P ERS P ECTI V E

DNA methylation: old dog, new tricks?

npg

© 2014 Nature America, Inc. All rights reserved.

Cornelia G Spruijt1 & Michiel Vermeulen1–3 DNA methylation is an epigenetic modification that is generally associated with repression of transcription initiation at CpG-island promoters. Here we argue that, on the basis of recent high-throughput genomic and proteomic screenings, DNA methylation can also have different outcomes, including activation of transcription. This is evidenced by the fact that transcription factors can interact with methylated DNA sequences. Furthermore, in certain cellular contexts, genes containing methylated promoters are highly transcribed. Interestingly, this uncoupling between methylated DNA and repression of transcription seems to be particularly evident in germ cells and pluripotent cells. Thus, contrary to previous assumptions, DNA methylation is not exclusively associated with repression of transcription initiation. Epigenetics refers to changes in gene expression and phenotype that occur without changes in DNA sequence. Epigenetic regulation of gene expression is achieved at least partially through post-translational modifications of histone proteins or by chemical modifications of the DNA itself. DNA methylation was the first epigenetic modification to be described. In higher eukaryotes, DNA is methylated mainly on the position 5 carbon of the cytosine base, and it occurs most frequently on symmetrical CpG dinucleotides, although in embryonic stem cells and aging brain cells, non-CpG methylation is quite abundant1,2. In vertebrates, DNA methylation in promoter regions is generally associated with gene silencing in cis, and it can thus affect the cell’s repertoire of expressed genes, thereby influencing cellular phenotype. DNA methylation is catalyzed by DNA methyltransferases (DNMTs). In mammals, three DNMTs have been identified: the ‘maintenance’ methyltransferase DNMT1 and the ‘de novo’ methyltransferases DNMT3A and DNMT3B3. After DNA replication, both daughter duplexes contain one methylated and one nonmethylated strand. DNMT1 methylates the newly synthesized strand of these hemimethylated duplexes, thereby ensuring faithful transfer of DNA methylation patterns during cell division3. DNMT3A and DNMT3B generally catalyze methylation of DNA in a DNA replication– independent manner, but recent data have indicated that this functional distinction between DNMT1 and DNMTs 3A and 3B is not as strict as previously thought4,5. 1Department

of Molecular Cancer Research, University Medical Center Utrecht, Utrecht, the Netherlands. 2Department of Molecular Biology, Radboud Institute for Molecular Life Sciences, Radboud University Nijmegen, Nijmegen, the Netherlands. 3Cancer Genomics Netherlands, the Netherlands. Correspondence should be addressed to M.V. ([email protected]). Received 10 July; accepted 30 September; published online 5 November 2014; doi:10.1038/nsmb.2910

Dynamic DNA methylation patterns are very important during early development. During lineage commitment, differentiating cells are thought to methylate promoters of nontranscribed genes specific to other lineages to permanently silence them. In contrast, genes that are essential for lineage specification are kept nonmethylated. DNA methylation–mediated gene silencing is thought to involve multiple mechanisms that are still not completely understood. Methylation can directly interfere with the binding of transcription factors to DNA6–9. Methylated DNA also recruits transcriptionally repressive methyl-CpG (mCpG)-binding proteins (MBPs)10–13. Furthermore, DNA methylation can affect nucleosome positioning14. Nonmethylated CpG islands, which are defined as DNA stretches containing a high density of CpG dinucleotides, however, are bound by CXXC domain–containing activating complexes15–17. Recently, a large number of putative new methylcytosine (mC)-binding proteins have been identified, mainly through high-throughput screening. Many of these proteins are transcription factors that are generally not known to function as repressors of transcription. Furthermore, large-scale DNA methylation profiling has revealed that genes with methylated promoters are highly transcribed in certain cellular contexts. Here we review these recent studies and summarize the new insights that they have generated (Fig. 1). DNA methylation as a repressive epigenetic mark Knowledge of the link between DNA methylation and repression of transcription originated from experiments in which methylated and nonmethylated reporter constructs were transfected into mammalian cells and Xenopus oocytes18,19. Because the nonmethylated reporter gene is expressed at much higher levels compared to the methylated reporter in these assays10,11, the biological function of mC was believed to be exerted through proteins that differentially interact with C in its methylated and nonmethylated form. Three so-called MBP families were identified in the 1990s: the mCpG-binding domain (MBD)-containing family (MeCP2 and MBD1–MBD6); the Kaiso family, which binds to mC via C2H2 zinc-finger domains (Kaiso, ZBTB38 and ZBTB4); and the SET- and RING-associated domain (SRA) family, which contains only two members, UHRF1 and UHRF2 (refs. 20,21). Some of these proteins, such as MeCP2, MBD2 and UHRF1, bind to mCpG with high affinity in vitro in a DNA sequence–independent manner. The Kaiso family proteins bind mCpG in a sequence-specific manner22. Several mammalian proteins (MBD3, MBD5 and MBD6) carry an MBD according to sequence homology but do not bind with high affinity to methylated DNA in vitro23. Both MeCP2 and MBD2 associate with multisubunit protein complexes containing histone deacetylases (HDACs). Because histone deacetylase activity is linked with gene silencing24,25, this implies a

