SWI/SNF complex is essential for NRSF-mediated suppression of neuronal genes in human nonsmall cell lung carcinoma cell lines.

Mammalian chromatin remodeling factor, SWI/SNF complex contains a single molecule of either Brm or BRG1 as the ATPase catalytic subunit. Here, we show that the SWI/SNF complex forms a larger complex with neuron-restrictive silencer factor (NRSF) and its corepressors, mSin3A and CoREST, in human nonsmall cell lung carcinoma cell lines. We also demonstrate that the strong transcriptional suppression of such neuron-specific genes as synaptophysin and SCG10 by NRSF in these non-neural cells requires the functional SWI/SNF complex; these neuronal genes were elevated in cell lines deficient in both Brm and BRG1, whereas retrovirus vectors expressing siRNAs targeting integral components of SWI/SNF complex (Brm/BRG1 or Ini1) induced expression of these neuronal genes in SWI/SNF-competent cell lines. In cell lines deficient in both Brm and BRG1, exogenous Brm or BRG1 suppressed expression of these neuronal genes in an ATP-dependent manner and induced efficient and specific deacetylation of histone H4 around the NRSF binding site present in the synaptophysin gene by a large complex containing the recruited functional SWI/SNF complex. Patients with Brm/BRG1-deficient lung carcinoma have been reported to carry poor prognosis; derepression of NRSF-regulated genes including these neuron-specific genes could contribute to enhance tumorigenicity and also would provide selective markers for Brm/BRG1-deficient tumors.

Lack of induction of neuroretinal cell proliferation by Rous sarcoma virus variants that carry the c-src gene.

Expression of p60v-src of Rous sarcoma virus in cultured chicken embryo neuroretinal cells was previously shown to result in the transformation and sustained proliferation of normally quiescent cell populations. We show here that Rous sarcoma virus variants that encode p60c-src, the cellular homolog of p60v-src, lack the ability to induce morphological transformation and cell proliferation of cultured neuroretinal cells. Neuroretinal cells infected with c-src-containing viruses, however, possess no less p60 protein kinase activity assayed in the immune complex than those infected with the transformation-defective Rous sarcoma virus mutants PA101 or PA104, which do stimulate the growth of these cells.

Low level of cellular protein phosphorylation by nontransforming overproduced p60c-src.

We have previously found that Rous sarcoma virus variants in which the viral src (v-src) gene is replaced by the cellular src (c-src) gene have no transforming activity. In this study, we analyzed the basis for the inability of the p60c-src overproduced by these variants to transform cells. Phosphorylations of tyrosine residues in total cell protein or in cellular 34K protein are known to be markedly enhanced upon infection with wild-type Rous sarcoma virus. We found that these tyrosine phosphorylations were only slightly increased in the c-src-containing virus-infected cells, whereas both levels were significantly increased by infection with wild-type Rous sarcoma virus, or transforming mutant viruses which are derived from c-src-containing viruses by spontaneous mutation. Phosphorylation at tyrosine 416 of p60 itself was also extremely low in overproduced p60c-src and high in p60s of transforming mutant viruses. In immunoprecipitates with monoclonal antibody, the overproduced p60c-src had much lower casein tyrosine kinase activity than did p60v-src. We previously showed that p60 myristylation and plasma membrane localization may be required for cell transformation. p60c-src was similar to transforming p60s in these properties. These results strongly suggest that the low level of tyrosine phosphorylation by overproduced p60c-src accounts for its inability to transform cells.

Amino acid substitutions sufficient to convert the nontransforming p60c-src protein to a transforming protein.

We have previously shown that Rous sarcoma virus variants that carry the cellular homolog (c-src) of the viral src gene (v-src) do not transform chicken embryo fibroblasts. We also have shown that replacement of sequences upstream or downstream from the BglI site of the cellular src gene with the corresponding regions of v-src restored transforming activity to the hybrid genes. Since there are only six amino acid changes between p60c-src and p60v-src within the sequences upstream from BglI, we constructed chimeric molecules involving v-src and c-src to determine the effect of each amino acid substitution on the biological activities of the gene product. We found that the change from Thr to Ile at position 338 or the replacement of a fragment of c-src containing Gly-63, Arg-95, and Thr-96 with a corresponding fragment of v-src containing Asp-63, Trp-95, and Ile-96 converted p60c-src into a transforming protein by the criteria of focus formation, anchorage-independent growth, and tumor formation in newborn chickens. These mutations also resulted in elevation of the protein kinase activity of p60c-src.

