CpG methylation of the CENP-B box reduces human CENP-B binding Yoshinori Tanaka 1,2, *, Hitoshi Kurumizaka 1,3 and Shigeyuki Yokoyama 1,2,4 1 Protein Research Group, RIKEN Genomic Sciences Center, Suehiro-cho, Tsurumi, Yokohama, Japan 2 Department of Biophysics and Biochemistry, Graduate School of Science, University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, Japan 3 Waseda University School of Science and Engineering, 3-4-1 Okubo, Shinjuku-ku, Tokyo, Japan 4 RIKEN Harima Institute at Spring-8, 1-1-1 Kohto, Mikazuki-cho, Sayo, Hyogo, Japan The centromere of eukaryotic chromosomes plays an essential role in the proper segregation of chromo- somes at mitosis and meiosis, and has a special hetero- chromatin structure, which is composed of a-satellite DNA repeats and their associated proteins. The human centromere proteins A, B and C (CENP-A, CENP-B and CENP-C, respectively) are such centro- mere-specific DNA-binding proteins [1–7]. Neither CENP-A nor CENP-C shows any sequence specificity in DNA binding. In contrast, CENP-B is known to specifically bind a 17 base-pair sequence (the CENP-B box), which appears in every other a-satellite repeat (171 base-pairs) in human centromeres [8–10]. CENP-B is an 80 kDa protein that contains DNA- binding and dimerization domains at the N-terminus and C-terminus, respectively [11–13]. Biochemical ana- lyses with nucleosomes reconstituted in vitro suggested that the CENP-B box sequence functions as a cis element for centromere-specific nucleosome assembly [14]. In vivo analyses with cultured human cells Keywords CENP-B; centromere; DNA methylation; chromatin; heterochromatin Correspondence 2 H. Kurumizaka, Waseda University School of Science and Engineering, 3-4-1 Okubo, Shinjuku-ku, Tokyo 169-8555, Japan Fax: +81 3 5292 9211 Tel: +81 3 5286 8189 E-mail: kurumizaka@waseda.jp S. Yokoyama, Protein Research Group, RIKEN Genomic Sciences Center, 1-7-22 Suehiro-cho, Tsurumi, Yokohama 230-0045, Japan Fax: +81 45 503 9195 Tel: +81 45 503 9196 E-mail: yokoyama@biochem.s.u-tokyo.ac.jp *Present address Toray Industries, Inc. New Frontiers Research Laboratories, 1111 Tebiro, Kamak- ura, Kanagawa 248–0036, Japan (Received 15 September 2004, accepted 1 October 2004) doi:10.1111/j.1432-1033.2004.04406.x In eukaryotes, CpG methylation is an epigenetic DNA modification that is important for heterochromatin formation. Centromere protein B (CENP- B) specifically binds to the centromeric 17 base-pair CENP-B box DNA, which contains two CpG dinucleotides. In this study, we tested complex formation by the DNA-binding domain of CENP-B with methylated and unmethylated CENP-B box DNAs, and found that CENP-B preferentially binds to the unmethylated CENP-B box DNA. Competition analyses revealed that the affinity of CENP-B for the CENP-B box DNA is reduced nearly to the level of nonspecific DNA binding by CpG methylation. Abbreviations CENP-B, centromere protein B; RNAi, RNA interference; siRNA, small interfering RNA. 282 FEBS Journal 272 (2005) 282–289 ª 2004 FEBS revealed that the presence of the CENP-B box is essen- tial for the formation of functional minichromosomes [15,16]. Therefore, the CENP-B box is required for the formation of a functional centromere. However, CENP-B null mice appeared to be normal [17–19]. This discrepancy about the CENP-B dispensibility for mouse development and the CENP-B box requirement for minichromosome maintenance in human cells may be explained by presuming the existence of functional homologues of CENP-B. In fact, CENP-B-like pro- teins have been