nature structural & molecular biology  VOLUME 21  NUMBER 11  NOVEMBER 2014

949

P ERS P ECTI V E a

Figure 1  Models of transcriptional regulation by DNA methylation. (a) The ‘old’ textbook model Silent gene Active gene describing how DNA methylation regulates transcription. Left, methylated CpG-island promoters (P) recruit transcriptionally repressive P Gene P Gene MBD proteins and prevent transcription-factor binding. Right, nonmethylated CpG islands are bound by transcription factors. (b) New models describing regulation of transcription by DNA methylation. Left, genes with methylated Silent gene Active gene CpG-island promoters are repressed by repressive MBD-containing complexes. In ? addition, methylation of an enhancer (E) can E P Gene E P Gene block binding of a transcription factor. Right, most active genes with nonmethylated CpGisland promoters are bound by CXXC domain– MBD containing activator complexes. In addition, H3K4me3 transcription factors bind to nonmethylated CXXC enhancers. Finally, gene bodies of active Methyl-sensitive transcription factor genes are highly methylated, and this serves to Methyl-CpG–binding transcription factor repress cryptic transcription. (c) Uncoupling E P Gene between DNA methylation and repression of Nonmethylated CpG transcription initiation. In some cases, such Methylated CpG as during early vertebrate development, some methylated promoters with low CpG density are actively transcribed. Transcriptionally repressive MBD proteins do not interact with these promoters, for yet-unknown reasons. Furthermore, some DNA sequences with low CpG density (including enhancers and promoters) can be bound by activating transcription factors. H3K4me3, trimethylated histone H3 Lys4, a promoter-associated histone mark associated with active transcription.

b

npg

© 2014 Nature America, Inc. All rights reserved.

c

role for MBD2 and MeCP2 in repression of transcription through recruitment of HDACs. MeCP2 has been shown to associate with the Sin3–HDAC and N-CoR–SMRT co-repressor complexes19,26. MBD2 and MBD3 interact, in a mutually exclusive manner, with the nucleosome-remodeling and histone deacetylase (NuRD) complex27. In luciferase reporter assays, recruitment of these proteins leads to transcriptional repression10,11,28. Many recent studies have clearly linked genome-wide promoter CpG-islands hypermethylation to gene silencing in various cell types29,30. Furthermore, reduction of DNA methylation levels in cells by 5-azadeoxycytidine treatment or by depleting DNMTs reactivates transcription in vivo31,32. These observations have led to a general textbook model in which promoter CpG-island methylation serves as a recruitment signal for transcriptionally repressive MBPs, which ultimately results in gene repression in cis. Diverse functions of proteins interacting with methylated DNA During the past decade, quantitative MS-based proteomics technology has emerged as a powerful tool to study important biological questions. Within the research field of epigenetics, this technology has been applied to identify specific interactions between nuclear proteins and epigenetic histone and DNA modifications, including mCpG. These studies uncovered a large number of proteins in different cell types, which show specificity for different mCpG-containing baits compared to their nonmethylated counterpart33–36. Although such in vitro MS-based approaches cannot discriminate direct from indirect interactors, a number of domains are consistently enriched in these experiments, thus suggesting that these domains directly read mCpG-containing DNA. In addition to the known mCpG-binding domains (MBD, zinc fingers and SRA), these include the homeobox and the winged-helix domains (including forkhead boxes). The interaction between the RFX5 winged-helix domain and mCpG-containing DNA has been characterized in detail by NMR, and the Kd of this interaction was determined to be in the low-micromolar range, indicating that this interaction may be biologically relevant36. 950

Although the above-mentioned MS-based proteomics studies are unbiased and can be applied to a variety of different cell lines or even tissues, most of the published mCpG interaction screens were performed with a small number of mCpG-containing baits. Recently, Hu and co-workers used a protein microarray consisting of 1,321 transcription factors and 210 cofactor proteins to investigate direct interactions with 154 different human promoter sequences, each of which contain at least one mCpG37. The authors identified 47 proteins that bind to one or more of the methylated promoter sequences. One of the identified mCpG- or CpG-binding transcription factors is KLF4, which was also described as a new mCpG-binding protein in a previously published MS-based study36. Hu et al. used conventional protein biochemistry, chromatin immunoprecipitation (ChIP)-bisulfite sequencing and luciferase reporter assays to further substantiate this finding37. Recently, the structure of the interaction between Klf4 and methylated DNA was solved38. KLF4 is one of the four Yamanaka reprogramming factors that can be used to generate induced pluripotent stem cells (iPSCs)39. Because many genomic KLF4-binding sites are close to binding sites for the other reprogramming factors (Oct4, Sox2 and Myc), KLF4 might have a pioneering role during the generation of iPSCs. When its expression is induced in somatic cells, KLF4 may bind to its methylated target genes and recruit factors to decondense chromatin and demethylate DNA. This may finally enable the other pluripotency factors to bind to their target genes to induce stem cell–specific gene expression. However, further studies are required to investigate this hypothesis. In addition to high-throughput screening methods, other more targeted studies have identified additional proteins with affinity for mCpG-containing DNA. Examples include ZFP57, C/EBPα and ZBTB4 (refs. 40–42). ZFP57 was originally discovered as a protein important for genomic imprinting43. In ZFP57-knockout mice, imprinted, silent alleles lose DNA methylation, and their expression is induced. These regions contain a consensus binding site for ZFP57, to which the protein can bind only when this binding site is methylated44. ZFP57 therefore seems to be essential for the transcriptional silencing of imprinted alleles via the recruitment of

