JBC Papers in Press. Published on June 11, 2010 as Manuscript M109.089433 The latest version is at http://www.jbc.org/cgi/doi/10.1074/jbc.M109.089433 THE DNMT3A PWWP DOMAIN READS HISTONE 3 LYSINE 36 TRIMETHYLATION AND GUIDES DNA METHYLATION Arunkumar Dhayalan1, Arumugam Rajavelu1, Philipp Rathert2, Raluca Tamas3, Renata Z. Jurkowska1, Sergey Ragozin1 & Albert Jeltsch1,* From 1Biochemistry Laboratory, School of Engineering and Science, Jacobs University Bremen, Campus Ring 1, 28759 Bremen, Germany, 2Active Motif Europe, Av F.D. Roosevelt, 104, BE- 1330 Rixensart, Belgium and the 3MoLife graduate program, School of Engineering and Science, Jacobs University Bremen, Campus Ring 1, 28759 Bremen, Germany Running title: Dnmt3a PWWP interacts with H3K36me3 Address correspondence to Albert Jeltsch, Prof. Dr., School of Engineering and Science, Jacobs University Bremen, Campus Ring 1, 28759 Bremen, Germany, +49 421 200 3247, Fax: +49 421 D o 200 3249 [email protected], http://www.jacobs-university.de/ses/ajeltsch/ w n lo a d e d The Dnmt3a DNA methyltransferase H3K36me3 interaction increases the fro m h contains in its N-terminal part a PWWP activity of Dnmt3a for methylation of ttp ://w domain, which is involved in chromatin nucleosomal DNA as observed using w w .jb targeting. Here, we have investigated the native nucleosomes isolated from human c .o rg interaction of the PWWP domain with cells after demethylation of the DNA with b/ y g u modified histone tails using peptide 5-Aza-2’-deoxycytidine as substrate for e s t o n arrays and show that it specifically methylation with Dnmt3a. These data D e c e recognizes the histone 3 lysine 36 suggest that the interaction of the PWWP m b e trimethylation mark. H3K36me3 is domain with H3K36me3 is involved in r 2 3 , 2 0 known to be a repressive modification targeting of Dnmt3a to chromatin 1 8 correlated with DNA methylation in carrying that mark, a model which is in mammals and heterochromatin in agreement with several studies on the Schizosaccharomyces pombe. These results genome wide distributions of DNA were confirmed by equilibrium peptide methylation and H3K36me3. binding studies and pull-down experiments with native histones and In mammals, DNA methylation plays purified native nucleosomes. The PWWP- important roles in differentiation, gene H3 K36me3 interaction is important for regulation, genomic imprinting, X the sub-nuclear localization of EYFP chromosome inactivation, and disease fused Dnmt3a. Furthermore, the PWWP- related processes (1-3). DNA methylation 1 Copyright 2010 by The American Society for Biochemistry and Molecular Biology, Inc. patterns are set during embryogenesis by the targeting of Dnmt3a (18). This mutation Dnmt3a and Dnmt3b DNA corresponds to the S282P mutation in the methyltransferases and their regulatory PWWP domain of human Dnmt3b, which factor Dnmt3-like (Dnmt3L) (4-6). However, has been identified in immunodeficiency, the mechanisms guiding these enzymes to centromeric heterochromatin instability, their target regions are not well understood. facial anomalies (ICF) syndrome patients Dnmt3a and 3b consist of a C-terminal (20). catalytic domain and an N-terminal part containing a PWWP domain and a PHD-like Here, we explore the possibility of the 3-Zn binding ADD domain (1,7,8). interaction of the Dnmt3a PWWP domain Biochemical studies provide evidence for a with histone peptides using modified histone direct interaction of Dnmt3a and 3b with tail peptide arrays. We discovered its D o w n native nucleosomes (9) which could be specific interaction with H3K36me3 and lo a d e mediated by the ADD domain or the PWWP show that this interaction is important for d fro m domain. The ADD domains of Dnmt3L and the sub-nuclear localization of Dnmt3a and h ttp Dnmt3a have been shown to interact with for its catalytic activity on native chromatin. ://w w w the H3 tail unmethylated at K4 (10-12), .jb c .o which can explain the anticorrelation of Experimental procedures rg b/ y DNA methylation and the activating g u e s H3K4me3 mark as observed in many Cloning, site directed mutagenesis, t o n D genome wide DNA methylation studies (13- expression and purification ec e m 16). However, the ADD domain is not The sequence encoding the PWWP domain be r 2 3 directly involved in heterochromatic of murine Dnmt3a (residues 279 – 420) was , 2 0 1 8 targeting of Dnmt3a (17,18). subcloned as GST fusion protein into the pGEX-6P2 vector (GE Healthcare) using PWWP domains belong to the Royal domain BamHI/XhoI sites. The GST-tagged PWWP super-family, members of which were domain and its variants were expressed and identified to interact with histone tails in purified as described (21). The full-length various modification states (19). The PWWP murine Dnmt3a2 cloned into pET28a domains of Dnmt3a and 3b are essential for (Novagen) vector was used for the heterochromatic targeting (17,18). An expression and purification of full-length S333P missense mutation in the Dnmt3a Dnmt3a2 (22), which is the predominant PWWP domain (numbering refers to murine isoform of Dnmt3a in embryonic stem cells Dnmt3a) led to the loss of chromatin (23). The catalytic domain of Dnmt3a 2 (Dnmt3a-CD) was expressed and purified as Binding of protein domains to peptides described (24). To prepare the EYFP- arrays Dnmt3a expression construct used for sub- The cellulose membrane was blocked by nuclear localization studies Dnmt3a was incubation in TTBS buffer (10 mM Tris/HCl cloned from a pET expression vector (25) pH 8.3, 0.05% Tween-20 and 150 mM into pcDNA3.1 yielding pcDNA-Dnmt3a. NaCl) containing 5% non-fat dried milk at Then, EYFP was amplified from pEYFP-C1 4 °C overnight. The membrane was then (Clonetech) and cloned into pcDNA- washed once with TTBS buffer, and Dnmt3a in front of the Dnmt3a gene to incubated with purified GST-tagged PWWP encode for an EGFP-Dnmt3a fusion protein. domain of Dnmt3a or mutant domains (10 nM) at room temperature for 1 hour in The E290A and D329A mutations in the interaction buffer (100 mM KCl, 20 mM D o w n PWWP domain of Dnmt3a and the D329A HEPES pH 7.5, 1 mM EDTA, 0.1 mM DTT lo a d e mutation in Dnmt3a2 and in EGFP-Dnmt3a and 10% glycerol). After washing in TTBS d fro m were introduced by using a PCR- buffer, the membrane was incubated with h ttp megaprimer mutagenesis method as goat anti-GST antibody (GE Healthcare ://w w w previously described (26). Successful #27- 4577-01 at 1:5000 dilution) in TTBS .jb c .o mutagenesis and cloning was confirmed by buffer for 1 h at room temperature. Then, the rg b/ y restriction marker site analysis and DNA membrane was washed three times with g u e s sequencing. TTBS and incubated with horseradish t o n D peroxidase conjugated anti-Goat antibody e c e m Synthesis of peptide SPOT arrays (Invitrogen #81-1620 1:12000) in TTBS for be r 2 3 Peptide arrays were synthesized on cellulose 1 h at room temperature. Finally, the , 2 0 1 8 membranes using the SPOT synthesis membrane was submerged in ECL method (27). Each spot had diameters of 2 developing solution (GE Healthcare) and mm and contained approximately 9 nmol of image was captured in X-ray film. peptide (Autospot Reference Handbook, Intavis AG). Successful synthesis of each Fluorescence depolarization peptide binding peptide was confirmed by bromophenol blue experiments staining of the membranes. H3 26-44 Fluorescence depolarization experiments K36me3 (with N-terminal fluorescein label) were carried out at 25 °C using a Varian and unmodified H3 26-44 peptides were Carry Eclipse fluorescence purchased from Intavis AG in purified form. spectrophotometer. For the determination of the binding constant of the PWWP domain 3 to the fluorophore coupled H3 K36me3 cell divisions under these conditions) and peptide, 0.1 µM of peptide was incubated the cells were allowed to recover for another with increasing concentrations of PWWP 48 h in medium without 5-aza-2’- domain (0.8 µM to 60 µM) in interaction deoxycytidine. From these cells, buffer (100 mM NaCl, 50 mM Tris/HCl pH nucleosomes were prepared as described 8.0, 2 mM EDTA, 0.1% Triton X 100) and above. fluorescence depolarization was determined (Excitaion at 494 nm, emission at 524 nm, GST pull-down experiments excitation and emission Slits at 5 nm). Data For GST pulldown experiments, 20 µl of were fitted to a binary binding equilibrium Glutathione Sepharose 4B beads (Amersham to determine the equilibrium binding Biosciences) was incubated with 20 µg of constant. For competition experiments, 30 GST-tagged PWWP domain of Dnmt3a or D o w n µM of fluorophore-coupled H3K36me3 was mutant domain for 1 hour at 4˚C in the lo a d e incubated with 30 µM of PWWP domain in interaction buffer (50 mM Tris/HCl pH 8.0, d fro m interaction buffer and their interaction was 100 mM NaCl, 2 mM EDTA and 0.1% h ttp competed by 2 and 5 fold molar excess of Triton X 100). Then, the beads were washed ://w w w unlabeled unmodified H3 26-44 peptide. once with interaction buffer and blocked .jb c .o with interaction buffer containing 5% BSA rg b/ y Preparation of native histones and for 1 hour at 4˚C. 35 µg of native histones or g u e s nucleosomes 100 µg native oligonucleosomes were added t o n D Native histones were isolated from HEK293 to the beads in interaction buffer and e c e m cells by acid extraction as described (28). incubated for 3 hours. Finally, the beads be r 2 3 Native nucleosomes were prepared from were washed 3 times with interaction buffer, , 2 0 1 8 HEK293 cells by partial micrococcal re-suspended in 2X SDS PAGE loading dye, nuclease (NEB) digestion of nuclei as boiled for 10 min and the supernatant described (29), except that after micrococcal containing the bound fractions was nuclease digestion the nuclei were removed separated on 16% SDS PAGE and by centrifugation and the supernatant transferred to nitrocellulose membrane. The containing nucleosomes (predominantly membrane was cut in two parts. The lower mononucleosomes) was used (Suppl. Fig. 3). part (<20 kDa) was probed with anti For the preparation of demethylated H3K36me3 antibody (Abcam ab9050) and nucleosomes, HEK293 cells were treated the upper part with anti-GST antibody (GE with 2 µM 5-aza-2’-deoxycytidine (Sigma- Healthcare) to provide a gel loading control Aldrich) for 72 h (corresponding to three (Supp. Fig 4A and 4B). 4 Confocal images were taken using a Zeiss For the specificity analysis, the beads were LSM 510 Meta instrument (software version loaded with GST-tagged PWWP domain 3.0) and 63x oil immersion objective. Argon and incubated with native oligonucleosomes laser line at 514 nm was used to excite in interaction buffer as described above, EYFP fluorescence and a BP530-550 filter washed 3 times with washing buffer (50 mM was used for image recording. Fluorescence Tris pH 8.0, 150 mM NaCl, 2 mM EDTA microscopy pictures were taken using an and 0.1% Triton X 100), re-suspended in 2X AXIOPLAN2 microscope equipped with SDS PAGE loading buffer and boiled for 10 AxioCam HRc camera and Achroplan 100x, min. The supernatant was split equally and oil immersion objective with filter sets loaded on 16% SDS PAGE gels. The DAPI FT395/LP420 and GFP BP450- separated proteins were transferred to 490/FT510 (all from Carl Zeiss). D o w n nitrocellulose membrane and probed with lo a d e different modification specific histone In vitro methylation of native chromatin d fro m antibodies: anti H3K36me3 (Abcam The methylation reactions were performed h ttp ab9050), anti H3K9me3 (Active Motif with either Dnmt3a2 enzyme, Dnmt3a2- ://w w w 39161), anti H3K4me3 (Active Motif D329A mutant or Dnmt3a-CD in .jb c .o 39159), and anti H3K27me3 (Active Motif methylation buffer (20 mM Hepes pH 7.5, rg b/ y 39535). Equal loading of bound fractions 2mM EDTA, 50 mM KCl) containing 0.75 g u e s was confirmed by Ponceau S staining of the µM radioactively labeled AdoMet (Perkin t o n D blot prior to the antibody staining (Suppl. Elmer) and 1 µg of demethylated native e c e m Fig. 4C). chromatin in 25 µL total reaction volume. be r 2 3 Additional methylation experiments were , 2 0 1 8 Cell culture, transfection and microscopy conducted in the presence of anti HEK293 and NIH3T3 cells were grown in H3K36me3 antibody (Abcam ab9050), anti DMEM with 5% fetal calf serum at 37°C at H3K9me3 antibody (Active Motif 39161), 5% CO . The NIH3T3 cells were seeded in anti H3K4me3 antibody (Active Motif 2 coverslips and transfected with EYFP- 39159), or anti H3K27me3 antibody (Active Dnmt3a or EYFP-Dnmt3a-D329A mutant Motif 39535) all in 1 in 5000 final dilution. constructs using FuGene6 (Roche) The methylation reactions were incubated at according to the manufacturer’s instructions. 