The sites of deposition of newly synthesized histone.

The chromosomal fragments produced by nuclease digestion of freshly replicated chromatin migrate more rapidly relative to bulk chromatin when analyzed in nucleoprotein gels. The cause of the anomalous migration has been studied and the evidence indicates that rather than reflecting a shorter nucleosomal repeat in vivo that it may be a consequence of nucleosome sliding during the digestion itself. The distinct electrophoretic characteristics of nucleosomal material containing newly replicated DNA have enabled us to examine their histone composition by two dimensional electrophoresis. We find that nucleosomes containing new DNA also contain newly synthesized histones H3 and H4. In contrast more than 50% of newly synthesized H2A and H2B, and essentially all of new H1, are deposited at sites on the bulk chromatin distinct from that material containing newly replicated DNA. In addition we show that newly synthesized histones H3 and H4 are bound unusually weakly when they first become associated with the chromatin.     

Replication-coupled chromatin assembly of newly synthesized histones: distinct functions for the histone tail domains

The maintenance of the genome during replication requires the assembly of nucleosomes with newly synthesized histones. Achieving the deposition of newly synthesized histones in chromatin implies their transport from the cytoplasm to the nucleus at the replication sites. Several lines of evidence have revealed critical functions of the histone tail domains in these conserved cellular processes. In this review, we discuss the role of the amino termini of the nucleosome building blocks, H2A/H2B and H3/H4, in different model systems. The experimental data showed that H2A/H2B tails and H3/H4 tails display distinct functions in nuclear import and chromatin assembly. Furthermore, we describe recent studies exploiting the unique properties of the slime mold, Physarum polycephalum , that have advanced understanding of the function of the highly conserved replication-dependent diacetylation of H4.     

Proximity and accessibility studies of histones in nuclei and free nucleosomes.

Histone proximity in chromatin was studied with the cleavable crosslinking reagent, dithiobissuccinimidyl propionate. Crosslinks between H4 and H2a, H4 and H2b, H4 and H3, H2a and H2b, H2b and H3 were found. H1 is also crosslinked to the nucleosomal histones. In nuclei, unsheared chromatin, and H1 depleted chromatin, the four nucleosomal histones are crosslinked at similar relative rates both in 5 mM salt and 100 mM salt. After micrococcal nuclease treatment to generate nucleosomes, H2a and H2b are crosslinked faster than H4 and H3. C14-NEM titration of thiopropionate residues bound to each histone shows that H2a and H2b are more accessible to this reagent after nuclease treatment but that the increased binding was not sufficient by itself to explain the increase in crosslinking. Bolton Hunter reagent was used to further study the accessibility of the four nucleosomal histones in whole chromatin and nuclease digested chromatin. These studies showed that salt increases the accessibility of all four histones while nuclease treatment decreases H4 accessibility.     

Separation of nucleosomes containing histones H1 and H5

Hen erythrocyte chromatin was digested with staphylococcal nuclease and fractionated by electrophoresis in polyacrylamide gels. Instead of the three bands described for mouse carcinoma chromatin, four main discrete components (MN1, MN2, MN2E and MN3) were resolved in the mononucleosome fraction of erythrocyte chromatin. MN2 contained all five histones and a DNA fragment of 165–180 base pairs. MN2E comprised four nucleosomal histones plus histone H5 (but not H1) and a DNA fragment of 170–190 base pairs. The relatively nuclease resistant MN3 fraction of erythrocyte nucleosomes contained H1 but no H5 histone. A more accurate analysis of the MN2 fraction in mouse carcinoma nucleosomes revealed some additional microheterogeneity depending on the presence of two different subfractions of H1.     

Multiacetylated forms of H4 are found in a putative transcriptionally competent chromatin fraction from trout testis.

We have examined the distribution of acetylated histones derived from various trout testis chromatin fractions of different composition. Our results indicate that a chromatin fraction, preferentially solubilized by micrococcal nuclease, containing the bulk of the HMG proteins and similar to a fraction released from intact trout nuclei and previously shown to be enriched in transcribed DNA sequences also possesses high levels of multiacetylated species of H4. Histones 2A, 2B and 3 are also acetylated in this particular chromatin fraction. Monoacetylated species of the 4 inner nucleosomal histones appear to be characteristic of the nucleohistone portion of trout testis chromatin.     

Differential association of linker histones H1 and H5 with telomeric nucleosomes in chicken erythrocytes.

Rat liver telomeric DNA is organised into nucleosomes characterised by a shorter and more homogeneous average nucleosomal repeat than bulk chromatin as shown by Makarov et al. (1). The latter authors were unable to detect the association of any linker histone with the telomeric DNA. We have confirmed these observations but show that in sharp contrast chicken erythrocyte telomeric DNA is organised into nucleosomes whose spacing length and heterogeneity are indistinguishable from those of bulk chromatin. We further show that chicken erythrocyte telomeric chromatin contains chromatosomes which are preferentially associated with histone H1 relative to histone H5. This contrasts with bulk chromatin where histone H5 is the more abundant species. This observation strongly suggests that telomeric DNA condensed into nucleosome core particles has a higher affinity for H1 than H5. We discuss the origin of the discrimination of the lysine rich histones in terms of DNA sequence preferences, telomere nucleosome preferences and particular constraints of the higher order chromatin structure of telomeres.     

Distribution of the core histones H2A.H2B.H3 and H4 during cell replication.

The distribution of newly synthesized core histones H2A, H2B, H3 and H4 relative to the DNA strand synthesized in the same generation has been examined in replicating Chinese Hamster ovary cells. Cells are grown for one generation in [14C]-lysine and thymidine, and then for one generation in [3H]-lysine and 5-bromodeoxyuridine (BrUdRib) and a further generation in unlabeled lysine and thymidine. This protocol produces equal amounts of unifilarly substituted and unsubstituted DNA. Monomer nucleosomes isolated from chromatin containing these two types of DNA can be distinguished by crosslinking with formaldehyde and banding to equilibrium in CsCl density gradients. The results indicate that the core histones are equally distributed between the two types of DNA. These findings are discussed in terms of current models for chromatin replication; they do not support any long term association of newly replicated histones with either the leading or lagging side of the replication fork.     

In vitro exchange of nucleosomal histones H2a and H2b

We have asked whether exogenous, radiolabeled histones can exchange with nucleosomal histones in an in vitro system. Using two different electrophoretic techniques, we were able to separate the histones contained in nucleosomes from those histones which were simply bound to the surface of the chromatin. Fluorography was used to determine which of the exogenous histones exchange with the nucleosomal histones. We observed substantial exchange of histones H1, H2a, and H2b when the chromatin and exogenous histones were incubated under approximately physiological conditions. We have also observed a small amount of exchange of H2a and H2b, as well as a substantial exchange of H1, from one chromatin fragment to another. Other conditions affecting the exchange of histones H2a and H2b are also reported.     

Contribution of histones H2A and H2B to the folding of nucleosomal DNA

We have studied the structural properties of nucleosomal particles deficient in histones H2A and H2B produced by modification of histone amino groups with dimethylmaleic anhydride [Jordano, J., Montero, F., & Palacián, E. (1984) Biochemistry (preceding paper in this issue)]. Digestion with DNase I of residual particles containing only 15% of the original H2A . H2B complement produces only discrete DNA fragments no longer than 70 nucleotides. As compared with the original nucleosomes, thermal denaturation of the residual particles shows a decrease from 140 to about 90 in the number of nucleotide base pairs per particle that melt at the highest temperature transition as well as a drop in the temperature of this transition. Circular dichroism spectra of the residual particles give ellipticity values around 275 nm, much higher than those corresponding to the control nucleosomes, which appears to indicate a loss in the compact DNA tertiary structure. When regeneration of the modified amino groups of the residual particles takes place in the presence of the complementary fraction containing histones H2A and H2B, but not in its absence, nucleosomal particles with the structural properties of the original nucleosomes are reconstituted. Therefore, the structural change observed in the residual particles can be assigned to the lack of histones H2A and H2B and not to the modified amino groups of the histones present in the residual particles. The results are consistent with the stabilization by histones H2A and H2B of a DNA length of 50-70 base pairs per nucleosome.     

Differential response of avian red blood cell nucleosomes to heparin

In avian erythrocyte chromatin, heparin interacts differentially with H1, H5 and the nucleosomal core histones. In non-erythroid cells, a partial extraction of H2A, H2B and H1 yields H3/H4/DNA complexes and particles of unchanged nucleosomal composition. The assay system for this heparin effect includes sucrose gradients, formaldehyde fixation and cesium chloride gradient centrifugation. A comparison of avian erythrocyte nucleosomes with chromatin subunits from precursor cells shows that H5 interferes with the heparin effect whereas a removal of H5 renders the core histones accessible to the polyanion.     

Exchange of histones H1, H2A, and H2B in vivo

We have asked whether histones synthesized in the absence of DNA synthesis can exchange into nucleosomal structures. DNA synthesis was inhibited by incubating hepatoma tissue culture cells in medium containing 5.0 mM hydroxyurea for 40 min. During the final 20 min, the cells were pulsed with [3H]lysine to radiolabel the histones (all five histones are substantially labeled under these conditions). By two electrophoretic techniques, we demonstrate that histones H1, H2A, and H2B synthesized in the presence of hydroxyurea do not merely associate with the surface of the chromatin but instead exchange with preexisting histones so that for the latter two histones there is incorporation into nucleosome structures. On the other hand, H3 and H4 synthesized during this same time period appear to be only weakly bound, if at all, to chromatin. These two histones have been isolated from postnuclear washes and purified. Some possible implications of in vivo exchange are discussed.     

Fate of newly synthesized histones shortly after interruption of DNA replication

The synthesis and association of histones with chromatin were studied using MH-134SC cells in suspension culture. Cultures containing approximately equal numbers of cells were pulse-labeled with [3H]lysine at various times after the interruption of DNA synthesis with hydroxyurea. Each culture was mixed with a fixed volume of a culture generally labeled with [14C]lysine at the time of harvesting. Acid-soluble proteins extracted from different subcellular fractions of cells labeled under various conditions were compared by electrophoresis on polyacrylamide gels containing acetic acid and urea. All types of chromatin histones were labeled nearly equally as [14C]marker histones by a 15 min pulse under normal conditions, except that a considerable portion of pulse-labeled H4 was in highly acetylated forms. Addition of hydroxyurea at the start of the pulse markedly reduced the labeling of H3 and H4, but affected the labeling of the other histones only slightly. When DNA synthesis was inhibited before the start of the pulse, labeling of all histones decreased significantly. The addition of hydroxyurea was found to cause transient accumulation of newly synthesized proteins in the cytoplasmic soluble fraction; these were characterized as H3 and H4 from their metabolic properties and their electrophoretic mobilities on sodium dodecyl sulfate-polyacrylamide gels. The results suggest that association of newly synthesized H3 and H4 histones is closely coupled with ongoing DNA replication. The implications of the results for the mechanism of formation of new nucleosomes are discussed.     

Core histones H2B and H4 are mobilized during infection with herpes simplex virus 1

The infecting genomes of herpes simplex virus 1 (HSV-1) are assembled into unstable nucleosomes soon after nuclear entry. The source of the histones that bind to these genomes has yet to be addressed. However, infection inhibits histone synthesis. The histones that bind to HSV-1 genomes are therefore most likely those previously bound in cellular chromatin. In order for preexisting cellular histones to associate with HSV-1 genomes, however, they must first disassociate from cellular chromatin. Consistently, we have shown that linker histones are mobilized during HSV-1 infection. Chromatinization of HSV-1 genomes would also require the association of core histones. We therefore evaluated the mobility of the core histones H2B and H4 as measures of the mobilization of H2A-H2B dimers and the more stable H3-H4 core tetramer. H2B and H4 were mobilized during infection. Their mobilization increased the levels of H2B and H4 in the free pools and decreased the rate of H2B fast chromatin exchange. The histones in the free pools would then be available to bind to HSV-1 genomes. The mobilization of H2B occurred independently from HSV-1 protein expression or DNA replication although expression of HSV-1 immediate-early (IE) or early (E) proteins enhanced it. The mobilization of core histones H2B and H4 supports a model in which the histones that associate with HSV-1 genomes are those that were previously bound in cellular chromatin. Moreover, this mobilization is consistent with the assembly of H2A-H2B and H3-H4 dimers into unstable nucleosomes with HSV-1 genomes.     

Disruption of the nucleosomes at the replication fork.

The fate of parental nucleosomes during chromatin replication was studied in vitro using in vitro assembled chromatin containing the whole SV40 genome as well as salt-treated and native SV40 minichromosomes. In vitro assembled minichromosomes were able to replicate efficiently in vitro, when the DNA was preincubated with T-antigen, a cytosolic S100 extract and three deoxynucleoside triphosphates prior to chromatin assembly, indicating that the origin has to be free of nucleosomes for replication initiation. The chromatin structure of the newly synthesized daughter strands in replicating molecules was analysed by psoralen cross-linking of the DNA and by micrococcal nuclease digestion. A 5- and 10-fold excess of protein-free competitor DNA present during minichromosome replication traps the segregating histones. In opposition to published data this suggests that the parental histones remain only loosely or not attached to the DNA in the region of the replication fork. Replication in the putative absence of free histones shows that a subnucleosomal particle is randomly assembled on the daughter strands. The data are compatible with the formation of a H3/H4 tetramer complex under these conditions, supporting the notion that under physiological conditions nucleosome core assembly on the newly synthesized daughter strands occurs by the binding of H2A/H2B dimers to a H3/H4 tetramer complex.