Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Enzymatic properties of de novo-type mouse DNA (cytosine-5) methyltransferases

Enzymatic properties of de novo-type mouse DNA (cytosine-5) methyltransferases 3506–3512 Nucleic Acids Research, 2001, Vol. 29, No. 17 © 2001 Oxford University Press Enzymaticpropertiesof de novo-type mouse DNA (cytosine-5) methyltransferases Asako Aoki, Isao Suetake, Junichi Miyagawa, Takayuki Fujio, Takahito Chijiwa , Hiroyuki Sasaki and Shoji Tajima* Institute for Protein Research, Osaka University, 3-2 Yamadaoka, Suita, Osaka 565-0871, Japan and National Institute of Genetics, Mishima, Shizuoka 411-8540, Japan Received May 29, 2001; Revised and Accepted July 16, 2001 ABSTRACT Dnmt1 is responsible for the latter activity. Recently, two DNA methyltransferases, Dnmt3a and Dnmt3b, were identified (9), We have purified GST-fused recombinant mouse and this Dnmt3 family is expected to be responsible for the Dnmt3a and three isoforms of mouse Dnmt3b to near creation of methylation patterns at an early stage of embryo- homogeneity. Dnmt3b3, an isoform of Dnmt3b, did genesis (10,11). A mutation in DNMT3b gene, a human not have DNA methylation activity. Dnmt3a, Dnmt3b1 homolog of Dnmt3b, has been proven to be the cause of ICF or Dnmt3b2 showed similar activity toward syndrome (10,12,13), a rare autosomal recessive disease that poly(dG-dC)-poly(dG-dC) for measuring de novo triggers chromosomal instability due to hypomethylation of the methylation activity, and toward poly(dI-dC)-poly(dI- satellite 2 and 3 regions of chromosomes 1, 9 and 16 (10,14). Although the Dnmt3 family is capable of creating a methyl- dC) for measuring total activity. This indicates that ation pattern in vivo (15), its ability to create a methylation the enzymes are de novo-type DNA methyltrans- pattern in vitro has only been reported for a crude extract over- ferases. The enzyme activity was inhibited by NaCl or expressed in insect cells (9). KCl at concentrations >100 mM. The kinetic para- It has been reported that non-CpG methylation occurs in AdoMet meter, K , for Dnmt3a, Dnmt3b1 and Dnmt3b2 integrated adenovirus DNA (16), an integrated plasmid (17), was 0.4, 1.2 and 0.9 µ M when poly(dI-dC)-poly(dI-dC) an L1 element from embryonic fibroblasts (18), and embryonic was used, and 0.3, 1.2 and 0.8 µ Mwhen stem cells (19). Using embryonic stem cells that lack Dnmt1, poly(dG-dC)-poly(dG-dC) was used, respectively. The Ramsahoye et al. showed that a DNA methyltransferase other DNA K values for Dnmt3a, Dnmt3b1 and Dnmt3b2 m than Dnmt1, possibly Dnmt3a, is responsible for the non-CpG were 2.7, 1.3 and 1.5 µ M when poly(dI-dC)-poly(dI-dC) methylation (19). was used, and 3.5, 1.0 and 0.9 µ M when poly(dG-dC)- To address the questions of whether (i) Dnmt3a and Dnmt3b poly(dG-dC) was used, respectively. For the methyl- are real de novo methylases, (ii) the three isoforms of Dnmt3b, i.e. 3b1, 3b2 and 3b3, exhibit an enzymatic difference, and ation specificity, Dnmt3a significantly methylated (iii) Dnmt3a and Dnmt3b can exhibit non-CpG methylation CpG >> CpA. On the other hand, Dnmt3b1 methylated activity in vitro, we have purified recombinant Dnmt3a and all CpG > CpT ≥ CpA. Immuno-purified Dnmt3a, three isoforms of Dnmt3b. Using the purified proteins, we Myc-tagged and overexpressed in HEK 293T cells, have shown that Dnmt3a and Dnmt3b exhibit similar de novo methylated CpG >> CpA > CpT. Neither Dnmt3a nor methylation activity with similar kinetic parameters, Dnmt3b1 Dnmt3b1 methylated the first cytosine of CpC. and Dnmt3b2 show almost identical enzymatic properties, and Dnmt3b3 has no DNA methylation activity. Interestingly, both Dnmt3a and Dnmt3b exhibit non-CpG methylating activity. In INTRODUCTION particular, Dnmt3b appears to have the ability to methylate In vertebrates, the fifth position of cytosine residues in CpG CpT and CpA more than does Dnmt3a. sequences in genomic DNA is often methylated (1). Dynamic regulation of DNA methylation is known to contribute to MATERIALS AND METHODS physiological or pathological phenomena such as tissue- specific gene expression (2–4), genomic imprinting (5), Plasmids X chromosome inactivation (6) and carcinogenesis (7). In vertebrates, two types of DNA methyltransferase activity have The coding sequences of mouse Dnmt3a, and Dnmt3b1, 2 and been reported, i.e. de novo- and maintenance-type DNA 3 cDNAs were subcloned into the BamHI site of the pGEX methyltransferase activities. In mouse, de novo-type DNA vector (Amersham Pharmacia Biotech, UK). The ATG coding methylation activity creates tissue-specific methylation initiation methionine was directly connected to the BamHI site patterns at the implantation stage of embryogenesis (8), and in the pGEX vector without any spacer sequence. For Dnmt3a maintenance-type activity ensures clonal transmission of and Dnmt3b1, the cDNA was also subcloned into the lineage-specific methylation patterns during replication. pCS2+MT vector, which was kindly provided by Drs H. B. *To whom correspondence should be addressed. Tel: +81 6 6879 8627; Fax: +81 6 6879 8629; Email: [email protected] Nucleic Acids Research, 2001, Vol. 29, No. 17 3507 Sucic and D. Turner (University of Michigan, USA) harboring bands electrophoresed on SDS–PAGE were immunodetected the Myc-tag sequence in the 5′ site. as described elsewhere (21). Purification of GST-fused Dnmt3 DNA methylation activity Poly(dI-dC)-poly(dI-dC) (dIdC) and poly(dG-dC)-poly(dG-dC) Dnmt3-harboring plasmids were expressed in BL21(DE3) at (dGdC) (Amersham Pharmacia Biotech) were used as the 20°C. After the OD at 650 nm reached 0.6, 0.5 mM IPTG was methyl acceptors for total and de novo methylation activity added to the culture medium, followed by further culturing for measurements, respectively (22,23). The methylation activity 18 h. All the purification steps thereafter were performed on was measured in 25 µ l of reaction (R) buffer (5 mM EDTA, ice or at 4°C. Harvested bacteria were suspended in 5 vol of 0.2 mM DTT, 26 mM NaCl, 20 mM Tris–HCl, pH 7.4) solubilization (S) buffer [0.33 M NaCl, 0.1% (w/v) Triton containing 50 ng (∼0.5 pmol) of the purified enzyme, 0.1 µ g X-100, 1/20 (v/v) protease inhibitor cocktail (Nakarai Tesque, DNA, which was ∼150 pmol CpI or CpG when 1 mol double- Japan) in Dulbecco’s phosphate-buffered saline (PBS)]. The stranded DNA with 1 CpI or CpG site was calculated to be cell suspension was sonicated and centrifuged at 10 000 g for 2 mol, and 134 pmol [ H]S-adenosyl-L-methione (AdoMet) 10 min, and the supernatant fraction recovered. For Dnmt3a, (15 Ci/mmol; Amersham Pharmacia Biotech), unless other- the solubilized fraction was loaded onto a DEAE–Sepharose wise stated. After the reaction, the mixtures were incubated (Amersham Pharmacia Biotech) column and the unbound frac- with 0.1 µ g of proteinase K (Nakalai Tesque, Japan) at 50°C tion was collected. For Dnmt3b, the solubilized fraction was for 20 min, and then the radioactivity was measured as precipitated and recovered by 20–30% saturation with described previously (21). Protein concentrations were deter- ammonium sulfate, and the precipitate was dissolved in 5 ml of mined by quantifying the CBB-stained SDS–PAGE gels with S buffer. Samples were loaded onto glutathione–Sepharose, an image analyzer (MCID; Imaging Research, Canada), using washed with 10 bedvol of Sbuffer, andthenelutedwith4bed bovine serum albumin as a standard. The specific activities vol of elution (E) buffer [0.33 M NaCl, 0.1% (w/v) Triton were expressed as moles of transferred methyl group per hour X-100, 1/200 (v/v) protease inhibitor cocktail, 10 mM per mole of each Dnmt3 (mol/h/mol Dnmt3). glutathione, reduced form, 1 mM DTT in 50 mM Tris–HCl, pH 8.0]. The eluted fractions were collected in tubes already Determination of the cytosine methylation containing a 1/10 vol of 1 M Tris–HCl, pH 7.4, and quickly As a methyl acceptor, the HindIII fragment of the myoD gene mixed. The main fractions were pooled and loaded onto a (accession no. X61655) subcloned into pUC19 (24) was used. Superdex 200 (Amersham Pharmacia Biotech) column equili- AdoMet (Sigma-Aldrich, MO) was purified on a Sep-Pack brated with E buffer minus reduced form glutathione. The Plus C18 column (Waters, Japan) before use. Dnmt3 at each step was checked by SDS–PAGE on a 7.5% For the methylation by GST-fused enzymes, the reaction polyacrylamide gel (20), and purity was monitored by mixture was incubated at 37°C for 1 h with 0.1 µ gpUC-myoD, Coomassie Brilliant Blue R-250 (CBB) staining. 