37°C for 2 hours and the reactions were Two days after transfection, the cells were stopped by the addition of excess of fixed with 4% paraformaldehyde and unlabeled AdoMet. Then, the reaction embedded with Mowiol (Carl Roth). mixtures were subjected to Proteinase K 5 (NEB) digestion at 42°C for 2 hours and the histone H3 K36 trimethylation mark with nucleic acids in the reaction mixture were high specificity (Fig. 1). purified using Nucleospin PCR purification (Machery Nagel). The amount of Binding to trimethylated lysine usually radioactivity incorporated inzo the DNA involves hydrophobic cages, which often was measured by liquid scintillation have an acidic residue nearby (19). Since the counting and normalized with respect to the PWWP domains from Dnmt3a and 3b share total amount of DNA as determined by UV high amino acid homology, we used the spectroscopy. Methylation of biotinylated structure of the Dnmt3b PWWP domain (31) oligonucleotide substrates by Dnmt3a2 and to model the structure of Dnmt3a’s PWWP its variants was studied by measuring the domain and identify such potential binding transfer of titrated methyl groups from pockets next to E290 and D329 (Suppl. Fig. D o w n labeled AdoMet into biotinylated 1). Trying to disrupt the histone interaction lo a d e oligonucleotides (Bt-GAG AAG CTG GGA of PWWP, we mutated E290 and D329 to d fro m CTT CCG GGA GGA GAG TGC / GCA alanine. These residues are both exposed on h ttp CTC TCC TCC CGG AAG TCC CAG CTT the surface of the Dnmt3a PWWP structural ://w w w CTC) using the avidin-biotin methylation model, suggesting their exchange will not .jb c .o assay as described (22,30). affect the overall domain structure. We did rg b/ y not mutate the buried hydrophobic residues g u e s Results of the predicted pockets in order to avoid t o n D effects on protein folding and stability. The ec e m Peptide array binding and structural mutant domain proteins could be expressed be r 2 3 modelling and purified with similar yields as the wild , 2 0 1 8 In order to investigate if the Dnmt3a type domain (Suppl. Fig. 2A) and CD PWWP domain interacts with modified spectroscopy confirmed the wild type-like folding of the D329A mutant domain (Suppl. histone peptides, we employed two Fig. 2B). Binding experiments on peptide different modified histone tail peptide arrays arrays showed that there was no significant containing in total more than 1000 peptides change in the peptide interaction of the in various combinations of known and E290A variant, but the D329A variant had hypothetical modification states of the H3, completely lost its ability to bind H4, H2A and H2B tails. We repeatedly specifically to the H3K36me3 peptide (Fig. observed that the GST-tagged PWWP 1B). domain of Dnmt3a interacted with the 6 Recently, the structure of the Eaf3 chromo- 118 µM (36). Peptide binding of the barrel domain was solved in complex with PWWP-D329A variant was at least 5 times an H3K36me2 analog (32). Over 144 weaker (data not shown). To assess the backbone atoms comprising the conserved binding specificity, binding was competed ß-barrel and adjacent α-helix fold, the Eaf3 with unmodified H3 26-44 peptide (Fig. 1D). chromo-barrel domain structure could be We observed that at 5 fold excess of superimposed with the mouse Dnmt3b competitor only 20% of binding was lost, PWWP domain structure with an RMSD of corresponding to a 25 fold preference for 1.37 Å (Suppl. Fig. 1). The overlay binding to the methylated peptide. positioned the methylated K36 residue immediately adjacent to D329 in the Nucleosome and histone pull-down Dnmt3a PWWP domain model. This experiments D o w n observation indicates that the biochemical The PWWP-H3K36m3 interaction was also lo a d e data obtained here are compatible with the confirmed by pull-down experiments. The d fro m structural results obtained with related GST-PWWP domain of Dnmt3a was able to h ttp domains. Furthermore, the D329 residue is pull-down native nucleosomes isolated from ://w w w close to the S333 residue (Suppl. Fig. 1), human cells, which contained the .jb c .o mutation of which leads to loss of H3K36me3 mark (Fig. 2A, Suppl. Figs. 3A rg b/ y heterochromatic localization (18). and 4A). In contrast, the PWWP-D329A g u e s mutant completely failed to interact with t o n D Equilibrium peptide binding experiments nucleosomes (Fig. 2A, Suppl. Fig. 4A). In ec e m The interaction of the PWWP domain with addition to H3K36me3, the nucleosome be r 2 3 H3K36me3 was confirmed by fluorescence fraction bound by the PWWP domain was , 2 0 1 8 depolarization using a fluorescently labeled also enriched with H3K9me3, but the H3 26-44 K36me3 peptide (Fig. 1C). The amounts of H3K4me3 and H3K27me3 were