Lysine-containing DNA-binding regions on the surface of the histone octamer in the nucleosome core particle

Homeotic gene regulation: a paradigm for epigenetic mechanisms underlying organismal development

The organization of eukaryotic genome into chromatin within the nucleus eventually dictates the cell type specific expression pattern of genes. This higher order of chromatin organization is established during development and dynamically maintained throughout the life span. Developmental mechanisms are conserved in bilaterians and hence they have body plan in common, which is achieved by regulatory networks controlling cell type specific gene expression. Homeotic genes are conserved in metazoans and are crucial for animal development as they specify cell type identity along the anterior-posterior body axis. Hox genes are the best studied in the context of epigenetic regulation that has led to significant understanding of the organismal development. Epigenome specific regulation is brought about by conserved chromatin modulating factors like PcG/trxG proteins during development and differentiation. Here we discuss the conserved epigenetic mechanisms relevant to homeotic gene regulation in metazoans.

DNA binding of centromere protein C (CENPC) is stabilized by single-stranded RNA

Centromeres are the attachment points between the genome and the cytoskeleton: centromeres bind to kinetochores, which in turn bind to spindles and move chromosomes. Paradoxically, the DNA sequence of centromeres has little or no role in perpetuating kinetochores. As such they are striking examples of genetic information being transmitted in a manner that is independent of DNA sequence (epigenetically). It has been found that RNA transcribed from centromeres remains bound within the kinetochore region, and this local population of RNA is thought to be part of the epigenetic marking system. Here we carried out a genetic and biochemical study of maize CENPC, a key inner kinetochore protein. We show that DNA binding is conferred by a localized region 122 amino acids long, and that the DNA-binding reaction is exquisitely sensitive to single-stranded RNA. Long, single-stranded nucleic acids strongly promote the binding of CENPC to DNA, and the types of RNAs that stabilize DNA binding match in size and character the RNAs present on kinetochores in vivo. Removal or replacement of the binding module with HIV integrase binding domain causes a partial delocalization of CENPC in vivo. The data suggest that centromeric RNA helps to recruit CENPC to the inner kinetochore by altering its DNA binding characteristics.

Here we address the issue of how genetic information is passed from one generation to the next without the involvement of specific DNA sequences. This type of inheritance is referred to as epigenetics. Centromeric sequences are highly variable and in many cases are not sufficient for centromere function. Rather, secondary features of the DNA, such as methylation or associated RNA molecules may serve to recruit key centromere binding proteins. Prior data from several species have established that single-stranded RNAs are surprisingly abundant on centromeric chromatin. Here we identified the DNA-binding domain of a key centromere binding protein in maize (CENPC) and showed that it requires single-stranded RNA to effectively bind DNA in vitro. When the DNA/RNA binding domain was deleted, the accuracy of CENPC targeting to centromeres was reduced but not abolished. The results bolster the view that centromere-bound RNA is one component of the epigenetic determination process that assures centromeres are stably inherited. In addition, our data suggest a general mechanism for how RNA can influence the binding of chromatin proteins to DNA.

Photoaffinity labelling of a DNA-binding site on the globular domain of histone H5

We have labelled a DNA-binding site on the globular domain of histone H5 (GH5) by ultraviolet-activated cross-linking of a self-complementary 5-bromodeoxyuridine (5BrU)-substituted oligonucleotide with the sequence 5'-AGCGA5BrUATCGCT-3'. Cross-linking was to His62, mainly to the protein backbone. This observation provides further support for the mode of binding of GH5 to DNA proposed on the basis of the similarity between the X-ray crystal structure of GH5 and the DNA-bound structures of catabolite activator protein and hepatic nuclear factor 3 gamma [Ramakrishnan, V. (1994) Curr. Opin. Struct. Biol. 4. 44-50].

