Nucleosomes inhibit the initiation of transcription but allow chain elongation with the displacement of histones

Promoters were assembled in nucleosomes or ligated to nucleosomes and transcribed with SP6 RNA polymerase or with mammalian RNA polymerase II and accessory factors. Neither polymerase would initiate transcription at a promoter in a nucleosome, but once engaged in transcription, both polymerases were capable of reading through a nucleosome. In the course of readthrough transcription, the histones were displaced from the DNA, as shown by the exposure of restriction sites and by a shift of the template to the position of naked DNA in a gel. It may be true, in general, that processive enzymes will traverse regions of DNA organized in nucleosomes and displace histones.

Isolation of a Saccharomyces cerevisiae centromere DNA-binding protein, its human homolog, and its possible role as a transcription factor

A protein that binds specifically to Saccharomyces cerevisiae centromere DNA element I was purified on the basis of a nitrocellulose filter-binding assay. This protein, termed centromere-binding protein 1 (CP1), was heat stable and renaturable from sodium dodecyl sulfate (SDS), and assays of eluates from SDS gels indicated a molecular weight of 57,000 to 64,000. An activity with similar specificity and stability was detected in human lymphocyte extracts, and analysis in SDS gels revealed a molecular weight of 39,000 to 49,000. CP1-binding sites occurred not only at centromeres but also near many transcription units, for example, adjacent to binding sites for the GAL4-positive regulatory protein upstream of the GAL2 gene in S. cerevisiae and adjacent to the TATA element of the adenovirus major late promoter. A factor (termed USF) that binds to the latter site and stimulates transcription has been isolated from HeLa cells by others.

Two-dimensional crystals of cholera toxin B-subunit-receptor complexes: projected structure at 17-A resolution

The B subunit of cholera toxin forms two-dimensional crystals when bound to its membrane receptor, ganglioside GM1, in phospholipid layers. A rectangular crystal lattice gives diffraction extending to 15-A resolution in negative stain, and image-processing of electron micrographs reveals a ring of five protein densities. The diameter of the central hole and the outer diameter of the ring are about 20 and 60 A, respectively. These data are consistent with a pentameric, doughnut-shaped structure of the B subunit that lies flat on a membrane surface. A hexagonal crystal lattice is obtained as well, and results of image processing and chemical crosslinking allow two interpretations: the B subunit may exist in both pentameric and hexameric forms or, more likely, the hexagonal lattice may represent a disordered or liquid crystalline form, in which a pentamer undergoes rotational averaging about its 5-fold axis.

Synthetic peptides as nuclear localization signals

The nuclear envelope defines a compartment boundary which is penetrated by pores that mediate a remarkable transport process. Precursor RNAs are retained in the nucleus, while processed messenger RNA, transfer RNA and ribosomal subunits are transported to the cytoplasm. Proteins destined for the nucleus become localized soon after synthesis and again following mitosis, while cytoplasmic proteins are excluded. The process is highly specific: a single base change in vertebrate initiator tRNAMet (tRNAiMet) reduces the rate of export 20-fold; a point mutation within the simian virus 40 (SV40) large-T antigen, converting Lys 128 to Thr or Asn, prevents import. Lys 128 lies within a short 'signal' sequence which, when fused to large non-nuclear proteins, causes their accumulation in nuclei. Regions of other eukaryotic proteins also seem to contain nuclear localization signals, although a single consensus sequence has not emerged. We report here that a synthetic peptide containing 10 residues of large-T antigen sequence serves as a nuclear localization signal when cross-linked to bovine serum albumin (BSA) or immunoglobulin G (IgG) and microinjected in Xenopus oocytes. Substitution of Thr at the position of Lys 128 in this peptide renders it six- to sevenfold less effective. The uptake of peptide-linked BSA is saturable, and the rate is diminished by co-injection of free peptide. These findings are indicative of a receptor-mediated uptake process. With the use of anti-peptide antibodies, a family of proteins is revealed in nuclear but not cytoplasmic extracts of human lymphocytes which contain large-T antigen-like sequences.

