Highly disordered histone H1-DNA model complexes and their condensates.

Disordered proteins play an essential role in a wide variety of biological processes, and are often posttranslationally modified. One such protein is histone H1; its highly disordered C-terminal tail (CH1) condenses internucleosomal linker DNA in chromatin in a way that is still poorly understood. Moreover, CH1 is phosphorylated in a cell cycle-dependent manner that correlates with changes in the chromatin condensation level. Here we present a model system that recapitulates key aspects of the in vivo process, and also allows a detailed structural and biophysical analysis of the stages before and after condensation. CH1 remains disordered in the DNA-bound state, despite its nanomolar affinity. Phase-separated droplets (coacervates) form, containing higher-order assemblies of CH1/DNA complexes. Phosphorylation at three serine residues, spaced along the length of the tail, has little effect on the local properties of the condensate. However, it dramatically alters higher-order structure in the coacervate and reduces partitioning to the coacervate phase. These observations show that disordered proteins can bind tightly to DNA without a disorder-to-order transition. Importantly, they also provide mechanistic insights into how higher-order structures can be exquisitely sensitive to perturbation by posttranslational modifications, thus broadening the repertoire of mechanisms that might regulate chromatin and other macromolecular assemblies.

Structural insights into the mechanism of negative regulation of single-box high mobility group proteins by the acidic tail domain.

The Drosophila and plant (maize) functional counterparts of the abundant vertebrate chromosomal protein HMGB1 (HMG-D and ZmHMGB1, respectively) differ from HMGB1 in having a single HMG box, as well as basic and acidic flanking regions that vary greatly in length and charge. We show that despite these variations, HMG-D and ZmHMGB1 exist in dynamic assemblies in which the basic HMG boxes and linkers associate with their intrinsically disordered, predominantly acidic, tails in a manner analogous to that observed previously for HMGB1. The DNA-binding surfaces of the boxes and linkers are occluded in "auto-inhibited" forms of the protein, which are in equilibrium with transient, more open structures that are "binding-competent." This strongly suggests that the mechanism of auto-inhibition may be a general one. HMG-D and ZmHMGB1 differ from HMGB1 in having phosphorylation sites in their tail and linker regions. In both cases, in vitro phosphorylation of serine residues within the acidic tail stabilizes the assembled form, suggesting another level of regulation for interaction with DNA, chromatin, and other proteins that is not possible for the uniformly acidic (hence unphosphorylatable) tail of HMGB1.

Characterization of the interaction between HMGB1 and H3-a possible means of positioning HMGB1 in chromatin.

High mobility group protein B1 (HMGB1) binds to the internucleosomal linker DNA in chromatin and abuts the nucleosome. Bending and untwisting of the linker DNA results in transmission of strain to the nucleosome core, disrupting histone/DNA contacts. An interaction between H3 and HMGB1 has been reported. Here we confirm and characterize the interaction of HMGB1 with H3, which lies close to the DNA entry/exit points around the nucleosome dyad, and may be responsible for positioning of HMGB1 on the linker DNA. We show that the interaction is between the N-terminal unstructured tail of H3 and the C-terminal unstructured acidic tail of HMGB1, which are presumably displaced from DNA and the HMG boxes, respectively, in the HMGB1-nucleosome complex. We have characterized the interaction by nuclear magnetic resonance spectroscopy and show that it is extensive for both peptides, and appears not to result in the acquisition of significant secondary structure by either partner.

HMGB1-facilitated p53 DNA binding occurs via HMG-Box/p53 transactivation domain interaction, regulated by the acidic tail.

Facilitated binding of p53 to DNA by high mobility group B1 (HMGB1) may involve interaction between the N-terminal region of p53 and the high mobility group (HMG) boxes, as well as HMG-induced bending of the DNA. Intramolecular shielding of the boxes by the HMGB1 acidic tail results in an unstable complex with p53 until the tail is truncated to half its length, at which point the A box, proposed to be the preferred binding site for p53(1-93), is exposed, leaving the B box to bind and bend DNA. The A box interacts with residues 38-61 (TAD2) of the p53 transactivation domain. Residues 19-26 (TAD1) bind weakly, but only in the context of p53(1-93) and not as a free TAD1 peptide. We have solved the structure of the A-box/p53(1-93) complex by nuclear magnetic resonance spectroscopy. The incipient amphipathic helix in TAD2 recognizes the concave DNA-binding face of the A box and may be acting as a single-stranded DNA mimic.