A novel histone H4 mutant defective in nuclear division and mitotic chromosome transmission

The histone proteins are essential for the assembly and function of th e eukaryotic chromosome. Here we report the first isolation of a temperature-sensitive lethal histone H4 mutant defective in mitotic chromosome transmission Saccharomyces cerevisiae. The mutant requires two amino acid substitutions in histone H4: a lethal Thr-to-Ile change at position 82, which lies within one of the DNA-binding surfaces of the protein, and a substitution of Ala to Val at position 89 that is an intragenic suppressor. Genetic and biochemical evidence shows that the mutant histone H4 is temperature sensitive for function but not for synthesis, deposition, or stability. The chromatin structure of 2 micrometer circle minichromosomes is temperature sensitive in vivo, consistent with a defect in H4-DNA interactions. The mutant also has defects in transcription, displaying weak Spt- phenotypes. At the restrictive temperature, mutant cells arrest in the cell cycle at nuclear division, with a large bud, a single nucleus with 2C DNA content, and a short bipolar spindle. At semipermissive temperatures, the frequency of chromosome loss is elevated 60-fold in the mutant while DNA recombination frequencies are unaffected. High-copy CSE4, encoding an H3 variant related to the mammalian CENP-A kinetochore antigen, was found to suppress the temperature sensitivity of the mutant without suppressing the Spt- transcription defect. These genetic, biochemical, and phenotypic results indicate that this novel histone H4 mutant defines one or more chromatin-dependent steps in chromosome segregation.

Rearrangement of the histone H2A C-terminal domain in the nucleosome

Using zero-length covalent protein-DNA crosslinking, we have mapped the histone-DNA contacts in nucleosome core particles from which the C- and N-terminal domains of histone H2A were selectively trimmed by trypsin or clostripain. We found that the flexible trypsin-sensitive C-terminal domain of histone H2A contacts the dyad axis, whereas its globular domain contacts the end of DNA in the nucleosome core particle. The appearance of the histone H2A contact at the dyad axis occurs only in the absence of linker DNA and does not depend on the absence of linker histones. Our results show the ability of the histone H2A C-terminal domain to rearrange. This rearrangement might play a biological role in nucleosome disassembly and reassembly and the retention of the H2A-H2B dimer (or the whole octamer) during the passing of polymerases through the nucleosome.

Cooperative binding of the globular domains of histones H1 and H5 to DNA

In view of the likely role of H1-H1 interactions in the stabilization of chromatin higher order structure, we have asked whether interactions can occur between the globular domains of the histone molecules. We have studied the properties of the isolated globular domains of H1 and the variant H5 (GH1 and GH5) and we have shown (by sedimentation analysis, electron microscopy, chemical cross-linking and nucleoprotein gel electrophoresis) that although GH1 shows no, and GH5 little if any, tendency to self-associate in dilute solution, they bind highly cooperatively to DNA. The resulting complexes appear to contain essentially continuous arrays of globular domains bridging 'tramlines' of DNA, similar to those formed with intact H1, presumably reflecting the ability of the globular domain to bind more than one DNA segment, as it is likely to do in the nucleosome. Additional (thicker) complexes are also formed with GH5, probably resulting from association of the primary complexes, possibly with binding of additional GH5. The highly cooperative nature of the binding, in close apposition, of GH1 and GH5 to DNA is fully compatible with the involvement of interactions between the globular domains of H1 and its variants in chromatin folding.

Succinylation of histone amino groups facilitates transcription of nucleosomal cores

Treatment of nucleosomal cores with succinic anhydride, which modifies preferentially the amino-terminal domains of core histones, takes place without dissociation of the particles. Low levels of modification, which cause small structural effects, are accompanied by substantial increases in the efficiency of the nucleosomal cores as in vitro transcription templates for RNA polymerase II. The transcriptional properties of the succinylated nucleosomal cores are similar to those of the acetylated particles (Piñeiro et al., (1991) Biochem. Biophys. Res. Commun. 177, 370-376), indicating that no specific blocking by acetyl residues is required to facilitate in vitro transcription. Moreover, to obtain a certain level of stimulation, a smaller number of groups has to be modified by succinic than by acetic anhydride.

Histone structure and function

The past year has seen major advances in our understanding of histone and nucleosome structure and function. Direct DNA mapping and thermodynamic experiments have finally provided conclusive evidence that the histones impose an altered helical pitch on the DNA as it is wrapped on the surface of the core histone octamer. Further, it is now clear that lysine acetylation in the amino-terminal domains of histones H3 and H4 can alter the topology of the DNA in chromatin and probably influence its higher-order folding. Genetic experiments reported in the past year have provided a wealth of new information on histone structure and function, including the identification of the peptide domain of histone H4 that is necessary for permanent gene repression, the confirmation that nucleosome structure is critical for centromere function, and evidence that histone acetylation plays a significant role in chromosome dynamics.

Sperm decondensation in Xenopus egg cytoplasm is mediated by nucleoplasmin

At fertilization, sperm chromatin decondenses in two stages, which can be mimicked in extracts of Xenopus eggs. Rapid, limited decondensation is followed by slower, membrane-dependent decondensation and swelling. Nucleoplasmin, an acidic nuclear protein, occurs at high concentration in Xenopus eggs and has a histone-binding role in nucleosome assembly. Immunodepleting nucleoplasmin from egg extracts inhibits the initial rapid stage of sperm decondensation, and also the decondensation of myeloma nuclei, relative to controls of mock depletion and TFIIIA depletion. Readdition of purified nucleoplasmin recues depleted extracts. A physiological concentration of purified nucleoplasmin alone decondenses both sperm and myeloma nuclei. We conclude that nucleoplasmin is both necessary and sufficient for the first stage of sperm decondensation in Xenopus eggs.

Molecular modelling study of changes induced by netropsin binding to nucleosome core particles

It is well known that certain sequence-dependent modulators in structure appear to determine the rotational positioning of DNA on the nucleosome core particle. That preference is rather weak and could be modified by some ligands as netropsin, a minor-groove binding antibiotic. We have undertaken a molecular modelling approach to calculate the relative energy of interaction between a DNA molecule and the protein core particle. The histones particle is considered as a distribution of positive charges on the protein surface that interacts with the DNA molecule. The molecular electrostatic potentials for the DNA, simulated as a discontinuous cylinder, were calculated using the values for all the base pairs. Computing these parameters, we calculated the relative energy of interaction and the more stable rotational setting of DNA. The binding of four molecules of netropsin to this model showed that a new minimum of energy is obtained when the DNA turns toward the protein surface by about 180 degrees, so a new energetically favoured structure appears where netropsin binding sites are located facing toward the histones surface. The effect of netropsin could be explained in terms of an induced change in the phasing of DNA on the core particle. The induced rotation is considered to optimize non-bonded contacts between the netropsin molecules and the DNA backbone.

Structure of nucleosomes and organization of internucleosomal DNA in chromatin

We have compared the mononucleosomal pattern produced by micrococcal nuclease digestion of condensed and unfolded chromatin and chromatin in nuclei from various sources with the repeat length varying from 165 to 240 base-pairs (bp). Upon digestion of isolated H1-containing chromatin of every tested type in a low ionic strength solution (unfolded chromatin), a standard series of mononucleosomes (MN) was formed: the core particle, MN145, and H1-containing, MN165, MN175, MN185, MN195, MN205 and MN215 (the indexes give an approximate length of the nucleosomal DNA that differs in these particles by an integral number of 10 bp). In addition to the pattern of unfolded chromatin, digestion of whole nuclei or condensed chromatin (high ionic strength of Ca2+) gave rise to nuclei-specific, H1-lacking MN155. Digestion of H1-lacking chromatin produced only MN145, MN155 and MN165 particles, indicating that the histone octamer can organize up to 165 bp of nucleosomal DNA. Although digestion of isolated sea urchin sperm chromatin (repeat length of about 240 bp) at a low ionic strength gave a typical "unfolded chromatin pattern", digests of spermal nuclei contained primarily MN145, MN155, MN235 and MN245 particles. A linear arrangement of histones along DNA (primary organization) of the core particle was found to be preserved in the mononucleosomes, with the spacer DNA length from 10 to 90 bp on one (in MN155) or both sides of core DNA being a multiple of about 10 bp. In MN235, the core particle occupies preferentially a central position with the length of the spacer DNA on both sides of the core DNA being usually about 30 + 60 or 40 + 50 bp. Histone H1 is localized at the ends of these particles, i.e. close to the centre of the spacer DNA. The finding that globular part of histones H3 and sea urchin sperm H2B can covalently bind to spacer DNA suggests their involvement in the organization of chromatin superstructure. Our data indicate that decondensation of chromatin is accompanied by rearrangement of histone H1 on the spacer DNA sites adjacent to the core particle and thus support a solenoid model for the chromatin superstructure in nuclei in which the core DNA together with the spacer DNA form a continuous superhelix.

Core histone-DNA interactions in sea urchin sperm chromatin. The N-terminal tail of H2B interacts with linker DNA

A three-stage chemical modification procedure [Lambert, S. F. & Thomas, J. O. (1986) Eur. J. Biochem. 160, 191-201; Thomas, J. O. & Wilson, C. M. (1986) EMBO J. 5, 3531-3537] for selectively radiolabelling lysine residues that interact with DNA has been used to investigate core histone--DNA interactions in sea urchin sperm chromatin, in particular to determine the binding site of the long N-terminal domain of sperm-specific H2B. Comparison of the patterns of radiolabelling of core histones from extended chromatin and nucleosome core particles (which lack linker DNA) reveals the regions of the histones involved in interactions with the linker. The results show that the N-terminal domain of H2B is bound to DNA outside the 146-bp nucleosome core, presumably to the linker DNA. H2A and H4 make no substantial contacts with the linker in extended chromatin; the N-terminal tail of H4 is bound within the core particle, but the N-terminal tail of H2A is not bound in core particles or in extended chromatin, and may therefore have a role in higher-order structure. H3, like H2B, makes contacts with DNA outside the 146-bp nucleosome core in its N-terminal region, as well as elsewhere, and probably interacts with the two 10-bp extensions that complete the two turns of DNA in the nucleosome and/or with the linker.

Dependence of the linking deficiency of supercoiled minichromosomes upon nucleosome distortion

The contribution from each nucleosome to the linking number of minichrosome DNA depends on two factors. These are the wrapping number, omega, which is the number of times the DNA wraps about the axis of the nucleosome; and the winding number, phi, which is the number of base pairs on the nucleosome divided by the helical repeat of the DNA. If the nucleosome is distorted with DNA surface contacts being preserved, phi remains unchanged. The wrapping number may still change, however, depending on the extent of the distortion. For example, if the usual cylindrical shape of the nucleosome is deformed into an ellipsoid while preserving the equatorial radius, then the wrapping number will increase. We apply these concepts to minichromosomes torsionally stressed by supercoiling with, for example, DNA gyrase. We analyze the experimental result that the maximum amount of supercoiling obtained by gyrase treatment of minichromosomes is the same as that of naked DNA. In particular, we show that this phenomenon can be explained by a relatively slight distortion of the nucleosome core while maintaining the surface contacts of the DNA on the core.

Effect of nucleosome distortion on the linking deficiency in relaxed minichromosomes

The wrapping of closed circular DNA on a protein surface, followed by relaxation with a topoisomerase and removal of proteins, produces a characteristic DNA linking deficiency, delta Lk. We show that the magnitude of delta Lk depends upon the surface shape, and we calculate changes in delta Lk caused by particular distortions of the protein wrapping surface. If the DNA remains attached to the surface during distortion, the DNA winding number, phi, is not altered. The change in delta Lk is then equal to the change in the surface linking number, SLk, which is a straightforward measure of the wrapping of the DNA around the surface. For left-handed wrapping, as in a nucleosome, SLk = -n, the number of times that the DNA axis winds around the axis of the protein complex. We calculate values of SLk for the helical wrapping of a constant length of DNA on protein surfaces having the shapes of cylinders and of ellipsoids and hyperboloids of revolution. If the equatorial radius of the protein is fixed, change in shape from a cylinder to a hyperboloid increases SLk, while the corresponding change to an ellipsoid reduces SLk. We apply the general results to the interpretation of experiments in which minichromosomes are relaxed with topoisomerase at various temperatures and delta Lk is determined. The result is that a distortion of the nucleosome core by at most 5% (the change in the radius at the axial extremity relative to the equator) is sufficient to explain the observed delta Lk changes.

Use of selectively trypsinized nucleosome core particles to analyze the role of the histone "tails" in the stabilization of the nucleosome