binding constant was determined as 64 µM, reduced in comparison with H3K36me3 (Fig. which is weak but in the range of results 2B). The relative reduction in amount of observed with other individual chromatin H3K4me3 and K27me3 indicates that the binding domains, like BHC80 PHD finger pull-down is specific. The presence of the binding to H3K4me0 peptide: 33 µM (33), H3K9me3 mark in the pull down was L3MBTL1 to the H3K9me1 peptide: 26 µM expected on the basis of the overlapping (34), BRD2 bromodomain 1 to the distribution profiles of the H3K36me3 and H4K5ac/K12ac peptide: 360 µM (35) or H3K9me3 marks across transcribed Brd4 bromodomain 1 binding to H3K14ac: chromatin (37). Using native histones 7 isolated from human cells (Suppl. Fig. 3B), treatment reduces DNA methylation at we observed that the GST-PWWP domain unique sequences and repeats by more than of Dnmt3a, but not its D329A mutant, 50% (15,38). The native nucleosomes carry interacted with histone H3K36me3 mark their endogenous histone modification (Fig. 2C, Suppl Fig 4B), indicating that the pattern and partially demethylated DNA. interaction does not require DNA. This experimental design allowed us to study the methylation activity of Dnmt3a Sub nuclear localization of Dnmt3a using a substrate that mimics the in vivo We investigated the sub-nuclear distribution situation. Demethylated nucleosomes were of the EYFP-fused Dnmt3a and its D329A methylated with Dnmt3a2, a shorter isoform mutant in NIH3T3 cells after transient of Dnmt3a lacking the N-terminal variable expression of the proteins. Both proteins region (23), the Dnmt3a2 D329A mutant D o w n were located in the nucleus, which was (Dnmt3a2-D329A) and the catalytic domain lo a d e expected since the nuclear localization of Dnmt3a (Dnmt3a-CD) using d fro m signals of Dnmt3a are outside of the PWWP radioactively labeled S-adenosyl-L- h ttp domain (17). Within the nucleus, Dnmt3a methionine (AdoMet) as cofactor. The ://w w w was found in large spots corresponding to amounts of Dnmt3a2, Dnmt3a2-D329A and .jb c .o DAPI stained heterochromatin as reported Dnmt3a-CD enzymes used in the rg b/ y earlier (17,18) (Fig. 3, Suppl. Fig. 5). In nucleosome methylation experiments were g u e s contrast, the mutant protein was found more normalized to have similar activity on t o n D homogenously distributed in the nucleus oligonucleotide substrates (Suppl. Fig. 6). e c e m indicating partial loss of heterochromatic Using the native chromatin as substrate, we be r 2 3 localization. This result suggests that the observed that Dnmt3a2 is approximately 2.5 , 2 0 1 8 interaction of PWWP with K36me3 is fold more active than Dnmt3a-CD (Fig. 4, important for the localization of Dnmt3a to Suppl. Fig. 6). The Dnmt3a2 enzyme heterochromatin. containing the mutated D329A-PWWP domain, which fails to interact with the DNA methylation experiments H3K36me3 mark, only showed a basal level The influence of the PWWP-H3K36me3 of activity similar to Dnmt3a-CD (Fig. 4). interaction on the methylation of nucleosomal DNA by Dnmt3a was In order to confirm that the increase in investigated using native nucleosomes from activity of the wild type Dnmt3a2 is due to human cells prepared after treatment of the the PWWP-H3K36me3 interaction, we cells with 5-aza-2’-deoxycytidine. This performed the chromatin methylation 8 experiments in the presence of an anti- PWWP domain, but not for Dnmt3b (data H3K36me3 antibody. While addition of the not shown), while DNA binding of the anti-H3K36me3 antibody had no effect on Dnmt3b-PWWP domain has been reported the activity of Dnmt3a2-D329A or Dnmt3a- to be stronger than that of Dnmt3a (17). CD, it decreased the activity of Dnmt3a2 to a basal level, similar to the level of activity We show that the H3K36me3-PWWP observed with the Dnmt3a2-D329A or interaction mediates chromatin binding, Dnmt3a-CD (Fig. 4). Additional unrelated because the heterochromatic localization of antibodies used as controls had no or very Dnmt3a was disturbed in the D329A mutant weak effect on the catalytic activity of which no longer binds to H3K36me3. Dnmt3a2 (Suppl. Fig. 6). Taken together, Targeting of Dnmt3a by an H3K36me3- these results demonstrate that the PWWP PWWP interaction is in agreement with D o w n domain-H3K36me3 interaction increases the previous results showing that the Dnmt3a lo a d e