VOLUME 21  NUMBER 11  NOVEMBER 2014  nature structural & Molecular biology

P ERS P ECTI V E Figure 2  Schematic representation of the different classes of DNA-binding domains that have been implicated in C, mC and hmC binding: CXXC15,34,36, MBD34,36,53,54, SRA34,36,54,58, forkhead box or winged helix34,36, C2H2 zinc finger22,34,36,37,40,44,83, helix-turn-helix (homeobox)34,36, basic leucine zipper36,41 and helix-loop-helix8,37,84. Both the sequence-dependent and the sequence-independent branches show domains with binding specificities for multiple DNA modifications.

Domain Specificity CXXC

MBD

Sequence independent

Forkhead box Winged helix ZBTB

Kdm2A, Kdm2B, Cxxc1, Cxxc5

15, 34, 36

mC

Mbd1 (5 nM), Mbd2 (60–100 nM), Mbd4 (220 nM)

34, 36, 54

36, 53, 54

mC + hmC Uhrf1 (hemi-mC, 110 nM; mC, 103 µM; hmC, 1.5 µM) 34, 36 54 36, 58 Uhrf2 C, mC

Rfx5 (mC, 3 µM), FoxK1, FoxK2

34, 36

C, mC

Kaiso, ZBTB38

22

Krüppel like

36, 37 C, mC Klf3, Klf4 (mC, 479 nM), Klf5 DNMT1 and Kap1–Setdb1 (ref. 44). C/EBPα zinc finger ZNF-ZFP 34, 84 C ZNF131, ZNF395 is a leucine zipper–­containing transcription 40, 44 C + mC ZFP57 (C, 16–80 nM; mC 8 nM) factor that is very important for lineage Helix-turn36 C Prox1 commitment of mammalian cells. Rishi and helix 34, 36 Hox2.4, Meis1, Dlx5, Zhx3, Pbx1, Pax6 mC co-workers recently showed that C/EBPα (homeobox) 34, 36 mC, hmC Zhx1, Zhx2 Sequence interacts with the cis-regulatory element 36 C Jun, Creb1, ATF6 dependent Basic (CRE; TGACGTCA) when it is methylated. leucine zipper C/EBPα 41 mC The authors further showed that in vivo bindHelix-looping of C/EBPα to such methylated elements C c-MYC, MAX 8, 83 helix is important for C/EBPα-mediated activaARNT2 mC 37 tion of target-gene transcription during ??? ??? differentiation of keratinocytes 41. These examples illustrate that a DNA binding event driven by methylation of CpG dinucleotides can result in a different silencing. Further in vivo studies, however, are clearly needed to functional outcome depending on the biological function of the substantiate this hypothesis. protein that binds this particular DNA sequence. These findings prompt a reevaluation of the original classifica- The plot thickens: hmC, fC and caC enter the stage tion of MBPs. The MBD and SRA families of MBPs predominantly Recently, the family of TET proteins was shown to convert mC to interact with mCpG in a DNA sequence–independent manner, and 5-hydroxymethylcytosine (hmC)46. HmC is particularly abundant their genome-wide binding profiles correlate with mCpG density12. in embryonic stem cells and in brain cells47,48. Further iterative Many zinc finger–containing transcription factors may, in addition to TET-meditated oxidation results in the formation of 5-formylcytosine binding their nonmethylated consensus DNA-binding sites, interact (fC) and 5-carboxycytosine (caC)49. Both fC and caC serve as subwith mCpGs in sequence contexts different from those at the known strates for thymine glycosylase (TDG)50, which, in combination with consensus DNA-binding sites. Three types of zinc-finger proteins the base-excision-repair machinery, forms an active DNA demethylawith putative mCpG-binding ability have been identified: the ZBTB- tion pathway (reviewed in ref. 51). POZ subfamily (including Kaiso), Krüppel-like zinc-finger proteins The function of oxidized mC is currently unclear. Several groups (including KLF4) and the ZNF-ZFP zinc finger–protein subfamily have therefore pursued the identification of readers for hmC, fC and (including ZFP57). Some of the mC-specific binding of these zinc caC36,52, and the number of DNA repair–associated proteins found to finger–containing proteins may be explained by the fact that a mC interact with hmC, fC and caC reinforces the proposed link between structurally resembles a T when viewed from the major groove of oxidized mC derivatives and active DNA demethylation. In addition the DNA helix. Methylation of a C in a particular DNA sequence can to DNA repair–associated proteins (including helicases and glycosythus reconstitute a TpG-containing consensus binding site for a zinc lases), transcription factors and chromatin-modifying enzymes were finger–containing transcription factor. This phenomenon has been found to interact specifically with mC derivatives. The number of shown, at least in vitro, for a number of proteins, including ZBTB4 identified readers for fC and caC greatly exceeds the number of readand RBP-J33,42. Recent work has revealed additional domains that ers for hmC. Furthermore, only limited overlap has been observed are capable of binding to mCpG in a sequence-dependent manner. with regard to proteins (in)directly interacting with each of the modiThese domains include the homeobox, forkhead box or winged helix fied C bases36. Most of the ‘classic’ MBD proteins have a lower affinity and basic leucine zipper (Fig. 2). Further studies are needed to char- for hmC compared to mC. The exception to this is MeCP2, which was acterize the Kd of these interactions and to determine their physi- reported to bind to hmC, albeit with a slightly lower affinity comological relevance, for example by using ChIP–bisulfite sequencing45. pared to mC53,54. Other groups, however, have not observed binding Furthermore, ChIP-sequencing experiments in DNMT-deficient cells of MeCP2 to hmC55,56. Mammalian MBD3, which does not bind mC or in cells in which DNA methylation is reduced by 5-­azadeoxycytidine with a high affinity, was reported to interact with hmC57, a finding treatment can reveal which proportion of the binding sites for a par- that other studies have failed to reproduce36,54. The SRA family of ticular transcription factor in the genome are DNA-methylation proteins, which consists of UHRF1 and UHRF2, binds to both mC driven. In any case, the repertoire of proteins that specifically inter- and hmC. Whereas UHRF1 binds with a similar affinity to mC and act with methylated DNA sequences clearly extends beyond the three hmC36,58, UHRF2 shows a clear preference for hmC36,59. classes of proteins that were originally reported some 20 years ago. Chemically, mC and hmC differ in both their polarity and their The known functions of these proteins are diverse, and this implies size. Whereas mC is hydrophobic, hmC can form hydrogen bonds. that the biology of DNA methylation, particularly on DNA stretches Furthermore, the hydroxymethyl group is more bulky than the methyl with relatively low CpG density, encompasses more than just gene group and needs to be accommodated in a larger binding pocket. domains