Suppression of cellular transformation by glucocorticoid receptor mutants.

Glucocorticoid receptor (GR) has been shown to suppress activator protein 1 (AP-1)-mediated transcription by several molecular mechanisms. We previously showed that activation of endogenous AP-1 is essential for cellular transformation induced by oncogenes, such as v-src and c-Ha-ras. In the present study, we have analyzed whether high levels of GR expression suppress cellular transformation caused by these oncogenes. To eliminate the ligand effects that induce the transcriptional stimulation via glucocorticoid response elements, we constructed two GR mutants: CD-GR-1, lacking the COOH-terminal portion, including both the ligand and Hsp90-binding domains, and tau1TDCD-GR-1, a derivative of CD-GR-1 lacking the tau1 transactivation domain. When these GR mutants were expressed in chicken embryonic fibroblasts by retroviral vectors, they translocated into the nucleus without addition of glucocorticoid to the culture medium, and they suppressed cellular transformation caused by v-src and c-Ha-ras, as well as by c-fos and c-jun. Cellular transformation by v-myc was not suppressed by these mutants. Such suppressive effects of these GR mutants have a very similar oncogene dependency to that of dominant-negative mutants of AP-1. This suggests that GR can be altered to suppress cellular transformation by inhibiting endogenous AP-1 activity without activating glucocorticoid response elements.

Inhibition of jun transformation by a mutated fos gene: design of an anti-oncogene.

The protein products of the fos and jun oncogenes (Fos and Jun) function as transcriptional regulators in the form of homo- or heterodimeric complexes that bind to DNA. Dimerization is mediated by a leucine zipper structure that serves to juxtapose alpha-helical regions of each protein, rich in basic amino acids, that form a bipartite DNA-binding domain. Although Fos participates exclusively in heterodimeric complexes, Jun can function either as a homodimer that has a low apparent affinity for DNA or as a more stable heterodimer with Fos that has a higher apparent affinity for DNA. We have used these properties of Fos and Jun to design a mutated fos gene, lacking a functional DNA-binding domain (supfos1), that suppresses the transforming activity of jun in trans. Here we show that chicken embryo fibroblasts transformed by jun revert to a normal phenotype after infection by a retroviral vector encoding supFos1. Furthermore, infection of normal cells with the supfos1 vector renders them resistant to subsequent transformation by jun. Inhibition of jun transformation was associated with the appearance of supFos1-Jun heterodimers and a reduction in the AP-1 DNA-binding activity contributed by Jun homodimers. These findings demonstrate that the function of leucine zipper-containing transcription factors can be investigated by the procedure of intracellular immunization.

Escape from redox regulation enhances the transforming activity of Fos.

Fos and Jun form dimeric complexes that bind to DNA sequences containing activator protein 1 (AP-1) sites and regulate gene expression. The in vitro DNA-binding activity of these proteins is sensitive to reduction-oxidation (redox). Reduction of a single conserved cysteine residue, located in the DNA-binding domain, either by reducing agents or by a nuclear redox factor (Ref-1), is required for AP-1 DNA-binding activity. Replacing the critical cysteine with serine results in a protein that can bind to DNA in vitro even under oxidizing conditions. To determine whether redox control affects the function of Fos in vivo, we have constructed, and compared the properties of, retroviral vectors expressing either a truncated Fos protein (F118-211) or a truncated Fos protein in which the critical cysteine was replaced by serine (FC154S). In infected chicken embryo fibroblasts (CEFs), both vectors expressed similar levels of Fos protein, which formed heterodimers with Jun at equivalent efficiencies. However, extracts from cells expressing FC154S exhibited a threefold increase in AP-1 DNA-binding activity compared with cells expressing F118-211. Furthermore, this enhanced binding activity was resistant to treatment with the oxidizing agent diamide. Infection of CEFs by virus expressing FC154S resulted in increased numbers of transformed colonies and an increase in colony size compared with those obtained following infection by virus expressing Fos 118-211. These results suggest that redox regulation may limit the total level of functional Fos-Jun complexes in vivo and that escape from this control enhances transforming activity.

Differential expression of fos and jun family members in the developing chicken gastrointestinal tract.

