Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Genomic profiling of DNA methyltransferases reveals a role for DNMT3B in genic methylation


DNA methylation is an epigenetic modification associated with transcriptional repression of promoters and is essential for mammalian development. Establishment of DNA methylation is mediated by the de novo DNA methyltransferases DNMT3A and DNMT3B, whereas DNMT1 ensures maintenance of methylation through replication1. Absence of these enzymes is lethal2, and somatic mutations in these genes have been associated with several human diseases3,4. How genomic DNA methylation patterns are regulated remains poorly understood, as the mechanisms that guide recruitment and activity of DNMTs in vivo are largely unknown. To gain insights into this matter we determined genomic binding and site-specific activity of the mammalian de novo DNA methyltransferases DNMT3A and DNMT3B. We show that both enzymes localize to methylated, CpG-dense regions in mouse stem cells, yet are excluded from active promoters and enhancers. By specifically measuring sites of de novo methylation, we observe that enzymatic activity reflects binding. De novo methylation increases with CpG density, yet is excluded from nucleosomes. Notably, we observed selective binding of DNMT3B to the bodies of transcribed genes, which leads to their preferential methylation. This targeting to transcribed sequences requires SETD2-mediated methylation of lysine 36 on histone H3 and a functional PWWP domain of DNMT3B. Together these findings reveal how sequence and chromatin cues guide de novo methyltransferase activity to ensure methylome integrity.

This is a preview of subscription content

Access options

Buy article

Get time limited or full article access on ReadCube.


All prices are NET prices.

Figure 1: DNMT3A2 and DNMT3B1 show similar localization to most parts of the genome, but are differently recruited to gene bodies.
Figure 2: De novo methylation relative to DNMT binding and chromatin features.
Figure 3: DNMT3B binding follows co-transcriptional deposition of H3K36me3.
Figure 4: The H3K36 methyltransferase SETD2 and the PWWP domain of DNMT3B are required for targeting de novo methylation to transcribed genes.

Accession codes

Primary accessions

Gene Expression Omnibus

Data deposits

Genome-wide datasets generated for this study are deposited at GEO under the accession number GSE57413.


  1. Goll, M. G. & Bestor, T. H. Eukaryotic cytosine methyltransferases. Annu. Rev. Biochem. 74, 481–514 (2005)

    CAS  PubMed  Google Scholar 

  2. Okano, M., Bell, D. W., Haber, D. A. & Li, E. DNA methyltransferases Dnmt3a and Dnmt3b are essential for de novo methylation and mammalian development. Cell 99, 247–257 (1999)

    CAS  Article  PubMed  Google Scholar 

  3. Xu, G. L. et al. Chromosome instability and immunodeficiency syndrome caused by mutations in a DNA methyltransferase gene. Nature 402, 187–191 (1999)

    CAS  ADS  Article  PubMed  Google Scholar 

  4. Yan, X.-J. et al. Exome sequencing identifies somatic mutations of DNA methyltransferase gene DNMT3A in acute monocytic leukemia. Nature Genet. 43, 309–315 (2011)

    CAS  Article  PubMed  Google Scholar 

  5. Smith, Z. D. & Meissner, A. DNA methylation: roles in mammalian development. Nature Rev. Genet. 14, 204–220 (2013)

    CAS  Article  PubMed  Google Scholar 

  6. Baubec, T., Ivanek, R., Lienert, F. & Schübeler, D. Methylation-dependent and -independent genomic targeting principles of the MBD protein family. Cell 153, 480–492 (2013)

    CAS  Article  PubMed  Google Scholar 

  7. Chen, T., Ueda, Y., Dodge, J. E., Wang, Z. & Li, E. Establishment and maintenance of genomic methylation patterns in mouse embryonic stem cells by Dnmt3a and Dnmt3b. Mol. Cell. Biol. 23, 5594–5605 (2003)

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  8. Stadler, M. B. et al. DNA-binding factors shape the mouse methylome at distal regulatory regions. Nature 480, 490–495 (2011)

    CAS  ADS  Article  PubMed  Google Scholar 

  9. Ooi, S. K. T. et al. DNMT3L connects unmethylated lysine 4 of histone H3 to de novo methylation of DNA. Nature 448, 714–717 (2007)

    CAS  ADS  Article  PubMed  PubMed Central  Google Scholar 

  10. Tsumura, A. et al. Maintenance of self-renewal ability of mouse embryonic stem cells in the absence of DNA methyltransferases Dnmt1, Dnmt3a and Dnmt3b. Genes Cells 11, 805–814 (2006)

    CAS  Article  PubMed  Google Scholar 

  11. Felle, M. et al. Nucleosomes protect DNA from DNA methylation in vivo and in vitro. Nucleic Acids Res. 39, 6956–6969 (2011)

