Contribution of nucleosome binding preferences and co-occurring DNA sequences to transcription factor binding.

BACKGROUND: Chromatin plays a critical role in regulating transcription factors (TFs) binding to their canonical transcription factor binding sites (TFBS). Recent studies in vertebrates show that many TFs preferentially bind to genomic regions that are well bound by nucleosomes in vitro. Co-occurring secondary motifs sometimes correlated with functional TFBS. RESULTS: We used a logistic regression to evaluate how well the propensity for nucleosome binding and co-occurrence of a secondary motif identify which canonical motifs are bound in vivo. We used ChIP-seq data for three transcription factors binding to their canonical motifs: c-Jun binding the AP-1 motif (TGA(C)/(G)TCA), GR (glucocorticoid receptor) binding the GR motif (G-ACA---(T)/(C)GT-C), and Hoxa2 (homeobox a2) binding the Pbx (Pre-B-cell leukemia homeobox) motif (TGATTGAT). For all canonical TFBS in the mouse genome, we calculated intrinsic nucleosome occupancy scores (INOS) for its surrounding 150-bps DNA and examined the relationship with in vivo TF binding. In mouse mammary 3134 cells, c-Jun and GR proteins preferentially bound regions calculated to be well-bound by nucleosomes in vitro with the canonical AP-1 and GR motifs themselves contributing to the high INOS. Functional GR motifs are enriched for AP-1 motifs if they are within a nucleosome-sized 150-bps region. GR and Hoxa2 also bind motifs with low INOS, perhaps indicating a different mechanism of action. CONCLUSION: Our analysis quantified the contribution of INOS and co-occurring sequence to the identification of functional canonical motifs in the genome. This analysis revealed an inherent competition between some TFs and nucleosomes for binding canonical TFBS. GR and c-Jun cooperate if they are within 150-bps. Binding of Hoxa2 and a fraction of GR to motifs with low INOS values suggesting they are not in competition with nucleosomes and may function using different mechanisms.

Overlapping ETS and CRE Motifs ((G/C)CGGAAGTGACGTCA) preferentially bound by GABPα and CREB proteins.

Previously, we identified 8-bps long DNA sequences (8-mers) that localize in human proximal promoters and grouped them into known transcription factor binding sites (TFBS). We now examine split 8-mers consisting of two 4-mers separated by 1-bp to 30-bps (X(4)-N(1-30)-X(4)) to identify pairs of TFBS that localize in proximal promoters at a precise distance. These include two overlapping TFBS: the ETS⇔ETS motif ((C/G)CCGGAAGCGGAA) and the ETS⇔CRE motif ((C/G)CGGAAGTGACGTCAC). The nucleotides in bold are part of both TFBS. Molecular modeling shows that the ETS⇔CRE motif can be bound simultaneously by both the ETS and the B-ZIP domains without protein-protein clashes. The electrophoretic mobility shift assay (EMSA) shows that the ETS protein GABPα and the B-ZIP protein CREB preferentially bind to the ETS⇔CRE motif only when the two TFBS overlap precisely. In contrast, the ETS domain of ETV5 and CREB interfere with each other for binding the ETS⇔CRE. The 11-mer (CGGAAGTGACG), the conserved part of the ETS⇔CRE motif, occurs 226 times in the human genome and 83% are in known regulatory regions. In vivo GABPα and CREB ChIP-seq peaks identified the ETS⇔CRE as the most enriched motif occurring in promoters of genes involved in mRNA processing, cellular catabolic processes, and stress response, suggesting that a specific class of genes is regulated by this composite motif.

Assessment of CASP6 predictions for new and nearly new fold targets.

