Molecular cloning and characterization of cellular genes whose expression is repressed by the adenovirus E1a gene products and growth factors in quiescent rat cells

Induction of Sp1 in differentiating human embryonal carcinoma cells triggers transcription of the fibronectin gene

Cells of the human embryonal carcinoma line NEC14 proliferate as densely packed clusters consisting of small, polygonal stem cells and do not express a detectable level of fibronectin (FN). Upon induction of differentiation by treatment with N,N'-hexamethylene bisacetamide (HMBA), the level of FN mRNA increased steeply within 24 h and FN began to be accumulated, along with the organization of actin filaments in the cells. The FN promoter elements required for the activation were analyzed in reference to a cluster of GC boxes by using the chloramphenicol acetyltransferase (CAT) gene fused to 5' sequential-deletion derivatives of the promoter and promoters carrying base substitutions in the GC boxes. Among four GC boxes, GC boxes 2 and 3 had the greatest effect on promoter activation, and base substitutions in these GC boxes resulted in 80% reduction in promoter activity. The pattern of DNA-protein complex formation with these GC boxes changed drastically after induction of differentiation. The extract prepared from undifferentiated NEC14 cells formed fast-migrating complexes (UnD complexes), while the extract prepared from NEC14 cells treated with HMBA for 24 h formed slow-migrating complexes containing Sp1. Both complexes were formed predominantly with GC box 2. Base substitutions within the GC boxes completely abolished the formation of both UnD and Sp1 complexes. Consistent with these changes, the Sp1 level increased steeply within 24 h. Induction of Sp1 expression in NEC14 cells effectively stimulated the promoter activity of the transfected FN promoter-CAT constructs. These results indicate that activation of the FN promoter in differentiating NEC14 cells occurs by the steep induction of Sp1, which prevents an undifferentiated cell factor from binding to the Sp1 sites.

Identification and characterization of novel genes modulated in the thyroid of dogs treated with methimazole and propylthiouracil

Induction of cell proliferation by mitogen or growth factor stimulation leads to the specific stimulation or repression of a large number of genes. To better understand differentiated epithelial cell growth regulation, we have initiated a study to identify genes which are regulated by the thyrotropin-dependent mitogenic pathway in dog thyroid cells. A thyroid cDNA library was prepared from a methimazole and propylthiouracil-treated dog and differentially screened with probes derived from control or stimulated thyroids. Among 19 clones isolated, 6 encode known proteins (inwardly rectifying potassium channel, nucleosome assembly protein, ribosomal protein L7, elongation factor 1alpha, non-muscle myosin light chain, and heat shock protein 90beta). The 13 others correspond to proteins whose function is unknown. Among them, 5 correspond to mRNAs whose expression was modulated by mitogenic stimulation of thyrocytes in primary culture. A preliminary characterization of two of these cDNAs is reported: clone 5, which might represent a novel, atypical protein kinase, and clone 3, which contains ankyrin-like repeats, suggesting that it might interact with other proteins.

Modulation of transcription of the rat fibronectin gene by cell density

The fibronectin (FN) gene is under complex regulatory control in vitro and in vivo. Sequences from the rat FN gene directed efficient expression of a lacZ reporter gene product, beta-galactosidase, in NIH/3T3 mouse fibroblasts. Stable transfectants were generated to facilitate studies of gene regulation by cell growth state. The expression of FN-lacZ constructs increased approximately twofold when cultures attained confluence, relative to total protein. The magnitude of this increase correlates well with that observed for FN mRNA levels and protein synthesis rate. Fragments containing 4.9, 0.9, or 0.3 kbp upstream of the transcription start site are equally responsive to cell density and/or cell contact. Deletion of a cAMP-responsive element enhanced the response, suggesting a negative role for this sequence motif and demonstrating that the FN gene is regulated by cell density at the transcriptional level. The effect of high cell density is apparently different from decreased growth rate, as incubation with low serum did not result in increased expression of the lacZ reporter. Finally, conditioned medium from dense cells did not enhance reporter gene expression in sparse cells, suggesting that the density signal is not transmitted via a soluble factor.

G10BP, an E1A-inducible negative regulator of Sp1, represses transcription of the rat fibronectin gene