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  12. Kelly, T. K. et al. Genome-wide mapping of nucleosome positioning and DNA methylation within individual DNA molecules. Genome Res. 22, 2497–2506 (2012)

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  13. Krogan, N. J. et al. Methylation of histone H3 by Set2 in Saccharomyces cerevisiae is linked to transcriptional elongation by RNA polymerase II. Mol. Cell. Biol. 23, 4207–4218 (2003)

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  14. Tippmann, S. C. et al. Chromatin measurements reveal contributions of synthesis and decay to steady-state mRNA levels. Mol. Syst. Biol. 8, 593 (2012)

    Article  PubMed  PubMed Central  Google Scholar 

  15. Dhayalan, A. et al. The Dnmt3a PWWP domain reads histone 3 lysine 36 trimethylation and guides DNA methylation. J. Biol. Chem. 285, 26114–26120 (2010)

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  16. Kungulovski, G. et al. Application of histone modification-specific interaction domains as an alternative to antibodies. Genome Res. 24, 1842–1853 (2014)

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  17. Shirohzu, H. et al. Three novel DNMT3B mutations in Japanese patients with ICF syndrome. Am. J. Med. Genet. 112, 31–37 (2002)

    Article  PubMed  Google Scholar 

  18. Qiu, C., Sawada, K., Zhang, X. & Cheng, X. The PWWP domain of mammalian DNA methyltransferase Dnmt3b defines a new family of DNA-binding folds. Nature Struct. Biol. 9, 217–224 (2002)

    CAS  PubMed  Google Scholar 

  19. Qin, S. & Min, J. Structure and function of the nucleosome-binding PWWP domain. Trends Biochem. Sci. 39, 536–547 (2014)

    CAS  Article  PubMed  Google Scholar 

  20. Smallwood, S. A. et al. Dynamic CpG island methylation landscape in oocytes and preimplantation embryos. Nature Genet. 43, 811–814 (2011)

    CAS  Article  PubMed  Google Scholar 

  21. Jin, B. et al. Linking DNA methyltransferases to epigenetic marks and nucleosome structure genome-wide in human tumor cells. Cell Reports 2, 1411–1424 (2012)

    CAS  ADS  Article  PubMed  Google Scholar 

  22. Yang, X. et al. Gene body methylation can alter gene expression and is a therapeutic target in cancer. Cancer Cell 26, 577–590 (2014)

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  23. Pastor, W. A. et al. Genome-wide mapping of 5-hydroxymethylcytosine in embryonic stem cells. Nature 473, 394–397 (2011)

    CAS  ADS  Article  PubMed  PubMed Central  Google Scholar 

  24. Gilsbach, R. et al. Dynamic DNA methylation orchestrates cardiomyocyte development, maturation and disease. Nature Commun. 5, 5288 (2014)

    CAS  ADS  Article  Google Scholar 

  25. Hovestadt, V. et al. Decoding the regulatory landscape of medulloblastoma using DNA methylation sequencing. Nature 510, 537–541 (2014)

    CAS  ADS  Article  PubMed  Google Scholar 

  26. Tweedie, S., Charlton, J., Clark, V. & Bird, A. Methylation of genomes and genes at the invertebrate–vertebrate boundary. Mol. Cell. Biol. 17, 1469–1475 (1997)

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  27. Zemach, A., McDaniel, I. E., Silva, P. & Zilberman, D. Genome-wide evolutionary analysis of eukaryotic DNA methylation. Science 328, 916–919 (2010)

    CAS  ADS  Article  PubMed  Google Scholar 

  28. Nanty, L. et al. Comparative methylomics reveals gene-body H3K36me3 in Drosophila predicts DNA methylation and CpG landscapes in other invertebrates. Genome Res. 21, 1841–1850 (2011)

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  29. Suzuki, M. M. & Bird, A. DNA methylation landscapes: provocative insights from epigenomics. Nature Rev. Genet. 9, 465–476 (2008)

    CAS  Article  PubMed  Google Scholar 

  30. Jones, P. A. Functions of DNA methylation: islands, start sites, gene bodies and beyond. Nature Rev. Genet. 13, 484–492 (2012)

    CAS  Article  PubMed  Google Scholar 

  31. Bibel, M., Richter, J., Lacroix, E. & Barde, Y.-A. Generation of a defined and uniform population of CNS progenitors and neurons from mouse embryonic stem cells. Nature Protocols 2, 1034–1043 (2007)

    CAS  Article  PubMed  Google Scholar 

  32. Cong, L. et al. Multiplex genome engineering using CRISPR/Cas systems. Science 339, 819–823 (2013)

    CAS  ADS  Article  PubMed  PubMed Central  Google Scholar 

  33. Hsu, P. D. et al. DNA targeting specificity of RNA-guided Cas9 nucleases. Nature Biotechnol. 31, 827–832 (2013)

    CAS  Article  Google Scholar 

  34. Vezzoli, A. et al. Molecular basis of histone H3K36me3 recognition by the PWWP domain of Brpf1. Nature Struct. Mol. Biol. 17, 617–619 (2010)