© 2014 Nature America, Inc. All rights reserved.

C2H2

References:

C

mC + hmC MeCP2 (mC, 10 nM; hmC, 260 nM) SRA

DNA-binding

npg

Examples (Kd)

nature structural & molecular biology  VOLUME 21  NUMBER 11  NOVEMBER 2014

951

P ERS P ECTI V E

npg

© 2014 Nature America, Inc. All rights reserved.

Apparently, the chemical differences between different oxidized mC derivatives result in quite distinct binding patterns. This is clear from examples such as RFX5, which potently binds mC and also interacts with C and fC but is strongly repelled by hmC and caC 36. Further biochemical and structural studies are required to decipher the molecular mechanisms underlying these observations. On the basis of the distinct biochemical properties of the oxidized mC derivatives and their different, mostly nonoverlapping readers, each of these modifications may have its own specific function(s). Given the seemingly very potent DNA-damage response that is triggered by fC and caC, we expect that these modifications mainly function as DNA demethylation intermediates, whereas hmC may also have a role in transcription regulation. Genome-wide methylation profiling: lots of data and surprises Numerous technologies can be used to profile DNA methylation patterns in cells. Of these methods, whole-genome bisulfite sequencing provides single-base-pair resolution and is therefore considered the gold standard for genome-wide DNA methylation profiling. The disadvantage of most bisulfite-based methods is that they cannot discriminate between mC and its oxidized derivatives. Additional methods, such as glucosyltransferase treatment of hmC to protect and enrich it, as well as oxidative bisulfite sequencing (oxBS-seq) and formyl chemically assisted bisulfite sequencing (fCAB-seq), have been developed to facilitate hmC and fC profiling60–62. A comprehensive review of genome-wide (hydroxy)methylation profiling results is beyond the scope of this perspective, but a number of major, surprising observations stand out. First of all, genome-wide profiling has revealed that gene bodies are methylated63–65. A correlation can be found between the density of gene-body methylation and gene expression, but the best correlation is between the density of gene-body methylation and replication timing 66. These results imply a possible link between DNA methylation in gene bodies and the regulation of gene expression or DNA replication. Evidence has shown that gene-body methylation inhibits transcription initiation from cryptic promoters67. Further studies are required to decipher the molecular mechanisms underlying these observations. Second, methylome profiling during early development has revealed that most of the differentially methylated regions (DMRs) in the genome map to enhancer sequences and not to promoters68–72. Furthermore, methylated promoters (i.e., with low CpG density) are not always silenced in germ cells and pluripotent cells but are sometimes actively transcribed73,74. These methylated genes are apparently not silenced by transcriptionally repressive MBD proteins but are instead bound by proteins that activate transcription or at least do not interfere with it74–76. Third, non-CpG methylation is prominent in embryonic stem cells and also in aging brain cells1,2. Non-CpG methylation appears to be not a random process but a regulated one that is abundant in certain cell types or tissues, thus indicating that this type of C methylation may have unique functions that are yet to be discovered. The overall conclusion of these studies is that the function of DNA methylation extends beyond promoter CpG methylation and includes regulation of enhancer activity and may include regulation of replication, among other things. Furthermore, even well-described methylation events such as promoter CpG methylation do not always result in gene silencing, especially when these promoters are characterized by a low CpG density. Outlook In this perspective, we have reviewed the increasingly complex biology of DNA methylation in light of results from recent high-throughput 952