3–5 µ g Dnmt3 protein, 100 µ M AdoMet in 500 µ lof R buffer. For the Myc-tagged Dnmt3 bound to Sepharose, the reaction Preparation of Myc-tagged Dnmt3 mixture contained 1 µ gpUC-myoD, 5 µ g Dnmt3 protein and Myc-tagged Dnmt3 was expressed in HEK 293T cells. Trans- 160 µ MAdoMetin 200 µ l of R buffer, and was then incubated fection with lipofectoamine (Gibco BRL, MD) was performed at 37°C for 3 h under agitation. After the reaction, the plasmid according to the manufacturer’s instructions. Cells transfected was recovered, and the methylation reaction was repeated with the plasmid were washed with PBS, and then solubilized twice more with fresh enzyme-coupled Sepharose and with 2 M NaCl, 0.3% Triton X-100, 1/50 (v/v) protease inhibitor AdoMet. cocktail in 20 mM Tris–HCl, pH 7.4. Dnmt3 thus extracted The sodium bisulfite reaction was performed basically as was immunoprecipitated with anti-Myc monoclonal antibody described elsewhere (25) with a slight modification. Briefly, (9E10)-coupled Sepharose at 4°C for 4 h. The matrix was HindIII-digested plasmids were treated with 0.3 M NaOH for washed with the extraction buffer and then used for the 30 min and then incubated in a mixture comprising 2.9 M methylation reaction. Na S O and 100 mM hydroquinone, with the pH adjusted to 5. 2 2 5 The reaction mixture was incubated for a total of 5 h at 55°C Antibodies with 1 min incubation at 95°Cevery1h. The recoveredDNA Antisera reacted with mouse Dnmt3a and Dnmt3b were raised was alkaline-treated and then used for PCR with three sets of against maltose binding protein fusion proteins, and an primers designed to amplify intron2 of the myoD gene as antiserum reacted with glutathione S-transferase (GST) was follows: upperF, GGTGGATTTAGGAGGATGAGTAAT- raised against GST. The antibodies raised in rabbits were GGAG; upperR, CCCCAAATACTAAAAACCAACCAC- immunoselected using GST fusion protein- or GST-coupled ACAC; lower1F: GGTTTTAGGATTGGGTTAGTTTTAGG- Sepharose as affinity matrices. To this end, a DNA [232 TGTTAGG; lower1R, CCCTCATACCTAAATACTCCTAA- (SmaI)–701 (SmaI), encoding amino acids 7–162] of mouse CATCTAACA; lower2F, GTGTTAGATGTTAGGAGTATT- Dnmt3a (accesion no. AF068625) and a DNA [1–662 (SalI), TAGGTATGAGGG; lower2R, CACCTAAACCACTAC- encoding amino acids 1–181] of mouse Dnmt3b1 (accession CCCCAAACC. Primer set upperF and upperR amplifies the no. AF068626) were ligated into pMAL-c2 (NEB, MA) and upper strand of modified intron2, and primer sets lower1F and pGEX vectors, and expressed in BL21(DE3) in the presence of lower1R, and lower2F and lower2R amplify the lower strand IPTG. GST and the fusion proteins were purified with of intron2. The amplification reaction comprised 35 cycles of glutathione–Sepharose or amylose resin according to the incubation of the reaction mixture for denaturation at 94°Cfor manufacturers’ instructions. The Dnmt3a, Dnmt3b and GST 1min,annealingat 60°C for 1 min and extension at 72°Cfor 3508 Nucleic Acids Research, 2001, Vol. 29, No. 17 Figure 1. Schematic illustration of Dnmt3a, Dnnmt3b1, Dnmt3b2 and Dnmt3b3. The regions of the conserved motifs in the C-terminal catalytic domain (I∼X) and the Cys-rich region (Cys-rich) are indicated. Dnmt3b2 lacks amino acid residues 363–382, and Dnmt3b3 lacks amino acid residues 363–382 and 731–813, of Dnmt3b1. GST in the N-terminus of each Dnmt3 indicates the fused GST moiety. The numbers above the bars are amino acid numbers. 1 min. The amplified fragment was subcloned into the SmaI site of pUC19 and the modified nucleotides were determined by the dideoxy method (26). RESULTS Expression and purification of Dnmt3a, Dnmt3b1, Dnmt3b2 and Dnmt3b3 in Escherichia coli Each Dnmt3 was expressed as a fusion protein with GST (GST-Dnmt3) in the pGEX vector (Fig. 1), and then expressed in E.coli as described in the Materials and Methods. The fractions at each purification step were electrophoresed and Figure 2. Purification of GST-fused Dnmt3s. An aliquot of a sample obtained stained (Fig. 2). The Dnmt3a expressed was purified with at each purification step was subjected to SDS–PAGE and then stained with CBB. (A) Solubilized (lane 1), DEAE–Sepharose unbound (lane 2), glutathione– DEAE–Sepharose (Fig. 2A, lane 2), glutathione–Sepharose Sepharose unbound (lane 3), washed (lane 4), eluted (lane 5) and Superdex (lane 5) and Superdex 200 (lane 6). Expressed Dnmt3b1, 3b2, 200 eluted (lane 6) fractions of GST-fused Dnmt3a. 1/20 000 (lanes 1–4), and 3b3 were fractionated with ammonium sulfate (Fig. 2B, 1/1000 (lane 5) and 1/100 (lane 6) of the recovered fractions were loaded. C and D, lanes 3), glutathione–Sepharose (lanes 7) and (B–D) Solubilized (lane 1), ammonium sulfate-precipitated with 20% satura- Superdex 200 (lanes 8). The apparent molecular weights of the tion (lane 2), 20–30%-saturation (lane 3), supernatant on 30% saturation (lane 4), glutathione–Sepharose unbound (lane 5), washed (lane 6), eluted (lane 7) GST-fused Dnmt3a, Dnmt3b1, Dnmt3b2 and Dnmt3b3 were and Superdex 200 eluted (lane 8) fractions of GST-fused Dnmt3b1 (B), 130, 120, 115 and 110 kDa, as calculated by SDS–PAGE (Fig. 2). Dnmt3b2 (C) and Dnmt3b3 (D). 1/20 000 (lanes 1–5), 1/1000 (lanes 6 and 7) Considering the size of GST (26 kDa), and the calculated and 1/100 (lane 8) of the recovered fractions were loaded. Arrowheads indicate molecular weights of Dnmt3a, Dnmt3b1, Dnmt3b2 and the position of Dnmt3. (E) Purified Dnmt3a (lane 1), Dnmt3b1 (lane 2), Dnmt3b2 (lane 3), Dnmt3b3 (lane 4) and GST (lane 5) were electrophoresed in Dnmt3b3 are 102, 97, 95 and 88 kDa, respectively (9), the the same gel and then CBB-stained. Purified Dnmt3a (lanes 6, 11 and 16), major bands (indicated by arrowheads) are expected to be Dnmt3b1 (lanes 7, 12 and 17), Dnmt3b2 (lanes 8, 13 and 18), Dnmt3b3 (lanes intact forms of the proteins. Since these bands reacted with 9, 14 and 19) and GST (lanes 10, 15 and 20) were immunodetected with both anti-GST and anti-Dnmt3 antibodies (Fig. 2E, lanes 6–19), specific antibodies reactive with Dnmt3a (lanes 6–10), Dnmt3b (lanes 11–15) all the purified Dnmt3 proteins possessed an intact N-terminal and GST (lanes 16–20), respectively. Molecular weight sizes (kDa) are indi- cated beside each gel. region. The final preparations were ∼80% pure, as judged on densitometry (Fig. 2E, lanes 1–4). DNA methylation activities of purified Dnmt3s C-terminal catalytic domain of Dnmt3b1 or Dnmt3b2, showed The DNA methylation activities were determined for the no DNA methylation activity. purified enzymes. Purified Dnmt3a, Dnmt3b1 and Dnmt3b2 We next examined the effect of salt on the methylation activities showed similar activities toward dI-dC compared to those of each Dnmt3. As shown in Figure 3, all the Dnmt3s were toward dG-dC when fixed amounts of AdoMet and DNA were inhibited in the presence of >100 mM NaCl. Even when the used (Table 1). The results clearly suggest that both Dnmt3a and Dnmt3b act as a de novo methylase. Interestingly, salt was changed to KCl, the profiles of inhibition of the Dnmt3b3, which lacks 62 amino acid residues of the enzymes were similar to those for NaCl. Nucleic Acids Research, 2001, Vol. 29, No. 17 3509 Table 1. Specific activities of GST-fused Dnmt3a, Dnmt3b1, Dnmt3b2 and Dnmt3b3. The methyltransferase activity was determined for each Dnmt3 Dnmt3a Dnmt3b1 Dnmt3b2 Dnmt3b3 a b dIdC 1.07 ± 0.34 (4) 0.68 ± 0.10 (5) 0.64 ± 0.08 (5) 0.00 ± 0.00 (3) dGdC 0.52 ± 0.09 (4) 0.43 ± 0.12 (3) 0.40 ± 0.07 (5) 0.00 ± 0.01 (3) DNA methylation activity was expressed as mol/h/mol Dnmt3 ± standard deviation. The numbers in parentheses are the numbers of DNA methylation activity measurements with different enzyme preparations. Figure 3. Effects of salts on the DNA methylation activity of Dnmt3s. The Figure 4. Effect of the AdoMet concentration on the activity of Dnmt3s. The effects of KCl (A)and NaCl (B) on the DNA methylation activities [specific DNA methylation activities [specific activity (s.a.) in mol/h/mol Dnmt3] of activity (s.a.) in mol/h/mol Dnmt3] of Dnmt3a (circles), Dnmt3b1 (open GST-fused Dnmt3a (A), Dnmt3b1 (B)and Dnmt3b2(C) were titrated with squares) and Dnmt3b2 (filled squares) was determined for methyl acceptors AdoMet (left). As methyl acceptors, dIdC (circles) and dGdC (squares) were dIdC (left) and dGdC (right). used. A typical example for each enzyme is demonstrated. Right panels show double reciprocal plots of the respective titration curves. Kinetic parameters of purified Dnmt3s To determine the K for AdoMet, a methyl donor for methyla- Typical titration profiles are shown in Figure 5, and the calculated DNA DNA K and V values obtained on three measurements tion activity, we measured the activity with AdoMet concentra- m max tions of 0.15–5.4 µ M. A typical titration curve for each enzyme with different preparations of enzymes are summarized in AdoMet AdoMet DNA is shown in Figure 4, and the calculated K and V Table 3. The K values of each enzyme for dI-dC and m max m values obtained on three measurements with different prepara- dG-dC expressed as the concentrations of CpI and CpG, AdoMet tions of enzymes are summarized in Table 2. The K respectively, were similar, suggesting that Dnmt3s fulfill the value for Dnmt3a was 2–4-fold smaller than those of Dnmt3b1 conditions for a de novo methylase. On comparison of their DNA DNA and Dnmt3b2, indicating that Dnmt3a has higher affinity for K values, Dnmt3a showed a relatively high K value m m AdoMet AdoMet than Dnmt3bs. For the calculated V values, compared with Dnmt3bs, indicating that Dnmt3a may have max Dnmt3a showed a higher value than those of Dnmt3bs. lower affinity for DNA than Dnmt3bs. In the presence of high K for DNA, a methyl acceptor, was also determined in the concentration of DNA, the activities were inhibited for both presence of dI-dC or dG-dC (0.15–18 µ Mof CpI or CpG). Dnmt3a and Dnmt3bs (Fig. 5). 3510 Nucleic Acids Research, 2001, Vol. 29, No. 17 Table 2. K values of Dnmt3a, Dnmt3b1 and Dnmt3b2 for AdoMet AdoMet AdoMet K (µ M) V (mol/h/mol Dnmt3) m max dIdC dGdC dIdC dGdC Dnmt3a 0.4 ± 0.1 0.3 ± 0.0 1.33 ± 0.31 1.22 ± 0.22 Dnmt3b1 1.2 ± 0.3 1.2 ± 0.3 1.00 ± 0.01 0.73 ± 0.11 Dnmt3b2 0.9 ± 0.2 0.8 ± 0.2 0.70 ± 0.17 0.69 ± 0.09 Three independent experiments similar to that shown in Figure 4 were performed using different Dnmt3 preparations. The values are means ± standard deviation. Table 3. K values of Dnmt3a, Dnmt3b1 and Dnmt3b2 for DNA DNA DNA K (µ M) V m max Z(mol/h/mol Dnmt3) dIdC dGdC dIdC dGdC Dnmt3a 2.7 ± 0.4 3.5 ± 1.2 1.7 ± 1.2 1.2 ± 0.7 Dnmt3b1 1.3 ± 0.4 1.0 ± 0.3 1.0 ± 0.1 0.8 ± 0.1 Dnmt3b2 1.5 ± 0.4 0.9 ± 0.3 1.1 ± 0.2 0.6 ± 0.1 Three independent experiments similar to that shown in Figure 5 were performed using different Dnmt3 preparations. Table 4. Methylation specificities of GST-fused Dnmt3a and Dnmt3b1 CpG CpA CpT CpC Dnmt3a 101.6 10.4 4.0 3.5 Dnmt3b1 47.1 15.7 24.3 1.4 Dnmt3b3 2.1 3.1 2.6 3.7 The methylation sites of the subcloned myoD gene were determined by the sodium bisulfite method. Totals of 1729, 700 and 1909 of CpN were analyzed after treatment with Dnmt3a, Dnmt3b1 and Dnmt3b3, respectively, and the numbers of methylated dinucleotides were determined and normalized as to 1000 CpN. Since Dnmt3b3 had no DNA methylation activity, the number of methylated CpN was taken as the background in this experiment. Figure 5. Effect of the DNA concentration on the activity of Dnmt3s. The DNA methylation activities [specific activity (s.a.) in mol/h/mol Dnmt3] of more than other dinucleotide sequences. Under identical GST-fused Dnmt3a (A), Dnmt3b1 (B)and Dnmt3b2(C) were titrated with conditions, Dnmt3a methylated 10.4, 4.0 and 3.5 cytosine dIdC (circles) or dGdC (squares). One mole double-stranded DNA with 1 CpI or CpG site was calculated to be 2 mol. A typical example for each enzyme is residues in CpA, CpT and CpC per 1000 CpN dinucleotides. demonstrated. Right panels show double reciprocal plot of the respective Considering the background (experimental error) with inactive titration curves. Dnmt3b3, which was 2.1, 3.1, 2.6 and 3.7 cytosine methylation for CpG, CpA, CpT and CpC per 1000 CpN dinucleotides, respectively, Dnmt3a significantly methylated CpA, although Sequence specificity of the Dnmt3 methylation site the level was very low. It has been postulated that Dnmt3a may methylate the cytosine Dnmt3b1 also methylated the CpG sequence most (47.1/1000) residues in sequences other than CpG sequences in vivo,as among the dinucleotide sequences; however, the specificity judged by analysis of Dnmt1-targeted embryonic stem cells was not as strict as in the case of Dnmt3a. Dnmt3b1 methylated (19). To directly analyze the sequence specificity of Dnmt3a CpA (15.7/1000) and CpT (24.3/1000) at relatively high and Dnmt3b, we used the myoD gene as the methyl acceptor frequencies compared to Dnmt3a. Neither Dnmt3a nor and determined the methylated sequences by the sodium Dnmt3b1 methylated the first cytosine in the CpC sequence. bisulfite method (25). The results are summarized in Table 4. We could not find any rule for the methylation sequence as to Dnmt3a methylated CpG (101.6 per 1000 CpN dinucleotides) the third nucleotide specificity (CpNpN). Nucleic Acids Research, 2001, Vol. 29, No. 17 3511 activity in vitro using purified enzymes. In the present study, we have expressed and purified GST-fused recombinant Dnmt3a and Dnmt3bs, and have shown that both Dnmt3a and Dnmt3bs exhibit de novo methylation activity in vitro.For Dnmt1 with dI-dC as the methyl acceptor, the activity is >10-fold higher than that toward dG-dC (27), and this property of Dnmt1 is the basis for its maintenance activity. On the other hand, although the specific activities were very low, Dnmt3a and Dnmt3bs showed similar activities toward both dI-dC and dG-dC, indicating that these enzymes have the ability to create a methylation pattern. Dnmt3a and Dnmt3bs each have a catalytic domain in the Figure 6. Expression and immuno-purification of Dnmt3s. Myc-tagged C-terminal half, in which the 10 motifs conserved in DNA Dnmt3a (lanes 1 and 2) and Dnmt3b1 (lanes 3 and 4) were expressed in HEK (cytosine-5) methyltransferases exist (28). Dnmt3b3 lacks 293T cells, extracted (lanes 1 and 3), and immunoprecipitated with anti-Myc 62 amino acid residues in the sequence between motifs VIII monoclonal antibody (9E10)-coupled Sepharose (lanes 2 and 4). Asterisks indicate the heavy chain of 9E10. Molecular weight sizes (kDa) are indicated and IX, called the TRD (target recognition domain) (28) in beside the gel. bacterial methylases, and the part of motif IX. It was not known whether Dnmt3b3 exhibits DNA methylation activity or not. In the present study, it has been shown that Dnmt3b3 has no DNA methylation activity. As Dnmt3b3 is expressed Table 5. Methylation specificities of Myc-tagged Dnmt3a and Dnmt3b1 in vivo under physiological conditions, the protein may play a role as a dominant negative factor. CpG CpA CpT CpC Similar to human maintenance-type DNA (cytosine-5) Dnmt3a 62.6 9.9 5.8 0.0 methyltransferase (DNMT1) (29), interestingly, all Dnmt3a, Dnmt3b1 11.1 1.1 1.1 1.1 Dnmt3b1 and Dnmt3b2 were inhibited at high concentrations –Dnmt3 0.0 0.0 0.0 0.0 of DNA (Fig. 5). Pradhan et al. proposed that DNMT1 may form a catalytically inactive complex with DNA at high The methylation sites of the subcloned myoD gene were determined by the concentrations of DNA through the domain other than its sodium bisulfite method. Totals of 6075, 4485 and 2793 of CpN were catalytic domain (29). The inhibitory effect of DNA on analyzed after treatment with Dnmt3a, Dnmt3b1, and without Dnmt3, respec- Dnmt3s could be an indication that Dnmt3s also have a DNA tively, and the numbers of methylated dinucleotides were determined and normalized as to 5000 CpN. binding domain other than their catalytic domain. Recently, the methylated cytosine residues in genomic DNA prepared from embryonic stem cells that lack Dnmt1 were analyzed (19). In these cells, the cytosine methylation frequencies In the recombinant Dnmt3, GST was in the N-terminus. To for CpA and CpT were reported to be ∼50 and 15% of that for eliminate the effect of GST, we changed the GST N-terminal the CpG sequence. In