    CAS  Article  Google Scholar 

  35. Wen, H. et al. ZMYND11 links histone H3.3K36me3 to transcription elongation and tumour suppression. Nature 508, 263–268 (2014)

    CAS  ADS  Article  PubMed  PubMed Central  Google Scholar 

  36. Wu, H. et al. Structural and histone binding ability characterizations of human PWWP domains. PLoS ONE 6, e18919 (2011)

    ADS  Article  PubMed  PubMed Central  Google Scholar 

  37. Lawrence, M. et al. Software for computing and annotating genomic ranges. PLOS Comput. Biol. 9, e1003118 (2013)

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  38. Hackenberg, M. et al. CpGcluster: a distance-based algorithm for CpG-island detection. BMC Bioinformatics 7, 446 (2006)

    Article  PubMed  PubMed Central  Google Scholar 

  39. Anders, S. & Huber, W. Differential expression analysis for sequence count data. Genome Biol. 11, R106 (2010)

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  40. Akalin, A., Franke, V., Vlahoviček, K., Mason, C. E. & Schübeler, D. genomation: a toolkit to summarize, annotate and visualize genomic intervals. Bioinformatics (2014)

  41. Teif, V. B. et al. Genome-wide nucleosome positioning during embryonic stem cell development. Nature Struct. Mol. Biol. 19, 1185–1192 (2012)

    CAS  Article  Google Scholar 

  42. Marks, H. et al. The transcriptional and epigenomic foundations of ground state pluripotency. Cell 149, 590–604 (2012)

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  43. Chen, T., Tsujimoto, N. & Li, E. The PWWP domain of Dnmt3a and Dnmt3b is required for directing DNA methylation to the major satellite repeats at pericentric heterochromatin. Mol. Cell. Biol. 24, 9048–9058 (2004)

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  44. Bachman, K. E., Rountree, M. R. & Baylin, S. B. Dnmt3a and Dnmt3b are transcriptional repressors that exhibit unique localization properties to heterochromatin. J. Biol. Chem. 276, 32282–32287 (2001)

    CAS  Article  PubMed  Google Scholar 

  45. Untergasser, A. et al. Primer3–new capabilities and interfaces. Nucleic Acids Res. 40, e115 (2012)

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  46. Jung, H. R., Pasini, D., Helin, K. & Jensen, O. N. Quantitative mass spectrometry of histones H3.2 and H3.3 in Suz12-deficient mouse embryonic stem cells reveals distinct, dynamic post-translational modifications at Lys-27 and Lys-36. Mol. Cell. Proteomics 9, 838–850 (2010)

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  47. Ferrari, K. J. et al. Polycomb-dependent H3K27me1 and H3K27me2 regulate active transcription and enhancer fidelity. Mol. Cell 53, 49–62 (2014)

    CAS  Article  PubMed  Google Scholar 

  48. Zhang, Y. et al. H3K36 histone methyltransferase Setd2 is required for murine embryonic stem cell differentiation toward endoderm. Cell Reports 8, 1989–2002 (2014)

    CAS  Article  PubMed  Google Scholar 

  49. Pei, J., Kim, B. H. & Grishin, N. V. PROMALS3D: a tool for multiple protein sequence and structure alignments. Nucleic Acids Res. 36, 2295–2300 (2008)

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  50. Jeong, S. et al. Selective anchoring of DNA methyltransferases 3A and 3B to nucleosomes containing Methylated DNA. Mol. Cell. Biol. 29, 5366–5376 (2009)

    CAS  Article  PubMed  PubMed Central  Google Scholar 

  51. Langmead, B., Trapnell, C., Pop, M. & Salzberg, S. L. Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biol. 10, R25 (2009)

    Article  PubMed  PubMed Central  Google Scholar 

Download references


We are grateful to S. Dessus-Babus, K. Jacobeit, T. Roloff (FMI), I. Nissen, and C. Beisel (ETH BSSE) for processing deep-sequencing samples. We thank R. Sack for processing histone H3K27/36 methylation measurements and analysis. We further thank L. Hoerner for technical assistance and M. Stadler and D. Gaidatzis for bioinformatic advice and providing tools to process and analyse deep-sequencing data. FMI imaging and FACS facilities for technical support. P. Ginno (FMI), N. Thomae (FMI), and M. Lorincz (UBC Vancouver) for advice and critical comments on the manuscript. Research in the laboratory of D.S. is supported by the Novartis Research Foundation, the European Union (NoE ‘‘EpiGeneSys’’ FP7- HEALTH-2010-257082 and the ‘‘Blueprint’’ consortium FP7-282510), the European Research Council (EpiGePlas), the SNF Sinergia program, and the Swiss initiative in Systems Biology (RTD Cell Plasticity). J.S. acknowledges support from the Boehringer Ingelheim Fonds. T.B. and A.R.K. acknowledge support from EMBO postdoctoral long-term fellowships.