Characterization of chromoshadow domain-mediated binding of heterochromatin protein 1α (HP1α) to histone H3.

The chromoshadow domain (CSD) of heterochromatin protein 1 (HP1) was recently shown to contribute to chromatin binding and transcriptional regulation through interaction with histone H3. Here, we demonstrate the structural basis of this interaction for the CSD of HP1α. This mode of H3 binding is dependent on dimerization of the CSD and recognition of a PxVxL-like motif, as for other CSD partners. NMR chemical shift mapping showed that the H3 residues that mediate the CSD interaction occur in and adjacent to the αN helix just within the nucleosome core. Access to the binding region would require some degree of unwrapping of the DNA near the nucleosomal DNA entry/exit site.

H1 and HMGB1: modulators of chromatin structure.

Histone H1 and HMGB1 (high-mobility group protein B1) are the most abundant chromosomal proteins apart from the core histones (on average, one copy per nucleosome and per ten nucleosomes respectively). They are both highly mobile in the cell nucleus, with high on/off rates for binding. In vivo and in vitro evidence shows that both are able to organize chromatin structure, with H1 binding resulting in a more stable structure and HMGB1 binding in a less stable structure. The binding sites for H1 and HMGB1 in chromatin are partially overlapping, and replacement of H1 by HMGB1 through the highly dynamic nature of their binding, possibly facilitated by interaction between them, could result in switching of chromatin states. Binding of HMGB1 to DNA or chromatin is regulated by its long and highly acidic tail, which is also involved in H1 binding. The present article focuses mainly on HMGB1 and its interaction with chromatin and H1, as well as its chaperone role in the binding of certain transcription factors (e.g. p53) to their cognate DNA.

Tail-mediated collapse of HMGB1 is dynamic and occurs via differential binding of the acidic tail to the A and B domains.

The architectural DNA-binding protein HMGB1 consists of two tandem HMG-box domains joined by a basic linker to a C-terminal acidic tail, which negatively regulates HMGB1-DNA interactions by binding intramolecularly to the DNA-binding faces of both basic HMG boxes. Here we demonstrate, using NMR chemical-shift mapping at different salt concentrations, that the tail has a higher affinity for the B box and that A box-tail interactions are preferentially disrupted. Previously, we proposed a model in which the boxes are brought together in a collapsed, tail-mediated assembly, which is in dynamic equilibrium with a more extended form. Small-angle X-ray scattering data are consistent with such a dynamic equilibrium between collapsed and extended structures and are best represented by an ensemble. The ensembles contain a significantly higher proportion of collapsed structures when the tail is present. (15)N NMR relaxation measurements show that full-length HMGB1 has a significantly lower rate of rotational diffusion than the tail-less protein, consistent with the loss of independent domain motions in an assembled complex. Mapping studies using the paramagnetic spin label MTSL [(1-oxyl-2,2,5,5-tetramethyl-3-pyrrolidin-3-yl)methyl methanethiosulfonate] placed at three locations in the tail confirm our previous findings that the tail binds to both boxes with some degree of specificity. The end of the tail lies further from the body of the protein and is therefore potentially free to interact with other proteins. MTSL labelling at a single site in the A domain (C44) causes detectable relaxation enhancements of B domain residues, suggesting the existence of a "sandwich"-like collapsed structure in which the tail enables the close approach of the basic domains. These intramolecular interactions are presumably important for the dynamic association of HMGB1 with chromatin and provide a mechanism by which protein-protein interactions or posttranslational modifications might regulate the function of the protein at particular sites, or at particular stages in the cell cycle.

The interaction of HMGB1 and linker histones occurs through their acidic and basic tails.

H1 and HMGB1 bind to linker DNA in chromatin, in the vicinity of the nucleosome dyad. They appear to have opposing effects on the nucleosome, H1 stabilising it by "sealing" two turns of DNA around the octamer, and HMGB1 destabilising it, probably by bending the adjacent DNA. Their presence in chromatin might be mutually exclusive. Displacement/replacement of one by the other as a result of their highly dynamic binding in vivo might, in principle, involve interactions between them. Chemical cross-linking and gel-filtration show that a 1:1 linker histone/HMGB1 complex is formed, which persists at physiological ionic strength, and that complex formation requires the acidic tail of HMGB1. NMR spectroscopy shows that the linker histone binds, predominantly through its basic C-terminal domain, to the acidic tail of HMGB1, thereby disrupting the interaction of the tail with the DNA-binding faces of the HMG boxes. A potential consequence of this interaction is enhanced DNA binding by HMGB1, and concomitantly lowered affinity of H1 for DNA. In a chromatin context, this might facilitate displacement of H1 by HMGB1.

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.

Mapping intramolecular interactions between domains in HMGB1 using a tail-truncation approach.

The mechanism underlying negative regulation of HMGB1-DNA interaction by the acidic C-terminal tail is ill defined. To address this issue, we have devised a novel NMR chemical-shift perturbation mapping strategy to elucidate interactions between the tail, which consists solely of aspartic acid and glutamic acid residues, and the two well characterized HMG-box DNA-binding domains. A series of HMGB1 tail-truncation mutants differing from each other by five residues was generated. Chemical-shift perturbation mapping using these mutants shows that tails of different lengths bind with different affinities. Nevertheless, the truncated tails bind along the same path on the HMG boxes as the full-length tail, differences in length being manifested in differences in the "reach". The tail makes extensive contacts with the DNA-binding surfaces of both HMG boxes, thus explaining the basis of negative regulation of HMGB1-DNA interaction by the tail.

Structure of a complex of tandem HMG boxes and DNA.

The high-mobility group protein HMGB1 contains two tandem DNA-binding HMG box domains, A and B, linked by a short flexible linker that allows the two domains to behave independently in the free protein. There is no structural information on how the linked domains and linker behave when bound to DNA, mainly due to the lack of any DNA-sequence preference of HMGB1. We report the structure determination, by NMR spectroscopy, of a well-defined complex of two tandem HMG boxes bound to a 16 bp oligonucleotide. The protein is an engineered version of the AB di-domain of HMGB1, in which the A box has been replaced by the HMG box of the sequence-specific transcription factor SRY, to give SRY.B. In the SRY.B/DNA complex, both HMG boxes bind in the minor groove and contribute to the overall DNA bending by intercalation of bulky hydrophobic residues between base-pairs; the bends reinforce each other, and the basic linker lies partly in the minor groove. As well as being the first structure of an HMG-box di-domain bound to DNA, this provides the first structure of the B domain of HMGB1 bound to DNA.

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.

Engineering the structural stability and functional properties of the GI domain into the intrinsically unfolded GII domain of the yeast linker histone Hho1p.

Yeast Hho1p contains two domains, GI and GII, that are homologous to the single globular domain of the linker histone H1 (GH1). We showed previously that the isolated GI and GII domains have different structural stabilities and functional properties. GI, like GH1 and the related GH5, is stably folded at low ionic strength (10 mM sodium phosphate) and gives strong protection of chromatosome-length DNA ( approximately 166 bp) during micrococcal nuclease digestion of chromatin. GII is intrinsically unfolded in 10 mM sodium phosphate and gives weak chromatosome protection, but in 250 mM sodium phosphate has a structure very similar to that of GI as determined by NMR spectroscopy. We now show that the loop between helices II and III in GII is the cause of both its instability and its inability to confer strong chromatosome protection. A mutant GII, containing the loop of GI, termed GII-L, is stable in 10 mM sodium phosphate and is as effective as GI in chromatosome protection. Two GII mutants with selected mutations within the original loop were also slightly more stable than GII. In GII, two of the four basic residues conserved at the second DNA binding site ("site II") on the globular domain of canonical linker histones, and in GI, are absent. Introduction of the two "missing" site II basic residues into GII or GII-L destabilised the protein and led to decreased chromatosome protection relative to the protein without the basic residues. In general, the ability to confer chromatosome protection in vitro is closely related to structural stability (the relative population of structured and unstructured states). We have determined the structure of GII-L by NMR spectroscopy. GII-L is very similar to GII folded in 250 mM sodium phosphate, with the exception of the substituted loop region, which, as in GI, contains a single helical turn.

Distinct properties of the two putative "globular domains" of the yeast linker histone, Hho1p.

The putative linker histone in Saccharomyces cerevisiae, Hho1p, has two regions of sequence (GI and GII) that are homologous to the single globular domains of linker histones H1 and H5 in higher eukaryotes. However, the two Hho1p "domains" differ with respect to the conservation of basic residues corresponding to the two putative DNA-binding sites (sites I and II) on opposite faces of the H5 globular domain. We find that GI can protect chromatosome-length DNA, like the globular domains of H1 and H5 (GH1 and GH5), but GII does not protect. However, GII, like GH1 and GH5, binds preferentially (and with higher affinity than GI) to four-way DNA junctions in the presence of excess linear DNA competitor, and binds more tightly than GI to linker-histone-depleted chromatin. Surprisingly, in 10 mM sodium phosphate (pH 7.0), GII is largely unfolded, whereas GI, like GH1 and GH5, is structured, with a high alpha-helical content. However, in the presence of high concentrations of large tetrahedral anions (phosphate, sulphate, perchlorate) GII is also folded; the anions presumably mimic DNA in screening the positive charge. This raises the possibility that chromatin-bound Hho1p may be bifunctional, with two folded nucleosome-binding domains.

Two homologous domains of similar structure but different stability in the yeast linker histone, Hho1p.

The Saccharomyces cerevisiae homologue of the linker histone H1, Hho1p, has two domains that are similar in sequence to the globular domain of H1 (and variants such as H5). It is an open question whether both domains are functional and whether they play similar structural roles. Preliminary structural studies showed that the two isolated domains, GI and GII, differ significantly in stability. In 10 mM sodium phosphate (pH 7), the GI domain, like the globular domains of H1 and H5, GH1 and GH5, was stably folded, whereas GII was largely unstructured. However, at high concentrations of large tetrahedral anions (phosphate, sulphate, perchlorate), which might mimic the charge-screening effects of DNA phosphate groups, GII was folded. In view of the potential significance of these observations in relation to the role of Hho1p, we have now determined the structures of its GI and GII domains by NMR spectroscopy under conditions in which GII (like GI) is folded. The backbone r.m.s.d. over the ordered residues is 0.43 A for GI and 0.97 A for GII. Both structures show the "winged-helix" fold typical of GH1 and GH5 and are very similar to each other, with an r.m.s.d. over the structured regions of 1.3 A, although there are distinct differences. The potential for GII to adopt a structure similar to that of GI when Hho1p is bound to chromatin in vivo suggests that both globular domains might be functional. Whether Hho1p performs a structural role by bridging two nucleosomes remains to be determined.

Methylation-dependent silencing at the H19 imprinting control region by MeCP2.

Methylation of CpG dinucleotides is correlated with transcriptional repression of genes, including imprinted genes. In the case of the imprinted H19 gene, a 2 kb imprinting control region (ICR) is subject to differential methylation, as it is methylated only on the silenced paternal allele. This region has previously been shown to act as a silencer element at the endogenous locus. The proteins that bind at the H19 differentially methylated domain (DMD) and mediate transcriptional silencing have yet to be identified, although a family of proteins containing a methyl-CpG-binding domain (MBD), of which MeCP2 is the best characterised, are obvious candidates. MeCP2 can bind to a single methylated CpG dinucleotide through its MBD and also contains a transcriptional repression domain (TRD). The TRD interacts with Sin3a and histone deacetylases (HDACs) in vivo, forming a repressive complex. Here we show that MeCP2 is recruited to the H19 DMD in vivo and can silence a reporter gene regulated by the H19 DMD in a methylation-dependent manner. This repression can be alleviated by deletion of the TRD from MeCP2 or by inhibition of HDAC activity. These data indicate that transcriptional silencing from the H19 ICR involves recruitment of MeCP2 and presumably an associated protein complex with deacetylase activity. This complex may also be recruited to the ICR in vivo, resulting in a compact, repressive chromatin structure capable of silencing the paternal H19 allele.