Higher order structure in a short repeat length chromatin.

Polynucleosomes from calf brain cortical neurone nuclei have an average repeat length of less than 168 base pairs. The ability of this material to adopt higher order structure has been assessed by various physical techniques. Although containing on average less DNA per nucleosome than is required to form a chromatosome, this short repeat length chromatin folded in an H1 dependent manner to a structure with properties similar to those observed for longer repeat length chromatins such as that of chicken erythrocyte (McGhee, J.D., D.C. Rau, E. Charney, and G. Felsenfeld, 1980, Cell, 22:87-96). These observations are discussed in the context of H1 location in the higher order chromatin fiber.

Participation of core histone "tails" in the stabilization of the chromatin solenoid.

We show here that the solenoid is maintained by the combination of linker histones and the nonglobular, highly basic "tails" of the core histones, which play only a minor part in the formation of the nucleosome core (Whitlock and Simpson, 1977. J. Biol. Chem. 252:6,516--6,520; Lilley and Tatchell, 1977. Nucleic Acids Res. 4:2,039--2,055; and Whitlock and Stein, 1978. J. Biol. Chem. 253:3,857--3,861). Polynucleosomes that contain core histones devoid of tails remain substantially unfolded under conditions otherwise favorable for the formation of solenoids. The tails can be replaced by extraneous basic polypeptides and in the presence of the linker histones the solenoid structure is then spontaneously recovered, as judged by a wide variety of structural criteria. The inference is that the core histone tail segments function by providing electrostatic shielding of the DNA charge and at the same time bridging adjacent nucleosomes in the solenoid. Our results carry the further implication that posttranscriptional modifications, such as acetylation of epsilon-amino groups, that reduce the positive charge of the core histone tails will tend to destabilize the higher-order structure and could thus render the DNA with which they are associated more readily available for transcription.

Regulation of the higher-order structure of chromatin by histones H1 and H5.

Chicken erythrocyte chromatins containing a single species of linker histone, H1 or H5, have been prepared, using reassembly techniques developed previously. The reconstituted complexes possess the conformation of native chicken erythrocyte chromatin, as judged by chemical and structural criteria; saturation is reached when two molecules of linker histone are bound per nucleosome, as in native erythrocyte chromatin, which the resulting material resembles in its appearance in the electron microscope and quantitatively in its linear condensation factor relative to free DNA. The periodicity of micrococcal nuclease-sensitive sites in the linker regions associated with histone H1 or H5 is 10.4 base pairs, suggesting that the spatial organization of the linker region in the higher-order structure of chromatin is similar to that in isolated nucleosomes. The susceptible sites are cut at differing frequencies, as previously found for the nucleosome cores, leading to a characteristic distribution of intensities in the digests. The scission frequency of sites in the linker DNA depends additionally on the identity of the linker histone, suggesting that the higher-order structure is subject to secondary modulation by the associated histones.

Control of RNA polymerase binding to chromatin by variations in linker histone composition.

We have measured the frequency of initiation sites in chromatin for RNA polymerase in vitro as a function of the composition of linker histones (H1 and its analogues). In linker histone-depleted chromatin, RNA chain initiation appears to be restricted to the exposed linker DNA. On titration with purified linker histones, initiation is further restricted to an extent determined by the amount and type of linker histone, and the source of depleted chromatin. The extent of repression is correlated with the capacity of linker histones to induce the formation of higher-order structure in the complex. The results suggest that the effects of linker histones are mediated through the higher-order structure of chromatin, which prevents access of polymerase to the linker DNA. Accordingly, we find that structures imposed by the linker histones after polymerase binding are not inhibitory. Microscopy reveals that the higher-order structure in partially condensed chromatin is discontinuous, with solenoidal units spaced by sections of unravelled nucleosomes. Since salt stimulation of linker histone exchange does not result in derepression of linkers in our assay, we conclude that the distribution of higher-order units in chromatin is static and that the linker histones exchange between high-affinity sites in established units. We have previously shown that the globin gene is selectively unfolded in tissues that express the gene. The present results suggest that the transcriptional activity of specific genes is maintained by differential linker histone binding within chromatin.

Roles of H1 domains in determining higher order chromatin structure and H1 location.