We have analysed the expression patterns of all the known fos/jun family genes, which encode the components of the transcription factor AP-1, in the chicken embryonic digestive tract that develops into the esophagus, proventriculus, gizzard, small intestine, ceca and large intestine. From soon after formation of the tubular structure, each gene transcript was localized in distinct domains of the epithelium and mesenchyme in all of these major gastrointestinal organs, independently of the anterior-posterior axis. fra-2 was expressed predominantly in epithelium, which also expressed junD, while low-level expression of junD was also detected in smooth muscle cell precursors in mesenchyme. Expression of c-jun and c-fos was detectable in both mesenchyme and epithelium through the whole tract. In the differentiated proventriculus, the developed glandular epithelium expressed c-jun and junD, but not fra-2, while luminal epithelium expressed fra-2 and junD, but not c-jun. These results suggest that distinct Fos/Jun protein heterodimers play important roles in maintaining the epithelial-mesenchymal interactions. Similar expression patterns to those of fra-2 and junD were established from earlier stages by Sonic hedgehog gene and the Indian hedgehog gene, respectively, both of which are important in forming the inductive network between epithelium and mesenchyme of the digestive tract.

fra-2 promoter can respond to serum-stimulation through AP-1 complexes.

fra-2 (fos-related antigen-2) expression is detected at a basal level even in growth-arrested chicken embryo fibroblasts (CEF), but upon serum-stimulation high levels of its transcripts are transiently observed. This induction is delayed and prolonged compared to that of c-fos. Transient expression experiments in CEF using a series of constructs of chicken fra-2 promoter region linked to the CAT reporter gene indicated previously that serum response element (SRE) is not required for full serum inducibility. In this report, we show that constructs in which the CRE-like sequence and both AP-1 binding sites are disrupted lack serum inducibility, suggesting that either of these enhancers is important in serum induction of fra-2. In growth-arrested CEF, small amounts of Fra-2/c-Jun complex bind to the AP-1 consensus sequences in fra-2 promoter, while a significant part of the enhanced AP-1 binding activity after 60-120 min of serum stimulation is attributable to c-Fos/c-Jun heterodimer. At later times Fra-2/c-Jun again becomes the main complex. Transient expression assays in F9 cells indicated that c-Fos/c-Jun heterodimers have strong stimulatory effects on fra-2 promoter activity, while Fra-2/c-Jun complex has lower transcriptional activity than that of c-Jun homodimer. These results suggest that c-Fos (induced at earlier times) and c-Jun proteins are at least partly responsible for serum-induced expression of fra-2.

Cloning and characterisation of the mouse fra-2 gene.

Transcription factor AP-1 is comprised of multiple protein complexes that include members of a family of genes related to the proto-oncogene c-fos. In this report, we have extended the analysis of one member of this family, fos-related antigen-2 (fra-2), by isolating and characterising genomic and cDNA clones encoding the mouse fra-2 homolog. The overall gene structure (number and positions of introns) was similar to that of both the chicken fra-2 gene and other members of the fos family, and the relative positions of putative enhancers in the 5' regulatory region were well conserved between the mouse and chicken fra-2 genes. High levels of fra-2 mRNA were detected in ovary, stomach, small and large intestine, brain, lung and heart. The mouse Fra-2 protein showed 94% and 87.5% conservation with human and chicken Fra-2, respectively, and mouse Fra-2, like the chicken homolog, induced transformation of chicken embryo fibroblasts. The characterisation of the mouse fra-2 gene provides a basis for analysis of Fra-2 function in the whole animal.

Phosphorylation and high level expression of Fra-2 in v-src transformed cells: a pathway of activation of endogenous AP-1.

Chicken embryo fibroblasts (CEF) transformed with v-src were previously reported to revert to normal phenotype after the introduction of dominant-negative mutants of Fos or Jun, indicating that endogenous AP-1 activity is essential for the cellular transformation. The major changes in the expression levels of fos and jun family genes induced by v-src were the elevation of fra-2 and c-jun transcripts. We show here that extensive phosphorylation of the AP-1 component Fra-2 is a major qualitative change in v-src transformed CEF and that several Ser and Thr residues in a C-terminal region of Fra-2 (amino acids 266-323) are phosphorylated specifically. The induced kinase activity was detected at the position of 42 kDa by in gel kinase assay using the Fra-2 C-terminal region as a substrate, and it was identified as chicken ERK2. JNK1 and JNK2, other members of the MAP kinase family, were not significantly activated in v-src transformed CEF and Fra-2 was not a good substrate for JNKs. fra-2 promoter analysis indicated that this promoter activity is elevated in v-src transformed CEF via two AP-1 binding sites and CRE-like sequence. We propose that phosphorylation of Fra-2 by ERK2 converts it from an inefficient transcriptional activator to an active one and further that fra-2 expression is autoregulated in response to the phosphorylation status of its gene product.