This is a report of the assessment of the predictions made for the CASP6 protein structure prediction experiment conducted in 2004 in the New Fold (NF) category. There were nine protein domains that were judged to have new folds (NF) and 16 for which a similar structure was known but the sequence similarity was judged to be too low for them to be easily recognized (FR/A). We selected all NF targets and eight of the 16 FR/A targets judged to be at the borderline between NF and FR/A for evaluation in the NF category. A total of 165 prediction groups submitted over 7400 structural models for these targets. The quality of these models was evaluated using the GDT_TS scores of the structural similarity detection program LGA and by visual inspection of the top-scoring models. The best models submitted bore an overall similarity to the target structure for three or four of the nine NF targets and for all but one of the FR/A targets. High-scoring models for the NF targets were submitted by several different groups. When both the NF and FR/A targets were considered, Baker group dominated by submitting best models for seven of the 17 targets, but 14 other groups also managed to submit best models for one or more targets.

PATE, a gene expressed in prostate cancer, normal prostate, and testis, identified by a functional genomic approach.

To identify target antigens for prostate cancer therapy, we have combined computer-based screening of the human expressed sequence tag database and experimental expression analysis to identify genes that are expressed in normal prostate and prostate cancer but not in essential human tissues. Using this approach, we identified a gene that is expressed specifically in prostate cancer, normal prostate, and testis. The gene has a 1.5-kb transcript that encodes a protein of 14 kDa. We named this gene PATE (expressed in prostate and testis). In situ hybridization shows that PATE mRNA is expressed in the epithelial cells of prostate cancers and in normal prostate. Transfection of the PATE cDNA with a Myc epitope tag into NIH 3T3 cells and subsequent cell fractionation analysis shows that the PATE protein is localized in the membrane fraction of the cell. Analysis of the amino acid sequence of PATE shows that it has structural similarities to a group of proteins known as three-finger toxins, which includes the extracellular domain of the type beta transforming growth factor receptor. Restricted expression of PATE makes it a potential candidate for the immunotherapy of prostate cancer.

Clp ATPases and their role in protein unfolding and degradation.

Although much has been learned about the structure and function of Clp chaperones and their role in proteolysis, the mechanism of protein unfolding catalyzed by Clp ATPases and the mechanism of translocation of the unfolded proteins from Clp ATPases to partner proteases remain unsolved puzzles. However, models in which mechanical force is used to destabilize the structure of the substrate in a processive and directional manner are probable. It also seems likely that when ClpA ATPases are associated with proteases, unfolding is coupled to extrusion of the unfolded protein into the proteolytic cavity. In summary, it is anticipated that the large family of Clp ATPases will accomplish their many important cellular functions by similar mechanisms and what has been learned by studying the prokaryotic members reviewed here will shed a great deal of light on all members of the family.

Probing contacts between the ribonuclease H domain of HIV-1 reverse transcriptase and nucleic acid by site-specific photocross-linking.

Cys(38) and Cys(280) of p66/p51 human immunodeficiency virus type 1 reverse transcriptase (HIV-1 RT) can be converted to Ser without affecting enzyme function. We have exploited this feature to construct and purify "monocysteine" RT derivatives for site-specific modification with the photoactivable cross-linking agent, p-azidophenacyl bromide. Acylation of a unique cysteine residue introduced at the extreme C terminus of the p66 subunit (C(561)) with an azidophenacyl group allowed us to probe contacts between residues C-terminal to alpha-helix E' of the RNase H domain and structurally divergent nucleic acid duplexes. In a binary complex of RT and template-primer, we demonstrate efficient cross-linking to primer nucleotides -21 to -24/-25, and template nucleotides -18 to -21. Cross-linking specificity was confirmed by an analogous evaluation following limited primer extension, where the profile is displaced by the register of DNA synthesis. Finally, contact with a DNA primer hybridized to an isogenic RNA or DNA template indicates subtle alterations in cross-linking specificity, suggesting differences in nucleic acid geometry between duplex DNA and RNA/DNA hybrids at the RNase H domain. These data exemplify how site-specific acylation of HIV-1 RT can be used to provide high resolution structural data to complement crystallographic studies.

Transgenic mice demonstrate AP-1 (activator protein-1) transactivation is required for tumor promotion.