identified in humans, and the func- tional redundancy of CENP-B homologues has also been found in the fission yeast Schizosaccharomyces pombe [20–22]. In eukaryotic cells, methylation of the cytosine within the CpG dinucleotide is an epigenetic DNA modification that is important for heterochromatin formation. The human a-satellite consensus sequence Fig. 1. Human a-satellite DNA and the CENP-B(1–129)–DNA complex structure. (A) Schematic diagram showing the organization of human a-satellite DNA and the CENP-B box. Arrows and circles in the middle row indicate a-satellite DNA repeats and CENP-B boxes, respectively. The a-satellite consensus sequence [23], containing a 17 bp CENP-B box, is shown in the bottom row. The three boxes, marked as sites 1, 2 and 3, indicate the essential bases for CENP-B binding to the CENP-B box DNA. The three CpG sequences in the a-satellite sequence are shown in red. (B) The CENP-B(1–129)–DNA complex structure [25]. The essential bases for CENP-B binding are coloured red in the model structure. The CpG sites that are sharply kinked by the CENP-B binding are encircled in red. The nucleotide sequences of sites 1 and 3 are indicated in the left column. (C) Model for the interaction between CENP-B and the CpG-methylated CENP-B box. (Left) Interaction between CENP-B and the methylated CENP-B box DNA at sites 1 and 3. The methyl groups, which are modelled on the C5 atoms of cytosines in the CpG dinucleotides, are coloured red. The van der Waals radii of the methyl groups (2.0 A ˚ ) are shown in orange circles. The van der Waals radii of the Thr44 and Arg125 side chains are shown in grey circles. Distances between the cytosine methyl groups and the nearest Thr44 and Arg125 side chain atoms are indicated by arrows. (Right) Interaction between CENP-B and the unmethylated CENP-B box DNA at sites 1 and 3. The pink dotted line in the right panel indicates the hydrogen bond between the side chain NH 2 group of Arg125 and the N7 atom of guanine. Y. Tanaka et al. CpG methylation reduces human CENP-B binding 1 FEBS Journal 272 (2005) 282–289 ª 2004 FEBS 283 contains only three CpG sequences within its 171 base- pair sequence [23]. Interestingly, two of the three CpG sequences in the a-satellite consensus sequence are located within site 1 (5¢-pTpTpCpG-3¢) and site 3 (5¢-pCpGpGpG-3¢) of the CENP-B box (Fig. 1A; [9]). Demethylation of satellite DNA sequences, accom- plished by growing cells in the presence of the DNA methyltransferase inhibitor, 5-aza-2¢-deoxycytidine, resulted in the redistribution of CENP-B [24], indica- ting that the CpG methylation affects the CENP- B–DNA interaction. Our structural analysis of the CENP-B DNA-binding domain [CENP-B(1–129)] complexed with the CENP-B box DNA revealed that CENP-B induced sharp kinks at the CpG sequences of sites 1 and 3 upon DNA binding (Fig. 1B; [25]). There- fore, CpG methylation may regulate CENP-B binding to the CENP-B box DNA within centromeric a-satel- lite repeats. In this study, we tested the complex formation of CENP-B(1-129) with methylated and unmethylated CENP-B box DNAs by a complex-reconstitution assay, and found that the affinity of CENP-B for CpG methylated CENP-B box DNA is significantly reduced. Results Model for the interaction between CENP-B and the CpG-methylated CENP-B box In eukaryotes, DNA methylation occurs at the C5 atom of cytosine by the action of methyltransferases. To evaluate the effect of CpG methylation on the CENP-B–DNA interaction, we modeled additional methyl groups at the C5 atoms of site 1 (cytosine 6 and