Downregulation of the fibronectin (FN) gene in a rat 3Y1 derivative cell line, XhoC, transformed by the adenovirus E1A and E1B genes seems to be caused by the induction of a negative regulator, G10BP, which binds to three G-rich sequences in the promoter (T. Nakamura, T. Nakajima, S. Tsunoda, S. Nakada, K. Oda, H. Tsurui, and A. Wada, J. Virol. 66:6436-6450, 1992). These are the G10 stretch and two GC boxes consisting of the G10 stretch with one internal C residue insertion. The recognition sequences of G10BP and Sp1 (GGGCGG) overlap in these GC boxes. To analyze the mechanism of the downregulation, G10BP was purified by DNA affinity chromatography, and its molecular mass was estimated to be about 30 kDa. The promoter was modified by substituting the sequence GGGG with ATCC or CTTA in these G-rich sequences, leaving the Sp1 motif intact, and by replacing the Sp1 motif by the T stretch. Transcription of FN promoter-chloramphenicol acetyltransferase fusion genes carrying the base substitution in one or more of these G-rich sequences both in vivo and in vitro revealed that the base substitution in any G-rich sequence results in reduction of promoter activity, although the downstream GC box (GCd) plays a primary role. The addition of G10BP severely inhibited the activities of the FN promoters carrying the wild-type GCd in vitro, while the promoters carrying the mutant GCd were unaffected. The binding affinity of G10BP and Sp1 to each of the G-rich sequences, analyzed by gel shift assays, indicated that G10BP binds strongly to the GCd, moderately to the G10 stretch, and weakly to GCu, while Sp1 binds strongly to GCu, moderately to GCd, and weakly to the G10 stretch. Sp1 binding to GCd and the G10 stretch was inhibited by G10BP, while binding to GCu was unaffected. These results indicate that FN gene transcription is inhibited in XhoC cells primarily by exclusion of Sp1 binding to GCd by G10BP and that G10BP is a new class of Sp1 negative regulator.

Induction of E1A-responsive negative factors for transcription of the fibronectin gene in adenovirus E1-transformed rat cells

The level of fibronectin (FN) gene expression is very high in resting rat 3Y1 cells but greatly decreased in adenovirus E1-transformed cells. To study the mechanism of this down-regulation, nuclear factors binding to the 5'-flanking region of the FN gene were analyzed by gel retardation assay and DNase I footprinting. Nuclear factors that were present in the transformed cells but nearly absent in resting 3Y1 cells interacted with multiple sites of the promoter region. Oligonucleotide competition with the FN promoter-chloramphenicol acetyltransferase (CAT) reporter constructs (pFCAT) for these factors in the transformed cells indicated that all of them had a negative effect on FN gene expression. Of them, a factor(s) (G10BP) binding to the G10 stretch from positions -239 to -230 and to two GC boxes consisting of the G10 stretch with one internal C residue insertion from positions -105 to -95 and -54 to -44 had the strongest repressive activity. Introduction of substitutive mutations into these G-rich sequences resulted in the increase in CAT activity of pFCAT in the transformed cells. The recognition sequences of G10BP and Sp1 overlap in two GC boxes. G10BP has stronger affinity for heparin and GC boxes than does Sp1, suggesting that G10BP may repress FN gene transcription by displacing Sp1.

Fibronectin gene expression in proliferating, quiescent, and SV40-infected mouse kidney cells

To study alterations in cellular gene expression in mouse kidney cell cultures infected with simian virus 40 (SV40) or polyomavirus, we performed a differential screening of a mouse kidney cDNA library with probes prepared from mRNAs of virus-infected and mock-infected cells. We isolated and characterized cDNA recombinant pKT13 which detected increased mRNA levels in infected cells. Sequence analysis of pKT13 revealed close to 100% homology with the 3'-end of mouse fibronectin (FN) mRNA. Since primary cultures of baby mouse kidney cells have been extensively characterized in our laboratories, we studied FN gene expression at different stages of uninfected and virus-infected cultures. High levels of FN and of its mRNA were found in the kidneys of suckling mice, while in primary cultures of proliferating epithelial kidney cells the expression of FN was very low until the cultures became confluent. Thereafter FN increased and reached high levels in cells which were irreversibly arrested in phase Go and which had apparently exhausted their finite division potential. Infection of confluent cultures with polyomavirus or SV40 resulted in a further stimulation of FN gene expression. However, during abortive infection with SV40, FN mRNA and FN levels decreased with emergence of transformed cells and were low in an established SV40-transformed mouse kidney cell line. These changes in FN gene expression suggest that high levels of FN might be indicative in vivo for terminal differentiation and in vitro for cellular senescence.

Regulation and expression of a growth arrest-specific gene (gas5) during growth, differentiation, and development

The growth arrest-specific gas5 gene was isolated from mouse genomic DNA and structurally characterized. The transcriptional unit is divided into 12 exons that span around 7 kb. An alternative splicing mechanism gives rise to two mature mRNAs which contain either 11 or 12 exons, and both are found in the cytoplasm of growth-arrested cells. In vivo, the gas5 gene is ubiquitously expressed in mouse tissues during development and adult life. In Friend leukemia and NIH 3T3 cells, the levels of gas5 gene mRNA were high in saturation density-arrested cells and almost undetectable in actively growing cells. Run-on experiments indicated that the gas5 gene is transcribed at the same level in both growing and arrested cells. On the other hand, in dimethyl sulfoxide-induced differentiating cells a sharp decrease in the rate of transcription was observed shortly before the cells reached the postmitotic stage. These results indicate that in density-arrested cells accumulation of gas5 mRNA is controlled at the posttranscriptional level while in differentiating cells expression is regulated transcriptionally.

