Solution structure of the matrix attachment region-binding domain of chicken MeCP2 Bjo¨ rn Heitmann 1 , Till Maurer 1 , Joachim M. Weitzel 2 , Wolf H. Stra¨ tling 2 , Hans Robert Kalbitzer 1 and Eike Brunner 1 1 Institut fu ¨ r Biophysik und physikalische Biochemie, Universita ¨ t Regensburg, Germany; 2 Institut fu ¨ r Medizinische Biochemie und Molekularbiologie, Universita ¨ tsklinikum Hamburg-Eppendorf, Germany Methyl-CpG-binding protein 2 (MeCP2) is a multifunc- tional protein involved in chromatin organization and silencing of methylated DNA. MAR-BD, a 125-amino- acid residue domain of chicken MeCP2 (cMeCP2, origin- ally named ARBP), is the minimal protein fragment required to recognize MAR elements and mouse satellite DNA. Here we report the solution structure of MAR-BD as determined by multidimensional heteronuclear NMR spectroscopy. The global fold of this domain is very simi- lar to that of rat MeCP2 MBD and MBD1 MBD (the methyl-CpG-binding domains of rat MeCP2 and methyl- CpG-binding domain protein 1, respectively), exhibiting a three-stranded antiparallel b-sheet and an a-helix a 1 . We show that the C-terminal portion of MAR-BD also contains an amphipathic helical coil, a 2 /a 3 . The hydrophilic residues of this coil form a surface opposite the DNA interface, available for interactions with other domains of MeCP2 or other proteins. Spectroscopic studies of the complex formed by MAR-BD and a 15-bp fragment of a high-affinity binding site from mouse satellite DNA indi- cates that the coil is also involved in proteinÆDNA inter- actions. These studies provide a basis for discussion of the consequences of six missense mutations within the helical coil found in Rett syndrome cases. Keywords: chicken methyl-CpG-binding protein 2 (cMeCP2); MAR-binding protein (ARBP); NMR spectro- scopy; proteinÆDNA interaction. Methylation of the DNA at cytosines in the dinucleotide sequence CpG plays an important role in the regulation of gene expression and imprinting as well as during develop- ment. The information laid down in the methylation pat- tern is read by a family of methyl-CpG-binding proteins: MeCP2, MBD1, MBD2, MBD3, and MBD4 [1]. The founding member of this family is MeCP2, methyl-CpG- binding protein 2. Rat MeCP2 was identified through its ability to recognize methylated DNA [2], and the chicken homolog (originally named ARBP) was identified through its ability to bind MAR elements, the putative bases of chromatin loop domains [3]. MeCP2 acts as a transcrip- tional repressor [4] and exerts this function through interaction with the corepressor mSin3A and targeting of histone deacetylases to methylated DNA [5]. An additional histone deacetylase-independent mode of repression may operate for a distinct set of promoters [6,7]. Targeting of histone deacetylases is also involved in transcriptional repression by MBD1 [8]. MBD3 is a component of a multisubunit remodeling complex, NuRD, containing his- tone deacetylase activities [9]. MBD2 interacts with the NuRD complex and directs it to methylated DNA. MeCP2 is expressed in all tissues of the human body and, at particularly high levels, in neurons of the postnatal brain [10,11]. This observation is in line with the fact that mutations in the MECP2 gene cause Rett syndrome, an X-linked, dominant neurological disorder that is one of the most common causes of mental retardation in females [12]. At 6–18 months of age, affected girls gradually lose any acquired speech and purposeful hand use. They also suffer from microcephaly, severe mental retardation, autistic behavior, seizures, gait apraxia, and breathing abnormali- ties. Studies on transgenic mice that mimic the Rett phenotype indicate that MeCP2 is required for the main- tenance of neuronal physiology rather than brain develop- ment [13,14]. MeCP2 is an abundant component of the pericentromeric heterochromatin of mouse chromosomes [2]. In methylated murine major satellite DNA, MeCP2 recognizes in vitro two sites (I and II) with high affinity: K d ¼ (2.2–5.7) · 10 )10 M [15]. In nonmethylated satellite DNA, MeCP2 binds to these sites with slightly reduced affinity [K d ¼ (6.2–13.2) · 10 )10 M ]. The DNA-binding region of MeCP2 is the most highly conserved portion of the protein. The minimal sequence necessary to recognize methylated DNA (named methyl-CpG-binding domain, MBD) comprises Correspondence to E. Brunner, Institut fu ¨ r Biophysik und physikalische Biochemie, Universita ¨ t Regensburg, D-93040 Regensburg, Germany. Fax: + 49 941943 2479, Tel.: + 49 941943 2492, E-mail: eike.brunner@biologie.uni-regensburg.de Abbreviations: MeCP2, methyl-CpG-binding protein 2; cMeCP2, chicken MeCP2; rMeCP2, rat MeCP2; MBD, methyl-CpG-binding domain; MBD1, 2, 3, and 4, methyl-CpG-binding domain protein 1, 2, 3, and 4; MAR-BD, matrix attachment region-binding domain; ARBP, attachment region-binding protein; mSin3A, a mammalian corepressor protein interacting with MeCP2; NuRD, a multisubunit complex including MBD3; HSQC, heteronuclear single-quantum coherence; HBHA(CO)NH and CC(CO)NH, names of 3D heteronuclear correlation NMR experiments. (Received 1 April 2003, revised 4 June 2003, accepted 10 June 2003) Eur. J. Biochem. 270, 3263–3270 (2003) Ó FEBS 2003 doi:10.1046/j.1432-1033.2003.03714.x amino-acid residues 78–162 [16], while the minimal frag- ment required to bind to a chicken MAR element (here called MAR-binding domain, MAR-BD) encompasses residues 71–195 of human MeCP2 [corresponding to residues 72–196 in chicken (c)MeCP2] [15]. The solution structure of the MBD of rat (r)MeCP2 has recently been determined [17]. The MBD adopts a wedge- shaped structure, composed of an antiparallel b-sheet on one face of the wedge and a three-turn a-helix with an antiparallel one-turn helix on the other face. It is thought that the two inner strands of the b-sheet lie within the major groove of the DNA and that a hydrophobic pocket formed by the side chains of Y123 and I125 contacts the methyl groups of methylated CpG. The DNA interface further- more contains several arginine and lysine side chains forming hydrogen bonds with the bases and contacting the DNA backbone. The solution structure of the MBD of MBD1 shows high similarity to that of MeCP2 except for the C-terminus [18]. At the C-terminus of the MBD, MeCP2 exhibits a one-turn helix, while MBD1 is folded into a hairpin loop. The MBD of MBD1 contacts the methyl groups of a methylated CpG through a hydrophobic patch formed by the side chains of five residues, V20, R22, Y34, R44, and S45 [19]. Although the solution structure of the core region of the MBD of MeCP2 has been determined and some essentials of its interaction with DNA are grossly understood, many questions remain to be solved. In particular, the structure and function of the sequences flanking the core region are not known. The importance of these flanking sequences is highlighted by several missense mutations causing Rett syndrome. Interestingly, six mutations cluster in the region C-terminal to the a-helix. Here we describe the solution structure of MAR-BD and show that its C-terminal portion contains an amphipathic helical coil, a 2 /a 3 . This helical coil mainly contributes to the surface opposite to the DNA interface, providing a platform for interactions with other domains of MeCP2 or other proteins [19]. The consequences of six missense mutations within the coil found in Rett syndrome cases are discussed. Experimental procedures Sample preparation 15 N-labeled and 15 N/ 13 C-labeled, His-tagged chicken MeCP2 72-196 ,namedMAR-BD,wasexpressedinEscheri- chia coli BL21(DE3)pLysS from plasmid pET-cARBP- Ex4.2 in isotope labeled Bio-Express media (Cambridge Isotope Laboratories/Promochem, Wesel, Germany). The labeling was nonspecific, i.e. no amino-acid-type selective labeling was used. Purification of the protein on Ni 2+ / agarose beads and on a Mono S FPLC column was performed as described previously [20]. MAR-BD contains the non-native sequence MGHHHHHH at its N-terminus. NMR spectroscopy NMR measurements on free MAR-BD, i.e. in the uncom- plexed state, were performed at 298 K and pH 6.8 on 1–2 m M samplesinNSbuffer[10m M sodium phosphate (46.3% Na 2 HPO 4 and 53.7% NaH 2 PO 4 ), 0.5 m M NaN 3 , with either 5% D 2 O or 100% D 2 O, and 0.1 m M 2,2- dimethyl-2-silapentanesulfonic acid] containing 154 m M NaCl. Complexes of MAR-BD with the unlabeled, non- methylated, double-stranded oligonucleotide 5 0 -ATGACG AAATCACTA-3 0 (MeCP2-binding site I in mouse satellite DNA [15]) were generated by mixing the protein with the oligonucleotide at molar ratios of 1 : 1.2 and 1 : 2.4 in NS buffer containing 10 m M NaCl. NMR data were collected on Bruker DRX-600 or DRX-800 NMR spectrometers equipped with four channels, pulsed-field gradient triple- resonance or quadruple-resonance probes with either z or xyz gradients. The 1 H-NMR chemical shifts were referenced using 2,2-dimethyl-2-silapentanesulfonic acid as an internal standard. The chemical shifts of 15 N and 13 C were referenced indirectly following the recommendations summarized in [21]. In addition to the spectra recorded for sequential assignment of the NMR signals of the backbone nuclei described in [20], HCCH-TOCSY (9 ms mixing time), HBHA(CO)NH, and CC(CO)NH spectra were measured to further advance the extent of assignment of the side chain nuclei. NOE distance restraints were obtained from 2D NOESY and 13 C-edited and 15 N-edited 3D NOESY-HSQC spectra measured in D 2 OandH 2 O, respectively, with a mixing time of 150 ms, except for the 15 N-edited spectrum (100 ms). These mixing times turned out to be optimal for obtaining the maximum number of NOE contacts of sufficient signal-to-noise ratio. NOESY spectra were acquired with 1024 (complex) data points in direct dimen- sion. For the 2D NOESY and 3D NOESY experiments, 1024 and 128 data points, respectively, were acquired in the indirect 1 H dimension. In the indirect 13 Cand 15 N dimension, 64 data points were acquired, and forward linear prediction was used. Zero-filling was applied in the direct and indirect spectral dimensions. ProteinÆDNA complexes were analyzed by comparison of 1 H- 15 N HSQC spectra of free MAR-BD with those of MAR-BD titrated with the 15- mer oligonucleotide duplex at molar ratios of 1 : 1.2 and of 1 : 2.4. Spectra analysis NMR data were analysed and processed with the computer programs XWINNMR , AURELIA [22], and AUREMOL [23] (Bruker, Karlsruhe, Germany). In the final part of the assignment of the NOESY spectra, the number of identified NOEs was increased by comparison of back-calculated spectra with the experimental data [24,25]. Based on a preliminary structure, the NOESY spectra were simulated using the relaxation matrix approach. Through comparison with the corresponding experimental spectra, new NOE restraints were obtained, which were used in subsequent structure calculations. These newly calculated structures were then used for the next step in the iteration process. This procedure was continued until the quality of the structure could not be further improved. Structure calculations and analysis Structure calculations were performed using the computer program DYANA [26]. Distance information from NOEs was included in the structure calculations assuming an error of 30%; 240 structures were calculated. The / and w angle 3264 B. Heitmann et al.(Eur. J. Biochem. 270) Ó FEBS 2003 restraints were obtained through a database search for backbone chemical shifts and sequence homology using the computer program TALOS [27]. Secondary structure ele- ments and root mean square deviations (rmsd) were determined using the program MOLMOL [28]. Rmsd values were calculated for the best 10 structures with respect to the value of the target function. Analysis of the / and w angles of the nonglycine and nonproline residues was carried out with the computer program PROCHECK - NMR [29]. The secondary structure was calculated with the Kabsch-Sander algorithm [30] implemented in the computer program MOL- MOL . The atomic co-ordinates were deposited in the Protein Data Bank under accession number PDB 1UB1. Results and discussion NMR signal assignment and secondary structure Sequential NMR signal assignment of MAR-BD of cMeCP2 has been described previously [20]. The relatively high number of proline residues (>10%) and the fact that only a core region of free MAR-BD appears to be well folded made the assignment of the NMR signals difficult. Some residues located in unfolded regions in the vicinity of proline residues give rise to at least two sets of signals (with one set having significantly higher intensities), indicating slow exchange between different conformations on the NMR time scale. Comparison of the 1 H a , 15 N, 13 C a , 13 C b , and 13 C 0 chemical shifts with random coil values [31] and the CSI plot [32], which is a consensus of the different shifts, predicts a three-stranded b-sheet [strand b 1 , residues 104–110 (GWTRKLK); b 2 , residues 120–127 (KYDVY LIN); b 3 , residues 131–135 (KAFRS)] immediately followed by the three turns of an a-helix [a 1 , residues 136–145 (KVELIAYFEK); numbering refers to chicken MeCP2; Fig. 1]. Interestingly, the 13 C 0 chemical shifts for the residues in the vicinity of V160 show indications of an additional helical structure. We therefore used the computer program TALOS [27] to further explore the secondary structure elements. This program predicts / and w angle values on the basis of a database search for chemical shifts of backbone nuclei and sequence homology. The program judges a prediction as ÔgoodÕ for (/,w) pairs, when nine or 10 matches occur with small dispersion of the angle values. Besides the regions already shown to contain elements of secondary structure by the CSI plot, two other regions are predicted to exhibit secondary structure elements, namely residues 96–101 and residues 152–163 (Fig. 2). Analysis of the NOESY spectra reveals that residues 96–101 do not show NOE contacts characteristic of b-sheet conformation or other secondary structure elements [20]. In contrast, residues 152–163 exhibit several NOE contacts indicative of helical regions. This prediction will be further corroborated below. Tertiary structure Structure calculations are based on a data set consisting of 891 different NMR-derived distance and torsion angle restraints. Among the distance restraints, 447 intraresidual, 186 sequential, and 196 medium-range and long-range restraints were found (Table 1). The number of NOE contacts identified is plotted in Fig. 3A as a function of the residue number. Obviously, the relatively high number of NOE contacts (>10 per residue) commonly expected for Fig. 1. Chemical shift analysis. Top, chemical shift index (CSI [32]) for MAR-BD. Bottom, the 1 H, 15 N and 13 C chemical shift differences (in 1 p.p.m.) relative to the random coil values [31] are given as a function of the residue number. Fig. 2. Database search for chemical shifts of backbone nuclei and sequence homology. Residues judged as ÔgoodÕ by the computer program TALOS [27] and number of database matches for these residues indi- cated by black bars. To keep the figure simple, the number of database matches is not given for residues not judged as ÔgoodÕ by TALOS . Ó FEBS 2003 Solution structure of chicken MeCP2 (Eur. J. Biochem. 270) 3265 structured regions is only observed for the central part of cMeCP2 MAR-BD, i.e. for the region between residues 101 and 163. This indicates that the remaining N-terminal and C-terminal regions of the domain are unstructured, as already predicted by the chemical shift and TALOS analyses. However, we note that the region between residues 152 and 163 is obviously structured, confirming our conclusions drawn from the TALOS analysis (Fig. 2). A superimposition of the best 10 structures with respect to the target function is shown in Fig. 3B together with a ribbon plot of one selected structure. As can be seen, the described secondary-structure elements are well defined, as also indicated by the rmsd values given in Table 1. Analysis of the Ramachandran plot shows that the dihedral angles / and w for the secondary structure elements are all found in the most favored or the additionally allowed region. A selected example of the 10 structures shown in Fig. 3B is compared in Fig. 4 (middle) with the structure of rMeCP2 MBD (left) and MBD1 MBD (right) [17,18]. The global fold of these three domains turns out to be identical, which is not surprising considering the high degree of sequence similarity. The core of MAR-BD consists of the above described three-stranded antiparallel b-sheet followed by a-helix a 1 . For these secondary structure elements, the Kabsch-Sander algorithm imple- mented in the computer program MOLMOL always identifies the following: b 1 , residues 106–110 (TRKLK); b 2 , residues 122–126 (DVYLI); b 3 , residues 132–133 (AF); and a 1 , residues 136–144 (KVELIAYFE). Strands b 1 and b 2 are separated by a flexible loop. The core of free MAR-BD is hydrophobic, consisting mainly of residues T106, K108, V123, L125, F133, L139, F143, F158, and T161. The C-terminus of helix a 1 isfollowedbyanextendedloop ending in a one-turn helix [a 2 , residues 153–155 (PND)]. After a short interruption by three residues, a third short helix [a 3 , residues 159–163 (TVTGR)] could be identified. This Ôhelical coilÕ, a 2 /a 3 , is arranged antiparallel to a 1 .The N-terminal (72–100) and C-terminal residues (164–196) exhibit a significantly reduced number of NOE contacts (compare with Fig. 3A). This behavior is characteristic of unfolded regions and agrees with the results of the chemical- shift analysis (see above). Additional efforts were made to confirm the helical structure of residues 153–155, as these residues were not always shown to exhibit a helical structure by the Kabsch-Sander algorithm implemented in the computer program MOLMOL [28,30]. The / and w angles of these residues, however, are clearly found in the region characteristic of residues located in a-helices. This observa- tion is made for all the other calculated structures, strongly supporting the observation that residues 153–155 form a short helix. DNA-binding site In murine metaphase chromosomes, MeCP2 preferentially localizes to the pericentromeric regions containing highly Table 1. Structural statistics and rmsd values. Type of restraint Number Total 891 Intraresidual NOEs 447 Sequential (i, i + 1) NOEs 186 Medium-range (i, i + j;1<j < 5) NOEs 65 Long-range (i, i + j;4<j) NOEs 131 Angle restraints a 62 Atoms used for the calculation of root mean square deviations rmsd/nm Backbone atoms (N, C a , and C 0 ) for residues 95–170 0.268 Heavy atoms for residues 95–170 0.329 Backbone (secondary structure elements, see text) 0.037 Heavy atoms (secondary structure elements, see text) 0.102 Ramachandran plot analysis % residues found in this region b Most favored region 89.7 (58.3) Additionally allowed region 10.3 (28.1) Generously allowed region – (8.3) Disallowed region – (5.2) a Using the computer program TALOS [27]. b These data were determined using the computer program PROCHECK - NMR [29] taking into account only residues located in secondary structure elements. The values determined for the entire molecule are given in parentheses. Fig. 3. NOE statistics and structure of cMeCP2 MAR-BD. (A) Number of NOE contacts as a function of residue number. Filled bars, intraresidual NOEs; dark grey bars, sequential NOEs; light grey bars, medium-range NOEs (sequential distance 2–4 residues); white bars, long-range NOEs (sequential distance >4). (B) Superimposition of the best 10 structures with respect to the target function (see text) and ribbon plot of one of these 10 structures. 3266 B. Heitmann et al.(Eur. J. Biochem. 270) Ó FEBS 2003 methylated satellite DNA [2]. Biochemical studies showed thatMeCP2recognizestwosites(IandII)inmethylated satellite DNA with high affinity (see above) [15]. Binding to these sites in nonmethylated satellite DNA occurs with slightly reduced affinity. When cMeCP2 MAR-BD was complexed with methylated DNA, only poorly resolved spectra could be obtained (data not shown). In contrast, spectra of satisfactory resolution were acquired if a nonmethylated 15-mer oligonucleotide duplex encompas- sing site I was used (Fig. 5B). Distinct chemical shift changes were observed for the signals in the 1 H- 15 N HSQC spectrum on complex formation (Fig. 5A). To evaluate the shift of a signal in 1 Hand 15 N dimension, we introduce the term Ôtotal induced chemical shiftÕ, D T , defined as the sum of the absolute values of the 1 H N and 15 N chemical shift changes measured in Hz. Obviously, the entire molecule is affected by DNA binding, as most of the residues exhibit a total induced chemical shift, D T , far beyond the experimen- tal error of  10 Hz. Strongly affected residues are located in the loop between strands a 1 and a 2 , in helix a 1 ,andinthe helical coil a 2 /a 3 . The involvement of residues located in the unfolded parts of unbound MAR-BD is hard to predict, because a considerable fraction of the NMR signals of those regions could not be assigned in the spectra of the proteinÆDNA complex. For such residues, the total induced chemical shift plotted in Fig. 5A is a minimum value. This minimum value was estimated from the difference between the chemical shift observed for the free protein and that of the next nearest unassigned signal. For a better understanding of this procedure, a section of the 1 H- 15 N HSQC spectrum of free MAR-BD (solid grey lines) is overlayed in Fig. 5B by the corresponding section of the spectrum of MAR-BD com- plexed with DNA (dashed black lines). Biological and medical implications The MAR-BD of MeCP2 extends the methyl-CpG- binding domain (MBD) by seven N-terminal and 33 C-terminal residues [15]. Consequently, we found that the core of MAR-BD folds into the same structure as the core of MBD, i.e. an antiparallel three-stranded a-sheet followed by an a-helix, a 1 [17]. This fold is also shared by the MBD of MBD1 [18]. Moreover, the structure of MAR-BD described here and the reported structure of its MBD [17] coincide in possessing the short helix a 2 (residues P153, N154, and D155 in chicken MAR-BD). This short helix is located at the side of the domain opposite to the DNA-binding face with solvent exposed residues N154 and D155. In addition to helix a 2 ,wehave identified a third helix, a 3 , in MAR-BD at T159 to R163. Helices a 2 and a 3 are connected by a stretch of three residues (F156, D157, and F158), which are conserved at comparable positions among all the other members of the MBD protein family [18]. The helical coil a 2 /a 3 is amphipathic. P153 and G162 are buried in the protein core. Also, F156, F158, and V160 are tightly packed into the hydrophobic core of the domain. On the other hand, residues D152, N154, D155, D157, and R163 are solvent exposed and cluster in a small patch located opposite to the DNA interface (Fig. 6). This patch is negatively charged through three aspartic acids, but also contains one positive charge through R163. It has been proposed that the negative charges on the surface of the domain Fig. 4. Comparison of cMeCP2 MAR-BD with rMeCP2 MBD and MBD1 MBD. Ribbon plots of rMeCP2 MBD [17] (left), the core region of cMeCP2 MAR-BD (residues 95–170) (middle; this study), and MBD1 MBD [18] (right). Ó FEBS 2003 Solution structure of chicken MeCP2 (Eur. J. Biochem. 270) 3267 opposite the DNA interface have a role in interactions with another protein or with another domain within MeCP2 [19, 1 33]. Notably, the C-terminal region of MAR- BD differs significantly from that of MBD1 MBD. In MBD1 MBD, helix a 1 is shortened by one turn, helix a 2 is lacking, and helix a 3 is replaced by a hairpin loop. First of all, these differences and the corresponding amino-acid changes generate a characteristic protein interaction site in each of the domains. Secondly, they cause differences in the mode of interaction with the periphery of the DNA target site. For example, a lysine residue at the tip of the hairpin loop that mediates a backbone contact with DNA is unique in MBD1. In MAR-BD, the significant chemical shift changes of the helical coil region found in titration experiments with a mouse satellite DNA-derived oligonucleotide duplex are noteworthy (Fig. 5A). They probably indicate that the helical coil is also involved in proteinÆDNA interactions. Chemical shift changes further- more indicate that residues close to the N-terminus and C-terminus also contribute to these interactions. In this study, we used a nonmethylated high-affinity binding site from mouse satellite DNA for complex formation with MAR-BD [15]. As the chemical shifts obtained with this DNA fragment closely resemble those obtained with a methylated CpG sequence [17], our data corroborate pre- vious findings that MeCP2 also recognizes nonmethylated sequences [3,15,34]. Considerable interest in the structure of MeCP2 was generated by the discovery that mutations in MECP2 cause Rett syndrome, an X-linked, dominant neurological disorder primarily affecting young girls [12]. Intriguingly, six missense mutations cluster in the helical coil a 2 /a 3 , emphasizing the importance of this region: P153(152)R; F156(155)I,S,C; D157(156)G,E; F158(157)I; T159(158) M,A; and G162(161)R,W (here and in the discussion below, human numbering is given in parentheses). F156(155), D157(156), and F158(157) are conserved Fig. 5. Effects of mouse satellite site I on chemical shifts of MAR-BD. (A) Total induced chemical shift, D T , observed for complex formation with a 15-mer oligonucleotide duplex from MeCP2 high-affinity binding site I of mouse satellite DNA at a molar protein/DNA ratio of 1 : 2.4 vs. the residue number. The dotted line indicates D T ¼ 100 Hz. White bars denote residues where the signals could also be assigned unambiguously for the proteinÆDNA complex. Black bars indicate residues with ambiguous signal assignment for the proteinÆDNA complex. In such cases, the minimum induced chemical shift, i.e. the distance to the next nearest unassigned signal in the spectrum is shown. Grey bars indicate proline residues. (B) Selected region from the 1 H- 15 N HSQC spectra of free MAR-BD (solid grey lines) and of the proteinÆDNA complex (nonmethylated DNA, dashed black lines). Fig. 6. Structural models depicting residues affected by DNA binding and residues mutated in Rett syndrome cases. Ribbon plot of the backbone of cMeCP2 MAR-BD for two orientations of the molecule (left) and total induced chemical shift, D T , projected on the surface of the MAR-BD (right). Blue, residues with D T <100 Hz.Red,residues with D T > 100 Hz. Orange, residues with D T < 100 Hz but sequen- tially neighbored to residues exhibiting D T > 100 Hz. Three residues at the surface opposite the DNA interface (bottom) and mutated in Rett syndrome cases (D98, D157, and T159) are marked by arrows (left) and by white boxes (right). 3268 B. Heitmann et al.(Eur. J. Biochem. 270) Ó FEBS 2003 among all MBD-containing proteins, P153(152) and T159(158) only in MeCP2 and MBD4, but G162(161) solely occurs in MeCP2 [18]. The structure of the helical coil a 2 /a 3 allows us to interpret the consequences of the six mutations. As P153(152) and G162(161) are buried in the protein core, replacement of each of these residues with positively charged arginines [or the bulky tryptophan in the case of G162(161)W] is predicted to generate gross structural disturbance of the fold. Likewise, as the side chains of F156(155) and F158(157) contribute to the hydrophobic core of MAR-BD, their replacement with isoleucine [or serine in the case of F156(155)S] probably causes unfolding of the domain. In fact, the Rett mutation F156(155)S has been previously shown to disrupt the domain and to cause severe reduction of the binding affinity to methylated DNA [33,35]. Residue F158(157) is equivalent to F64 in MBD1; mutation of this residue, F64A, has been shown to disrupt the tertiary structure of the domain, resulting in total loss of binding to methy- lated DNA [18]. Residue D157(156) is a Rett mutation site of considerable interest, because its replacement with glutamic acid is only a minimal change with conservation of the negative charge. Continuing with the hypothesis that the negatively charged surface opposite the DNA interface serves as a protein interaction site, we have to infer that insertion of the small methylene group by the D157(156)E mutation disrupts such interactions. D98(97), which is located in close vicinity to D157(156) (Fig. 6, bottom), is also mutated in Rett syndrome cases. In one patient, D98(97) is replaced with glutamic acid, reminiscent of the D157(156)E mutation which causes the same minimal change. Thus, the negatively charged surface critical for interactions with another domain of MeCP2 or another protein probably includes D98(97). T159(158), another target residue at a Rett mutation site, is located adjacent to D157(156) (Fig. 6, bottom). As the putative interaction of D157(156) with another protein or domain is disrupted by the insertion of a methylene group, it follows that replacement of the neighboring T159(158) with alanine or methionine may compromise this interaction as well. Consistent with the location of T159(158) at the surface opposite the DNA interface, Rett mutation T159(158)M, the most common mutation in MeCP2 [36], was previously shown to cause little reduction in the affinity for methylated DNA [33,35]. Collectively, these observations suggest that D98(97), D157(156), and T159(158) form a region critical for the interaction between MAR-BD and another domain of MeCP2 or another protein (see Fig. 6, bottom). Candidate interacting sequences within MeCP2 are evolutionarily conserved basic regions, such as residues 249–271 (human numbering), which contains the nuclear localization signal, and residues 284–309, the terminal portion of the transcriptional repression domain [6]. Intriguingly, several Rett missense mutations that affect basic residues cluster in these two regions. Know- ledge of the structure of a larger portion of MeCP2 including the transcriptional repression domain would resolve these speculations and clarify the role of the helical coil as well as of residues 164–195 in the DNA- recognition process. Acknowledgements We thank Susanne Giehler for excellent technical assistance, and Ingrid Cuno for carefully proofreading the manuscript. Financial support from the Deutsche Forschungsgemeinschaft is gratefully acknowledged (grants SFB 545-B2, Str145/12-3, and Br 1278/8). References 1. Hendrich, B. & Bird, A. (1998) Identification and characterization of a family of mammalian methyl-CpG binding proteins. Mol. Cell. Biol. 18, 6538–6547. 2. Lewis, J.D., Meehan, R.R., Henzel, W.J., Maurer-Fogy, I., Jeppesen, P., Klein, F. & Bird, A. (1992) Purification, sequence, and cellular localization of a novel chromosomal protein that binds to methylated DNA. Cell 69, 905–914. 3. von Kries, J.P., Buhrmester, H. & Stra ¨ tling, W.H. (1991) A matrix/scaffold attachment region binding protein: identification, purification, and mode of binding. Cell 64, 123–135. 4. Nan, X., Campoy, F.J. & Bird, A. (1997) MeCP2 is a transcrip- tional repressor with abundant binding sites in genomic chroma- tin. Cell 88, 471–481. 5. Nan, X., Ng, H H., Johnson, C.A., Laherty, C.D., Turner, B.M., Eisenmann, R.N. & Bird, A. (1998) Transcriptional repression by the methyl-CpG-binding protein MeCP2 involves a histone dea- cetylase complex. Nature (London) 393, 386–389. 6. Yu, F., Thiesen, J. & Stra ¨ tling, W.H. (2000) Histone deacetylase- independent transcriptional repression by methyl-CpG-binding protein 2. Nucleic Acids Res. 28, 2201–2206. 7. Kaludow, N.K. & Wolffe, A.P. (2000) MeCP2 driven transcrip- tional repression in vitro: selectivity for methylated DNA, action at a distance and contacts with the basal transcription machinery. Nucleic Acids Res. 28, 1921–1928. 8. Ng, H H., Jeppesen, P. & Bird, A. (2000) Active repression of methylated genes by the chromosomal protein MBD1. Mol. Cell. Biol. 20, 1394–1406. 9. Zhang, Y., Ng, H H., Erdjument-Bromage, H., Tempst, P., Bird, A. & Reinberg, D. (1999) Analysis of the NuRD subunits reveals a histone deacetylase core complex and a connection with DNA methylation. Genes Dev. 13, 1924–1935. 10. Reichwald, K., Thiesen, J., Wiehe, T., Weitzel, J., Stra ¨ tling, W.H., Kiochis, P., Poustka, A., Rosenthal, A. & Platzer, M. (2000) Comparative sequence analysis of the MECP2-locus in human and mouse reveals new transcribed regions. Mamm. Genome 11, 182–190. 11. LaSalle, J.M., Goldstine, J., Balmer, D. & Greco, C.M. (2001) Quantitative localization of heterologous methyl-CpG-binding protein 2 (MeCP2) expression phenotypes in normal and Rett syndrome brain by laser scanning cytometry. Hum. Mol. Genet. 10, 1729–1740. 12. Amir, R.E., Van den Veyver, L.B., Wan, M., Tran, C.Q., Francke, U. & Zoghbi, H.Y. (1999) Rett syndrome is caused by mutation in X-linked MECP2, encoding methyl-CpG-binding protein 2. Nat. Genet. 23, 185–188. 13. Chen, R.Z., Akbarian, S., Tudor, M. & Jaenisch, R. (2001) Deficiency of methyl-CpG binding protein-2 in CNS neurons results in a Rett-like phenotype in mice. Nat. Genet. 27, 327–331. 14. Guy, J., Hendrich, B., Holmes, M., Martin, J.E. & Bird, A. (2001) AmouseMecp2-null mutation causes neurological symptoms that mimic Rett syndrome. Nat. Genet. 27, 322–326. 15. Weitzel, J.M., Buhrmester, H. & Stra ¨ tling, W.H. (1997) Chicken MAR-binding protein ARBP is homologous to rat methyl-CpG- binding protein MeCP2. Mol. Cell. Biol. 17, 5656–5666. 16. Nan, X.S., Meehan, R.R. & Bird, A. (1993) Dissection of the methyl-CpG binding domain from the chromosomal protein MeCP2. Nucleic Acids Res. 21, 4886–4892. Ó FEBS 2003 Solution structure of chicken MeCP2 (Eur. J. Biochem. 270) 3269 17. Wakefield, R.I.D., Smith, B.O., Nan, X., Free, A., Soteriou, A., Uhrin, D., Bird, A.P. & Barlow, P.N. (1999) The solution struc- ture of the domain from MeCP2 that binds to methylated DNA. J. Mol. Biol. 291, 1055–1065. 18. Ohki, I., Shimotake, N., Fujita, N., Nakao, M. & Shirakawa, M. (1999) Solution structure of the methyl-CpG-binding domain of the methylation-dependent transcriptional repressor MBD1. EMBO J. 18, 6653–6661. 19. Ohki, I., Shimotake, N., Fujita, N., Jee, J G., Ikegami, T., Nakao, M. & Shirakawa, M. (2001) Solution structure of the methyl-CpG binding domain of human MBD1 in complex with methylated DNA. Cell 105, 487–497. 20. Brunner, E., Weitzel, J., Heitmann, B., Maurer, T., Stra ¨ tling, W.H. & Kalbitzer, H.R. (2000) Sequence-specific 1 H, 13 C, and 15 N assignments of the MAR-binding domain of chicken MeCP2/ ARBP. J. Biomol. NMR 17, 175–176. 21. Markley, J.L., Bax, A., Arata, Y., Hilbers, C.W., Kaptein, R., Sykes, B.D., Wright, P.E. & Wu ¨ thrich, K. (1998) Recommenda- tions for the presentation of NMR structures of proteins and nucleic acids. Pure Appl. Chem. 70, 117–142. 22. Neidig, K P., Geyer, M., Go ¨ rler, A., Antz, C., Saffrich, R., Beneicke, W. & Kalbitzer, H.R. (1995) AURELIA , a program for computer-aided analysis of multidimensional NMR spectra. J. Biomol. NMR 6, 255–270. 23. Gronwald, W., Moussa, S., Elsner, R., Jung, A., Ganslmeier, B.,Trenner,J.,Kremer,W.,Neidig,K P.&Kalbitzer,H.R. (2002) Automated assignment NOESY NMR spectra using a knowledge based method (KNOWNOE). J. Biomol. NMR 23, 271–287. 24. Go ¨ rler, A. & Kalbitzer, H.R. (1997) Relax, a flexible pro- gram for the back calculation of NOESY spectra based on complete-relaxation-matrix-formalism. J. Magn. Reson. 124, 177–188. 25. Go ¨ rler, A., Gronwald, W., Neidig, K P. & Kalbitzer, H.R. (1999) Computer assisted assignment of 13 Cand 15 N edited 3D-NOESY- HSQC spectra using back calculated and experimental spectra. J. Magn. Reson. 137, 39–45. 26. Gu ¨ ntert, P., Mumenthaler, C. & Wu ¨ thrich, K. (1997) Torsion angle dynamics for NMR structure calculation with the new program DYANA . J. Mol. Biol. 273, 283–298. 27. Cornilescu, G., Delagio, F. & Bax, A. (1999) Protein backbone angle restraints from searching a database for chemical shift and sequence homology. J. Biomol. NMR 13, 289–302. 28. Koradi, R., Billeter, M. & Wu ¨ thrich, K. (1996) MOLMOL :apro- gram for display and analysis of macromolecular structures. J. Mol. Graphics 14, 51–55. 29. Laskowski, R.A., Rullmann, J.A., MacArthur, M.W., Kaptein, R. & Thornton, J.M. (1996) AQUA and PROCHECK - NMR :programs for checking the quality of protein structures solved by NMR. J. Biomol. NMR 5, 67–81. 30. Kabsch, W. & Sander, C. (1983) Dictionary of protein secondary structure: pattern recognition of hydrogen-bonded and geo- metrical features. Biopolymers 22, 2577–2637. 31. Wishart, D.S., Bigam, C.G., Hodges, R.S. & Sykes, B.D. (1995) 1 H, 13 C, and 15 N random coil NMR chemical shifts of the com- mon amino acids. I. Investigations of nearest-neighbor effects. J. Biomol. NMR 5, 67–81. 32. Wishart, D.S. & Sykes, B.D. (1994) The 13-C chemical shift index: a simple method for the identification of protein secondary structure using 13-C chemical shift data. J. Biomol. NMR 4, 171–180. 33. Free, A., Wakefield, R.I.D., Smith, B.O., Dryden, D.T.F., Barlow, P.N.&Bird,A.P.(2001)DNArecognitionbythemethyl-CpG binding domain of MeCP2. J. Biol. Chem. 276, 3353–3360. 34. Buhrmester,H.,vonKries,J.P.&Stra ¨ tling, W.H. (1995) Nuclear matrix protein ARBP recognizes a novel DNA sequence motif with high affinity. Biochemistry 34, 4108–4117. 35. Ballestar, E., Yusufzai, T.M. & Wolffe, A.P. (2000) Effects of Rett syndrome mutations of the methyl-CpG binding domain of the transcriptional repressor MeCP2 on selectivity for association with methylated DNA. Biochemistry 39, 7100–7106. 36. Dragich, J., Houwink-Manville, I. & Schanen, C. (2000) Rett syndrome: a surprising result of mutation in Mecp2. Hum. Mol. Genet. 9, 2365–2375. 3270 B. Heitmann et al.(Eur. J. Biochem. 270) Ó FEBS 2003 . contacting the DNA backbone. The solution structure of the MBD of MBD1 shows high similarity to that of MeCP2 except for the C-terminus [18]. At the C-terminus of. 72–196 in chicken (c )MeCP2] [15]. The solution structure of the MBD of rat (r )MeCP2 has recently been determined [17]. The MBD adopts a wedge- shaped structure,