Author information

Authors and Affiliations



T.B. initiated the study; T.B. and D.S. designed experiments; cell lines were generated by T.B. and D.F.C.; genomics experiments were performed by T.B.; interaction experiments were performed by T.B. and C.W.; genome-wide binding and methylation data was analysed by T.B. and L.B.; T.B., A.R.K., J.S., A.A. and L.B. designed/provided experimental and analytical tools; the manuscript was prepared by T.B. and D.S. All authors discussed the results and commented on the manuscript.

Corresponding authors

Correspondence to Tuncay Baubec or Dirk Schübeler.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Extended data figures and tables

Extended Data Figure 1 Biotin tagging of DNMT3A2 and DNMT3B in mouse embryonic stem cells.

a, Outline of the recombinase-assisted mapping of biotin-tagged proteins (RAMBiO) approach6. Expression constructs are inserted at a defined genomic site using recombinase-mediated cassette exchange (RMCE). Tagged proteins are biotinylated in vivo by the constitutively expressed biotin ligase Bir-A. ChIP is performed using streptavidin-based immunoprecipitation. Protein variants, inserted at the same genomic locus and under the same promoter can be compared to their wild-type counterparts in a controlled manner. b, Immunoblot for DNMT3A2 and DNMT3B1 to monitor expression levels of endogenous and biotinylated DNMT proteins. Detection was performed with antibodies against DNMT3A and DNMT3B using whole ES cell extracts loaded in two different concentrations (1× and 3×). LaminB1 served as a loading control. c, Nuclear localization of biotin-tagged DNMT3A2 and DNMT3B1 detects localization at DAPI-dense chromocentres, as previously reported for their endogenous counterparts43,44. Detection of biotinylated proteins was performed by streptavidin-AF568. d, Scatterplots showing genome-wide correlations between replicates of DNMT3A2 and DNMT3B1 calculated based on log2-transformed read counts in 1-kb sized windows covering the entire genome but excluding highly repetitive regions (n = 1,995,777 windows). Pearson’s correlation is shown. e, Dendrogram shows hierarchical clustering of genome-wide DNMT3A2 and DNMT3B1 enrichments over input in 1-kb sized windows covering the entire genome. f, Immunoblot showing protein levels of biotin-tagged DNMT3A2, expressed from the RMCE site using CAG or CMV promoters. g, Scatterplot showing correlation in genome-wide binding between DNMT3A2 expressed at different levels (CAG and CMV) calculated based on log2-transformed read counts in 1-kb sized windows covering the entire genome (n = 1,995,777 windows). This indicates similar binding, despite different expression levels.

Extended Data Figure 2 Genomic binding preferences of biotin-tagged DNMTs.

a, DNMT3A2 and DNMT3B1 enrichment follows methyl-CpG density. Shown are methyl-CpG density per 100 bp (grey) and enrichments for DNMT3A2 (red) or DNMT3B1 (blue) at 1-kb windows of the entire genome ranked by CpG density (n = 1,995,777 windows). Dashed line displays log2-transformed CpG density. Smoothing was applied using 200 consecutive windows. Genomic windows overlapping with unmethylated CpG islands are indicated by a green square. b, Enrichment of genomic features at DNMT3A2 and DNMT3B1-bound sites indicate an increased preference for gene bodies by DNMT3B. Shown are enrichments of different genomic features within DNMT-enriched sites (DNMT3A2, n = 6,499; DNMT3B1, n = 4,799) compared to their average genomic distribution. c, Boxplots showing the distribution of various genomic and epigenomic features at sites preferentially bound by DNMT3A2 (red, n = 6,499) or DNMT3B1 (blue, n = 4,799). Genome-wide distribution is shown in grey for comparison. Box indicates the interquartile range (IQR) and whiskers denote the 1.5× IQR range. Statistical differences between DNMT3A2 and DNMT3B1 are calculated using Mann–Whitney–Wilcoxon (MWW) tests (P < 2.2 × 10−16). Strongest differences between DNMT3A2 and DNMT3B1-enriched sites are found for H3K36me3 and mRNA. d, Unsupervised clustering highlights similarities and individual preferences of the de novo DNMTs to genomic sites with unique chromatin and DNA methylation features extracted from publicly available data sets. Cross-correlation matrix is calculated based on Pearson’s correlation coefficient derived from 1-kb windows covering chromosome 19 (n = 100,477). e, Genome-wide correlations between DNMT3A2 or DNMT3B1 enrichment over input (log2) and H3K36me3 enrichment over input (log2) using 1-kb windows covering the entire genome (n = 1,995,777). Pearson’s correlation is shown.