Using immobilized trypsin and an appropriate fractionation procedure, we have been able to prepare, for the first time, nucleosome core particles containing selectively trypsinized histone domains. The particles thus obtained: [(H3T-H4T)2-2(H2AT-H2BT)].DNA; [(H3-H4)2-2(H2AT-H2BT)].DNA; [H3T-H4T)2-2(H2A-H2B)].DNA (where T means trypsinized), together with the non-trypsinized controls have been characterized using the following techniques: analytical ultracentrifugation, circular dichroism, thermal denaturation and DNAse I digestion. The major aim of this study was to analyze the role of the amino-terminal regions (the histone "tails") on the stability of the nucleosome in solution. The data obtained from this analysis clearly show that stability of the nucleosome core particle to dissociation (below a salt concentration of 0.7 M-NaCl) is not affected by the presence or the absence of any of the N-terminal regions of the histones. Furthermore, these histone regions make very little contribution, if any, to the conformational transition that nucleosomes undergo in this range of salt concentrations. They play, however, a very important role in determining the thermal stability of the particle, as reflected in the dramatic alterations exhibited by the melting profiles upon selective removal of these tails by trypsinization. The melting data can be explained by a simple hypothesis that ascribes interaction of H2A/H2B and H3/H4 tails to particular regions of the nucleosomal DNA.

Studies of histidine phosphorylation by a nuclear protein histidine kinase show that histidine-75 in histone H4 is masked in nucleosome core particles and in chromatin

Histone H4 is a good substrate in vitro for the protein histidine kinase activity found both in Physarum polycephalum nuclear extracts and in Saccharomyces cerevisiae cell extracts. However, histone H4 in nucleosome core particles is not a substrate for these kinases. Isolated chromatin was also not a substrate for the protein histidine kinase. The results significantly limit possible interpretations of histidine phosphorylation on histone H4 in vivo and provide a new, sharper focus for future work. In addition, a polynucleotide kinase activity was identified in the Physarum extracts.

Histone accessibility determined by lysine-specific acetylation in chicken erythrocyte nuclei

N-Hydroxysulfosuccinimidyl [3H]acetate was synthesized and, following the determination of the optimal reaction conditions, was used to acetylate histones in chicken erythrocyte nuclei at 4 degrees C, pH 8. The histones were extracted from the labelled nuclei and the distribution of the acetyl groups determined from the amount of tritiated acetate in isolated peptides. The relative degree of acetylation of molecules was H1 1.0, H5 0.81, H2B 0.48, H2A 0.24, H3 0.24, H4 0.16. Histone H1 is the most exposed histone followed by H5. The core histones are much less accessible to chemical modification than the linker histones by a factor of 4-5. Histones H2A, H2B and H5 appear to be labelled at random along the entire polypeptide chain, while histones H3 and H4 are labelled almost exclusively in the first 30 residues from the N terminus. Control and acetylated chicken erythrocyte nuclei were digested with DNase I and the resulting DNA hybridized to globin and ovalbumin cDNAs. Acetylation, at 14 molecules acetate/core nucleosome or 20 molecules acetate/chromatosome, increased the DNase I sensitivity of the ovalbumin gene to that of the globin sequences in the control sample, while the globin sequences became even more nuclease-sensitive. Our results suggest that increased sensitivity of chromatin towards nuclease digestion might be due to increased solubility of the chromatin fibre.

A highly basic histone H4 domain bound to the sharply bent region of nucleosomal DNA

A nucleosomal core particle is composed of two each of histones H2A, H2B, H3 and H4 located inside the particle with approximately 47 base pairs (bp) of DNA wrapped around the octamer in about 1.8 turns of a left-handed superhelix. The path of the superhelix is not smooth; the DNA is sharply bent, or kinked, at positions symmetrically disposed at a distance of about one and four double-helical turns in both directions from the nucleosomal dyad axis (designated as sites +/- 1 and +/- 4 respectively). This non-uniform bending is considered archetypal to other DNA-protein complexes, but its mechanism is not clear (reviewed in ref. 4). DNA-histone chemical cross-linking within the core particle has revealed strong binding of each of the two histone H4 molecules to DNA at a distance of 1.5 helical turns either side of the nucleosomal dyad axis (sites +/- 1.5). In each of these sites, a single flexible domain of H4 was previously shown to contact three points, at about nucleotides 55 and 65 on one strand and nucleotide 88 on the complementary strand, numbering from the 5' terminus of each 147-base strand; these three locations are closely juxtaposed across the highly compressed minor and major grooves (Fig. 1). Here we report that the amino-acid residue of histone H4 cross-linked at the 1.5 site is histidine-18, embedded in a highly basic cluster Lys-Arg-His-Arg-Lys-Val-Leu-Arg which is probably involved in the sharp bending of the DNA double helix at the +/- 1 sites.