our study, with Dnmt3a, after subtracting tag to a Myc-tag. Myc-tagged Dnmt3a and Dnmt3b1 were the background, CpA and CpT were methylated to ∼7and 1%, expressed in HEK 293T cells under the CMV promoter and respectively, of the methylated CpG level. On the other hand, then immuno-purified with anti-Myc monoclonal antibody- Dnmt3b1 methylated CpA and CpT to ∼28 and 46%, respec- coupled Sepharose. The expressed and purified Dnmt3a and tively, of the CpG methylation level. The ability of Dnmt3a to Dnmt3b1 are shown in Figure 6. The enzymes bound to a methylate CpA and CpT itself cannot explain the levels of CpA matrix were used directly for the in vitro methylation reaction. andCpT methylationinDnmt1-targetedembryonic stem cells The results are summarized in Table 5. Although the activity (19). Dnmt3b is expected to contribute mainly to non-CpG was low compared to that of the GST-fused enzymes, the methylation, especially of the CpT sequence. Although we dinucleotide specificity of Dnmt3a was shown to be similar to added an equal amount of Dnmt3b1 as Dnmt3a protein to the that of GST-fused Dnmt3a (Table 4). Interestingly, Myc-tagged methylation reaction mixture, the myc-tagged Dnmt3b1 bound Dnmt3a significantly methylated CpT, which was not obvious to the matrix showed extremely low activity. The low activity when GST-fused Dnmt3a was used. The activity of Dnmt3b1 of Dnmt3b1 could be due to the instability of Dnmt3b1 was significantly low, only the CpG sequence specificity being compared to Dnmt3a. significant (11.1/5000 CpN dinucleotides). ACKNOWLEDGEMENTS DISCUSSION This work was supported in part by the Program for the Promo- De novo DNA methyltransferase activity plays a crucial role in tion of Fundamental Studies in Health Sciences of the Organi- the creation of methylation patterns in genomic DNA during zation for Pharmaceutical Safety and Research of Japan, the embryogenesis and germ line cells. Dnmt3a and Dnmt3b are Program for the Promotion of Basic Research Activities for the most likely candidates responsible for the de novo DNA Innovative Biosciences, and Grants-in-Aid from the Ministry methyltransferase activity (10). However, it has not been proven of Education, Culture, Sports, Science, and Technology of that Dnmt3a and Dnmt3bs exhibit de novo DNA methylation Japan. 3512 Nucleic Acids Research, 2001, Vol. 29, No. 17 methylation of constitutive heterochromatin in adult and fetal tissues. NOTE ADDED IN PROOF Hum. Genet., 99, 738–745. In the course of revision, Gowher and Jeltsch published work 15. Hsieh,C.L. (1999) In vivo activity of murine de novo methyltransferases, on purified recombinant Dnmt3a reporting a similar K for Dnmt3a and Dnmt3b. Mol. Cell. Biol., 19, 8211–8218. 16. Toth,M., Müeller,U. and Doerfler,W. (1990) Establishment of a de novo DNA and an ability for non-CpG methylation activity (30). methylation pattern. Transcription factor binding and deoxycytidine methylation at CpG and non-CpG sequences in an integrated adenovirus promoter. J. Mol. Biol., 214, 673–683. REFERENCES 17. Clark,S.J., Harrison,J. and Frommer,M. (1995) CpNpG methylation in 1. Antequera,F. and Bird,A. (1993) CpG islands in DNA methylation. In mammalian cells. Nature Genet., 10, 20–27. Jost,J.P. and Saluz,H.P. (eds), Molecular Biology and Biological 18. Woodcock,D.M., Lawler,C.B., Linsenmeyer,M.E., Doherty,J.P. and Significance. Birkhäuser Verlag, Basel, pp. 169–185. Warren,W.D. (1997) Asymmetric methylation in the hypermethylated 2. Shen,C.K.J. and Maniatis,T. (1980) Tissue-specific DNA methylation in a CpG promoter region of the human L1 retrotransposon. J. Biol. Chem., clusterofrabbit β-like globin genes. Proc. Natl Acad. Sci. USA, 77, 272, 7810–7816. 6634–6638. 19. Ramsahoye,B.H., Biniszkiewicz,D., Lyko,F., Clark,V., Bird,A. and 3. Razin,A. and Cedar,H. (1991) DNA methylation and gene expression. Jaenisch,R. (2000) Non-CpG methylation is prevalent in embryonic stem Microbiol. Rev., 55, 451–458. cells and may be mediated by DNA methyltransferase 3a. Proc. Natl 4. Tajima,S. and Suetake,I. (1998) Regulation and function of DNA Acad. Sci. USA, 97, 5237–5242. methylationinvertebrates. J. Biochem., 123, 993–999. 20. Laemmli,U.K. (1970) Cleavage of structural proteins during the assembly 5. Jaenisch,R. (1997) DNA methylation and imprinting: Why bother? of the head of bacteriophage T4. Nature, 227, 75–78. Trends Genet., 13, 323–329. 21. Kimura,H., Suetake,I. and Tajima,S. (1999) Xenopus maintenance-type 6. Riggs,A.D. and Porter,T.N. (1996) X-chromosome inactivation and DNA methyltransferase is accumulated in and translocated into germinal epigenetic mechanisms. In Russo,V.E.A., Martinssen,R.A. and vesicles of oocytes. J. Biochem., 125, 1175–1182. Riggs,A.D. (eds), Epigenetic Mechanisms of Gene Regulation. Cold 22. Pedrali-Noy,G. and Weissbach,A. (1986) Mammalian DNA Spring Harbor Press, Cold Spring Harbor, NY, pp. 231–248. methyltransferases prefer poly(dI-dC) as a substrate. J. Biol. Chem., 261, 7. Laird,P.W. and Jaenisch,R. (1996) The role of DNA methylation in cancer 7600–7602. genetics and epigenetics. Annu. Rev. Genet., 30, 441–464. 23. Pfeifer,G.P. and Drahvosky,D. (1986) Preferential binding of DNA 8. Monk,M. (1990) Changes in DNA methylation during mouse embryonic methyltransferase and increased de novo methylation of deoxyinosine- development in relation to X-chromosome activity and imprinting. containing DNA. FEBS Lett., 207, 75–78. Philos. Trans. R. Soc. Lond. B Biol. Sci., 326, 299–312. 24. Takagi,H., Tajima,S. and Asano,A. (1995) Overexpression of DNA 9. Okano,M., Xie,S. and Li,E. (1998) Cloning and characterization of a methyltransferase in myoblast cells accelerates myotube formation. family of novel mammalian DNA (cytosine-5) methyltransferases. Eur. J. Biochem., 231, 282–291. Nature Genet., 19, 219–220. 25. Clark,S.J., Harrison,J., Paul,C.L. and Frommer,M. (1994) High sensitivity 10. Okano,M., Bell,D.W., Haber,D.A. and Li,E. (1999) DNA mapping of methylated cytosine. Nucleic Acids Res., 22, 2990–2997. methyltransferases Dnmt3a and Dnmt3b are essential for de novo 26. Sanger,F., Nicklen,S. and Coulson,A.R. (1977) DNA sequencing with methylation and mammalian development. Cell, 99, 247–257. chain-terminating inhibitors. Proc. Natl Acad. Sci. USA, 74, 5463–5467. 11. Lyko,F., Ramsahoye,B.H., Kashevsky,H., Tudor,M., Mastrangelo,M.A., 27. Tajima,S., Tsuda,H., Wakabayashi,N., Asano,A., Mizuno,S. and Orr-Weaver,T.L. and Jaenisch,R. (1999) Mammalian (cytosine-5) Nishimori,K. (1995) Isolation and expression of a chicken DNA methyltransferases cause genomic DNA methylation and lethality in methyltransferase cDNA.J.Biochem., 117, 1050–1057. Drosophila. Nature Genet., 23, 363–366. 28. Kumar,S., Cheng,X., Klimasauskas,S., Mi,S., Posfai,J., Roberts,R.J. and 12. Xu,G.L., Bestor,T.H., Bourc’his,D., Hsieh,C.L., Tommerup,N., Wilson,G.G. (1994) The DNA (cytosine-5) methyltransferase. Bugge,M., Hulten,M., Qu,X., Russo,J.J. and Viegas-Pequignot,E. (1999) Nucleic Acids Res., 22, 1–10. Chromosome instability and immunodeficiency syndrome caused by 29. Pradhan,S., Bacolla,A., Wells,R.D. and Roberts,R.J. (1999) Recombinant mutations in a DNA methyltransferase gene. Nature, 402, 187–191. human DNA (cytosine-5) methyltranasferase. I. Expression, purification 13. Hansen,R.S., Wijmenga,C., Luo,P., Stanek,A.M., Canfield,T.K., and comparison of de novo and maintenance methylation. J. Biol. Chem., Weemaes,C.M.R. and Gartler,S.M. (1999) The DNMT3B DNA 274, 33002–33010. methyltransferase gene is mutated in the ICF immunodeficiency syndrome. Proc. Natl Acad. Sci. USA, 96, 14412–14417. 30. Gowher,H. and Jeltsch,A. (2001) Enzymatic properties of recombinant 14. Miniou,P., Jeanpierre,M., Bourc’his,D., Coutinho Barbosa,A.C., Dnmt3a DNA methyltransferase from mouse: the enzyme modifies DNA Blanquet,V. and Viegas-Pequignot,E. (1997) Alpha-satellite DNA in a non-processive manner and also methylates non-CpA sites. methylation in normal individuals and in ICF patients: heterogeneous J. Mol. Biol., 309, 1201–1208. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Nucleic Acids Research Oxford University Press

Enzymatic properties of de novo-type mouse DNA (cytosine-5) methyltransferases

Download PDF
7 pages

Loading next page...
 
/lp/oxford-university-press/enzymatic-properties-of-de-novo-type-mouse-dna-cytosine-5-uP04yZBEo4

References (34)

Publisher
Oxford University Press
ISSN
0305-1048
eISSN
1362-4962
DOI
10.1093/nar/29.17.3506
Publisher site
See Article on Publisher Site

Abstract

3506–3512 Nucleic Acids Research, 2001, Vol. 29, No. 17 © 2001 Oxford University Press Enzymaticpropertiesof de novo-type mouse DNA (cytosine-5) methyltransferases Asako Aoki, Isao Suetake, Junichi Miyagawa, Takayuki Fujio, Takahito Chijiwa , Hiroyuki Sasaki and Shoji Tajima* Institute for Protein Research, Osaka University, 3-2 Yamadaoka, Suita, Osaka 565-0871, Japan and National Institute of Genetics, Mishima, Shizuoka 411-8540, Japan Received May 29, 2001; Revised and Accepted July 16, 2001 ABSTRACT Dnmt1 is responsible for the latter activity. Recently, two DNA methyltransferases, Dnmt3a and Dnmt3b, were identified (9), We have purified GST-fused recombinant mouse and this Dnmt3 family is expected to be responsible for the Dnmt3a and three isoforms of mouse Dnmt3b to near creation of methylation patterns at an early stage of embryo- homogeneity. Dnmt3b3, an isoform of Dnmt3b, did genesis (10,11). A mutation in DNMT3b gene, a human not have DNA methylation activity. Dnmt3a, Dnmt3b1 homolog of Dnmt3b, has been proven to be the cause of ICF or Dnmt3b2 showed similar activity toward syndrome (10,12,13), a rare autosomal recessive disease that poly(dG-dC)-poly(dG-dC) for measuring de novo triggers chromosomal instability due to hypomethylation of the methylation activity, and toward poly(dI-dC)-poly(dI- satellite 2 and 3 regions of chromosomes 1, 9 and 16 (10,14). Although the Dnmt3 family is capable of creating a methyl- dC) for measuring total activity. This indicates that ation pattern in vivo (15), its ability to create a methylation the enzymes are de novo-type DNA methyltrans- pattern in vitro has only been reported for a crude extract over- ferases. The enzyme activity was inhibited by NaCl or expressed in insect cells (9). KCl at concentrations >100 mM. The kinetic para- It has been reported that non-CpG methylation occurs in AdoMet meter, K , for Dnmt3a, Dnmt3b1 and Dnmt3b2 integrated adenovirus DNA (16), an integrated plasmid (17), was 0.4, 1.2 and 0.9 µ M when poly(dI-dC)-poly(dI-dC) an L1 element from embryonic fibroblasts (18), and embryonic was used, and 0.3, 1.2 and 0.8 µ Mwhen stem cells (19). Using embryonic stem cells that lack Dnmt1, poly(dG-dC)-poly(dG-dC) was used, respectively. The Ramsahoye et al. showed that a DNA methyltransferase other DNA K values for Dnmt3a, Dnmt3b1 and Dnmt3b2 m than Dnmt1, possibly Dnmt3a, is responsible for the non-CpG were 2.7, 1.3 and 1.5 µ M when poly(dI-dC)-poly(dI-dC) methylation (19). was used, and 3.5, 1.0 and 0.9 µ M when poly(dG-dC)- To address the questions of whether (i) Dnmt3a and Dnmt3b poly(dG-dC) was used, respectively. For the methyl- are real de novo methylases, (ii) the three isoforms of Dnmt3b, i.e. 3b1, 3b2 and 3b3, exhibit an enzymatic difference, and ation specificity, Dnmt3a significantly methylated (iii) Dnmt3a and Dnmt3b can exhibit non-CpG methylation CpG >> CpA. On the other hand, Dnmt3b1 methylated activity in vitro, we have purified recombinant Dnmt3a and all CpG > CpT ≥ CpA. Immuno-purified Dnmt3a, three isoforms of Dnmt3b. Using the purified proteins, we Myc-tagged and overexpressed in HEK 293T cells, have shown that Dnmt3a and Dnmt3b exhibit similar de novo methylated CpG >> CpA > CpT. Neither Dnmt3a nor methylation activity with similar kinetic parameters, Dnmt3b1 Dnmt3b1 methylated the first cytosine of CpC. and Dnmt3b2 show almost identical enzymatic properties, and Dnmt3b3 has no DNA methylation activity. Interestingly, both Dnmt3a and Dnmt3b exhibit non-CpG methylating activity. In INTRODUCTION particular, Dnmt3b appears to have the ability to methylate In vertebrates, the fifth position of cytosine residues in CpG CpT and CpA more than does Dnmt3a. sequences in genomic DNA is often methylated (1). Dynamic regulation of DNA methylation is known to contribute to MATERIALS AND METHODS physiological or pathological phenomena such as tissue- specific gene expression (2–4), genomic imprinting (5), Plasmids X chromosome inactivation (6) and carcinogenesis (7). In vertebrates, two types of DNA methyltransferase activity have The coding sequences of mouse Dnmt3a, and Dnmt3b1, 2 and been reported, i.e. de novo- and maintenance-type DNA 3 cDNAs were subcloned into the BamHI site of the pGEX methyltransferase activities. In mouse, de novo-type DNA vector (Amersham Pharmacia Biotech, UK). The ATG coding methylation activity creates tissue-specific methylation initiation methionine was directly connected to the BamHI site patterns at the implantation stage of embryogenesis (8), and in the pGEX vector without any spacer sequence. For Dnmt3a maintenance-type activity ensures clonal transmission of and Dnmt3b1, the cDNA was also subcloned into the lineage-specific methylation patterns during replication. pCS2+MT vector, which was kindly provided by Drs H. B. *To whom correspondence should be addressed. Tel: +81 6 6879 8627; Fax: +81 6 6879 8629; Email: [email protected] Nucleic Acids Research, 2001, Vol. 29, No. 17 3507 Sucic and D. Turner (University of Michigan, USA) harboring bands electrophoresed on SDS–PAGE were immunodetected the Myc-tag sequence in the 5′ site. as described elsewhere (21). Purification of GST-fused Dnmt3 DNA methylation activity Poly(dI-dC)-poly(dI-dC) (dIdC) and poly(dG-dC)-poly(dG-dC) Dnmt3-harboring plasmids were expressed in BL21(DE3) at (dGdC) (Amersham Pharmacia Biotech) were used as the 20°C. After the OD at 650 nm reached 0.6, 0.5 mM IPTG was methyl acceptors for total and de novo methylation activity added to the culture medium, followed by further culturing for measurements, respectively (22,23). The methylation activity 18 h. All the purification steps thereafter were performed on was measured in 25 µ l of reaction (R) buffer (5 mM EDTA, ice or at 4°C. Harvested bacteria were suspended in 5 vol of 0.2 mM DTT, 26 mM NaCl, 20 mM Tris–HCl, pH 7.4) solubilization (S) buffer [0.33 M NaCl, 0.1% (w/v) Triton containing 50 ng (∼0.5 pmol) of the purified enzyme, 0.1 µ g X-100, 1/20 (v/v) protease inhibitor cocktail (Nakarai Tesque, DNA, which was ∼150 pmol CpI or CpG when 1 mol double- Japan) in Dulbecco’s phosphate-buffered saline (PBS)]. The stranded DNA with 1 CpI or CpG site was calculated to be cell suspension was sonicated and centrifuged at 10 000 g for 2 mol, and 134 pmol [ H]S-adenosyl-L-methione (AdoMet) 10 min, and the supernatant fraction recovered. For Dnmt3a, (15 Ci/mmol; Amersham Pharmacia Biotech), unless other- the solubilized fraction was loaded onto a DEAE–Sepharose wise stated. After the reaction, the mixtures were incubated (Amersham Pharmacia Biotech) column and the unbound frac- with 0.1 µ g of proteinase K (Nakalai Tesque, Japan) at 50°C tion was collected. For Dnmt3b, the solubilized fraction was for 20 min, and then the radioactivity was measured as precipitated and recovered by 20–30% saturation with described previously (21). Protein concentrations were deter- ammonium sulfate, and the precipitate was dissolved in 5 ml of mined by quantifying the CBB-stained SDS–PAGE gels with S buffer. Samples were loaded onto glutathione–Sepharose, an image analyzer (MCID; Imaging Research, Canada), using washed with 10 bedvol of Sbuffer, andthenelutedwith4bed bovine serum albumin as a standard. The specific activities vol of elution (E) buffer [0.33 M NaCl, 0.1% (w/v) Triton were expressed as moles of transferred methyl group per hour X-100, 1/200 (v/v) protease inhibitor cocktail, 10 mM per mole of each Dnmt3 (mol/h/mol Dnmt3). glutathione, reduced form, 1 mM DTT in 50 mM Tris–HCl, pH 8.0]. The eluted fractions were collected in tubes already Determination of the cytosine methylation containing a 1/10 vol of 1 M Tris–HCl, pH 7.4, and quickly As a methyl acceptor, the HindIII fragment of the myoD gene mixed. The main fractions were pooled and loaded onto a (accession no. X61655) subcloned into pUC19 (24) was used. Superdex 200 (Amersham Pharmacia Biotech) column equili- AdoMet (Sigma-Aldrich, MO) was purified on a Sep-Pack brated with E buffer minus reduced form glutathione. The Plus C18 column (Waters, Japan) before use. Dnmt3 at each step was checked by SDS–PAGE on a 7.5% For the methylation by GST-fused enzymes, the reaction polyacrylamide gel (20), and purity was monitored by mixture was incubated at 37°C for 1 h with 0.1 µ gpUC-myoD, Coomassie Brilliant Blue R-250 (CBB) staining. 3–5 µ g Dnmt3 protein, 100 µ M AdoMet in 500 µ lof R buffer. For the Myc-tagged Dnmt3 bound to Sepharose, the reaction Preparation of Myc-tagged Dnmt3 mixture contained 1 µ gpUC-myoD, 5 µ g Dnmt3 protein and Myc-tagged Dnmt3 was expressed in HEK 293T cells. Trans- 160 µ MAdoMetin 200 µ l of R buffer, and was then incubated fection with lipofectoamine (Gibco BRL, MD) was performed at 37°C for 3 h under agitation. After the reaction, the plasmid according to the manufacturer’s instructions. Cells transfected was recovered, and the methylation reaction was repeated with the plasmid were washed with PBS, and then solubilized twice more with fresh enzyme-coupled Sepharose and with 2 M NaCl, 0.3% Triton X-100, 1/50 (v/v) protease inhibitor AdoMet. cocktail in 20 mM Tris–HCl, pH 7.4. Dnmt3 thus extracted The sodium bisulfite reaction was performed basically as was immunoprecipitated with anti-Myc monoclonal antibody described elsewhere (25) with a slight modification. Briefly, (9E10)-coupled Sepharose at 4°C for 4 h. The matrix was HindIII-digested plasmids were treated with 0.3 M NaOH for washed with the extraction buffer and then used for the 30 min and then incubated in a mixture comprising 2.9 M methylation reaction. Na S O and 100 mM hydroquinone, with the pH adjusted to 5. 2 2 5 The reaction mixture was incubated for a total of 5 h at 55°C Antibodies with 1 min incubation at 95°Cevery1h. The recoveredDNA Antisera reacted with mouse Dnmt3a and Dnmt3b were raised was alkaline-treated and then used for PCR with three sets of against maltose binding protein fusion proteins, and an primers designed to amplify intron2 of the myoD gene as antiserum reacted with glutathione S-transferase (GST) was follows: upperF, GGTGGATTTAGGAGGATGAGTAAT- raised against GST. The antibodies raised in rabbits were GGAG; upperR, CCCCAAATACTAAAAACCAACCAC- immunoselected using GST fusion protein- or GST-coupled ACAC; lower1F: GGTTTTAGGATTGGGTTAGTTTTAGG- Sepharose as affinity matrices. To this end, a DNA [232 TGTTAGG; lower1R, CCCTCATACCTAAATACTCCTAA- (SmaI)–701 (SmaI), encoding amino acids 7–162] of mouse CATCTAACA; lower2F, GTGTTAGATGTTAGGAGTATT- Dnmt3a (accesion no. AF068625) and a DNA [1–662 (SalI), TAGGTATGAGGG; lower2R, CACCTAAACCACTAC- encoding amino acids 1–181] of mouse Dnmt3b1 (accession CCCCAAACC. Primer set upperF and upperR amplifies the no. AF068626) were ligated into pMAL-c2 (NEB, MA) and upper strand of modified intron2, and primer sets lower1F and pGEX vectors, and expressed in BL21(DE3) in the presence of lower1R, and lower2F and lower2R amplify the lower strand IPTG. GST and the fusion proteins were purified with of intron2. The amplification reaction comprised 35 cycles of glutathione–Sepharose or amylose resin according to the incubation of the reaction mixture for denaturation at 94°Cfor manufacturers’ instructions. The Dnmt3a, Dnmt3b and GST 1min,annealingat 60°C for 1 min and extension at 72°Cfor 3508 Nucleic Acids Research, 2001, Vol. 29, No. 17 Figure 1. Schematic illustration of Dnmt3a, Dnnmt3b1, Dnmt3b2 and Dnmt3b3. The regions of the conserved motifs in the C-terminal catalytic domain (I∼X) and the Cys-rich region (Cys-rich) are indicated. Dnmt3b2 lacks amino acid residues 363–382, and Dnmt3b3 lacks amino acid residues 363–382 and 731–813, of Dnmt3b1. GST in the N-terminus of each Dnmt3 indicates the fused GST moiety. The numbers above the bars are amino acid numbers. 1 min. The amplified fragment was subcloned into the SmaI site of pUC19 and the modified nucleotides were determined by the dideoxy method (26). RESULTS Expression and purification of Dnmt3a, Dnmt3b1, Dnmt3b2 and Dnmt3b3 in Escherichia coli Each Dnmt3 was expressed as a fusion protein with GST (GST-Dnmt3) in the pGEX vector (Fig. 1), and then expressed in E.coli as described in the Materials and Methods. The fractions at each purification step were electrophoresed and Figure 2. Purification of GST-fused Dnmt3s. An aliquot of a sample obtained stained (Fig. 2). The Dnmt3a expressed was purified with at each purification step was subjected to SDS–PAGE and then stained with CBB. (A) Solubilized (lane 1), DEAE–Sepharose unbound (lane 2), glutathione– DEAE–Sepharose (Fig. 2A, lane 2), glutathione–Sepharose Sepharose unbound (lane 3), washed (lane 4), eluted (lane 5) and Superdex (lane 5) and Superdex 200 (lane 6). Expressed Dnmt3b1, 3b2, 200 eluted (lane 6) fractions of GST-fused Dnmt3a. 1/20 000 (lanes 1–4), and 3b3 were fractionated with ammonium sulfate (Fig. 2B, 1/1000 (lane 5) and 1/100 (lane 6) of the recovered fractions were loaded. C and D, lanes 3), glutathione–Sepharose (lanes 7) and (B–D) Solubilized (lane 1), ammonium sulfate-precipitated with 20% satura- Superdex 200 (lanes 8). The apparent molecular weights of the tion (lane 2), 20–30%-saturation (lane 3), supernatant on 30% saturation (lane 4), glutathione–Sepharose unbound (lane 5), washed (lane 6), eluted (lane 7) GST-fused Dnmt3a, Dnmt3b1, Dnmt3b2 and Dnmt3b3 were and Superdex 200 eluted (lane 8) fractions of GST-fused Dnmt3b1 (B), 130, 120, 115 and 110 kDa, as calculated by SDS–PAGE (Fig. 2). Dnmt3b2 (C) and Dnmt3b3 (D). 1/20 000 (lanes 1–5), 1/1000 (lanes 6 and 7) Considering the size of GST (26 kDa), and the calculated and 1/100 (lane 8) of the recovered fractions were loaded. Arrowheads indicate molecular weights of Dnmt3a, Dnmt3b1, Dnmt3b2 and the position of Dnmt3. (E) Purified Dnmt3a (lane 1), Dnmt3b1 (lane 2), Dnmt3b2 (lane 3), Dnmt3b3 (lane 4) and GST (lane 5) were electrophoresed in Dnmt3b3 are 102, 97, 95 and 88 kDa, respectively (9), the the same gel and then CBB-stained. Purified Dnmt3a (lanes 6, 11 and 16), major bands (indicated by arrowheads) are expected to be Dnmt3b1 (lanes 7, 12 and 17), Dnmt3b2 (lanes 8, 13 and 18), Dnmt3b3 (lanes intact forms of the proteins. Since these bands reacted with 9, 14 and 19) and GST (lanes 10, 15 and 20) were immunodetected with both anti-GST and anti-Dnmt3 antibodies (Fig. 2E, lanes 6–19), specific antibodies reactive with Dnmt3a (lanes 6–10), Dnmt3b (lanes 11–15) all the purified Dnmt3 proteins possessed an intact N-terminal and GST (lanes 16–20), respectively. Molecular weight sizes (kDa) are indi- cated beside each gel. region. The final preparations were ∼80% pure, as judged on densitometry (Fig. 2E, lanes 1–4). DNA methylation activities of purified Dnmt3s C-terminal catalytic domain of Dnmt3b1 or Dnmt3b2, showed The DNA methylation activities were determined for the no DNA methylation activity. purified enzymes. Purified Dnmt3a, Dnmt3b1 and Dnmt3b2 We next examined the effect of salt on the methylation activities showed similar activities toward dI-dC compared to those of each Dnmt3. As shown in Figure 3, all the Dnmt3s were toward dG-dC when fixed amounts of AdoMet and DNA were inhibited in the presence of >100 mM NaCl. Even when the used (Table 1). The results clearly suggest that both Dnmt3a and Dnmt3b act as a de novo methylase. Interestingly, salt was changed to KCl, the profiles of inhibition of the Dnmt3b3, which lacks 62 amino acid residues of the enzymes were similar to those for NaCl. Nucleic Acids Research, 2001, Vol. 29, No. 17 3509 Table 1. Specific activities of GST-fused Dnmt3a, Dnmt3b1, Dnmt3b2 and Dnmt3b3. The methyltransferase activity was determined for each Dnmt3 Dnmt3a Dnmt3b1 Dnmt3b2 Dnmt3b3 a b dIdC 1.07 ± 0.34 (4) 0.68 ± 0.10 (5) 0.64 ± 0.08 (5) 0.00 ± 0.00 (3) dGdC 0.52 ± 0.09 (4) 0.43 ± 0.12 (3) 0.40 ± 0.07 (5) 0.00 ± 0.01 (3) DNA methylation activity was expressed as mol/h/mol Dnmt3 ± standard deviation. The numbers in parentheses are the numbers of DNA methylation activity measurements with different enzyme preparations. Figure 3. Effects of salts on the DNA methylation activity of Dnmt3s. The Figure 4. Effect of the AdoMet concentration on the activity of Dnmt3s. The effects of KCl (A)and NaCl (B) on the DNA methylation