Extended Data Figure 3 De novo methylation follows binding preferences of DNMTs and indicates processivity and substrate accessibility.

a, Immunoblot showing expression levels of biotinylated DNMT3A2 and DNMT3B1 proteins in wild-type and TKO cells. Detection was performed using antibodies against DNMT3A or DNMT3B on whole cell extracts. Lamin B1 serves as loading control. Empty indicates wild-type Bir-A cells without biotin-tagged DNMTs. b, Scatterplots showing correlations between wild-type ES cells and TKO ES cells for DNMT3A2 and DNMT3B1 binding, respectively. Pearson’s correlation coefficients are calculated based on log2-transformed read counts in 1-kb windows covering the genome (n = 1,995,777). c, Global 5-methylcytosine levels measured by HPLC-MS in Dnmt-TKO cells, TKO cells with reintroduced DNMT3 isoforms and wild-type ES cells. Displayed are the fractions of methylated cytosines from two biological replicates for DNMT3A2 and DNMT3B1, respectively. DNMT3B6, a catalytic inactive isoform served as a negative control. Error bars indicate standard deviation calculated from three technical replicate measurements. d, Percentage of methylated cytosines (CpG only) in TKO cells expressing DNMT3A2 (red) or DNMT3B1 (blue) respectively. Calculation is based on whole-genome bisulphite sequencing using CpGs covered at least ten times (DNMT3A2, n = 7.94 × 106 CpGs; DNMT3B1, n = 7.35 × 106 CpGs). Dashed line indicates non-conversion rates based on spiked-in, unmethylated lambda phage DNA. e, De novo methylation follows genomic binding of DNMT3A2 and DNMT3B1. Shown is median DNA methylation calculated using 1-kb sized windows ranked by DNMT3A2 or DNMT3B1 enrichments. Equal numbers of genomic windows were binned (433 per bin) after DNMT3-enrichment ranking, and median methylation was calculated per bin (n = 1,000 bins). Only windows where at least 50% CpGs were covered more than 10 times in both experiments were used. f, Distance-dependent correlation of methylation levels between neighbouring CpGs indicates similar methylation for pairs of CpGs in close proximity and those at a 180 bp distance. g, Single-molecule analysis shows distance-dependent methylation of neighbouring CpGs on the same DNA strand. Shown is the fraction of neighbouring CpGs at a given distance, where both consecutive CpGs on the same DNA molecule are methylated. Grey dots indicate the same analysis but using randomized methylation states between reads as a control. h, Average density plots centred around strong CTCF binding sites (n = 4,126 sites) showing CTCF binding in wild-type (green) and TKO cells (orange)8. Nucleosome occupancy surrounding these CTCF sites is shown in black and was calculated based on MNaseI-seq in mouse ES cells41. i, De novo methylation in TKO cells occurs preferentially at linker regions between positioned nucleosomes, here illustrated around genomic sites bound by CTCF. Shown is average DNA methylation by DNMT3B1 in TKO cells.

Source data

Extended Data Figure 4 De novo DNA methylation and binding of DNMT3B to gene bodies correlates with transcriptional activity.

a, Boxplots showing the distribution of de novo DNA methylation at genomic regions bound by DNMT3A2 (n = 6,499), DNMT3B1 (n = 4,799) and genomic windows (n = 4,000, size = 1 kb) enriched for various chromatin and sequence-based features. Regions were selected as top 1% genomic windows enriched for the indicated feature and sufficiently covered in both bisulphite sequencing experiments. Grey boxplot indicates de novo methylation using a set of randomized genomic windows covered in both bisulphite sequencing experiments (n = 4,000 windows, size = 1 kb). The median of this randomized sample sets the baseline (dashed grey line). b, Relationship between transcriptional activity and de novo DNA methylation at gene bodies. Scatterplots comparing de novo DNA methylation in TKO cells expressing DNMT3A2 or DNMT3B1 with H3K36me3, RNA pol II enrichment or mRNA levels at gene bodies. Only gene bodies, sufficiently covered in both bisulphite sequencing experiments were used (n = 5,686). c, Scatterplot indicating binding of DNMT3B to actively transcribed gene bodies (n = 10,257). Shown is DNMT3 enrichment compared to H3K36me3 enrichment at gene bodies in log2 scale. Pearson’s correlation coefficient and Loess curves are shown (b and c).