Activator protein-1 (AP-1) is a transcription factor that consists of either a Jun-Jun homodimer or a Jun-Fos heterodimer. Transactivation of AP-1 is required for tumor promoter-induced transformation in mouse epidermal JB6 cells and for progression in mouse and human keratinocytes. Until now, the question of whether AP-1 transactivation is required for carcinogenesis in vivo has remained unanswered, as has the issue of functionally significant target genes. To address these issues we have generated a transgenic mouse in which transactivation mutant c-jun (TAM67), under the control of the human keratin-14 promoter, is expressed specifically in the basal cells of the epidermis where tumor induction is initiated. The keratin-14-TAM67 transgene was expressed in the epidermis, tongue, and cervix, with no apparent abnormalities in any tissue or organ. TAM67 expression blocked 12-O-tetradecanoylphorbol 13-acetate (TPA, phorbol 12-tetradecanoate 13-acetate) induction of the AP-1-regulated luciferase in AP-1 luciferase/TAM67 mice, but did not inhibit induction of candidate AP-1 target genes, collagenase-1 or stromelysin-3. More interestingly, TAM67 expression did not inhibit TPA-induced hyperproliferation. In two-stage skin carcinogenesis experiments, the transgenic animals showed a dramatic inhibition of papilloma induction. We conclude that transactivation of a subset of AP-1-dependent genes is required for tumor promotion and may be targeted for cancer prevention.

Signaling requirements for oncogenic forms of the Met tyrosine kinase receptor.

The Met tyrosine kinase receptor has been implicated in human cancer. Here we have examined the signaling requirements of three oncogenic forms of this molecule: wild type Met in response to ligand/autocrine stimulation, Met which has been mutationally activated, and Tpr-Met (a constitutively active truncated Met fusion protein). Previous studies have demonstrated the importance of a Grb2 binding site, and of specific tyrosine residues (i.e. Y8,9 and Y14,15) for Met function, and we have now explored the relevance of these and other sites for oncogenic Met signaling. Following substitution of various intracellular tyrosines for phenylalanine, we find that the transforming activity of each Met oncogene is dependent upon tyrosines Y8,9 and Y14,15, in addition to two novel tyrosines (Y6 and Y10) not previously implicated in Met signaling. Tyrosines Y6 and Y10 influence a variety of Met-mediated responses both in vitro (transformation, mitogenicity and invasion), and in vivo (tumorigenicity and metastasis). We also show that Tpr-Met is much more dependent on its Grb2 binding site for biological activity than are the other oncogenic forms of the Met receptor. Thus, although the three Met oncogenes examined are similar in their dependency on a number of specific tyrosines for activity, the signaling strategy employed by Tpr-Met can be differentiated from that of the other two.

Mutagenic analysis of functional domains of the mos proto-oncogene and identification of the sites important for MAPK activation and DNA binding.

We constructed in-frame deletion/replacement mutations in the Xenopus mos proto-oncogene that lie within conserved Mos-specific codons, but outside of the regions that are conserved among the src kinase family of genes. All gene products were assayed in vitro for kinase activity and in vivo for their ability to induce oocyte maturation, embryonic cleavage arrest and cellular transformation. Most mutations in Mos eliminated both kinase and biological activity. However, a mutation in Mos that removed two basic amino acid residues (R94 and K97) downstream from the lysine at the ATP binding site (K90) markedly enhanced autophosphorylation activity. Moreover, this mutant displayed markedly reduced biological activity, lacked transforming activity, and failed to activate mitogen activated protein kinase (MAPK). A second mutant Mos product, lacking amino acids R45-A54, displayed a five-fold increase in cellular transforming activity. This Mos mutant specifically localized to the cytoplasm; in contrast to wild-type (wt) Mos that localized to both the nucleus and the cytoplasm. These data indicate that Mos transforming activity is mediated via signalling exerted in the cytoplasm, presumably through MAPK, and that nuclear localization of the oncogene product interferes with transforming activity. We also show that amino acids R45-A54 are important for Mos DNA binding activity.

Intramolecular interactions in pancreatic ribonucleases.

A detailed analysis of the composition and properties of hydrophobic nuclei and microclusters in pancreatic ribonuclease A (RNase A) has been carried out. Distance calculations for all noncovalently bonded atoms revealed that the average number of nonpolar contacts between a side chain of an amino acid and its neighbors is substantially larger if it involves hydrophobic residues rather than nonhydrophobic ones. However, the difference decreased when the number of contacts per nonpolar group and/or atom were calculated. Three main nuclei and five microclusters were identified, and their quantitative parameters were calculated. These nuclei include hydrophobic residues with a substantial number of nonpolar contacts with the environment (Phe 8, Phe 120, Phe 46, Tyr 25, Tyr 97, Ile 107, Leu 35, Ile 81, Val 54, Val 108, Met 29, Met 30). Hydrophobic nuclei of RNase A differ in shape and in composition, in the number of intranuclear contacts and of associated residues, as well as in their internal mobility. All eight cysteine residues are involved in nonpolar interactions with amino acid residues of hydrophobic nuclei. Active site amino acid residues of RNase A form a noncovalent contact network comprised of themselves, as well as of many conserved residues from hydrophobic nuclei. Sequence alignment with some other members of the RNase A family of proteins shows remarkable similarity in positions and in conservation of the main nonpolar residues, comprising cores of two (out of three) hydrophobic nuclei. A correlation was shown to exist between the average density of contacts for side-chain atoms and the number of amino acids to be found in the appropriate positions in the sequences of related mammalian ribonucleases. However, there are certain amino acid positions in the third, smaller nucleus, which are highly variable within the family. Taking into account that this nucleus is composed of residues belonging to different elements of the secondary structure, it is likely that the mutual orientation of these elements can be somehow different for these proteins.

Structure of complex of synthetic HIV-1 protease with a substrate-based inhibitor at 2.3 A resolution.

The structure of a complex between a peptide inhibitor with the sequence N-acetyl-Thr-Ile-Nle-psi[CH2-NH]-Nle-Gln-Arg.amide (Nle, norleucine) with chemically synthesized HIV-1 (human immunodeficiency virus 1) protease was determined at 2.3 A resolution (R factor of 0.176). Despite the symmetric nature of the unliganded enzyme, the asymmetric inhibitor lies in a single orientation and makes extensive interactions at the interface between the two subunits of the homodimeric protein. Compared with the unliganded enzyme, the protein molecule underwent substantial changes, particularly in an extended region corresponding to the "flaps" (residues 35 to 57 in each chain), where backbone movements as large as 7 A are observed.

Conserved folding in retroviral proteases: crystal structure of a synthetic HIV-1 protease.

The rational design of drugs that can inhibit the action of viral proteases depends on obtaining accurate structures of these enzymes. The crystal structure of chemically synthesized HIV-1 protease has been determined at 2.8 angstrom resolution (R factor of 0.184) with the use of a model based on the Rous sarcoma virus protease structure. In this enzymatically active protein, the cysteines were replaced by alpha-amino-n-butyric acid, a nongenetically coded amino acid. This structure, in which all 99 amino acids were located, differs in several important details from that reported previously by others. The interface between the identical subunits forming the active protease dimer is composed of four well-ordered beta strands from both the amino and carboxyl termini and residues 86 to 94 have a helical conformation. The observed arrangement of the dimer interface suggests possible designs for dimerization inhibitors.

A semiempirical theory of the optical activity of saccharides.

A semiempirical theory of optical activity of saccharides is developed that (a) goes beyond previous empirical treatments, (b) yields calculated NaD rotations which correlate well with experimental data, and (c) accounts for a characteristic c.d. band observed for polysaccharides in the vacuum u.v. spectrum.

Conformation of (1----6)-beta-D-glucan by vacuum-ultraviolet circular dichroism.

Gelation of pustulan [(1----6)-beta-D-glucan] likely involves some chain ordering. To characterize the most probable conformation, we previously carried out energy calculations and identified the local energy minimums. Here, we analyze the (previously measured) vacuum-ultraviolet circular dichroism of pustulan by using a quadrant rule recently proposed for the 175-nm c.d. band and our earlier quantum-mechanical model for the 185-nm c.d. band. Most of the local-energy minimum conformations are shown to be inconsistent with the circular dichroism observed. Of the two distinct conformations consistent with the observed circular dichroism, one includes the global-energy minimum calculated earlier, i.e., (phi, psi, omega) = (20 degrees, -90 degrees, gt).