activities [specific DNA methylation activities [specific activity (s.a.) in mol/h/mol Dnmt3] of activity (s.a.) in mol/h/mol Dnmt3] of Dnmt3a (circles), Dnmt3b1 (open GST-fused Dnmt3a (A), Dnmt3b1 (B)and Dnmt3b2(C) were titrated with squares) and Dnmt3b2 (filled squares) was determined for methyl acceptors AdoMet (left). As methyl acceptors, dIdC (circles) and dGdC (squares) were dIdC (left) and dGdC (right). used. A typical example for each enzyme is demonstrated. Right panels show double reciprocal plots of the respective titration curves. Kinetic parameters of purified Dnmt3s To determine the K for AdoMet, a methyl donor for methyla- Typical titration profiles are shown in Figure 5, and the calculated DNA DNA K and V values obtained on three measurements tion activity, we measured the activity with AdoMet concentra- m max tions of 0.15–5.4 µ M. A typical titration curve for each enzyme with different preparations of enzymes are summarized in AdoMet AdoMet DNA is shown in Figure 4, and the calculated K and V Table 3. The K values of each enzyme for dI-dC and m max m values obtained on three measurements with different prepara- dG-dC expressed as the concentrations of CpI and CpG, AdoMet tions of enzymes are summarized in Table 2. The K respectively, were similar, suggesting that Dnmt3s fulfill the value for Dnmt3a was 2–4-fold smaller than those of Dnmt3b1 conditions for a de novo methylase. On comparison of their DNA DNA and Dnmt3b2, indicating that Dnmt3a has higher affinity for K values, Dnmt3a showed a relatively high K value m m AdoMet AdoMet than Dnmt3bs. For the calculated V values, compared with Dnmt3bs, indicating that Dnmt3a may have max Dnmt3a showed a higher value than those of Dnmt3bs. lower affinity for DNA than Dnmt3bs. In the presence of high K for DNA, a methyl acceptor, was also determined in the concentration of DNA, the activities were inhibited for both presence of dI-dC or dG-dC (0.15–18 µ Mof CpI or CpG). Dnmt3a and Dnmt3bs (Fig. 5). 3510 Nucleic Acids Research, 2001, Vol. 29, No. 17 Table 2. K values of Dnmt3a, Dnmt3b1 and Dnmt3b2 for AdoMet AdoMet AdoMet K (µ M) V (mol/h/mol Dnmt3) m max dIdC dGdC dIdC dGdC Dnmt3a 0.4 ± 0.1 0.3 ± 0.0 1.33 ± 0.31 1.22 ± 0.22 Dnmt3b1 1.2 ± 0.3 1.2 ± 0.3 1.00 ± 0.01 0.73 ± 0.11 Dnmt3b2 0.9 ± 0.2 0.8 ± 0.2 0.70 ± 0.17 0.69 ± 0.09 Three independent experiments similar to that shown in Figure 4 were performed using different Dnmt3 preparations. The values are means ± standard deviation. Table 3. K values of Dnmt3a, Dnmt3b1 and Dnmt3b2 for DNA DNA DNA K (µ M) V m max Z(mol/h/mol Dnmt3) dIdC dGdC dIdC dGdC Dnmt3a 2.7 ± 0.4 3.5 ± 1.2 1.7 ± 1.2 1.2 ± 0.7 Dnmt3b1 1.3 ± 0.4 1.0 ± 0.3 1.0 ± 0.1 0.8 ± 0.1 Dnmt3b2 1.5 ± 0.4 0.9 ± 0.3 1.1 ± 0.2 0.6 ± 0.1 Three independent experiments similar to that shown in Figure 5 were performed using different Dnmt3 preparations. Table 4. Methylation specificities of GST-fused Dnmt3a and Dnmt3b1 CpG CpA CpT CpC Dnmt3a 101.6 10.4 4.0 3.5 Dnmt3b1 47.1 15.7 24.3 1.4 Dnmt3b3 2.1 3.1 2.6 3.7 The methylation sites of the subcloned myoD gene were determined by the sodium bisulfite method. Totals of 1729, 700 and 1909 of CpN were analyzed after treatment with Dnmt3a, Dnmt3b1 and Dnmt3b3, respectively, and the numbers of methylated dinucleotides were determined and normalized as to 1000 CpN. Since Dnmt3b3 had no DNA methylation activity, the number of methylated CpN was taken as the background in this experiment. Figure 5. Effect of the DNA concentration on the activity of Dnmt3s. The DNA methylation activities [specific activity (s.a.) in mol/h/mol Dnmt3] of more than other dinucleotide sequences. Under identical GST-fused Dnmt3a (A), Dnmt3b1 (B)and Dnmt3b2(C) were titrated with conditions, Dnmt3a methylated 10.4, 4.0 and 3.5 cytosine dIdC (circles) or dGdC (squares). One mole double-stranded DNA with 1 CpI or CpG site was calculated to be 2 mol. A typical example for each enzyme is residues in CpA, CpT and CpC per 1000 CpN dinucleotides. demonstrated. Right panels show double reciprocal plot of the respective Considering the background (experimental error) with inactive titration curves. Dnmt3b3, which was 2.1, 3.1, 2.6 and 3.7 cytosine methylation for CpG, CpA, CpT and CpC per 1000 CpN dinucleotides, respectively, Dnmt3a significantly methylated CpA, although Sequence specificity of the Dnmt3 methylation site the level was very low. It has been postulated that Dnmt3a may methylate the cytosine Dnmt3b1 also methylated the CpG sequence most (47.1/1000) residues in sequences other than CpG sequences in vivo,as among the dinucleotide sequences; however, the specificity judged by analysis of Dnmt1-targeted embryonic stem cells was not as strict as in the case of Dnmt3a. Dnmt3b1 methylated (19). To directly analyze the sequence specificity of Dnmt3a CpA (15.7/1000) and CpT (24.3/1000) at relatively high and Dnmt3b, we used the myoD gene as the methyl acceptor frequencies compared to Dnmt3a. Neither Dnmt3a nor and determined the methylated sequences by the sodium Dnmt3b1 methylated the first cytosine in the CpC sequence. bisulfite method (25). The results are summarized in Table 4. We could not find any rule for the methylation sequence as to Dnmt3a methylated CpG (101.6 per 1000 CpN dinucleotides) the third nucleotide specificity (CpNpN). Nucleic Acids Research, 2001, Vol. 29, No. 17 3511 activity in vitro using purified enzymes. In the present study, we have expressed and purified GST-fused recombinant Dnmt3a and Dnmt3bs, and have shown that both Dnmt3a and Dnmt3bs exhibit de novo methylation activity in vitro.For Dnmt1 with dI-dC as the methyl acceptor, the activity is >10-fold higher than that toward dG-dC (27), and this property of Dnmt1 is the basis for its maintenance activity. On the other hand, although the specific activities were very low, Dnmt3a and Dnmt3bs showed similar activities toward both dI-dC and dG-dC, indicating that these enzymes have the ability to create a methylation pattern. Dnmt3a and Dnmt3bs each have a catalytic domain in the Figure 6. Expression and immuno-purification of Dnmt3s. Myc-tagged C-terminal half, in which the 10 motifs conserved in DNA Dnmt3a (lanes 1 and 2) and Dnmt3b1 (lanes 3 and 4) were expressed in HEK (cytosine-5) methyltransferases exist (28). Dnmt3b3 lacks 293T cells, extracted (lanes 1 and 3), and immunoprecipitated with anti-Myc 62 amino acid residues in the sequence between motifs VIII monoclonal antibody (9E10)-coupled Sepharose (lanes 2 and 4). Asterisks indicate the heavy chain of 9E10. Molecular weight sizes (kDa) are indicated and IX, called the TRD (target recognition domain) (28) in beside the gel. bacterial methylases, and the part of motif IX. It was not known whether Dnmt3b3 exhibits DNA methylation activity or not. In the present study, it has been shown that Dnmt3b3 has no DNA methylation activity. As Dnmt3b3 is expressed Table 5. Methylation specificities of Myc-tagged Dnmt3a and Dnmt3b1 in vivo under physiological conditions, the protein may play a role as a dominant negative factor. CpG CpA CpT CpC Similar to human maintenance-type DNA (cytosine-5) Dnmt3a 62.6 9.9 5.8 0.0 methyltransferase (DNMT1) (29), interestingly, all Dnmt3a, Dnmt3b1 11.1 1.1 1.1 1.1 Dnmt3b1 and Dnmt3b2 were inhibited at high concentrations –Dnmt3 0.0 0.0 0.0 0.0 of DNA (Fig. 5). Pradhan et al. proposed that DNMT1 may form a catalytically inactive complex with DNA at high The methylation sites of the subcloned myoD gene were determined by the concentrations of DNA through the domain other than its sodium bisulfite method. Totals of 6075, 4485 and 2793 of CpN were catalytic domain (29). The inhibitory effect of DNA on analyzed after treatment with Dnmt3a, Dnmt3b1, and without Dnmt3, respec- Dnmt3s could be an indication that Dnmt3s also have a DNA tively, and the numbers of methylated dinucleotides were determined and normalized as to 5000 CpN. binding domain other than their catalytic domain. Recently, the methylated cytosine residues in genomic DNA prepared from embryonic stem cells that lack Dnmt1 were analyzed (19). In these cells, the cytosine methylation frequencies In the recombinant Dnmt3, GST was in the N-terminus. To for CpA and CpT were reported to be ∼50 and 15% of