Extended Data Figure 5 Methylation analysis at selected sites indicates a role for DNMT3B in methylation fidelity at transcribed genes.

a, Scheme illustrating the targeted bisulphite sequencing approach. Primers for selected genomic regions (b) were designed using Primer3 (ref. 45), genomic DNA from wild-type and mutant ES cells was PCR-amplified after bisulphite conversion. PCR products were pooled and sequenced on an Illumina MiSeq machine. b, Selected genomic regions analysed by targeted bisulphite sequencing (Supplementary Table 1). The following sites were selected: methylated regions with reduced binding of DNMT3A2 and DNMT3B1 (unbound, n = 25, grey), shared binding by DNMT3A2 and DNMT3B1 (shared 3A2/3B1, n = 35, red), and preferential binding of DNMT3B1 (DNMT3B bound, n = 21, blue). Shown are log2-transformed reads indicating DNMT3A2/3B1 binding at the selected regions. c, DNA methylation at a subset of transcriptionally active, DNMT3B1-bound sites is reduced in Dnmt3b-KO, but not in Dnmt3a-KO ES cells. Shown are average methylation values from bisulphite PCR analysis at the genomic sites defined in b and sufficiently covered in bisulphite sequencing experiments. Analysis was performed in wild-type ES cells (grey) and ES cells lacking DNMT3A (red) or DNMT3B (blue). P values were calculated using MWW tests. d, Heatmap showing DNA methylation at single CpGs within analysed regions for different mutants including Dnmt-TKO cells as control (two technical replicates each). Clustering was performed based on methylation of single CpGs. e, Genomic examples for sites with reduced DNA methylation in Dnmt3b-KO cells. Top track shows endogenous methylation in wild-type ES cells (red dots) and CpG density per 100 bp (grey). H3K36me3 and DNMT3B1 binding are shown as library-normalized read counts per 100 bp. Percent DNA methylation per individual CpGs within amplified regions in Dnmt3a-KO and Dnmt3b-KO cells is shown below. f, Average DNA methylation at analysed sites in Dnmt3a or Dnmt3b KO cells compared to H3K36me3 enrichment or mRNA levels (log2-transformed) in wild-type ES cells. Black line indicates average as Loess curves.

Extended Data Figure 6 H3K36me3 guides binding of DNMT3B to active gene bodies.

a, Boxplots displaying enrichment for DNMT3A2 and DNMT3B1 at exons and introns that are grouped according to mRNA levels. Boxes denote the IQR and the lowest/highest values within 1.5× IQR are shown as whiskers. b, Scatterplots comparing H3K36me3 enrichment (log2) to DNMT3A2 and DNMT3B1 enrichments (log2) separately for exons and introns. Pearson’s correlation coefficient is shown. c, Boxplots showing distribution of CpG densities (calculated per 100 bp) for exons and introns with lowest and highest expression. Shown are the first and fifth quantiles from a, based on mRNA levels. d, Adjusted r2 indicates prediction performance for DNMT3A2 and DNMT3B1 binding based on various genomic and epigenomic features at selected genomic sites. Prediction was performed using linear models on individual and combined values over a defined set of genomic regions (60,000 genomic sites of 1 kb, see Methods). e, Scatterplot comparing H3K36me3 enrichment at gene bodies (n = 10,257) with RNA pol II enrichment or mRNA levels (log2). Pearson’s correlation coefficient is indicated. f, Boxplots showing distribution of mRNA levels, H3K36me3 enrichment over input and CpG density per 100 bp at introns that are not transcribed (mRNA (−) and H3K36me3 (−), n = 41,656), transcribed but are not H3K36me3 positive (mRNA (+) and H3K36me3 (−), n = 267) or transcribed and H3K36me3 positive (mRNA (+) and H3K36me3 (+), n = 11,670). Boxes denote the IQR and lowest/highest values within 1.5× IQR are shown as whiskers. g, Boxplots for same introns fractions as in f, but showing DNMT3A2 and DNMT3B1 enrichments over input (log2). P values were calculated using MWW tests. h, Boxplots indicating de novo DNA methylation in Dnmt-TKO cells expressing either DNMT3A2 (red) or DNMT3B1 (blue) for the same intron fractions as in f and g. i, Average profiles around TSS, grouped based on transcriptional activity in neuronal progenitor cells (NP cells, 2,051 genes per bin). Shown is average DNA methylation in wild-type NP cells and average read counts indicating bio-DNMT3B1 binding in NP cells.