proteomic and genomic approaches. The observations made in these studies indicate that the previously assumed strict correlation between DNA methylation and repression of transcription is in fact context dependent77,78. This is obvious from proteomic screens, which revealed that many transcription factors interact with mCpG-containing DNA sequences. This implies that in certain cases methylation of a CpG dinucleotide can induce activation of transcription initiation rather than repression, depending on the nature of the reader that interacts with this particular DNA sequence79,80. We hypothesize that transcription-factor binding to methylated DNA may be particularly relevant on enhancers or other DNA elements with a low CpG density. On DNA stretches with a high CpG density, we expect that repressive MBDs will usually be the dominant mCpG readers (Fig. 1c). Interestingly, interactions with methylated DNA are highly dynamic during cellular differentiation. For example, the transcriptionally repressive MBD2–NuRD complex does not interact with methylated DNA in nuclear extracts derived from mouse embryonic stem cells grown in 2i medium but does interact with methylated DNA in neuronal precursor cells or adult mouse brain nuclear extracts36. The molecular mechanism underlying this observation is not clear yet, but it may involve an isoform switch for MBD2, which was recently shown to be important for lineage commitment and differentiation81. Other explanations could be lower expression levels of MBD2 or posttranslational modifications that inhibit mCpG binding, as previously demonstrated for MeCP2 (ref. 82). Also relevant in this context are the previously mentioned observations from a recent methylome profiling study in adult male germ cells, which revealed that many methylated promoters are in fact highly transcribed74. It should be noted, however, that these highly transcribed methylated promoters are characterized by a relatively low CpG content. In any case, a (partial) uncoupling between DNA methylation and transcriptional repression apparently exists in male germ cells; this may be explained by a low abundance of transcriptionally repressive readers or their inability to interact with methylated DNA in those cells. Interestingly, this (partial) uncoupling between DNA methylation and repression of transcription is also evident during early vertebrate development (ref. 75 and M.V., unpublished observations). As more putative mCpG-interacting proteins are identified, the question is raised of how binding specificity among different mCpG binders is achieved. Eventually, this may come down to affinity and protein abundance. Thus, in order to understand the biology of mCpG dynamics and its interactors, global quantitative methods are required to determine direct interactions and their Kd values for genomic mCpG-containing sequences. An important method to also use in this context is ChIP–bisulfite sequencing, to prove that a protein of interest binds to methylated DNA sequences in vivo. These approaches have to be complemented with global profiling of absolute protein abundance. Eventually, such approaches will allow quantitative modeling of mCpG and its functions in different sequence contexts during cellular differentiation and transformation. Acknowledgments We would like to thank O. Bogdanovic, A. Brinkman, S. Kloet and A. Smits for critical reading of the manuscript. The Vermeulen laboratory is supported by grants from the Netherlands Organisation for Scientific Research (NWO-VIDI (no. 864.09.003) and Cancer Genomics Netherlands) and a European Research Council starting grant (no. 309384). COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests. Reprints and permissions information is available online at http://www.nature.com/ reprints/index.html.

VOLUME 21  NUMBER 11  NOVEMBER 2014  nature structural & Molecular biology

npg

© 2014 Nature America, Inc. All rights reserved.