activity of Dnmt3a2 for methylation of PWWP domain is necessary for d fro m native nucleosomal DNA. heterochromatic localization of Dnmt3a h ttp (17,18). In contrast, the ADD domain of ://w w w Discussion Dnmt3a which binds to H3 tails .jb c .o unmethylated at K4 (11,12) does not have a rg b/ y We show here that the Dnmt3a PWWP role in heterochromatic localization (17,18). g u e s domain specifically interacts with H3 tails t o n D containing the K36me3 mark. Peptide We show that the PWWP-H3K36me3 e c e m binding is weak but in the range of binding interaction stimulates the methylation be r 2 3 constants seen with other individual activity of Dnmt3a on chromatin bound , 2 0 1 8 domains interacting with histone peptides. In DNA isolated from human cells, because the the cell binding may be further supported by catalytic activity of the D329A mutant an interaction of the PWWP domain with dropped to levels similar to the isolated DNA, because the K36 position of the H3 catalytic domain. This result suggests an tail is positioned very close to the DNA, important role of the PWWP domain in such that the PWWP domain could interact guiding DNA methylation to chromatin with both epitopes simultaneously. carrying the H3K36me3 mark, which is in Interestingly, the DNA binding and peptide agreement with the observation that the binding properties of the PWWP domains PWWP domain is required for the activity of from Dnmt3a and 3b seem to be inverse: we Dnmt3a in the cell (17,39). This model is detected peptide binding of the Dnmt3a- further supported by the observation that the 9 genome wide distribution of DNA been proposed that in mammals Dnmt3a methylation is very similar to that of contributes to the maintenance of the DNA K36me3 methylation. Many studies have methylation at highly methylated regions found that H3K36me3 accumulates in (like repeats) (50). H3K36 methylation is an euchromatin in the body of active genes and evolutionary conserved mark suggesting that its distribution is anticorrelated with a transient methylation of H3K36 might H3K4me3 (37,40-43). Similarly, active occur in mammalian genomes as well and genes showed high DNA methylation in the transient H3K36 methylation might be one gene bodies, while inactive genes did not of the modifications that recruit Dnmt3a to (44,45). The strong correlation of DNA heterochromatin. methylation, absence of H3K4me3 and presence of H3K36me3 was also observed Dnmt3a has been shown to bind tightly to D o w n in two recent genome wide studies (14,16). chromatin (9). When combined with the lo a d e Additionally, a correlation of H3K36me3 previous findings, our data indicate that d fro m and DNA methylation was observed at there are several mechanisms that contribute h ttp exon-intron boundaries where exons were to the interaction of Dnmt3a with chromatin: ://w w w shown to have increased levels of both Dnmt3a is engaged in two interactions with .jb c .o H3K36me3 (46) and DNA methylation (16). the H3 tail, mediated by its ADD domain, rg b/ y which binds to the end of the H3 tail g u e s Like DNA methylation, H3K36me3 is a unmodified at K4 (11,12), and by its PWWP t o n D silencing mark, which recruits histone domain, which interacts with H3K36me3 e c e m deacetylation and represses intragenic located at more basal region of the H3 tail be r 2 3 transcriptional initiation (47). Consequently, (Fig. 4B). The anticorrelation of the , 2 0 1 8 the Smyd2 histone methlyltransferase H3K4me3 and H3K36me3 marks suggests (which generates K36me2) has a repressive that in the cell most H3 tails will either bind effect on reporter gene expression (48). In to both ADD and PWWP or none of them Schizosaccharomyces pombe the Set2 such that the targeting effects of both Histone methyltransferase methylates domains are synergistic. In addition, the histone H3 lysine 36 at repeat loci during S catalytic domain of Dnmt3a binds to DNA phase (49). Consequently, heterochromatic and polymerizes on the DNA (22,51) and it H3K36me3 transiently peaks in S phase and interacts with many other proteins including has an important role in the re-establishment Dnmt3L and Dnmt3b (52) which provide of heterochromatin after DNA replication additional contact points to chromatin that acts in a pathway parallel to Clr4. It has (10,53) suggesting that a complex network 10
Description: