Crystal structure of human methyl-binding domain IV glycosylase bound to abasic DNA

The mammalian repair protein MBD4 (methyl-CpG-binding domain IV) excises thymine from mutagenic G·T mispairs generated by deamination of 5-methylcytosine (mC), and downstream base excision repair proteins restore a G·C pair. MBD4 is also implicated in active DNA demethylation by initiating base excision repair of G·T mispairs generated by a deaminase enzyme. The question of how mismatch glycosylases attain specificity for excising thymine from G·T, but not A·T, pairs remains largely unresolved. Here, we report a crystal structure of the glycosylase domain of human MBD4 (residues 427-580) bound to DNA containing an abasic nucleotide paired with guanine, providing a glimpse of the enzyme-product complex. The mismatched guanine remains intrahelical, nestled into a recognition pocket. MBD4 provides selective interactions with the mismatched guanine (N1H, N2H(2)) that are not compatible with adenine, which likely confer mismatch specificity. The structure reveals no interactions that would be expected to provide the MBD4 glycosylase domain with specificity for acting at CpG sites. Accordingly, we find modest 1.5- to 2.7-fold reductions in G·T activity upon altering the CpG context. In contrast, 37- to 580-fold effects were observed previously for thymine DNA glycosylase. These findings suggest that specificity of MBD4 for acting at CpG sites depends largely on its methyl-CpG-binding domain, which binds preferably to G·T mispairs in a methylated CpG site. MBD4 glycosylase cannot excise 5-formylcytosine (fC) or 5-carboxylcytosine (caC), intermediates in a Tet (ten eleven translocation)-initiated DNA demethylation pathway. Our structure suggests that MBD4 does not provide the electrostatic interactions needed to excise these oxidized forms of mC.     

Towards an understanding of DNA recognition by the methyl-CpG binding domain 1

CpG methylation determines a variety of biological functions of DNA. The methylation signal is interpreted by proteins containing a methyl-CpG binding domain (MBDs). Based on the NMR structure of MBD1 complexed with methylated DNA we analysed the recognition mode by means of molecular dynamics simulations. As the protein is monomeric and recognizes a symmetrically methylated CpG step, the recognition mode is an asymmetric one. We find that the two methyl groups do not contribute equally to the binding energy. One methyl group is associated with the major part of the binding energy and the other one nearly does not contribute at all. The contribution of the two cytosine methyl groups to binding energy is calculated to be -3.6 kcal/mol. This implies a contribution of greater than two orders of magnitude to the binding constant. The conserved amino acid Asp32 is known to be essential for DNA binding by MBD1, but so far no direct contact with DNA has been observed. We detected a direct DNA base contact to Asp32. This could be the main reason for the importance of this amino acid. MBD contacts DNA exclusively in the major groove, the minor groove is reserved for histone contacts. We found a deformation of the minor groove shape due to complexation by MBD1, which indicates an information transfer between the major and the minor groove.     

Towards an Understanding of DNA Recognition by the Methyl-CpG Binding Domain 1

Abstract

CpG methylation determines a variety of biological functions of DNA. The methylation signal is interpreted by proteins containing a methyl-CpG binding domain (MBDs). Based on the NMR structure of MBD1 complexed with methylated DNA we analysed the recognition mode by means of molecular dynamics simulations.

As the protein is monomeric and recognizes a symmetrically methylated CpG step, the recognition mode is an asymmetric one. We find that the two methyl groups do not contribute equally to the binding energy. One methyl group is associated with the major part of the binding energy and the other one nearly does not contribute at all. The contribution of the two cytosine methyl groups to binding energy is calculated to be ?3.6 kcal/mol. This implies a contribution of greater than two orders of magnitude to the binding constant. The conserved amino acid Asp32 is known to be essential for DNA binding by MBD1, but so far no direct contact with DNA has been observed. We detected a direct DNA base contact to Asp32. This could be the main reason for the importance of this amino acid. MBD contacts DNA exclusively in the major groove, the minor groove is reserved for histone contacts. We found a deformation of the minor groove shape due to complexation by MBD1, which indicates an information transfer between the major and the minor groove.     

Examination of the specificity of DNA methylation profiling techniques towards 5-methylcytosine and 5-hydroxymethylcytosine

DNA cytosine-5 methylation is a well-studied epigenetic pathway implicated in gene expression control and disease pathogenesis. Different technologies have been developed to examine the distribution of 5-methylcytosine (5mC) in specific sequences of the genome. Recently, substantial amounts of 5-hydroxymethylcytosine (5hmC), most likely derived from enzymatic oxidation of 5mC by TET1, have been detected in certain mammalian tissues. Here, we have examined the ability of several commonly used DNA methylation profiling methods to distinguish between 5mC and 5hmC. We show that techniques based on sodium bisulfite treatment of DNA are incapable of distinguishing between the two modified bases. In contrast, techniques based on immunoprecipitation with anti-5mC antibody (methylated DNA immunoprecipitation, MeDIP) or those based on proteins that bind to methylated CpG sequences (e.g. methylated-CpG island recovery assay, MIRA) do not detect 5hmC and are specific for 5mC unless both modified bases occur in the same DNA fragment. We also report that several methyl-CpG binding proteins including MBD1, MBD2 and MBD4 do not bind to sequences containing 5hmC. Selective mapping of 5hmC will require the development of unique tools for the detection of this modified base.     

Position-specific suppression and enhancement of HIV-1 integrase reactions by minor groove benzo[a]pyrene diol epoxide deoxyguanine adducts: implications for molecular interactions between integrase and substrates

The viral protein HIV-1 integrase is required for insertion of the viral genome into human chromosomes and for viral replication. Integration proceeds in two consecutive integrase-mediated reactions: 3'-processing and strand transfer. To investigate the DNA minor groove interactions of integrase relative to known sites of integrase action, we synthesized oligodeoxynucleotides containing single covalent adducts of known absolute configuration derived from trans-opening of benzo-[a]pyrene 7,8-diol 9,10-epoxide by the exocyclic 2-amino group of deoxyguanosine at specific positions in a duplex sequence corresponding to the terminus of the viral U5 DNA. Because the orientations of the hydrocarbon in the minor groove are known from NMR solution structures of duplex oligonucleotides containing these deoxyguanosine adducts, a detailed analysis of the relationship between the position of minor groove ligands and integrase interactions is possible. Adducts placed in the DNA minor groove two or three nucleotides from the 3'-processing site inhibited both 3'-processing and strand transfer. Inosine substitution showed that the guanine 2-amino group is required for efficient 3'-processing at one of these positions and for efficient strand transfer at the other. Mapping of the integration sites on both strands of the DNA substrates indicated that the adducts both inhibit strand transfer specifically at the minor groove bound sites and enhance integration at sites up to six nucleotides away from the adducts. These experiments demonstrate the importance of position-specific minor groove contacts for both the integrase-mediated 3'-processing and strand transfer reactions.     

DNA sequence recognition by bispyrazinonaphthalimides antitumor agents

Bifunctional DNA intercalating agents have long attracted considerable attention as anticancer agents. One of the lead compounds in this category is the dimeric antitumor drug elinafide, composed of two tricyclic naphthalimide chromophores separated by an aminoalkyl linker chain optimally designed to permit bisintercalation of the drug into DNA. In an effort to optimize the DNA recognition capacity, different series of elinafide analogues have been prepared by extending the surface of the planar drug chromophore which is important for DNA sequence recognition. We report here a detailed investigation of the DNA sequence preference of three tetracyclic monomeric or dimeric pyrazinonaphthalimide derivatives. Melting temperature measurements and surface plasmon resonance (SPR) studies indicate that the dimerization of the tetracyclic planar chromophore considerably augments the affinity of the drug for DNA, polynucleotides, or hairpin oligonucleotides and promotes selective interaction with G.C sites. The (CH(2))(2)NH(CH(2))(3)NH(CH(2))(2) connector stabilizes the drug-DNA complexes. The methylation of the two nitrogen atoms of this linker chain reduces the binding affinity and increases the dissociation rates of the drug-DNA complexes by a factor of 10. DNase I footprinting experiments were used to investigate the sequence selectivity of the drugs, demonstrating highly preferential binding to G.C-rich sequences. It also served to select a high-affinity site encompassing the sequence 5'-GACGGCCAG which was then introduced into a biotin-labeled hairpin oligonucleotide to accurately measure the binding parameters by SPR. The affinity constant of the unmethylated dimer for this sequence is 500 times higher than that of the monomer compound and approximately 10 times higher than that of the methylated dimer. The DNA groove accessibility was also probed with three related oligonucleotides carrying G --> c(7)G, G --> I, and C --> M substitutions. The level of drug binding to the two hairpin oligonucleotides containing 7-deazaguanine (c(7)G) or 5-methylcytosine (M) residues is unchanged or only slightly reduced compared to that of the unmodified target. In contrast, incorporation of inosine (I) residues considerably decreases the extent of drug binding or even abolishes the interaction as is the case with the monomer. The pyrazinonaphthalimide derivatives are thus much more sensitive to the deletion of the exocyclic guanine 2-amino group exposed in the minor groove of the duplex than to the modification of the major groove elements. The complementary SPR footprinting methodology combining site selection and quantitative DNA affinity analysis constitutes a reliable method for dissecting the DNA sequence selectivity profile of reversible DNA binding small molecules.     

Selective DNA cleavage by elsamicin A and switch function of its amino sugar group.

We report guanine-specific recognition and selective cleavage of DNA by the antitumor antibiotic elsamicin A equipped with an amino sugar and compare these results with cleavage by chartarin and chartreusin antibiotics. The preferential cutting sites of DNA strand scission with elsamicin A are on the bases adjacent to the 3'-side of guanine residues such as 5'-GN sites, in particular 5'-GG sites. The present results also indicate that (1) the aglycon portion binds intercalatively to the 3'-side of guanine in host DNA, (2) the guanine 2-amino group has an important effect on selective DNA binding of elsamicin A, and (3) the amino sugar residue of elsamicin A facilitates the drug binding into the minor groove of B-DNA. In addition, we found that an acetylation of the amino group on the elsamicin A sugar portion plays an interesting switch function for the activity of elsamicin A. The biological implication of this switch has also been discussed.     

Recognition and potential mechanisms for replication and erasure of cytosine hydroxymethylation

Cytosine residues in mammalian DNA occur in at least three forms, cytosine (C), 5-methylcytosine (M; 5mC) and 5-hydroxymethylcytosine (H; 5hmC). During semi-conservative DNA replication, hemi-methylated (M/C) and hemi-hydroxymethylated (H/C) CpG dinucleotides are transiently generated, where only the parental strand is modified and the daughter strand contains native cytosine. Here, we explore the role of DNA methyltransferases (DNMT) and ten eleven translocation (Tet) proteins in perpetuating these states after replication, and the molecular basis of their recognition by methyl-CpG-binding domain (MBD) proteins. Using recombinant proteins and modified double-stranded deoxyoligonucleotides, we show that DNMT1 prefers a hemi-methylated (M/C) substrate (by a factor of >60) over hemi-hydroxymethylated (H/C) and unmodified (C/C) sites, whereas both DNMT3A and DNMT3B have approximately equal activity on all three substrates (C/C, M/C and H/C). Binding of MBD proteins to methylated DNA inhibited Tet1 activity, suggesting that MBD binding may also play a role in regulating the levels of 5hmC. All five MBD proteins generally have reduced binding affinity for 5hmC relative to 5mC in the fully modified context (H/M versus M/M), though their relative abilities to distinguish the two varied considerably. We further show that the deamination product of 5hmC could be excised by thymine DNA glycosylase and MBD4 glycosylases regardless of context.     

DNA sequence recognition by the antitumor drug ditercalinium

The antitumor drug ditercalinium is a rare example of a noncovalent DNA-binding ligand that forms bisintercalation complexes via the major groove of the double helix. Previous structural studies have revealed that the two connected pyridocarbazolium chromophores intercalate into DNA with the positively charged bis(ethylpiperidinium) linking chain oriented to the wide groove side of the helix. Although the interaction of ditercalinium with short oligonucleotides containing 4-6 contiguous GC base pairs has been examined in detail by biophysical and theoretical approaches, the sequence preference for ditercalinium binding to long DNA fragments that offer a wide variety of binding sites has been investigated only superficially. Here we have investigated both sequence preferences and possible molecular determinants of selectivity in the binding of ditercalinium to DNA, primarily using methods based upon DNase I footprinting. A range of multisite DNA substrates, including several natural restriction fragments and different PCR-generated fragments containing unconventional bases (2,6-diaminopurine, inosine, uridine, 5-fluoro- and 5-methylcytosine, 7-deazaguanine, 7-deazaadenine, and N(7)-cyanoboranoguanine), have been employed to show that ditercalinium selectively recognizes certain GC-rich sequences in DNA and to identify some of the factors which affect its DNA-binding sequence selectivity. Specifically, the footprinting data have revealed that the 2-amino group on the purines or the 5-methyl group on the pyrimidines is not essential for the formation of ditercalinium-DNA complexes whereas the major groove-oriented N(7) of guanine does appear as a key element in the molecular recognition process. The loss of N(7) at guanines but not adenines is sufficient to practically abolish sequence-selective binding of ditercalinium to DNA. Thus, as expected for a major groove binding drug, the N(7) of guanine is normally required for effective complex formation with GC base pairs, but interestingly the substitution of the N(7) with a relatively bulky cyanoborane group does not markedly affect the sequence recognition process. Therefore, the hydrogen bond accepting capability at N(7) of guanines is not sufficient to explain the GC-selective drug-DNA association, and the implications of these findings are considered.     

Control of DNA minor groove width and Fis protein binding by the purine 2-amino group

The width of the DNA minor groove varies with sequence and can be a major determinant of DNA shape recognition by proteins. For example, the minor groove within the center of the Fis–DNA complex narrows to about half the mean minor groove width of canonical B-form DNA to fit onto the protein surface. G/C base pairs within this segment, which is not contacted by the Fis protein, reduce binding affinities up to 2000-fold over A/T-rich sequences. We show here through multiple X-ray structures and binding properties of Fis–DNA complexes containing base analogs that the 2-amino group on guanine is the primary molecular determinant controlling minor groove widths. Molecular dynamics simulations of free-DNA targets with canonical and modified bases further demonstrate that sequence-dependent narrowing of minor groove widths is modulated almost entirely by the presence of purine 2-amino groups. We also provide evidence that protein-mediated phosphate neutralization facilitates minor groove compression and is particularly important for binding to non-optimally shaped DNA duplexes.     

Excision of thymine and 5-hydroxymethyluracil by the MBD4 DNA glycosylase domain: structural basis and implications for active DNA demethylation

The mammalian DNA glycosylase-methyl-CpG binding domain protein 4 (MBD4)-is involved in active DNA demethylation via the base excision repair pathway. MBD4 contains an N-terminal MBD and a C-terminal DNA glycosylase domain. MBD4 can excise the mismatched base paired with a guanine (G:X), where X is uracil, thymine or 5-hydroxymethyluracil (5hmU). These are, respectively, the deamination products of cytosine, 5-methylcytosine (5mC) and 5-hydroxymethylcytosine (5hmC). Here, we present three structures of the MBD4 C-terminal glycosylase domain (wild-type and its catalytic mutant D534N), in complex with DNA containing a G:T or G:5hmU mismatch. MBD4 flips the target nucleotide from the double-stranded DNA. The catalytic mutant D534N captures the intact target nucleotide in the active site binding pocket. MBD4 specifically recognizes the Watson-Crick polar edge of thymine or 5hmU via the O(2), N(3) and O(4) atoms, thus restricting its activity to thymine/uracil-based modifications while excluding cytosine and its derivatives. The wild-type enzyme cleaves the N-glycosidic bond, leaving the ribose ring in the flipped state, while the cleaved base is released. Unexpectedly, the C(1)' of the sugar has yet to be hydrolyzed and appears to form a stable intermediate with one of the side chain carboxyl oxygen atoms of D534, via either electrostatic or covalent interaction, suggesting a different catalytic mechanism from those of other DNA glycosylases.     

Probing the Dynamic Distribution of Bound States for Methylcytosine-binding Domains on DNA

Although highly homologous to other methylcytosine-binding domain (MBD) proteins, MBD3 does not selectively bind methylated DNA, and thus the functional role of MBD3 remains in question. To explore the structural basis of its binding properties and potential function, we characterized the solution structure and binding distribution of the MBD3 MBD on hydroxymethylated, methylated, and unmethylated DNA. The overall fold of this domain is very similar to other MBDs, yet a key loop involved in DNA binding is more disordered than previously observed. Specific recognition of methylated DNA constrains the structure of this loop and results in large chemical shift changes in NMR spectra. Based on these spectral changes, we show that MBD3 preferentially localizes to methylated and, to a lesser degree, unmethylated cytosine-guanosine dinucleotides (CpGs), yet does not distinguish between hydroxymethylated and unmethylated sites. Measuring residual dipolar couplings for the different bound states clearly shows that the MBD3 structure does not change between methylation-specific and nonspecific binding modes. Furthermore, residual dipolar couplings measured for MBD3 bound to methylated DNA can be described by a linear combination of those for the methylation and nonspecific binding modes, confirming the preferential localization to methylated sites. The highly homologous MBD2 protein shows similar but much stronger localization to methylated as well as unmethylated CpGs. Together, these data establish the structural basis for the relative distribution of MBD2 and MBD3 on genomic DNA and their observed occupancy at active and inactive CpG-rich promoters.     

The Human Proteins MBD5 and MBD6 Associate with Heterochromatin but They Do Not Bind Methylated DNA

Background

MBD5 and MBD6 are two uncharacterized mammalian proteins that contain a putative Methyl-Binding Domain (MBD). In the proteins MBD1, MBD2, MBD4, and MeCP2, this domain allows the specific recognition of DNA containing methylated cytosine; as a consequence, the proteins serve as interpreters of DNA methylation, an essential epigenetic mark. It is unknown whether MBD5 or MBD6 also bind methylated DNA; this question has interest for basic research, but also practical consequences for human health, as MBD5 deletions are the likely cause of certain cases of mental retardation.

Principal Findings

Here we report the first functional characterization of MBD5 and MBD6. We have observed that the proteins colocalize with heterochromatin in cultured cells, and that this localization requires the integrity of their MBD. However, heterochromatic localization is maintained in cells with severely decreased levels of DNA methylation. In vitro, neither MBD5 nor MBD6 binds any of the methylated sequences DNA that were tested.

Conclusions

Our data suggest that MBD5 and MBD6 are unlikely to be methyl-binding proteins, yet they may contribute to the formation or function of heterochromatin. One isoform of MBD5 is highly expressed in oocytes, which suggests a possible role in epigenetic reprogramming after fertilization.     

5-Methylcytosine DNA glycosylase activity is also present in the human MBD4 (G/T mismatch glycosylase) and in a related avian sequence

A 1468 bp cDNA coding for the chicken homolog of the human MBD4 G/T mismatch DNA glycosylase was isolated and sequenced. The derived amino acid sequence (416 amino acids) shows 46% identity with the human MBD4 and the conserved catalytic region at the C-terminal end (170 amino acids) has 90% identity. The non-conserved region of the avian protein has no consensus sequence for the methylated DNA binding domain. The recombinant proteins from human and chicken have G/T mismatch as well as 5-methylcytosine (5-MeC) DNA glycosylase activities. When tested by gel shift assays, human recombinant protein with or without the methylated DNA binding domain binds equally well to symmetrically, hemimethylated DNA and non-methylated DNA. However, the enzyme has only 5-MeC DNA glycosylase activity with the hemimethylated DNA. Footprinting of human MBD4 and of an N-terminal deletion mutant with partially depurinated and depyrimidinated substrate reveal a selective binding of the proteins to the modified substrate around the CpG. As for 5-MeC DNA glycosylase purified from chicken embryos, MBD4 does not use oligonucleotides containing mCpA, mCpT or mCpC as substrates. An mCpG within an A+T-rich oligonucleotide is a much better substrate than an A+T-poor sequence. The K_m of human MBD4 for hemimethylated DNA is ~10^–7 M with a V_max of ~10^–11 mol/h/µg protein. Deletion mutations show that G/T mismatch and 5-MeC DNA glycosylase are located in the C-terminal conserved region. In sharp contrast to the 5-MeC DNA glycosylase isolated from the chicken embryo DNA demethylation complex, the two enzymatic activities of MBD4 are strongly inhibited by RNA. In situ hybridization with antisense RNA indicate that MBD4 is only located in dividing cells of differentiating embryonic tissues.     

Structure-specific binding of MeCP2 to four-way junction DNA through its methyl CpG-binding domain

MeCP2, whose methylated DNA-binding domain (MBD) binds preferentially to DNA containing 5Me-CpG relative to linear unmethylated DNA, also binds preferentially, and with similar affinity, to unmethylated four-way DNA junctions through the MBD. The Arg133Cys (R133C) mutation in the MBD, a Rett syndrome mutation that abolishes binding to methylated DNA, leads to only a slight reduction in the affinity of the MBD for four-way junctions, suggesting distinct but partially overlapping modes of binding to junction and methylated DNA. Binding to unmethylated DNA junctions is likely to involve a subset of the interactions that occur with methylated DNA. High-affinity, methylation-independent binding to four-way junctions is consistent with additional roles for MeCP2 in chromatin, beyond recognition of 5Me-CpG.     

Differential binding of Escherichia coli McrA protein to DNA sequences that contain the dinucleotide m5CpG

The Escherichia coli McrA protein, a putative C⁵-methylcytosine/C⁵-hydroxyl methylcytosine-specific nuclease, binds DNA with symmetrically methylated HpaII sequences (Cm5CGG), but its precise recognition sequence remains undefined. To determine McrA’s binding specificity, we cloned and expressed recombinant McrA with a C-terminal StrepII tag (rMcrA-S) to facilitate protein purification and affinity capture of human DNA fragments with m5C residues. Sequence analysis of a subset of these fragments and electrophoretic mobility shift assays with model methylated and unmethylated oligonucleotides suggest that N(Y > R) m5CGR is the canonical binding site for rMcrA-S. In addition to binding HpaII-methylated double-stranded DNA, rMcrA-S binds DNA containing a single, hemimethylated HpaII site; however, it does not bind if A, C, T or U is placed across from the m5C residue, but does if I is opposite the m5C. These results provide the first systematic analysis of McrA’s in vitro binding specificity.     

Effect of site-specific DNA methylation and mutagenesis on recognition by methylated DNA-binding protein from human placenta.

Methylated DNA-binding protein (MDBP) from human placenta is the first protein shown to bind specifically to certain DNA sequences only when they are methylated at cytosine residues. Among the sites recognized by MDBP is pB site 1, a pBR322-derived sequence which has a high affinity for MDBP when methylated at all CpG positions. We have substituted pB site 1 with 5-methyl-cytosine (m5C) residues at one to three of its CpG dinucleotides on one strand by the use of m5C-containing oligonucleotides. MDBP binds best when all three CpG dinucleotides in the region 5'-ATCGTCACGGCGAT-3' are methylated. Even more binding is obtained when both strands are methylated. Alteration of various residues in this binding site by oligonucleotide-directed mutagenesis decreased the binding. However, two mutations which increased the dyad symmetry of part of the binding site yielded ligands with a higher affinity for MDBP.