Extended Data Figure 7 Binding of DNMT3B to gene bodies is dependent on SETD2.

a, Immunoblot illustrating absence of RNA pol II phosphorylation at serine 2 (S2P) upon treatment with 75 µM DRB for 10 h. H3K36me3 is still present after this block in elongation. b, DNMT3B1 binds active genes in the absence of acute transcription. Scatterplots comparing H3K36me3 and DNMT3B1 enrichments (log2) at gene bodies in ES cells after 10 h treatment with 75 µM DRB. c, Relative quantification by MS shows reduction of H3K36me3 in Setd2-KO ES cells. Data are shown relative to the corresponding wild-type background (Dnmt-wild type or Dnmt-TKO cells). Measurements for H3K36me3 in combination with H3K27 methylation on the same tryptic peptide are shown. Previous quantitative MS46 and genome-wide measurements47 indicate that H3K36me3/K27me0 and H3K36me3/K27me1 overlap in mouse ES cells. d, CRISPR/Cas9 mediated deletion in exon 3 of the Setd2 gene, identified by mRNA-seq in wild type and Setd2-KO cells. e, mRNA levels of highly expressed genes are not severely altered in absence of SETD2, as previously observed48. Scatterplot shows log2-transformed mRNA read counts mapping to transcripts in wild type and Setd2-KO cells. Histograms indicate distribution of mRNA levels. Pearson’s correlation coefficient is shown. f, Localization of endogenous DNMT3B to centromeric regions is not altered in absence of SETD2. Dnmt-TKO is shown as control. Detection was performed using antibodies specific to DNMT3B. g, Chromatin immunoprecipitation of biotinylated DNMT3B1 followed by histone modification analysis by immunoblot. H3K36me3-chromatin is precipitated only in Setd2 wild-type cells, whereas lack of SETD2 does not alter global chromatin precipitation (measured by total H3 recovery). h, Deletion of Setd2 leads to loss of preferential binding of DNMT3B1 to active genes, but does not alter DNMT3A2 binding (see Extended Data Fig. 4c). Pearson’s correlation coefficients and trend lines calculated using Loess fit are shown. i, Scatterplots comparing DNMT3B with DNMT3A2 binding to gene bodies in wild-type and Setd2-KO cells. This indicates similarities in binding in absence of H3K36me3. Highlighted gene bodies indicate H3K36me3-dependent binding of DNMT3B in wild-type cells. DNMT3B enrichment in wild-type and Setd2-KO cells are shown at similar scale for better comparison. j, DNMT3B1 binding at H3K36me3-positive exons and introns depends on SETD2 (compare to Fig. 3a). Pearson’s correlation coefficient between DNMT3A2 (red) or DNMT3B1 (blue) enrichments and H3K36me3 or CpG densities at exons and introns is calculated for wild-type (+) and for Setd2-KO (−) cells. Note that CpG density-dependent binding is not altered. k, Heatmaps of mRNA levels, Pol II occupancy, DNA methylation density, H3K36me3 and DNMT3B1 levels at all non-overlapping genes longer than 4 kb (n = 10,257). Genes were separated in five clusters based on k-means clustering. l, Boxplots indicate properties of gene bodies clustered based on k. DNMT3B1 binding to individual transcribed gene bodies (clusters 1 and 2) is detectable in wild-type and DRB-treated cells, but lost in Setd2-KO cells. Cluster 3 contains gene bodies with elevated CpG densities that are bound by DNMT3A2 and DNMT3B1 independent of H3K36me3.

Source data

Extended Data Figure 8 Targeting of de novo methylation to H3K36me3 sites by DNMT3B is impaired in Setd2-KO cells.

a, Genomic regions illustrating reduced binding of DNMT3B1 to active genes in absence of SETD2 and H3K36me3, whereas DNMT3A2 remains unchanged. Shown are read counts per 100 bp. Location of CpG islands (UCSC) and genes (RefSeq) are displayed below. b, Deletion of Setd2 in Dnmt-TKO cells (Dnmt, Setd2-KO) that express DNMT3B1 results in loss of de novo methylation of sites that were positive for H3K36me3 (Compare to Extended Data Fig. 4b). Scatterplot comparing de novo DNA methylation in Dnmt, Setd2-KO cells expressing DNMT3B1 to H3K36me3 enrichments (log2) at gene bodies in wild-type cells. Pearson’s correlation and trend line (Loess) are shown. c, Boxplots illustrating the distribution of de novo methylation by DNMT3B1 in Dnmt, Setd2-KO cells at different genomic regions (compare to Extended Data Fig. 4a). d, Boxplots indicating requirement for SETD2 in DNMT3B-dependent de novo methylation at H3K36me3 sites, but not at CpG-dense sites. Shown are H3K36me3 enrichments and CpG densities at sites preferentially methylated in Dnmt-TKO cells expressing DNMT3A2 or DNMT3B1 and Dnmt, Setd2-KO cells expressing DNMT3B1. Grey box denotes genome-wide distribution of H3K36me3 and CpG density for comparison. Boxes denote IQR and whiskers 1.5× IQR. e, Protection from de novo methylation of LMRs and UMRs is also observed in Dnmt, Setd2-KO cells expressing DNMT3B1. Shown are average density plots centred around methylation segments8. Black line indicates DNMT3B1 enrichments in Dnmt-TKO cells. f, Preferential de novo methylation at linker regions around sites bound by CTCF in DNMT3B1-expressing Dnmt, Setd2-KO cells. Shown are average density plots for DNA methylation in Dnmt, Setd2-KO cells (blue) and positioned nucleosomes in wild-type ES cells (black) centred at strong CTCF sites (n = 4,126).