A single gene from yeast for both nuclear and cytoplasmic polyadenylate-binding proteins: domain structure and expression

Nuclear and cytoplasmic poly(A)-binding proteins have been purified from Saccharomyces cerevisiae, and antisera have been used to isolate a gene that encodes them. The gene occurs in a single copy on chromosome 5 and gives rise to a unique, unspliced 2.1 kb transcript. The nuclear protein appears to be derived from the cytoplasmic one by proteolytic cleavage into 53 and 17 kd polypeptides that remain associated during isolation. DNA sequence determination reveals four tandemly arrayed 90 amino acid regions of homology that probably represent poly(A)-binding domains. A 55 residue A-rich region upstream of the initiator methionine codon in the mRNA shows an affinity for poly(A)-binding protein comparable to that of poly(A)180-220, raising the possibility of feedback regulation of translation.

A GAL family of upstream activating sequences in yeast: roles in both induction and repression of transcription

Binding sites for the GAL4-positive regulatory protein have been identified upstream of six galactose-inducible genes of Saccharomyces cerevisiae on the basis of (i) protection in DNAse I footprints, (ii) loss of protection when excess GAL4-binding oligonucleotide is added and (iii) homology with a 23-bp dyad-symmetric consensus sequence. Many of the binding sites have been shown to function as upstream activating sequences. The number of binding sites upstream of the various genes ranges from one to four, but a feature is conserved: in cases of multiple sites there is a pair with highest binding affinity located at dyad--dyad distances of 82--87 bp. We suggest that a pair of sites facilitates repression by the GAL80-negative regulatory protein, on the basis of (i) a correlation of a pair of sites (or only one) with full (or only partial) repression and (ii) the introduction of a second site abolishing transcription occurring with one.

A region flanking the GAL7 gene and a binding site for GAL4 protein as upstream activating sequences in yeast

A region of DNA 116 to 271 base-pairs upstream from the GAL7 gene of Saccharomyces cerevisiae activates transcription from a heterologous promoter and does so in either orientation, showing that the GAL7 upstream region contains an upstream activating sequence (UAS). The level of transcription obtained with two GAL7 UAS's in tandem was only 1.3 times that with one. Previous studies of the GAL1-GAL10 intergenic region were indicative of two binding sites for the GAL4 positive regulatory protein; we find that a single (synthetic) site is capable of gene activation. The level of transcription obtained with the intact GAL1-GAL10 UAS was five times that with the single site.

Nuclear polyadenylate-binding protein

Polyadenylate-binding activity can be detected in eluates from sodium dodecyl sulfate gels by a nitrocellulose filter-binding assay. Nuclear extracts from rat liver show a single peak of binding activity at 50 to 55 kilodaltons; cytoplasmic extracts show a single peak at 70 to 80 kilodaltons, corresponding to a 75-kilodalton protein previously described. Similar results are obtained with yeast and mouse fibroblasts, indicating a high degree of conservation of both nuclear and cytoplasmic polyadenylate-binding proteins. The activity from rat liver nuclei has been purified 125-fold on the basis of specific binding to polyadenylate and shows two main bands in sodium dodecyl sulfate gels at 53 and 55 kilodaltons.

Abundant nuclear ribonucleoprotein form of CAD RNA

Transcripts of the CAD gene in Syrian hamster cells are as abundant in the nucleus as in the cytoplasm. This was shown by in situ hybridization of whole cells and by solution and blot hybridization of subcellular fractions. Similar results were obtained both for wild-type cells and for a mutant containing amplified CAD genes in which the level of CAD RNA is 150-fold greater. CAD nuclear RNA is indistinguishable from mature mRNA by gel electrophoresis and blot hybridization. Discrete higher-molecular-weight precursors are undetectable, although the persistence of a short length of intervening sequence in the otherwise fully processed RNA is not excluded. CAD RNA is released from nuclei by sonication in physiological conditions in a ribonucleoprotein form that sediments as a broad peak at about 200S in a sucrose gradient. CAD sequences extracted from nuclei by treatment with EDTA and RNase are found in the 30S particles previously described.