Peptides derived from calf thymus H1 and rat liver H1, comprising only the globular and COOH-terminal domains of the intact molecule and therefore lacking NH2-terminal domains, have been shown by reconstitution to be as effective as the complete H1 molecule in inducing higher-order-chromatin structure. As the globular domain of H1 alone cannot induce chromatin folding, our results demonstrate that this function is primarily controlled by the COOH-terminal domain of the molecule. Surprisingly, these peptides do not locate correctly with respect to the nucleosome. This is demonstrated by their failure to confer upon reconstitutes the ability to protect DNA fragments of chromatosome length when digested with micrococcal nuclease. The precise placement of the H1 molecule (globular domain) with respect to the nucleosome is shown to be influenced by the "tail" domains of both H1 and the core histones.

Modulation of the relative trypsin sensitivities of the core histone 'tails'.

The order in which the core histone tails in chicken erythrocyte chromatin are attacked by trypsin has been reinvestigated. Results are presented to demonstrate that in the absence of linker histones H1 and H5 the relative order of core histone degradation by trypsin can be altered by changing the salt environment. In native chromatin, the presence of linker histones H1 and H5 inhibits this salt-dependent transition.

Transcriptional control, translation and function of the products of the five open reading frames of the Escherichia coli nir operon.

Five open reading frames designated nirB, nirD, nirE, nirC and cysG have been identified from the DNA sequence of the Escherichia coli nir operon. Complementation experiments established that the NirB, NirD and CysG polypeptides are essential and sufficient for NADH-dependent nitrite reductase activity (EC 1.6.6.4). A series of plasmids has been constructed in which each of the open reading frames has been fused in-phase with the beta-galactosidase gene, lacZ. Rates of beta-galactosidase synthesis during growth in different media revealed that nirB, -D, -E and -C are transcribed from the FNR-dependent promoter, p-nirB, located just upstream of the nirB gene: expression is co-ordinately repressed by oxygen and induced during anaerobic growth. Although the nirB, -D and -C open reading frames are translated into protein, no translation of nirE mRNA was detected. The cysG gene product is expressed from both p-nirB and a second, FNR-independent promoter, p-cysG, located within the nirC gene. No NADH-dependent nitrite reductase activity was detected in extracts from bacteria lacking either NirB or NirD, but a mixture of the two was as active as an extract from wild-type bacteria. Reconstitution of enzyme activity in vitro required stoichiometric quantities of NirB and NirD and was rapid and independent of the temperature during mixing. NirD remained associated with NirB during the initial stages of purification of the active enzyme, suggesting that NirD is a second structural subunit of the enzyme.

An examination of models for chromatin transcription.

Structural studies have revealed that chromatin is composed of repeating units or nucleosomes having two distinct domains, the nucleosome core and the linker region. The nucleosome core comprises 146 base pairs of DNA wound in one and three quarter turns around an octamer of histones made up of two symmetrical tetramers (1). It may be inferred on topological grounds that this structure must be perturbed during chromatin transcription and replication since the histone core bridges the supercoil which blocks the passage of polymerase along the template and prevents the unwinding of DNA required for enzymatic copying. A number of mechanisms for freeing the DNA template may be envisaged, and one detailed model, based on symmetrical dissociation of the histone tetramers, has been proposed (2). Here we present evidence against such unpairing or indeed any detachment of histones from the octamer during chromatin transcription, and we give reasons for favouring a transcriptional mechanism based upon the separation of the octamer from at least one of the DNA.

Nucleotide sequence, organisation and structural analysis of the products of genes in the nirB-cysG region of the Escherichia coli K-12 chromosome.

The DNA sequence and derived amino-acid sequence of a 5618-base region in the 74-min area of the Escherichia coli chromosome has been determined in order to locate the structural gene, nirB, for the NADH-dependent nitrite reductase and a gene, cysG, required for the synthesis of the sirohaem prosthetic group. Three additional open reading frames, nirD, nirE and nirC, were found between nirB and cysG. Potential binding sites on the NirB protein for NADH and FAD, as well as conserved central core and interface domains, were deduced by comparing the derived amino-acid sequence with those of database proteins. A directly repeated sequence, which includes the motif -Cys-Xaa-Xaa-Cys-, is suggested as the binding site for either one [4Fe-4S] or two [2Fe-2S] clusters. The nirD gene potentially encodes a soluble, cytoplasmic protein of unknown function. No significant similarities were found between the derived amino-acid sequence of NirD and either NirB or any other protein in the database. If the nirE open reading frame is translated, it would encode a 33-amino-acid peptide of unknown function which includes 8 phenylalanyl residues. The product of the nirC gene is a highly hydrophobic protein with regions of amino-acid sequence similar to cytochrome oxidase polypeptide 1.