cytosine 7¢) and site 3 (cytosine 15 and cytosine 16¢)in the CENP-B(1–129)–DNA complex structure (Fig. 1C). In the crystal structure of the CENP-B(1–129)–DNA complex [25], the CpG sequences in sites 1 and 3 of the CENP-B box DNA were sharply kinked by CENP- B(1–129) binding (16° for site 1 and 43° for site 3; Fig. 1B). As shown in Fig. 1C, additional methyl groups on cytosine 7¢ and cytosine 15 caused steric cla- shes with the side chains of Thr44 and Arg125, respect- ively. Therefore, the CENP-B(1–129)–methylated DNA complex model suggested that the methylations of cyto- sine 7¢ and cytosine 15 may sterically interfere with CENP-B binding to the CENP-B box DNA. The complex-reconstitution assay We tested whether the CpG methylations of the CENP-B box DNA actually affect CENP-B binding to the DNA, as suggested by the CENP-B(1–129)– methylated DNA complex model. To do so, we employed a complex-reconstitution assay with recom- binant CENP-B(1–129) and the methylated and un- methylated CENP-B box DNAs (21 bp; Fig. 2A). As the recombinant CENP-B(1–129) was only detected in the insoluble fraction when it was expressed in Escherichia coli cells, CENP-B(1–129) was purified under denaturing conditions in the presence of 6 m urea (Fig. 2B). In the complex-reconstitution assay, CENP-B(1–129) formed the complex with the CENP- B box DNA during the refolding process by dialysis against buffer without urea. The complex-formation efficiency was monitored by a gel-shift assay (Fig. 2C). A BC Fig. 2. The complex-reconstitution assay with CENP-B(1–129) and the CENP-B box DNA. (A) Schematic presentation of the complex- reconstitution assay. CENP-B(1–129) and the CENP-B box DNA strands are shown in blue and red, respectively. CENP-B(1–129) and the CENP-B box DNA were mixed under denaturing conditions in the presence of 6 M urea. The CENP-B(1–129)–DNA complex was reconstituted during the refolding process by dialysis against buffer without urea. (B) The purified recombinant CENP-B(1–129) protein was analyzed by 15–25% SDS/PAGE. Lane 1 indicates molecular mass markers (M), and lane 2 indicates CENP-B(1–129) purified under denaturing conditions in the presence of 6 M urea. (C) CENP-B(1–129)–DNA complex formation. The CENP-B(1–129)– DNA complex, reconstituted by the method shown in panel A, was analyzed by 20% nondenaturing PAGE. Bands were visualized by ethidium bromide staining. CpG methylation reduces human CENP-B binding 1 Y. Tanaka et al. 284 FEBS Journal 272 (2005) 282–289 ª 2004 FEBS CENP-B(1–129) preferentially binds to the unmethylated CENP-B box DNA rather than the methylated form Using the complex-reconstitution assay, we tested the binding of CENP-B(1–129) to the unmethylated and methylated CENP-B box DNAs (Fig. 3A). As shown in Fig. 3B, CENP-B(1–129) efficiently formed the com- plex with the unmethylated CENP-B box DNA in a concentration-dependent manner (lanes 1–5). CENP- B(1–129) also formed the complex with the methylated CENP-B box DNA (Fig. 3B, lanes 6–10), but with slightly reduced efficiency, as compared with the unmethylated CENP-B box DNA (Fig. 3C). Therefore, CENP-B has the potential to bind to the methylated CENP-B box DNA. A B CE D Fig. 3. CENP-B(1–129) preferentially binds to unmethylated CENP-B box DNA. (A) The 21-mer CENP-B box DNA and the 21-mer nonspecific DNA (DnaA box DNA [30]), used in this study. The methylated cytosine residues in the methylated CENP-B box DNA are labeled by CH 3 in the middle row. (B) Gel-shift analysis of complex formation between CENP-B(1–129) and the CENP-B box DNA, complexed