Specific protein binding to far upstream activating sequences in polymerase II promoters

A binding activity specific for the upstream activating sequence of the GAL1-GAL10 promoter of Saccharomyces cerevisiae has been purified 220-fold on the basis of a nitrocellulose filter-binding assay. The binding activity is enriched in a nuclear preparation and is likely to be the GAL4 gene product. DNase I-protection mapping patterns reveal binding to two 30-base-pair regions at the boundaries of the sequence. A nearly identical mapping pattern is obtained with the coordinately regulated GAL7 promoter. The four 30-base-pair regions of binding in the two promoters are closely homologous, with a core consensus sequence of C-G-CG-TG-C-A-A-C-A-G-T-G-C-T-C-C-G-A-A- GC-G-A-T. A synthetic oligonucleotide with such a sequence competes with the upstream activating sequence in the binding reaction.

Genes and pseudogenes for mouse U1 and U2 small nuclear RNAs

Regions of the mouse genome homologous to U1 and U2 small nuclear RNAs have been isolated. The U1 regions are identified as pseudogenes on the basis of: 1) insertions, deletions, and substitutions in the sequence of the DNA as compared with that of U1 RNA; 2) absence of complementary RNA in mouse tissues; and 3) lack of expression upon microinjection in Xenopus oocyte nuclei. A U2 region is identified as a gene on the basis of near identity of sequence with U2 RNA and capacity to direct the synthesis of U2-like RNA in Xenopus oocyte nuclei. Both U1 and U2 sequences are repeated in the mouse genome, the U2 sequences analyzed here in an inverted repeat arrangement.

The protein responsible for the repeating structure of cytoplasmic poly(A)-ribonucleoprotein

A 75,000-dalton protein has been purified approximately 1,000-fold from rat liver, based on its capacity to organize poly(A) in a 27-residue repeating structure. This protein may be identified with the major polypeptide component of cytoplasmic poly(A)-ribonucleoprotein (RNP) previously described. The poly(A)-organizing activity of the protein is detected only in cytoplasmic fractions. Upon nuclease digestion of the 75,000-dalton protein-poly(A) complex, monomers, and higher multimers of RNP subunits can be resolved in a sucrose gradient. The sedimentation rate of the monomer is compatible with a composition of one 75,000-dalton protein molecule and one 27-residue segment of poly(A).

Two-dimensional crystallization technique for imaging macromolecules, with application to antigen--antibody--complement complexes

Two-dimensional crystals are formed from macromolecules bound on the surface of a lipid monolayer. A ligand linked to the lipid orientates the binding, and lateral diffusion of the lipids facilitates crystallization. The crystals are suitable for structural analysis by image processing of electron micrographs. An example is the formation of ordered arrays of antibodies on a monolayer of a lipid hapten, and subsequent decoration of these arrays with the first component of complement. Image processing indicates the arrangement of antibodies and the site of complement binding. This approach should be widely applicable to molecular complexes, such as those in replication, protein synthesis, hormone-receptor interaction and metabolic processes.

Repeating structure of cytoplasmic poly(A)-ribonucleoprotein

A repeating structure of cytoplasmic poly(A)-ribonucleoprotein is revealed by digestion with T2 RNase. A pattern of fragments that are multiples of about 27 residues is obtained. The repeating structure is readily reconstituted from purified poly(A) and cytoplasmic factors. Reconstitution is specific for poly(A), as shown by the lack of competition by poly(G), poly(C), poly(dA), and tRNA. The repeating structure is absent from the nucleus, and so appears to be formed upon transport to the cytoplasm.

Random location of nucleosomes on genes for 5 S rRNA

The location of nucleosomes on genes for 5 S rRNA in rat liver was determined by the preparation of nucleosome core DNA fragments, hybridization with 5 S rRNA, RNase digestion, and gel electrophoresis. The resulting size distribution of RNA fragments was essentially the same as that found when the experiment was carried out with random DNA fragments.

Periodicity of deoxyribonuclease I digestion of chromatin

Two methods have been used to measure the single-strand lengths of the DNA fragments produced by deoxyribonuclease I digestion of chromatin. The average lengths obtained are muliples of about 10.4 bases, significantly different from the value of 10 previously reported. This periodicity in fragment lengths is closely related to the periodicity of the DNA double helix in chromatin, but the two values need not be exactly the same.

Relation of nucleosomes to nucleotide sequences in the rat

The relation of nucleosomes to nucleotide sequences is random for most single copy sequences in rat liver. This could be due to variation in the DNA content of nucleosomes, and a procedure for detecting such variation is described.