P ERS P ECTI V E 1. Ramsahoye, B.H. et al. Non-CpG methylation is prevalent in embryonic stem cells and may be mediated by DNA methyltransferase 3a. Proc. Natl. Acad. Sci. USA 97, 5237–5242 (2000). 2. Lister, R. et al. Global epigenomic reconfiguration during mammalian brain development. Science 341, 1237905 (2013). 3. Jurkowska, R.Z., Jurkowski, T.P. & Jeltsch, A. Structure and function of mammalian DNA methyltransferases. ChemBioChem 12, 206–222 (2011). 4. Arand, J. et al. In vivo control of CpG and non-CpG DNA methylation by DNA methyltransferases. PLoS Genet. 8, e1002750 (2012). 5. Liang, G. et al. Cooperativity between DNA methyltransferases in the maintenance methylation of repetitive elements. Mol. Cell. Biol. 22, 480–491 (2002). 6. Iguchi-Ariga, S.M. & Schaffner, W. CpG methylation of the cAMP-responsive enhancer/promoter sequence TGACGTCA abolishes specific factor binding as well as transcriptional activation. Genes Dev. 3, 612–619 (1989). 7. Campanero, M.R., Armstrong, M.I. & Flemington, E.K. CpG methylation as a mechanism for the regulation of E2F activity. Proc. Natl. Acad. Sci. USA 97, 6481–6486 (2000). 8. Prendergast, G.C. & Ziff, E.B. Methylation-sensitive sequence-specific DNA binding by the c-Myc basic region. Science 251, 186–189 (1991). 9. Blattler, A. & Farnham, P.J. Cross-talk between site-specific transcription factors and DNA methylation states. J. Biol. Chem. 288, 34287–34294 (2013). 10. Bakker, J., Lin, X. & Nelson, W.G. Methyl-CpG binding domain protein 2 represses transcription from hypermethylated -class glutathione S-transferase gene promoters in hepatocellular carcinoma cells. J. Biol. Chem. 277, 22573–22580 (2002). 11. Jiang, C.L. et al. MBD3L1 and MBD3L2, two new proteins homologous to the methyl-CpG-binding proteins MBD2 and MBD3: characterization of MBD3L1 as a testis-specific transcriptional repressor. Genomics 80, 621–629 (2002). 12. Baubec, T. et al. Methylation-dependent and -independent genomic targeting principles of the MBD protein family. Cell 153, 480–492 (2013). In this paper, a correlation between genome-wide MBD binding and mCpG density is shown. 13. Curradi, M. et al. Molecular mechanisms of gene silencing mediated by DNA methylation. Mol. Cell. Biol. 22, 3157–3173 (2002). 14. Huff, J.T. & Zilberman, D. Dnmt1-independent CG methylation contributes to nucleosome positioning in diverse eukaryotes. Cell 156, 1286–1297 (2014). 15. Thomson, J.P. et al. CpG islands influence chromatin structure via the CpG-binding protein Cfp1. Nature 464, 1082–1086 (2010). This is the first paper showing that nonmethylated CpG-rich sequences recruit CXXC domain–containing activator proteins. 16. Farcas, A.M. et al. KDM2B links the Polycomb Repressive Complex 1 (PRC1) to recognition of CpG islands. Elife 1, e00205 (2012). 17. Blackledge, N.P. et al. Variant PRC1 complex-dependent H2A ubiquitylation drives PRC2 recruitment and Polycomb domain formation. Cell 157, 1445–1459 (2014). 18. Kass, S.U., Landsberger, N. & Wolffe, A.P. DNA methylation directs a timedependent repression of transcription initiation. Curr. Biol. 7, 157–165 (1997). 19. Jones, P.L. et al. Methylated DNA and MeCP2 recruit histone deacetylase to repress transcription. Nat. Genet. 19, 187–191 (1998). 20. Defossez, P.A. & Stancheva, I. Biological functions of methyl-CpG-binding proteins. Prog. Mol. Biol. Transl. Sci. 101, 377–398 (2011). 21. Bogdanovic´  , O. & Veenstra, G.J. DNA methylation and methyl-CpG binding proteins: developmental requirements and function. Chromosoma 118, 549–565 (2009). 22. Filion, G.J. et al. A family of human zinc finger proteins that bind methylated DNA and repress transcription. Mol. Cell. Biol. 26, 169–181 (2006). 23. Laget, S. et al. The human proteins MBD5 and MBD6 associate with heterochromatin but they do not bind methylated DNA. PLoS ONE 5, e11982 (2010). 24. Laherty, C.D. et al. Histone deacetylases associated with the mSin3 corepressor mediate Mad transcriptional repression. Cell 89, 349–356 (1997). 25. Nagy, L. et al. Nuclear receptor repression mediated by a complex containing SMRT, mSin3A, and histone deacetylase. Cell 89, 373–380 (1997). 26. Lyst, M.J. et al. Rett syndrome mutations abolish the interaction of MeCP2 with the NCoR/SMRT co-repressor. Nat. Neurosci. 16, 898–902 (2013). 27. Le Guezennec, X. et al. MBD2/NuRD and MBD3/NuRD, two distinct complexes with different biochemical and functional properties. Mol. Cell. Biol. 26, 843–851 (2006). 28. Ng, H.H. et al. MBD2 is a transcriptional repressor belonging to the MeCP1 histone deacetylase complex. Nat. Genet. 23, 58–61 (1999). 29. Dong, S.M. et al. Promoter hypermethylation of multiple genes in carcinoma of the uterine cervix. Clin. Cancer Res. 7, 1982–1986 (2001). 30. Kang, S.H. et al. Transcriptional repression of the transforming growth factor- type I receptor gene by DNA methylation results in the development of TGF- resistance in human gastric cancer. Oncogene 18, 7280–7286 (1999). 31. Chiurazzi, P. et al. In vitro reactivation of the FMR1 gene involved in fragile X syndrome. Hum. Mol. Genet. 7, 109–113 (1998). 32. Robert, M.F. et al. DNMT1 is required to maintain CpG methylation and aberrant gene silencing in human cancer cells. Nat. Genet. 33, 61–65 (2003). 33. Bartels, S.J. et al. A SILAC-based screen for methyl-CpG binding proteins identifies RBP-J as a DNA methylation and sequence-specific binding protein. PLoS ONE 6, e25884 (2011). 34. Bartke, T. et al. Nucleosome-interacting proteins regulated by DNA and histone methylation. Cell 143, 470–484 (2010). 35. Mittler, G., Butter, F. & Mann, M. A SILAC-based DNA protein interaction screen that identifies candidate binding proteins to functional DNA elements. Genome Res. 19, 284–293 (2009).