Extended Data Figure 9 The N-terminal part of DNMT3B interacts with H3K36me3-modified nucleosomes and requires a functional PWWP domain.

a, Protein model for DNMT3B indicating known domains and mutations introduced in this study. b, Superimposition of published PWWP domain structures from DNMT3B with Bis-Tris in blue (Bis-Tris not shown)36, ZMYND11 in presence of K36me3 in green35 and BRPF1 in presence of K36me3 in orange34 shows structural similarities between the analysed PWWP domains. c, Close-up of the aromatic cage involved in K36me3 interactions. Same structures as in b. d, Comparison of amino acid sequences of DNMT3B, BRPF1 and ZMYND11 PWWP domains using PROMALS3D49 based on structure PDB IDs (1KHC, 2X4W and 4N4H). Conservation index is shown above (max score = 9), amino acid (AA) and secondary structure (SS) consensus is shown below. Conserved amino acids are indicated in bold, aromatic amino acids are indicated as the @ symbol. Consensus structures: h = α-helix, e = β-sheet. Residues of the aromatic cage involved in K36me3 recognition and one mutated in ICF syndrome patients (S277P) are highlighted by black and red rectangles, respectively. e, HeLa mononucleosome interaction assays using GST-fused PWWP domain alone (amino acids 227 to 343) and N-terminal part of DNMT3B containing the PWWP domain amino acids 1 to 427 (N-PWWP) (see a). Interactions with H3K36me3 nucleosomes are observed only for the extended PWWP domain. In line with this finding, deletion of the N-terminal part of DNMT3B, preceding the PWWP domain has previously been shown to disrupt anchoring to nucleosomes50. GST alone is used as control. f, The PWWP domain in the N-terminal DNMT3B construct is required for interactions with H3K36me3-modified nucleosomes. The ICF (S277P) mutation and substitutions of residues of the aromatic cage (WP-ST) disrupt interactions with H3K36me3 modified nucleosomes. Nucleosomes modified at H3K27me3 or H2AK119ub were not enriched in the pull-downs.

Extended Data Figure 10 Mutations in the PWWP domain of DNMT3B1 disrupt chromosomal binding and preference for H3K36me3.

a, Immunoblot of wild-type and PWWP mutant DNMT3B1 proteins after bio-ChIP indicates that a functional PWWP domain is required for interactions with chromatin modified at H3K36me3. Mutations introduced in the PWWP domain (S277P/ICF and substitutions in the conserved residues VW) abolish interactions with H3K36me3. b, Cellular fractionation into cytosolic (cy), soluble nuclear (nu) extracted at 100 mM NaCl and chromatin (ch) extracted at 500 mM NaCl indicates that introduced mutations do not severely affect nuclear localization or chromatin association of the biotin-tagged DNMT3B protein. Detection was performed using streptavidin. Asterisk indicates endogenous biotinylated proteins. Tubulin and LaminB1 served as loading controls. c, Substitutions in conserved PWWP domain residues (VW-RR, orange) result in loss of binding to active gene bodies (here shown at genes with high intragenic H3K36me3 levels, n = 2,051). A mutation within a flexible domain (S280P, green), three amino acids downstream of the ICF mutation, has no effect on H3K36me3 interactions, suggesting that the S277 residue, mutated in ICF patients is critical for this interaction (compare to Fig. 4d). d, Scatterplots showing the relation between H3K36me3 enrichments and enrichment of DNMT3B1 mutant proteins at gene bodies in wild-type ES cells. Pearson’s correlation coefficient is shown. e, Pearson’s correlation coefficient for DNMT3B1 mutant protein enrichments versus H3K36me3 levels at exons or introns (compare to Fig. 3a). ‘DRB’ indicates wild-type DNMT3B in cells treated with the pol II elongation inhibitor DRB (10 h, 75 µM).

Supplementary information

Supplementary Table 1

This spreadsheet contains information relevant to Extended Data Figure 5. Sheet 1 contains detailed bisulphite PCR primer information, genomic location, chromatin and DNA methylation measurements for regions of interest (panels: b, c and f of Extended Data Figure 5). Sheet 2 contains measured DNA methylation for individual CpGs in regions of interest (panel d). (XLSX 134 kb)

PowerPoint slides

Source data

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Baubec, T., Colombo, D., Wirbelauer, C. et al. Genomic profiling of DNA methyltransferases reveals a role for DNMT3B in genic methylation. Nature 520, 243–247 (2015).

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI:

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing