A role of basic residues and the putative intercalating phenylalanine of the HMG-1 box B in DNA supercoiling and binding to four-way DNA junctions

HMG (high mobility group) 1 is a chromosomal protein with two homologous DNA-binding domains, the HMG boxes A and B. HMG-1, like its individual HMG boxes, can recognize structural distortion of DNA, such as four-way DNA junctions (4WJs), that are very likely to have features common to their natural, yet unknown, cellular binding targets. HMG-1 can also bend/loop DNA and introduce negative supercoils in the presence of topoisomerase I in topologically closed DNAs. Results of our gel shift assays demonstrate that mutation of Arg(97) within the extended N-terminal strand of the B domain significantly (>50-fold) decreases affinity of the HMG box for 4WJs and alters the mode of binding without changing the structural specificity for 4WJs. Several basic amino acids of the extended N-terminal strand (Lys(96)/Arg(97)) and helix I (Arg(110)/Lys(114)) of the B domain participate in DNA binding and supercoiling. The putative intercalating hydrophobic Phe(103) of helix I is important for DNA supercoiling but dispensable for binding to supercoiled DNA and 4WJs. We conclude that the B domain of HMG-1 can tolerate substitutions of a number of amino acid residues without abolishing the structure-specific recognition of 4WJs, whereas mutations of most of these residues severely impair the topoisomerase I-mediated DNA supercoiling and change the sign of supercoiling from negative to positive.     

Nature of full-length HMGB1 binding to cisplatin-modified DNA

HMGB1, a highly conserved non-histone DNA-binding protein, interacts with specific DNA structural motifs such as those encountered at cisplatin damage, four-way junctions, and supercoils. The interaction of full-length HMGB1, containing two tandem HMG box domains and a C-terminal acidic tail, with cisplatin-modified DNA was investigated by hydroxyl radical footprinting and electrophoretic gel mobility shift assays. The full-length HMGB1 protein binds to DNA containing a 1,2-intrastrand d(GpG) cross-link mainly through domain A, as revealed by footprinting, with a dissociation constant K(d) of 120 nM. Site-directed mutagenesis of intercalating residues in both HMG domains A and B in full-length HMGB1 further supports the conclusion that only one HMG box domain is bound to the site of cisplatin damage. Interaction of the C-terminal tail with the rest of the HMGB1 protein was examined by EDC cross-linking experiments. The acidic tail mainly interacts with domain B and linker regions rather than domain A in HMGB1. These results illuminate the respective roles of the tandem HMG boxes and the C-terminal acidic tail of HMGB1 in binding to DNA and to the major DNA adducts formed by the anticancer drug cisplatin.     

A critical role in structure-specific DNA binding for the acetylatable lysine residues in HMGB1

The structure-specific DNA-binding protein HMGB1 (high-mobility group protein B1) which comprises two tandem HMG boxes (A and B) and an acidic C-terminal tail, is acetylated in vivo at Lys(2) and Lys(11) in the A box. Mutation to alanine of both residues in the isolated A domain, which has a strong preference for pre-bent DNA, abolishes binding to four-way junctions and 88 bp DNA minicircles. The same mutations in full-length HMGB1 also abolish its binding to four-way junctions, and binding to minicircles is substantially impaired. In contrast, when the acidic tail is absent (AB di-domain) there is little effect of the double mutation on four-way junction binding, although binding to minicircles is reduced approximately 15-fold. Therefore it appears that in AB the B domain is able to substitute for the non-functional A domain, whereas in full-length HMGB1 the B domain is masked by the acidic tail. In no case does single substitution of Lys(2) or Lys(11) abolish DNA binding. The double mutation does not significantly perturb the structure of the A domain. We conclude that Lys(2) and Lys(11) are critical for binding of the isolated A domain and HMGB1 to distorted DNA substrates.     

Intercalating residues determine the mode of HMG1 domains A and B binding to cisplatin-modified DNA

Cisplatin exerts its anticancer activity by forming covalent adducts with DNA. High-mobility group (HMG)-domain proteins recognize the major 1,2-intrastrand cisplatin-DNA cross-links and can mediate cisplatin cytotoxicity. The crystal structure of HMG1 domain A bound to cisplatin-modified DNA, further analyzed here, reveals intercalation of a key Phe37 residue. Other published structures of HMG domains bound to DNA, including NHP6A and HMG-D, similarly indicate amino acid side chains intercalating into linear DNA to form a bend. To delineate the importance of such side chain intercalations and further to explore the binding modes of different HMG domains toward prebent DNA structures, site-directed mutagenesis was used to generate HMG1 domain A and domain B mutants. The affinities of these mutant proteins for cisplatin-modified DNA were determined in gel electrophoresis mobility shift assays. The results indicate that intercalating residues at positions 16 or 37 can both contribute to the binding affinity. The data further reveal that the length of the loop between helices I and II is not critical for binding affinity. Footprinting analyses indicate that the position of the intercalating residue dictates the binding mode of the domain toward platinated DNA. Both congruent and offset positioning of the HMG domain with respect to the locus of the cisplatin-induced bend in the DNA were encountered. Packing interactions in the crystal structure suggest how full-length HMG1 might bind to DNA by contacting more than one duplex simultaneously. Taken together, these results demonstrate that cisplatin modification of DNA provides an energetically favorable, prebent target for HMG domains, which bind to these targets through one or more side chain and favorable hydrophobic surface interactions.     

Laser-induced photo-cross-linking of cisplatin-modified DNA to HMG-domain proteins.

Laser-induced photo-cross-linking was investigated for DNA, modified with cisplatin at specific sites, bound to structure-specific recognition domains of proteins in the high-mobility group (HMG) class. The efficiency of photo-cross-linking depends on the wavelength and power of the laser, the nature of the protein domain, and the oligodeoxyribonucleotide sequences flanking the platinated site. Introduction of 5-iodouridine at thymine sites of the oligodeoxyribonucleotide as an additional photoreactive group did not increase the photo-cross-linking yield. Formation of platinum-mediated DNA-DNA interstrand cross-linking observed previously upon irradiation with 302 nm light [Kane, S. A., and Lippard, S. J. (1996) Biochemistry 35, 2180-2188] was significantly reduced with laser irradiation. HMG1 domain B is superior to domain A for platinum-mediated photo-cross-linking, a result attributed to the different positioning of the proteins with respect to the platinum adduct and the greater ability of domain B to access photolabilized platinum in the major groove. Studies with proteins containing specifically mutated amino acids, and with DNA probes in which the sequences flanking the platinum cross-link site were varied, suggest that the most effective photo-cross-linking occurs for protein domains bound symmetrically and flexibly to cisplatin-modified DNA. The thermodynamic equilibrium between the protein-platinated DNA complex and its components, revealed in gel electrophoretic mobility shift assays (EMSAs), is significantly shifted to the right upon irreversible photo-cross-linking. Thus, only upon photo-cross-linking can the interaction of cisplatin-DNA 1,3-intrastrand d(GpTpG) or interstrand cross-links with HMG1 domain B protein be detected. Photo-cross-linking is thus an effective tool for investigating the interaction of cisplatin-modified DNA with damage-recognition proteins under heterogeneous conditions such those in cell extracts or living cells.     

Linker histones do not interact with DNA containing a single interstrand cross-link created by cisplatin

During our earlier investigations we have observed a prominent preference of the linker histone H1 for binding to a cis-platinated DNA (a synthetic fragment with global type of platination in respect to targets for cisplatin) comparing with unmodified and trans-Pt-modified DNA. In the present work we report our recent experimental results on the binding of the linker histones H1 and H5 to a cisplatin-modified synthetic DNA fragment containing a single nucleotide target d(GC/CG) for inter-platination. Surprisingly, no preferential binding of linker histones to cis-inter-platinated DNA was observed by means of the electromobility-shift assay. The same negative results were obtained with a part of the linker histone molecule suggested to be responsible for DNA-binding--its globular domain. Contrary, the data with another nuclear protein with similar DNA-binding properties as linker histones--HMGB1--showed a strong afinity for interaction with DNA containing interstrand cross-links.     

Molecular modeling of the chromatosome particle

In an effort to understand the role of the linker histone in chromatin folding, its structure and location in the nucleosome has been studied by molecular modeling methods. The structure of the globular domain of the rat histone H1d, a highly conserved part of the linker histone, built by homology modeling methods, revealed a three-helical bundle fold that could be described as a helix–turn–helix variant with its characteristic properties of binding to DNA at the major groove. Using the information of its preferential binding to four-way Holliday junction (HJ) DNA, a model of the domain complexed to HJ was built, which was subsequently used to position the globular domain onto the nucleosome. The model revealed that the primary binding site of the domain interacts with the extra 20 bp of DNA of the entering duplex at the major groove while the secondary binding site interacts with the minor groove of the central gyre of the DNA superhelix of the nucleosomal core. The positioning of the globular domain served as an anchor to locate the C-terminal domain onto the nucleosome to obtain the structure of the chromatosome particle. The resulting structure had a stem-like appearance, resembling that observed by electron microscopic studies. The C-terminal domain which adopts a high mobility group (HMG)-box-like fold, has the ability to bend DNA, causing DNA condensation or compaction. It was observed that the three S/TPKK motifs in the C-terminal domain interact with the exiting duplex, thus defining the path of linker DNA in the chromatin fiber. This study has provided an insight into the probable individual roles of globular and the C-terminal domains of histone H1 in chromatin organization.     

Redox state-dependent interaction of HMGB1 and cisplatin-modified DNA

HMGB1, one of the most abundant nuclear proteins, has a strong binding affinity for cisplatin-modified DNA. It has been proposed that HMGB1 enhances the anticancer efficacy of cisplatin by shielding platinated DNA lesions from repair. Two cysteine residues in HMGB1 domain A form a reversible disulfide bond under mildly oxidizing conditions. The reduced domain A protein binds to a 25-bp DNA probe containing a central 1,2-d(GpG) intrastrand cross-link, the major platinum-DNA adduct, with a 10-fold greater binding affinity than the oxidized domain A. The binding affinities of singly and doubly mutated HMGB1 domain A, respectively deficient in one or both cysteine residues that form the disulfide bond, are unaffected by changes in external redox conditions. The redox-dependent nature of the binding of HMGB1 domain A to cisplatin-modified DNA suggests that formation of the intradomain disulfide bond induces a conformational change that disfavors binding to cisplatin-modified DNA. Hydroxyl radical footprinting analyses of wild-type domain A bound to platinated DNA under different redox conditions revealed identical cleavage patterns, implying that the asymmetric binding mode of the protein across from the platinated lesion is conserved irrespective of the redox state. The results of this study reveal that the cellular redox environment can influence the interaction of HMGB1 with the platinated DNA and suggest that the redox state of the A domain is a potential factor in regulating the role of the protein in modulating the activity of cisplatin as an anticancer drug.     

Recognition of DNA interstrand cross-link of antitumor cisplatin by HMGB1 protein

Several proteins that specifically bind to DNA modified by cisplatin, including those containing HMG-domains, mediate antitumor activity of this drug. Oligodeoxyribonucleotide duplexes containing a single, site-specific interstrand cross-link of cisplatin were probed for recognition by the rat chromosomal protein HMGB1 and its domains A and B using the electrophoretic mobility-shift assay. It has been found that the full-length HMGB1 protein and its domain B to which the lysine-rich region (seven amino acid residues) of the A/B linker is attached at the N-terminus (the domain HMGB1b7) specifically recognize DNA interstrand cross-linked by cisplatin. The affinity of these proteins to the interstrand cross-link of cisplatin is not very different from that to the major 1,2-GG intrastrand cross-link of this drug. In contrast, no recognition of the interstrand cross-link by the domain B lacking this region or by the domain A with or without this lysine-rich region attached to its C-terminus is noticed under conditions when these proteins readily bind to 1,2-GG intrastrand adduct. A structural model for the complex formed between the interstrand cross-linked DNA and the domain HMGB1b7 was constructed and refined using molecular mechanics and molecular dynamics techniques. The calculated accessible areas around the deoxyribose protons correlate well with the experimental hydroxyl radical footprint. The model suggests that the only major adaptation necessary for obtaining excellent surface complementarity is extra DNA unwinding (approximately 40 degrees ) at the site of the cross-link. The model structure is consistent with the hypothesis that the enhancement of binding affinity afforded by the basic lysine-rich A/B linker is a consequence of its tight binding to the sugar-phosphate backbone of both DNA strands.     

The Saccharomyces cerevisiae linker histone Hho1p, with two globular domains, can simultaneously bind to two four-way junction DNA molecules

Saccharomyces cerevisiae encodes a single linker histone, Hho1p, with two globular domains. This raised the possibility that Hho1p could bind to two nucleosome cores simultaneously. To evaluate this idea, we studied the ability of a four-way junction, immobilized on the surface of a magnetic bead, to pull down a radiolabeled four-way junction in the presence of different Hho1 proteins. Four-way junctions are known to bind to H1, presumably due to structure similarities to the DNA at the nucleosomal entry/exit point. We found a significant increase in the ability of full-length Hho1p to pull down radiolabeled four-way junction DNA under ionic conditions where both globular domains could bind. The binding was structure specific, since the use of double-stranded DNA, or a mutant Hho1p in which the second DNA binding site of globular domain 1 was abolished, resulted in a significant decrease in bridged binding. Additionally, bridged binding required a covalent attachment between the two globular domains, since factor Xa protease treatment of the complex formed by a modified Hho1p that contained a factor Xa cleavage site between the two globular domains resulted in a significant release of radiolabeled four-way junction. These findings demonstrated that the two globular domains independently associated with two different four-way junction molecules in a manner that required amino acid residues implicated in structure-specific binding in the nucleosome. We discuss the implication of these findings on the chromatin structure of yeast and propose a model where a single Hho1 protein binds to two serially adjacent nucleosomes.     

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.     

Competition between HMG-I(Y), HMG-1 and histone H1 on four-way junction DNA.

High mobility group proteins HMG-I(Y) and HMG-1, as well as histone H1, all share the common property of binding to four-way junction DNA (4H), a synthetic substrate commonly used to study proteins involved in recognizing and resolving Holliday-type junctions formed during in vivo genetic recombination events. The structure of 4H has also been hypothesized to mimic the DNA crossovers occurring at, or near, the entrance and exit sites on the nucleosome. Furthermore, upon binding to either duplex DNA or chromatin, all three of these nuclear proteins share the ability to significantly alter the structure of bound substrates. In order to further elucidate their substrate binding abilities, electrophoretic mobility shift assays were employed to investigate the relative binding capabilities of HMG-I(Y), HMG-1 and H1 to 4H in vitro. Data indicate a definite hierarchy of binding preference by these proteins for 4H, with HMG-I(Y) having the highest affinity (Kd approximately 6.5 nM) when compared with either H1 (Kd approximately 16 nM) or HMG-1 (Kd approximately 80 nM). Competition/titration assays demonstrated that all three proteins bind most tightly to the same site on 4H. Hydroxyl radical footprinting identified the strongest site for binding of HMG-I(Y), and presumably for the other proteins as well, to be at the center of 4H. Together these in vitro results demonstrate that HMG-I(Y) and H1 are co-dominant over HMG-1 for binding to the central crossover region of 4H and suggest that in vivo both of these proteins may exert a dominant effect over HMG-1 in recognizing and binding to altered DNA structures, such as Holliday junctions, that have conformations similar to 4H.     

Helical phasing between DNA bends and the determination of bend direction.

The presence and location of bends in DNA can be inferred from the anomalous mobility of DNA fragments or protein-DNA complexes during electrophoresis in polyacrylamide gels. Direction of bending is not so easily determined. We show here that a protein-induced bend, when linked to a protein-independent DNA bend by a segment of variable length, exhibits an electrophoretic mobility that varies in a sinusoidal manner with the length of the linker. Mobility minima occur once for each addition to the linker of one helical turn of DNA. Since minima should occur when two bends reinforce one another, the direction of one bend relative to the other can be determined from the distances between the two centers of bending at which minima occur. Our results strongly support the idea that the A5-6 tracts in kinetoplast DNA bend towards the minor groove while the bend at the recombination site of the gamma delta resolvase (binding site I of the gamma delta res site) bends towards the major groove.     

A functional linker in human topoisomerase I is required for maximum sensitivity to camptothecin in a DNA relaxation assay.

Human topoisomerase I is composed of four major domains: the highly charged NH(2)-terminal region, the conserved core domain, the positively charged linker domain, and the highly conserved COOH-terminal domain. Near complete enzyme activity can be reconstituted by combining recombinant polypeptides that approximate the core and COOH-terminal domains, although DNA binding is reduced somewhat for the reconstituted enzyme (Stewart, L., Ireton, G. C., and Champoux, J. J. (1997) J. Mol. Biol. 269, 355-372). A reconstituted enzyme comprising the core domain plus a COOH-terminal fragment containing the complete linker region exhibits the same biochemical properties as a reconstituted enzyme lacking the linker altogether, and thus detachment of the linker from the core domain renders the linker non-functional. The rate of religation by the reconstituted enzyme is increased relative to the forms of the enzyme containing the linker indicating that in the intact enzyme the linker slows religation. Relaxation of plasmid DNA by full-length human topoisomerase I or a 70-kDa form of the enzyme that is missing only the non-essential NH(2)-terminal domain (topo70) is inhibited approximately 16-fold by the anticancer compound, camptothecin, whereas the reconstituted enzyme is nearly resistant to the inhibitory effects of the drug despite similar affinities for the drug by the two forms of the enzyme. Based on these results and in light of the crystal structure of human topoisomerase I, we propose that the linker plays a role in hindering supercoil relaxation during the normal relaxation reaction and that camptothecin inhibition of DNA relaxation depends on a direct effect of the drug on DNA rotation that is also dependent on the linker.     

The A.T-DNA-binding domain of mammalian high mobility group I chromosomal proteins. A novel peptide motif for recognizing DNA structure.

We have determined the domains of the mammalian high mobility group (HMG)I chromosomal proteins necessary and sufficient for binding to the narrow minor groove of stretches of A.T-rich DNA. Three highly conserved regions within each of the known HMG-I proteins is closely related to the consensus sequence T-P-K-R-P-R-G-R-P-K-K. A synthetic oligopeptide corresponding to this consensus "binding domain" (BD) sequence specifically binds to substrate DNA in a manner similar to the intact HMG-I proteins. Molecular Corey-Pauling-Koltun model building and computer simulations employing energy minimization programs to predict structure suggest that the consensus BD peptide has a secondary structure similar to the antitumor and antiviral drugs netropsin and distamycin and to the dye Hoechst 33258. In vitro these ligands, which also preferentially bind to A.T-rich DNA, have been demonstrated to effectively compete with both the BD peptide and the HMG-I proteins for DNA binding. The BD peptide also contains novel structural features such as a predicted Asx bend or "hook" at its amino-terminal end and laterally projecting cationic Arg/Lys side chains or "bristles" which may contribute to the binding properties of the HMG-I proteins. The predicted BD peptide structure, which we refer to as the "A.T-hook," represents a previously undescribed DNA-binding motif capable of binding to the minor groove of stretches of A.T base pairs.