that for eliminate the effect of GST, we changed the GST N-terminal the CpG sequence. In our study, with Dnmt3a, after subtracting tag to a Myc-tag. Myc-tagged Dnmt3a and Dnmt3b1 were the background, CpA and CpT were methylated to ∼7and 1%, expressed in HEK 293T cells under the CMV promoter and respectively, of the methylated CpG level. On the other hand, then immuno-purified with anti-Myc monoclonal antibody- Dnmt3b1 methylated CpA and CpT to ∼28 and 46%, respec- coupled Sepharose. The expressed and purified Dnmt3a and tively, of the CpG methylation level. The ability of Dnmt3a to Dnmt3b1 are shown in Figure 6. The enzymes bound to a methylate CpA and CpT itself cannot explain the levels of CpA matrix were used directly for the in vitro methylation reaction. andCpT methylationinDnmt1-targetedembryonic stem cells The results are summarized in Table 5. Although the activity (19). Dnmt3b is expected to contribute mainly to non-CpG was low compared to that of the GST-fused enzymes, the methylation, especially of the CpT sequence. Although we dinucleotide specificity of Dnmt3a was shown to be similar to added an equal amount of Dnmt3b1 as Dnmt3a protein to the that of GST-fused Dnmt3a (Table 4). Interestingly, Myc-tagged methylation reaction mixture, the myc-tagged Dnmt3b1 bound Dnmt3a significantly methylated CpT, which was not obvious to the matrix showed extremely low activity. The low activity when GST-fused Dnmt3a was used. The activity of Dnmt3b1 of Dnmt3b1 could be due to the instability of Dnmt3b1 was significantly low, only the CpG sequence specificity being compared to Dnmt3a. significant (11.1/5000 CpN dinucleotides). ACKNOWLEDGEMENTS DISCUSSION This work was supported in part by the Program for the Promo- De novo DNA methyltransferase activity plays a crucial role in tion of Fundamental Studies in Health Sciences of the Organi- the creation of methylation patterns in genomic DNA during zation for Pharmaceutical Safety and Research of Japan, the embryogenesis and germ line cells. Dnmt3a and Dnmt3b are Program for the Promotion of Basic Research Activities for the most likely candidates responsible for the de novo DNA Innovative Biosciences, and Grants-in-Aid from the Ministry methyltransferase activity (10). However, it has not been proven of Education, Culture, Sports, Science, and Technology of that Dnmt3a and Dnmt3bs exhibit de novo DNA methylation Japan. 3512 Nucleic Acids Research, 2001, Vol. 29, No. 17 methylation of constitutive heterochromatin in adult and fetal tissues. NOTE ADDED IN PROOF Hum. Genet., 99, 738–745. In the course of revision, Gowher and Jeltsch published work 15. Hsieh,C.L. (1999) In vivo activity of murine de novo methyltransferases, on purified recombinant Dnmt3a reporting a similar K for Dnmt3a and Dnmt3b. Mol. Cell. Biol., 19, 8211–8218. 16. Toth,M., Müeller,U. and Doerfler,W. (1990) Establishment of a de novo DNA and an ability for non-CpG methylation activity (30). methylation pattern. Transcription factor binding and deoxycytidine methylation at CpG and non-CpG sequences in an integrated adenovirus promoter. J. Mol. Biol., 214, 673–683. REFERENCES 17. Clark,S.J., Harrison,J. and Frommer,M. (1995) CpNpG methylation in 1. Antequera,F. and Bird,A. (1993) CpG islands in DNA methylation. In mammalian cells. Nature Genet., 10, 20–27. Jost,J.P. and Saluz,H.P. (eds), Molecular Biology and Biological 18. Woodcock,D.M., Lawler,C.B., Linsenmeyer,M.E., Doherty,J.P. and Significance. Birkhäuser Verlag, Basel, pp. 169–185. Warren,W.D. (1997) Asymmetric methylation in the hypermethylated 2. Shen,C.K.J. and Maniatis,T. (1980) Tissue-specific DNA methylation in a CpG promoter region of the human L1 retrotransposon. J. Biol. Chem., clusterofrabbit β-like globin genes. Proc. Natl Acad. Sci. USA, 77, 272, 7810–7816. 6634–6638. 19. Ramsahoye,B.H., Biniszkiewicz,D., Lyko,F., Clark,V., Bird,A. and 3. Razin,A. and Cedar,H. (1991) DNA methylation and gene expression. Jaenisch,R. (2000) Non-CpG methylation is prevalent in embryonic stem Microbiol. Rev., 55, 451–458. cells and may be mediated by DNA methyltransferase 3a. Proc. Natl 4. Tajima,S. and Suetake,I. (1998) Regulation and function of DNA Acad. Sci. USA, 97, 5237–5242. methylationinvertebrates. J. Biochem., 123, 993–999. 20. Laemmli,U.K. (1970) Cleavage of structural proteins during the assembly 5. Jaenisch,R. (1997) DNA methylation and imprinting: Why bother? of the head of bacteriophage T4. Nature, 227, 75–78. Trends Genet., 13, 323–329. 21. Kimura,H., Suetake,I. and Tajima,S. (1999) Xenopus maintenance-type 6. Riggs,A.D. and Porter,T.N. (1996) X-chromosome inactivation and DNA methyltransferase is accumulated in and translocated into germinal epigenetic mechanisms. In Russo,V.E.A., Martinssen,R.A. and vesicles of oocytes. J. Biochem., 125, 1175–1182. Riggs,A.D. (eds), Epigenetic Mechanisms of Gene Regulation. Cold 22. Pedrali-Noy,G. and Weissbach,A. (1986) Mammalian DNA Spring Harbor Press, Cold Spring Harbor, NY, pp. 231–248. methyltransferases prefer poly(dI-dC) as a substrate. J. Biol. Chem., 261, 7. Laird,P.W. and Jaenisch,R. (1996) The role of DNA methylation in cancer 7600–7602. genetics and epigenetics. Annu. Rev. Genet., 30, 441–464. 23. Pfeifer,G.P. and Drahvosky,D. (1986) Preferential binding of DNA 8. Monk,M. (1990) Changes in DNA methylation during mouse embryonic methyltransferase and increased de novo methylation of deoxyinosine- development in relation to X-chromosome activity and imprinting. containing DNA. FEBS Lett., 207, 75–78. Philos. Trans. R. Soc. Lond. B Biol. Sci., 326, 299–312. 24. Takagi,H., Tajima,S. and Asano,A. (1995) Overexpression of DNA 9. Okano,M., Xie,S. and Li,E. (1998) Cloning and characterization of a methyltransferase in myoblast cells accelerates myotube formation. family of novel mammalian DNA (cytosine-5) methyltransferases. Eur. J. Biochem., 231, 282–291. Nature Genet., 19, 219–220. 25. Clark,S.J., Harrison,J., Paul,C.L. and Frommer,M. (1994) High sensitivity 10. Okano,M., Bell,D.W., Haber,D.A. and Li,E. (1999) DNA mapping of methylated cytosine. Nucleic Acids Res., 22, 2990–2997. methyltransferases Dnmt3a and Dnmt3b are essential for de novo 26. Sanger,F., Nicklen,S. and Coulson,A.R. (1977) DNA sequencing with methylation and mammalian development. Cell, 99, 247–257. chain-terminating inhibitors. Proc. Natl Acad. Sci. USA, 74, 5463–5467. 11. Lyko,F., Ramsahoye,B.H., Kashevsky,H., Tudor,M., Mastrangelo,M.A., 27. Tajima,S., Tsuda,H., Wakabayashi,N., Asano,A., Mizuno,S. and Orr-Weaver,T.L. and Jaenisch,R. (1999) Mammalian (cytosine-5) Nishimori,K. (1995) Isolation and expression of a chicken DNA methyltransferases cause genomic DNA methylation and lethality in methyltransferase cDNA.J.Biochem., 117, 1050–1057. Drosophila. Nature Genet., 23, 363–366. 28. Kumar,S., Cheng,X., Klimasauskas,S., Mi,S., Posfai,J., Roberts,R.J. and 12. Xu,G.L., Bestor,T.H., Bourc’his,D., Hsieh,C.L., Tommerup,N., Wilson,G.G. (1994) The DNA (cytosine-5) methyltransferase. Bugge,M., Hulten,M., Qu,X., Russo,J.J. and Viegas-Pequignot,E. (1999) Nucleic Acids Res., 22, 1–10. Chromosome instability and immunodeficiency syndrome caused by 29. Pradhan,S., Bacolla,A., Wells,R.D. and Roberts,R.J. (1999) Recombinant mutations in a DNA methyltransferase gene. Nature, 402, 187–191. human DNA (cytosine-5) methyltranasferase. I. Expression, purification 13. Hansen,R.S., Wijmenga,C., Luo,P., Stanek,A.M., Canfield,T.K., and comparison of de novo and maintenance methylation. J. Biol. Chem., Weemaes,C.M.R. and Gartler,S.M. (1999) The DNMT3B DNA 274, 33002–33010. methyltransferase gene is mutated in the ICF immunodeficiency syndrome. Proc. Natl Acad. Sci. USA, 96, 14412–14417. 30. Gowher,H. and Jeltsch,A. (2001) Enzymatic properties of recombinant 14. Miniou,P., Jeanpierre,M., Bourc’his,D., Coutinho Barbosa,A.C., Dnmt3a DNA methyltransferase from mouse: the enzyme modifies DNA Blanquet,V. and Viegas-Pequignot,E. (1997) Alpha-satellite DNA in a non-processive manner and also methylates non-CpA sites. methylation in normal individuals and in ICF patients: heterogeneous J. Mol. Biol., 309, 1201–1208.

Journal

Nucleic Acids Research – Oxford University Press

Published: Sep 1, 2001

There are no references for this article.