Regulation and expression of a growth arrest-specific gene (gas5) during growth, differentiation, and development

The growth arrest-specific gas5 gene was isolated from mouse genomic DNA and structurally characterized. The transcriptional unit is divided into 12 exons that span around 7 kb. An alternative splicing mechanism gives rise to two mature mRNAs which contain either 11 or 12 exons, and both are found in the cytoplasm of growth-arrested cells. In vivo, the gas5 gene is ubiquitously expressed in mouse tissues during development and adult life. In Friend leukemia and NIH 3T3 cells, the levels of gas5 gene mRNA were high in saturation density-arrested cells and almost undetectable in actively growing cells. Run-on experiments indicated that the gas5 gene is transcribed at the same level in both growing and arrested cells. On the other hand, in dimethyl sulfoxide-induced differentiating cells a sharp decrease in the rate of transcription was observed shortly before the cells reached the postmitotic stage. These results indicate that in density-arrested cells accumulation of gas5 mRNA is controlled at the posttranscriptional level while in differentiating cells expression is regulated transcriptionally.

E1A-responsive elements for repression of rat fibronectin gene transcription

The level of fibronectin (FN) gene transcription in resting rat 3Y1 cells is very high but decreases steeply after growth stimulation by serum or by the induction of E1A expression. To study the mechanism of this E1A-mediated down-regulation, the 5' flanking regions of the FN gene with various deletions and substitutions were fused to the Escherichia coli chloramphenicol acetyltransferase (CAT) gene and introduced into resting 3Y1 cells with E1A expression plasmids. The results indicate that the G10 stretch located from nucleotide position -239 to -230 and two GC boxes from position -105 to -95 and position -54 to -44 are the primary E1A-responsive elements for repression of the FN gene. Two GC boxes also contain a G10 stretch that is interrupted by the presence of an internal C residue. These sequences overlap with the Sp1 motif GGGCGG. Substitution of the sequence GGGG with ATCC or CTTA in these G-rich sequences, leaving the Sp1 motif intact, completely abolished the E1A sensitivity of the promoter. Analysis of the E1A domains by using various E1A deletion mutants indicated that the domain for binding to the retinoblastoma susceptibility gene product (RB) is essential for efficient repression. These results suggest that the gene encoding a negative factor(s) binding to the three G-rich sequences in the FN promoter is repressed by RB in resting 3Y1 cells and derepressed by expression of E1A.

E1A-responsive elements for repression of rat fibronectin gene transcription

The level of fibronectin (FN) gene transcription in resting rat 3Y1 cells is very high but decreases steeply after growth stimulation by serum or by the induction of E1A expression. To study the mechanism of this E1A-mediated down-regulation, the 5' flanking regions of the FN gene with various deletions and substitutions were fused to the Escherichia coli chloramphenicol acetyltransferase (CAT) gene and introduced into resting 3Y1 cells with E1A expression plasmids. The results indicate that the G10 stretch located from nucleotide position -239 to -230 and two GC boxes from position -105 to -95 and position -54 to -44 are the primary E1A-responsive elements for repression of the FN gene. Two GC boxes also contain a G10 stretch that is interrupted by the presence of an internal C residue. These sequences overlap with the Sp1 motif GGGCGG. Substitution of the sequence GGGG with ATCC or CTTA in these G-rich sequences, leaving the Sp1 motif intact, completely abolished the E1A sensitivity of the promoter. Analysis of the E1A domains by using various E1A deletion mutants indicated that the domain for binding to the retinoblastoma susceptibility gene product (RB) is essential for efficient repression. These results suggest that the gene encoding a negative factor(s) binding to the three G-rich sequences in the FN promoter is repressed by RB in resting 3Y1 cells and derepressed by expression of E1A.

Subtractive cDNA cloning using oligo(dT)30-latex and PCR: isolation of cDNA clones specific to undifferentiated human embryonal carcinoma cells

The human embryonal carcinoma cell line NEC14 can be induced to differentiate by the addition of 10(-2)M N,N'-hexamethylene-bis-acetamide (HMBA). A subtractive cDNA library specific to undifferentiated NEC14 cells was constructed using oligo(dT)30-Latex and polymerase chain reaction (PCR). The method was designed to improve the efficiency of subtraction and the enrichment of cDNA clones corresponding to low abundance mRNAs. The single strand of cDNA was made from mRNA prepared from the HMBA-treated NEC14 cells using an oligo(dT)30 primer covalently linked to Latex particles. After removal of the mRNA template by heat-denaturation and centrifugation, the subtractive hybridization was carried out between the cDNA-oligo(dT)30-Latex and mRNA from untreated NEC14 cells. Unhybridized mRNA collected by centrifugation was hybridized repeatedly to the cDNA-oligo(dT)30-Latex and subtractive mRNA was converted to cDNA. The subtractive cDNA was then amplified by PCR and cloned into pBluescript II KS-. The cDNA library thus constructed consisted of approximately 10,000 independent clones with cDNA inserts of 1.7 Kb on average. Differential hybridization of these transformants indicated that approximately 3% of them contained cDNA inserts specific to the undifferentiated EC cells, some of which were derived from low abundance mRNAs.