoads,R.E.(2013). mRNAs containing the histone30stem-loop are degraded primarily by de-capping mediated by oligouridylation of the30end.RNA19,1–16. Tomecki,R.,Dmochowska, A.,Gewartowski,K.,Dziembowski, A.,and Stepien,P.P.(2004).Identi?cation of a novel human nuclear-encoded mito-chondrial poly(A)polymerase.Nucleic Acids Res.32,6001–6014. Ustianenko,D.,Hrossova,D.,Potesil,D.,Chalupnikova,K.,Hrazdilova,K., Pachernik,J.,Cetkovska,K.,Uldrijan,S.,Zdrahal,Z.,and Vanacova,S. (2013).Mammalian DIS3L2exoribonuclease targets the uridylated precursors of let-7miRNAs.RNA19,1632–1638.

Vilfan,I.D.,Tsai,Y.C.,Clark,T.A.,Wegener,J.,Dai,Q.,Yi,C.,Pan,T.,Turner, S.W.,and Korlach,J.(2013).Analysis of RNA base modi?cation and structural rearrangement by single-molecule real-time detection of reverse transcription. J.Nanobiotechnology11,8.

Wang,L.,Eckmann,C.R.,Kadyk,L.C.,Wickens,M.,and Kimble,J.(2002).A regulatory cytoplasmic poly(A)polymerase in Caenorhabditis elegans.Nature 419,312–316.

Wang,X.,Lu,Z.,Gomez,A.,Hon,G.C.,Yue,Y.,Han,D.,Fu,Y.,Parisien,M., Dai,Q.,Jia,G.,et al.(2014a).N6-methyladenosine-dependent regulation of messenger RNA stability.Nature505,117–120.Wang,Y.,Li,Y.,Toth,J.I.,Petroski,M.D.,Zhang,Z.,and Zhao,J.C.(2014b). N6-methyladenosine modi?cation destabilizes developmental regulators in embryonic stem cells.Nat.Cell Biol.16,191–198.

Wei,C.M.,Gershowitz,A.,and Moss,B.(1975).Methylated nucleotides block 50terminus of HeLa cell messenger RNA.Cell4,379–386.

Yeom,K.H.,Heo,I.,Lee,J.,Hohng,S.,Kim,V.N.,and Joo,C.(2011).Single-molecule approach to immunoprecipitated protein complexes:insights into miRNA uridylation.EMBO Rep.12,690–696.

Zhang,Z.,Theler,D.,Kaminska,K.H.,Hiller,M.,de la Grange,P.,Pudimat,R., Rafalska,I.,Heinrich,B.,Bujnicki,J.M.,Allain,F.H.,and Stamm,S.(2010). The YTH domain is a novel RNA binding domain.J.Biol.Chem.285,14701–14710.

Zheng,G.,Dahl,J.A.,Niu,Y.,Fedorcsak,P.,Huang,C.M.,Li,C.J.,Va?gb?, C.B.,Shi,Y.,Wang,W.L.,Song,S.H.,et al.(2013a).ALKBH5is a mammalian RNA demethylase that impacts RNA metabolism and mouse fertility.Mol.Cell 49,18–29.

Zheng,G.,Dahl,J.A.,Niu,Y.,Fu,Y.,Klungland,A.,Yang,Y.G.,and He,C. (2013b).Sprouts of RNA epigenetics:the discovery of mammalian RNA deme-thylases.RNA Biol.10,915–918.

Zhong,S.,Li,H.,Bodi,Z.,Button,J.,Vespa,L.,Herzog,M.,and Fray,R.G. (2008).MTA is an Arabidopsis messenger RNA adenosine methylase and interacts with a homolog of a sex-speci?c splicing factor.Plant Cell20, 1278–1288.

Cell158,August28,2014a2014Elsevier Inc.987