Kinetics of v-src-induced epithelial-mesenchymal transition in developing glandular stomach.

The oncogene function in primary epithelial cells is largely unclear. Recombination organ cultures in combination with the stable and transient gene transfer techniques by retrovirus and electroporation, respectively, enable us to transfer oncogenes specifically into primary epithelial cells of the developing avian glandular stomach (proventriculus). In this system, the epithelium and mesenchyme are mutually dependent on each other for their growth and differentiation. We report here that either stable or transient expression of v-src in the epithelium causes budding and migration of epithelial cells into mesenchyme. In response to the transient expression of v-Src or a constitutive active mutant of MEK, we observed immediate downregulation of the Sonic hedgehog gene and subsequent elimination of E-cadherine expression in migrating cells, suggesting the involvement of MAP kinase signaling pathway in these processes. v-src-expressing cells that were retained in the epithelium underwent apoptosis (anoikis) and detached from the culture. Continuous expression of v-src by, for example, Rous sarcoma virus (RSV) was required for the epithelial cells to acquire the ability to express type I collagen and fibronectin genes (mesenchymal markers), and finally to establish the epithelial-mesenchymal transition. These observations would partly explain why RSV does not apparently cause carcinoma formation, but induces sarcomas exclusively.

The chromosome structure and the cell cycle of Caulobacter crescentus. Isolation and analysis of envelope-free nucleoids.

Envelope-free nucleoids were isolated from an asymmetrically dividing bacterium, Caulobacter crescentus. In the nucleoid fraction, most of the DNA and nascent RNA in the cell and about 2% of the total cellular proteins were recovered. The sedimentation coefficient of the nucleoid was constant (1260 S) during the G1 period of the swarmer cell cycle and increased to 1940 S during the S period. Since both replicating (S period) and non-replicating (G1 period) chromosomes had a similar superhelical concentration, the increase in the sedimentation coefficient was simply explained by duplication of the nucleoid structure. The duplicated nucleoid was shown to segregate prior to the cell division. The pulse-labeled proteins recovered in the nucleoid fraction contained several stage-specific species, most of which were detected at the beginning of S period.

High-level expression of exogenous genes by replication-competent retrovirus vectors with an internal ribosomal entry site.

We report the construction of two types of Rous sarcoma virus (RSV)-based replication-competent avian retrovirus vectors, IR1 and IR2 to express an exogenous gene at a very high level. In these vectors, the internal ribosomal entry site (IRES) derived from encephalomyocarditis virus (EMCV) was inserted between the env gene and an exogenous gene. The IR1 vector retains the splicing acceptor site that is present in the downstream of the env gene while the IR2 vector lacks it. Using a v-fos mutant (v-fos-CD3) as an example of exogenous genes, we show here that both IR1 and IR2 vectors expressed the gene product, CD3, at expression levels 5- and 8-fold higher than that of their parental vector without IRES, respectively. These vectors were moderately stable and kept a high-level expression of CD3 for at least three passages through the cells. Analysis of viral transcripts indicate that exogenous genes carried by both IR vectors were translated exclusively from the IRES that is present in all the species of the viral transcripts. High-level expression of exogenous genes was also observed in the case of the Hoxa-13 gene in the IR1 vector or the fra-2 gene in the IR2 vector, indicating that the extremely high-level expression characteristic of these vectors is applicable to several exogenous genes.

Virion-associated RNA polymerase of bacteriophage phi 6 synthesizes three complete transcripts of double-stranded RNA genome in vitro.

Three single-stranded RNA transcripts synthesized in vitro by a virion-associated RNA polymerase of bacteriophage phi 6 were sequenced at their 5'- and 3'-termini. The sequences agreed with those of the + strands of the 3 double-stranded RNA segments [FEBS Lett. (1982) 141,111-115]. The results show that the transcription by phi 6 RNA polymerase initiates exactly at the 3'-ends of the template RNAs (-strands of the genomic RNA) and terminates exactly at the 5'-ends.