36. Spruijt, C.G. et al. Dynamic readers for 5-(hydroxy)methylcytosine and its oxidized derivatives. Cell 152, 1146–1159 (2013). In this paper, a comprehensive catalog of readers for mC and hmC in different cell types is presented. 37. Hu, S. et al. DNA methylation presents distinct binding sites for human transcription factors. Elife 2, e00726 (2013). This paper reveals that a large number of transcription factors interact with DNA sequences containing methylated CpGs. 38. Liu, Y. et al. Structural basis for Klf4 recognition of methylated DNA. Nucleic Acids Res. 42, 4859–4867 (2014). 39. Lewitzky, M. & Yamanaka, S. Reprogramming somatic cells towards pluripotency by defined factors. Curr. Opin. Biotechnol. 18, 467–473 (2007). 40. Liu, Y. et al. An atomic model of Zfp57 recognition of CpG methylation within a specific DNA sequence. Genes Dev. 26, 2374–2379 (2012). 41. Rishi, V. et al. CpG methylation of half-CRE sequences creates C/EBPα binding sites that activate some tissue-specific genes. Proc. Natl. Acad. Sci. USA 107, 20311–20316 (2010). This paper shows that the transcription factor C/EBP binds to a methylated CRE sequence to activate transcription during differentiation. 42. Sasai, N., Nakao, M. & Defossez, P.A. Sequence-specific recognition of methylated DNA by human zinc-finger proteins. Nucleic Acids Res. 38, 5015–5022 (2010). 43. Mackay, D.J. et al. Hypomethylation of multiple imprinted loci in individuals with transient neonatal diabetes is associated with mutations in ZFP57. Nat. Genet. 40, 949–951 (2008). 44. Quenneville, S. et al. In embryonic stem cells, ZFP57/KAP1 recognize a methylated hexanucleotide to affect chromatin and DNA methylation of imprinting control regions. Mol. Cell 44, 361–372 (2011). 45. Brinkman, A.B. et al. Sequential ChIP-bisulfite sequencing enables direct genomescale investigation of chromatin and DNA methylation cross-talk. Genome Res. 22, 1128–1138 (2012). 46. Tahiliani, M. et al. Conversion of 5-methylcytosine to 5-hydroxymethylcytosine in mammalian DNA by MLL partner TET1. Science 324, 930–935 (2009). 47. Kriaucionis, S. & Heintz, N. The nuclear DNA base 5-hydroxymethylcytosine is present in Purkinje neurons and the brain. Science 324, 929–930 (2009). 48. Münzel, M. et al. Quantification of the sixth DNA base hydroxymethylcytosine in the brain. Angew. Chem. Int. Edn Engl. 49, 5375–5377 (2010). 49. Ito, S. et al. Tet proteins can convert 5-methylcytosine to 5-formylcytosine and 5-carboxylcytosine. Science 333, 1300–1303 (2011). 50. Maiti, A. et al. TDG excision of fC may be a predominant element of pathways for active DNA demethylation. FASEB J. 27, 758.6 (2013). 51. Wu, H. & Zhang, Y. Reversing DNA methylation: mechanisms, genomics, and biological functions. Cell 156, 45–68 (2014). 52. Iurlaro, M. et al. A screen for hydroxymethylcytosine and formylcytosine binding proteins suggests functions in transcription and chromatin regulation. Genome Biol. 14, R119 (2013). 53. Mellén, M. et al. MeCP2 binds to 5hmC enriched within active genes and accessible chromatin in the nervous system. Cell 151, 1417–1430 (2012). 54. Hashimoto, H. et al. Recognition and potential mechanisms for replication and erasure of cytosine hydroxymethylation. Nucleic Acids Res. 40, 4841–4849 (2012). 55. Valinluck, V. et al. Oxidative damage to methyl-CpG sequences inhibits the binding of the methyl-CpG binding domain (MBD) of methyl-CpG binding protein 2 (MeCP2). Nucleic Acids Res. 32, 4100–4108 (2004). 56. Khrapunov, S. et al. Unusual characteristics of the DNA binding domain of epigenetic regulatory protein MeCP2 determine its binding specificity. Biochemistry 53, 3379–3391 (2014). 57. Yildirim, O. et al. Mbd3/NURD complex regulates expression of 5hydroxymethylcytosine marked genes in embryonic stem cells. Cell 147, 1498–1510 (2011). 58. Frauer, C. et al. Recognition of 5-hydroxymethylcytosine by the Uhrf1 SRA domain. PLoS ONE 6, e21306 (2011). 59. Zhou, T. et al. Structural basis for hydroxymethylcytosine recognition by the SRA domain of UHRF2. Mol. Cell 54, 879–886 (2014). 60. Booth, M.J. et al. Quantitative sequencing of 5-methylcytosine and 5-hydroxymethylcytosine at single-base resolution. Science 336, 934–937 (2012). 61. Li, Y., Song, C.X., He, C. & Jin, P. Selective capture of 5-hydroxymethylcytosine from genomic DNA. J. Vis. Exp. 68, e4441 (2012). 62. Song, C.X. et al. Genome-wide profiling of 5-formylcytosine reveals its roles in epigenetic priming. Cell 153, 678–691 (2013). 63. Ogoshi, K. et al. Genome-wide profiling of DNA methylation in human cancer cells. Genomics 98, 280–287 (2011). 64. Aran, D. et al. Replication timing-related and gene body-specific methylation of active human genes. Hum. Mol. Genet. 20, 670–680 (2011). 65. Hellman, A. & Chess, A. Gene body-specific methylation on the active X chromosome. Science 315, 1141–1143 (2007). 66. Jjingo, D. et al. On the presence and role of human gene-body DNA methylation. Oncotarget 3, 462–474 (2012). 67. Maunakea, A.K. et al. Conserved role of intragenic DNA methylation in regulating alternative promoters. Nature 466, 253–257 (2010). 68. Hogart, A. et al. Genome-wide DNA methylation profiles in hematopoietic stem and progenitor cells reveal overrepresentation of ETS transcription factor binding sites. Genome Res. 22, 1407–1418 (2012).