Theoretical pi-pi absorption and circular dichroic spectra of beta-turn model peptides.

The dipole interaction model, treated by the partially dispersive normal mode method, is used to calculate pi-pi absorption and circular dichroic spectra of beta-turn model peptides in certain conformations. These include Ac-Gly-Gly-NHMe, Ac-L-Ala-L-Ala-NHMe, and Ac-L-Ala-Gly-NHMe in the standard beta-turn conformations I, II, and III of Venkatachalam and cyclo(L-Ala-Gly-epsilon-aminocaproyl), cyclo(L-Ala-L-Ala-epsilon-aminocaproyl), and cyclo(L-Ala-D-Ala-epsilon-aminocaproyl) in the minimum-energy conformations of Nemethy et al. Boltzmann average circular dichroic spectra of the cyclic compounds agree with experimental spectra in most respects. The results are compared with previous theoretical CD spectra for these molecules and with conformational assignments based on other evidence. Absorption spectra in the pi-pi band are predicted to be moderately sensitive to conformation.

Theoretical pi-pi* absorption and circular dichroic spectra of cyclic dipeptides.

The dipole interaction model, treated by the partially dispersive normal mode method, is used to calculate circular dichroic spectra of cyclo(Gly-Gly), cyclo (Ala-Gly), cyclo(Ala-Ala), cyclo(Pro-Gly), cyclo(Pro-Ala), cyclo(Pro-Val), cyclo (Pro-D-Val), and cyclo(Pro-Pro) in the amide pi-pi* absorption band near 190 nm. Assuming a standard backbone geometry, spectra which are in fair to good agreement with experiment are obtained for these molecules. The spectra are predicted to be sensitive to conformations of Pro and Val side chains. The effects of dipeptide ring folding on calculated CD spectra are mostly consistent with those found by other workers, except that it is found that a planar ring conformation of cyclo (Ala-Ala) and cyclo (Ala-Gly) gives predicted spectra comparable to experiment. The same model gives theoretical absorption spectra consistent with available experimental data.

Conformational studies of dextran [alpha-(1----6)-D-glucan].

A theoretical conformational study of dextran, a (1----6)-linked alpha-D-glucan polysaccharide, has been made to allow n explicit comparison with earlier results on pustulan, the corresponding (1----6)-linked beta-D-glucan. The nonbonded, torsional and hydrogen bond contributions to potential energy were calculated as a function of rotational angles phi, psi, and omega. The (phi, psi, omega)-space of the disaccharide and of helices contain many local energy minima with very small energy differences. A comparison of (1----6)-alpha-D-glucans with (1----6)-beta-D-glucans indicates significantly differences in conformational behavior. Specifically, our results shed light on the fact that dextran does not gel, whereas pustulan does. The difference in tendency to gel may be related to the fact that dextran has no particularly favored conformations with structural regularity whereas pustulan does.

Theoretical study of the conformations of pustulan [(1----6)-beta-D-glucan].

The total potential energy including nonbonded, torsional and hydrogen bond contributions has been computed for pustulan, a (1----6) linked beta-D-glucan polysaccharide, as a function of rotational angles phi, psi, and omega. The (phi,psi,omega)-space contains many local minima and at least three distinct deep minima. Two minima at (phi,psi,omega) = (25 degrees, 190 degrees,gg) and (phi,psi,omega) = 65 degrees, 150 degrees,gg) of almost equal energies have helical parameters (n = 5.2, h = 1.0A) and (n = 3.2, h = 1.5A), respectively. A third minimum at (phi,psi,omega) = (40 degrees, 70 degrees gt) leads to an extended zig-zag structure (n = 2.2, h = 2.2A). Energy maps obtained for gentiobiose, the disaccharide of pustulan, also reveal many local minima and the small energy differences among them indicate that gentiobiose is extremely flexible. Gentiodextrins, a family of cyclic molecules of (1----6)-beta-D-glucose residues, were also studied. Conformations free from steric hindrance were found for cyclic molecules with three to six glucose residues.