Biochemical and functional analysis of highly phosphorylated forms of c-Jun protein.

We report here that, upon UV irradiation or growth stimulation, endogenous c-Jun (40 kDa) in chicken embryo fibroblasts (CEF) is converted into several forms with apparently higher molecular weights in SDS-polyacrylamide gel electrophoresis (45, 44, 42 kDa). Two of the bands (44 and 45 kDa) were transient after growth stimulation, but were much more persistent after UV irradiation. In both cases, the drastic mobility shifts were accompanied with the activation of endogenous JNK activity but not of MAPK activity, and the bands were shown to represent different phosphorylation states of c-Jun rather than ubiquitinated c-Jun. Biochemical analysis indicated that phosphorylation at Ser63 and Ser73 was not sufficient to produce these drastic mobility shifts, which additionally required phosphorylation at Thr91 and Thr93. Substitution of both Ser63 and Ser73 with either Ala or Asp had no significant effect on the transforming activity of c-Jun, but the mutants failed to show drastic mobility shifts even after UV irradiation. These results indicate that Ser63 and Ser73 are essential for the drastic mobility shifts and further suggest that the highly phosphorylated forms of c-Jun are not directly involved in cellular transformation.

Semi-conservative transcription of double-stranded RNA catalyzed by bacteriophage phi 6 RNA polymerase.

Treatment of Pseudomonas phaseolicola double-stranded RNA bacteriophage phi 6 with sodium deoxycholate converted the virions to nucleocapsids, which had in vitro RNA polymerase activity. The incorporation of [3H]UMP continued for at least 7 h. The initial incorporation was detected as intermediate RNA. Radioactivity was chased first into three segments of double-stranded RNA, and then into small, medium, and large species of single-stranded RNA successively via the intermediate RNA. Several copies of single-stranded RNA at least were synthesized from a template. The RNA synthesis clearly took place by a semi-conservative mechanism with respect to templates. That is, 5-bromo UTP was incorporated into one strand of double-stranded RNA to make a hybrid RNA of brominated and unbrominated strands. Furthermore, one strand of the 3H-labeled parental double-stranded RNA was shown to be released as single-stranded RNA.

T-cell subsets in drug-induced toxic epidermal necrolysis. Possible pathogenic mechanism induced by CD8-positive T cells.

A patient with bromisovalum-induced toxic epidermal necrolysis showed pronounced delayed hypersensitivity to bromisovalum by patch testing. Biopsy specimens from the cutaneous lesion and the site of the positive patch test reaction were analyzed and compared immunohistologically. The findings were similar: most of the mononuclear cells disposed along the dermoepidermal junction and migrating into the epidermis were CD8-positive lymphocytes, whereas the dermal inflammatory infiltrates were composed predominantly of CD4-positive lymphocytes. This case showed the potential usefulness of patch testing in evaluating cases of toxic epidermal necrolysis. We believe that delayed hypersensitivity plays a crucial role in the development of drug-induced toxic epidermal necrolysis. Furthermore, potential effector cells with phenotypic characteristics of CD8-positive lymphocytes (suppressor/cytotoxic T cells) seem to represent important mediators of the epidermal damage of the cutaneous lesion in our case.

Selecting informative genes with parallel genetic algorithms in tissue classification.

Recent advances in biotechnology offer the ability to measure the levels of expression of thousands of genes in parallel. Analysis of such data can provide understanding and insight into gene function and regulatory mechanisms. Several machine learning approaches have been used to aid to understand the functions of genes. However, these tasks are made more difficult due to the noisy nature of array data and the overwhelming number of gene features. In this paper, we use the parallel genetic algorithm to filter out the informative genes relative to classification. By combing with the classification method proposed by Golub et al. and Slonim et al., we classify the data sets with tissues of different classes, and the preliminary results are presented in this paper.

Learning polynomial feedforward neural networks by genetic programming and backpropagation.

This paper presents an approach to learning polynomial feedforward neural networks (PFNNs). The approach suggests, first, finding the polynomial network structure by means of a population-based search technique relying on the genetic programming paradigm, and second, further adjustment of the best discovered network weights by an especially derived backpropagation algorithm for higher order networks with polynomial activation functions. These two stages of the PFNN learning process enable us to identify networks with good training as well as generalization performance. Empirical results show that this approach finds PFNN which outperform considerably some previous constructive polynomial network algorithms on processing benchmark time series.