nature structural & molecular biology  VOLUME 21  NUMBER 11  NOVEMBER 2014

953

P ERS P ECTI V E 76. Fouse, S.D. et al. Promoter CpG methylation contributes to ES cell gene regulation in parallel with Oct4/Nanog, PcG complex, and histone H3 K4/K27 trimethylation. Cell Stem Cell 2, 160–169 (2008). 77. Baubec, T. & Schubeler, D. Genomic patterns and context specific interpretation of DNA methylation. Curr. Opin. Genet. Dev. 25, 85–92 (2014). 78. Jones, P.A. Functions of DNA methylation: islands, start sites, gene bodies and beyond. Nat. Rev. Genet. 13, 484–492 (2012). 79. Boyes, J. & Bird, A. Repression of genes by DNA methylation depends on CpG density and promoter strength: evidence for involvement of a methyl-CpG binding protein. EMBO J. 11, 327–333 (1992). 80. Hsieh, C.L. Dependence of transcriptional repression on CpG methylation density. Mol. Cell. Biol. 14, 5487–5494 (1994). 81. Lu, Y. et al. Alternative splicing of MBD2 supports self-renewal in human pluripotent stem cells. Cell Stem Cell 15, 192–101 (2014). 82. Tao, J. et al. Phosphorylation of MeCP2 at Serine 80 regulates its chromatin association and neurological function. Proc. Natl. Acad. Sci. USA 106, 4882–4887 (2009). 83. Donaldson, N.S. et al. Kaiso regulates Znf131-mediated transcriptional activation. Exp. Cell Res. 316, 1692–1705 (2010). 84. Solomon, D.L., Amati, B. & Land, H. Distinct DNA binding preferences for the c-Myc/Max and Max/Max dimers. Nucleic Acids Res. 21, 5372–5376 (1993).

npg

© 2014 Nature America, Inc. All rights reserved.

69. Hon, G.C. et al. Epigenetic memory at embryonic enhancers identified in DNA methylation maps from adult mouse tissues. Nat. Genet. 45, 1198–1206 (2013). 70. Ziller, M.J. et al. Charting a dynamic DNA methylation landscape of the human genome. Nature 500, 477–481 (2013). 71. Xie, W. et al. Epigenomic analysis of multilineage differentiation of human embryonic stem cells. Cell 153, 1134–1148 (2013). 72. Kulis, M. et al. Epigenomic analysis detects widespread gene-body DNA hypomethylation in chronic lymphocytic leukemia. Nat. Genet. 44, 1236–1242 (2012). 73. Seisenberger, S. et al. The dynamics of genome-wide DNA methylation reprogramming in mouse primordial germ cells. Mol. Cell 48, 849–862 (2012). 74. Hammoud, S.S. et al. Chromatin and transcription transitions of mammalian adult germline stem cells and spermatogenesis. Cell Stem Cell 15, 239–253 (2014). This paper shows that many methylated CpG-island promoters in male germ cells are actively transcribed. 75. Bogdanovic, O. et al. Temporal uncoupling of the DNA methylome and transcriptional repression during embryogenesis. Genome Res. 21, 1313–1327 (2011). This paper reveals a temporal uncoupling between CpG-island methylation and repression of transcription during early Xenopus laevis development.

954

VOLUME 21  NUMBER 11  NOVEMBER 2014  nature structural & Molecular biology

DNA methylation: old dog, new tricks?

DNA methylation is an epigenetic modification that is generally associated with repression of transcription initiation at CpG-island promoters. Here w...
421KB Sizes 0 Downloads 6 Views