with increasing amounts of CENP-B(1–129), by 20% PAGE. The unmethylated CENP-B box DNA (lanes 1–5; 1 l M) and the methylated CENP-B box DNA (lanes 6–10; 1 l M) were used in this study. The CENP-B(1–129) concentrations were 0 lM (lanes 1 and 6), 0.5 lM (lanes 2 and 7), 1 lM (lanes 3 and 8), 3 lM (lanes 4 and 9), and 5 lM (lanes 5 and 10). (C) Graphic representation of the complex formation rates shown in panel B. Unmethylated (d) and methylated (s) CENP-B box DNAs, respectively. (D) Competition analysis for CENP-B binding between the un- methylated and methylated CENP-B box DNAs. The 32 P-labeled unmethylated CENP-B box DNA (1 lM)orthe 32 P-labeled methylated CENP- B box DNA (1 l M) was mixed with the indicated amounts of the unlabeled competitor methylated or unmethylated CENP-B box DNAs, respectively, in the presence of CENP-B(1–129) (3 l M). The complexes were analyzed by 20% PAGE. (E) Graphic representation of the com- plex formation rates shown in panel D. Closed and open circles indicate experiments with the 32 P-labeled unmethylated (d) and methylated (s) CENP-B box DNAs. Complex formation rates (%) are plotted against the concentrations of the competitor DNAs. Y. Tanaka et al. CpG methylation reduces human CENP-B binding 1 FEBS Journal 272 (2005) 282–289 ª 2004 FEBS 285 Next, we tested the binding affinity of CENP-B(1– 129) to the methylated CENP-B box DNA. To do so, we performed a competition assay. In this assay, complex formation between CENP-B(1–129) and the 32 P-labeled unmethylated or methylated CENP-B box DNAs was tested in the presence of unlabeled compet- itor DNA. As shown in Fig. 3D, complex formation between CENP-B(1–129) and the 32 P-labeled methyla- ted CENP-B box DNA was significantly inhibited in the presence of only one-third of the amount of unlabe- led unmethylated CENP-B box DNA (0.33 lm), relat- ive to the 32 P-labeled methylated form (1 lm) (lanes 7–12, and E). In contrast, complex formation between CENP-B(1–129) and the 32 P-labeled unmethylated CENP-B box DNA was not affected, even in the pres- ence of a threefold excess of unlabeled methylated CENP-B box DNA (3 lm) (Fig. 3D, lanes 1–6, and E). These results indicate that CENP-B preferentially forms a complex with the unmethylated CENP-B box DNA, rather than the methylated CENP-B box DNA. The CENP-B binding affinity to the methylated CENP-B box DNA is similar to the level of nonspecific DNA binding In order to compare the CENP-B binding to the CENP-B box DNA and a nonspecific DNA, we per- formed a competition assay with the methylated or unmethylated CENP-B box DNAs and a nonspecific DNA. The DnaA box sequence, which is not specific- ally bound by CENP-B, was used as nonspecific DNA (Fig. 3A, bottom row). The complex formation between CENP-B(1–129) and the 32 P-labeled unmethy- lated CENP-B box DNA was not affected when it was titrated with unlabeled nonspecific DNA (Fig. 3A and 4A, lanes 1–6, and B). On the other hand, the complex between CENP-B(1–129) and the 32 P-labeled methyla- ted CENP-B box DNA was significantly dissociated in the presence of an equal amount of unlabeled nonspe- cific competitor DNA (1 lm) (Fig. 4A, lanes 7–12, and B). Therefore, the CpG methylation at sites 1 and 3 of the CENP-B box DNA reduces the CENP-B binding affinity almost to the level of nonspecific DNA binding. Discussion In this study, we found that CpG methylation of the CENP-B box sequence reduces the binding affinity between CENP-B and the CENP-B box DNA nearly to the level of nonspecific binding. In the CENP-B(1– 129)–methylated DNA complex model, the additional methyl groups on cytosine 7¢ and cytosine 15 caused steric clashes with the side chains of Thr44 and Arg125, respectively (Fig. 1C). In site 3, the CENP-B a-helix 8 (120–129 amino acid residues), containing four Arg residues (Arg125, 127, 128 and 129), penet- rates perpendicularly into the major groove around the CpG sequence. Interestingly, the Arg125 side chain directly formed a hydrogen bond with the N7 atom of guanine 16, and specifically recognized the site 3 sequence (Fig. 1C, right). In contrast, the Arg127, 128 and 129 residues bound to the backbone phosphates, and did not directly interact with the DNA bases. Steric hindrance of the specific interaction between Arg125 and the N7 atom of guanine 16 by the CpG methylation may be a reason for the reduced specificity of CENP-B to the CENP-B box sequence. What is the functional meaning of the CpG methyla- tion of the CENP-B box DNA? Recently, a link between centromeric heterochromatin formation and the RNA interference (RNAi) machinery was discov- ered. In fission yeast, the RNAi machinery is required for chromosome segregation, gene silencing and nor- mal centromere function [ 3 26–28]. In chicken–human A B Fig. 4. CpG methylation reduces the affinity between CENP-B and the CENP-B box DNA to nearly the level of nonspecific DNA bind- ing. (A) The 32 P-labeled CENP-B box DNA (1 lM) was mixed with the indicated amounts of the unlabeled nonspecific DNA competitor in the presence of CENP-B(1–129) (3 l M). The complexes were analyzed by 20% PAGE. (B) Graphic representation of the complex formation rates shown in panel A. 32 P-labeled unmethylated (d) and methylated (s) CENP-B box DNAs. Formation rates (%) are plotted against the concentrations of the competitor DNAs. CpG methylation reduces human CENP-B binding 1 Y. Tanaka et al. 286 FEBS Journal 272 (2005) 282–289 ª 2004 FEBS hybrid DT40 cells, the loss of Dicer, which cleaves double-stranded RNAs into 21- to 23-nucleotide small interfering RNAs (siRNAs), resulted in cell death with the accumulation of abnormal mitotic cells, and caused the accumulation of transcripts from the human a-sat- ellite [29]. In Dicer-proficient cells, the a-satellite RNA transcripts are quickly processed to siRNAs, which may induce heterochromatin formation in the chromo- somal regions with the same sequences as the siRNAs. These findings suggest that the a-satellite RNA tran- scripts may be involved in functional centromere for- mation. CENP-B has the potential to induce nucleosome assembly in the vicinity of the CENP-B box sequence [14]. This CENP-B-induced nucleosome assembly may inhibit the production of RNA tran- scripts from the a-satellite DNA (Fig. 5). Therefore, the CpG methylations of the CENP-B box sequence may function in RNAi-dependent heterochromatin for- mation by regulating CENP-B-binding to the CENP-B box sequence in the a-satellite repeats. The relationship between the DNA binding of CENP-B and the pro- duction of the a-satellite RNA transcripts is an intrigu- ing subject for future studies. In eukaryotes, the CpG methylation patterns are epigenetically conserved. In this study, we found that CENP-B preferentially binds to the unmethylated CENP-B box DNA, rather than the methylated form. Stable binding of CENP-B to the unmethylated CENP-B box sequence may limit the access of methyl- transferases, and maintain the unmethylated region within the centromere. It has been reported that deme- thylation of satellite DNA sequences by the DNA methyltransferase inhibitor, 5-aza-2¢-deoxycytidine, resulted in the redistribution of CENP-B [24]. There- fore, the CENP-B localization, which depends on the CpG methylation, may function as an epigenetic marker to form a functional centromeric chromatin structure during cell division. Experimental procedures The complex-reconstitution assay The recombinant CENP-B(1–129) protein was purified under denaturing conditions in the presence of 6 m urea, as des- cribed previously [25]. The synthesized 21-mer oligonucleo- tides were purchased from Espec-Oligo (Ibaraki, Japan) 4 . The indicated amounts of the purified CENP-B(1–129) protein and the CENP-B box DNA strands (1 lm for each strand) were mixed in the presence of 6 m urea, 500 mm NaCl, 5 mm dithiothreitol and 0.1 mgÆmL )1 BSA, and the volume of the reaction was adjusted to 80 lL. The reaction mixtures were first dialyzed against 10 mm Tris ⁄ HCl buffer (pH 7.5), con- taining 500 mm NaCl, 5 mm dithiothreitol and 6 m urea, for 4 h at room temperature, and then for 4 h at 4 °C. Then, the samples were dialyzed against 10 mm Tris ⁄ HCl buffer (pH 7.5), containing 100 mm NaCl and 1 mm 2-mercapto- ethanol, for 16 h at 4 °C without urea. A 10 lL aliquot of the reaction mixture was mixed with 4 lL of 20% (v ⁄ v) gly- cerol, and was analyzed on a 20% (w ⁄ v) polyacrylamide gel in 0.5· TBE buffer (45 mm Tris base, 45 mm boric acid and 1.25 mm EDTA). The gel (20 cm · 20 cm · 0.1 cm) was run at 10 mA for 5 h. Bands were visualized by ethidium bro- mide staining or by autoradiography, if the 32 P-labeled DNA was used as a substrate, and were quantitated with a BAS2500 image analyzer (Fuji, Tokyo, Japan) 5 . Competition analysis The purified CENP-B(1-129) protein (3 lm) and the 32 P-labe- led CENP-B box DNA strands (1 lm for each strand) were mixed in the presence of 6 m urea, 500 mm NaCl, 5 mm di- thiothreitol and 0.1 mgÆmL )1 BSA. Then, various amounts Fig. 5. Model for a link between CENP-B box CpG methylation and RNAi-dependent heterochromatin formation. Arrows and purple cir- cles indicate a-satellite DNA repeats and CENP-B boxes, respect- ively. The methylated CENP-B boxes are labelled ‘Me’. Red and yellow circles indicate CENP-B and CENP-C, respectively, and green circles indicate nucleosomes. Blue bars indicate RNA tran- scripts, which are produced from the a-satellite DNA for the RNA- i-dependent heterochromatin formation. In this model, the CpG methylation pattern of the CENP-B boxes in the a-satellite repeats are epigenetically maintained, and CENP-B preferentially binds to the unmethylated CENP-B boxes. CENP-B may induce centromere- specific nucleosome assembly with CENP-A, CENP-C and histones, and the nucleosomes may inhibit the production of RNA transcripts from the a-satellite DNA. The a-satellite RNA transcripts are quickly processed by Dicer to siRNAs, which may be involved in centro- meric heterochromatin formation [29]. Y. Tanaka et al. CpG methylation reduces human CENP-B binding 1 FEBS Journal 272 (2005) 282–289 ª 2004 FEBS 287 of unlabeled competitor DNA (21-mer) were added to the reaction mixture, and the volume was adjusted to 80 lL. The reaction mixtures were first dialyzed against 10 mm Tris ⁄ HCl buffer (pH 7.5), containing 500 mm NaCl, 5 mm dithiothrei- tol and 6 m urea, for 4 h at room temperature, and then for 4 h at 4 °C. Then, the samples were dialyzed against 10 mm Tris ⁄ HCl buffer (pH 7.5), containing 100 mm NaCl and 1mm 2-mercaptoethanol, for 16 h at 4 °C without urea. A 10 lL aliquot of the reaction mixture was mixed with 4 lL of 20% (v ⁄ v) glycerol, and was analyzed on a 20% (w ⁄ v) polyacrylamide gel in 0.5· TBE buffer (45 mm Tris base, 45 mm boric acid and 1.25 mm EDTA). The gel (20 cm · 20 cm · 0.1 cm) was run at 10 mA for 5 h. Bands were visualized by autoradiography, and were quantitated with a BAS2500 image analyzer (Fuji). Acknowledgements This work was supported by the Bioarchitect Research Program (RIKEN), the RIKEN Structural Geno- mics ⁄ Proteomics Initiative (RSGI), the National Project on Protein Structural and Functional Analyses, and Grants-in-Aid from the Japanese Society for the Promo- tion of Science (JSPS) and the Ministry of Education, Sports, Culture, Science and Technology, Japan. References 1 Earnshaw WC & Rothfield N (1985) Identification of a family of human centromere proteins using autoimmune sera from patients with scleroderma. Chromosoma 91, 313–321. 2 Palmer DK, O’Day K, Trong HL, Charbonneau H & Margolis RL (1991) Purification of the centromere- specific protein CENP-A and demonstration that it is a distinctive histone. Proc Natl Acad Sci USA 88, 3734– 3738. 3 Sullivan KF, Hechenberger M & Masri K (1994) Human CENP-A contains a histone H3 related histone fold domain that is required for targeting to the centro- mere. J Cell Biol 127, 581–592. 4 Earnshaw WC, Sullivan KF, Machlin PS, Cooke CA, Kaiser DA, Pollard TD, Rothfield NF & Cleveland DW (1987) Molecular cloning of cDNA for CENP-B, the major human centromere autoantigen. J Cell Biol 104, 817–829. 5 Saitoh, H, Tomkiel J, Cooke CA, Ratrie H 3rd, Maurer M, Rothfield NF & Earnshaw WC (1992) CENP-C, an autoantigen in scleroderma, is a component of the human inner kinetochore plate. Cell 70, 115–125. 6 Sugimoto K, Yata H, Muro Y & Himeno M (1994) Human centromere protein C (CENP-C) is a DNA- binding protein which possesses a novel DNA-binding motif. J Biochem (Tokyo) 116, 877–881. 7 Politi V, Perini G, Trazzi S, Pliss A, Raska I, Earnshaw WC & Valle GD (2002) CENP-C binds the alpha-satel- lite DNA in vivo at specific centromere domains. J Cell Sci 115, 2317–2327. 8 Masumoto H, Masukata H, Muro Y, Nozaki N & Okazaki T (1989) A human centromere antigen (CENP-B) interacts with a short specific sequence in alphoid DNA, a human centromeric satellite. J Cell Biol 109, 1963–1973. 9 Masumoto H (1993) Properties of CENP-B and its Target Sequence in a Satellite DNA. Springer–Verlag, Berlin, Germany. 10 Ikeno M, Masumoto H & Okazaki T (1994) Distribu- tion of CENP-B boxes reflected in CREST centromere antigenic sites on long-range alpha-satellite DNA arrays of human chromosome 21. Hum Mol Genet 3, 1245–1257. 11 Yoda K, Kitagawa K, Masumoto H, Muro Y & Okazaki T (1992) A human centromere protein, CENP-B, has a DNA binding domain containing four potential alpha helices at the NH2 terminus, which is separable from dimerizing activity. J Cell Biol 119, 1413–1427. 12 Kitagawa K, Masumoto H, Ikeda M & Okazaki T (1995) Analysis of protein–DNA and protein–protein interactions of centromere protein B (CENP-B) and properties of the DNA-CENP-B complex in the cell cycle. Mol Cell Biol 15, 1602–1612. 13 Tawaramoto MS, Park SY, Tanaka Y, Nureki O, Kurumizaka H & Yokoyama S (2003) Crystal structure of the human CENP-B dimerization domain at 1.65 A resolution. J Biol Chem 278, 51454–51461. 14 Yoda K, Ando S, Okuda A, Kikuchi A & Okazaki T (1998) In vitro assembly of the CENP-B ⁄ alpha-satellite DNA ⁄ core histone complex: CENP-B causes nucleo- some positioning. Genes Cells 3, 533–548. 15 Ikeno M, Grimes B, Okazaki T, Nakano M, Saitoh K, Hoshino H, McGill NI, Cooke H & Masumoto H (1998) Construction of YAC-based mammalian artificial chromosomes. Nat Biotechnol 16, 431–439. 16 Ohzeki J, Nakano M, Okada T & Masumoto H (2002) CENP-B box is required for de novo centromere chro- matin assembly on human alphoid DNA. J Cell Biol 159, 765–775. 17 Fowler KJ, Hudson DF, Salamonsen LA, Edmondson SR, Earle E, Sibson MC & Choo KH (2000) Uterine dysfunction and genetic modifiers in centromere protein B-deficient mice. Genome Res 10, 30–41. 18 Hudson DF, Fowler KJ, Earle E, Saffery R., Kalitsis P, Trowell H, Hill J, Wreford NG, de Kretser DM, Cancilla MR et al. (1998) Centromere protein B null mice are mitotically and meiotically normal but have lower body and testis weights. J Cell Biol 141, 309–319. CpG methylation reduces human CENP-B binding 1 Y. Tanaka et al. 288 FEBS Journal 272 (2005) 282–289 ª 2004 FEBS 19 Perez-Castro AV, Shamanski FL, Meneses JJ, Lovato TL, Vogel KG, Moyzis RK & Pedersen R (1998) Centromeric protein B null mice are viable with no apparent abnormalities. Dev Biol 201, 135–143. 20 Toth M, Grimsby J, Buzsaki G & Donovan GP (1995) Epileptic seizures caused by inactivation of a novel gene, jerky, related to centromere binding protein-B in trans- genic mice. Nat Genet 11, 71–75. 21 Smit AF & Riggs AD (1996) Tiggers and DNA transpo- son fossils in the human genome. Proc Natl Acad Sci USA 93, 1443–1448. 22 Irelan JT, Gutkin GI & Clarke L (2001) Functional redundancies, distinct localizations and interactions among three fission yeast homologs of centromere pro- tein-B. Genetics 157, 1191–1203. 23 Choo KH, Vissel B, Nagy A, Earle E & Kalitsis P (1991) A survey of the genomic distribution of alpha satellite DNA on all the human chromosomes, and deri- vation of a new consensus sequence. Nucleic Acids Res 19, 1179–1182. 24 Mitchell AR, Jeppesen P, Nicol L, Morrison H & Kipling D (1996) Epigenetic control of mammalian cen- tromere protein binding: does DNA methylation have a role? J Cell Sci 109, 2199–2206. 25 Tanaka Y, Nureki O, Kurumizaka H, Fukai S, Kawag- uchi S, Ikuta M, Iwahara J, Okazaki T & Yokoyama S (2001) Crystal structure of the CENP-B protein-DNA complex: the DNA-binding domains of CENP-B induce kinks in the CENP-B box DNA. EMBO J 20, 6612– 6618. 26 Provost P (2002) Dicer is required for chromosome seg- regation and gene silencing in fission yeast. Proc Natl Acad Sci USA 99, 16648–16653. 27 Hall IM, Moma K & Grewal SIS (2003) RNA interfer- ence machinery regulates chromosome dynamics during mitosis and meiosis in fission yeast. Proc Natl Acad Sci USA 100, 193–198. 28 Volpe T, Schramke V, Hamilton GL, White SA, Teng G, Martienssen RA & Allshire RC (2003) RNA interfer- ence is required for normal centromere function in fis- sion yeast. Chromosome Res 11, 137–146. 29 Fukagawa T, Nogami M, Yoshikawa M, Ikeno M, Okazaki T, Takami Y, Nakayama T & Oshimura M (2004) Dicer is essential for formation of the hetero- chromatin structure in vertebrate cells. Nat Cell Biol 6, 784–791. 30 Fujikawa N, Kurumizaka H, Nureki O, Terada T, Shirouzu M, Katayama T & Yokoyama S (2003) Structural basis of replication origin recognition by the DnaA protein. Nucleic Acids Res 31, 2077–2086. Y. Tanaka et al. CpG methylation reduces human CENP-B binding 1 FEBS Journal 272 (2005) 282–289 ª 2004 FEBS 289 . with CENP-B binding to the CENP-B box DNA. The complex-reconstitution assay We tested whether the CpG methylations of the CENP-B box DNA actually affect CENP-B. regulating CENP-B- binding to the CENP-B box sequence in the a-satellite repeats. The relationship between the DNA binding of CENP-B and the pro- duction of the