Chromatin structure and protein binding in the putative regulatory region of the c-myc gene in Burkitt lymphoma

Modular sequence elements associated with origin regions in eukaryotic chromosomal DNA

We have postulated that chromosomal replication origin regions in eukaryotes have in common clusters of certain modular sequence elements (Benbow, Zhao, and Larson, BioEssays 14, 661-670, 1992). In this study, computer analyses of DNA sequences from six origin regions showed that each contained one or more potential initiation regions consisting of a putative DUE (DNA unwinding element) aligned with clusters of SAR (scaffold associated region), and ARS (autonomously replicating sequence) consensus sequences, and pyrimidine tracts. The replication origins analyzed were from the following loci: Tetrahymena thermophila macronuclear rDNA gene, Chinese hamster ovary dihydrofolate reductase amplicon, human c-myc proto-oncogene, chicken histone H5 gene, Drosophila melanogaster chorion gene cluster on the third chromosome, and Chinese hamster ovary rhodopsin gene. The locations of putative initiation regions identified by the computer analyses were compared with published data obtained using diverse methods to map initiation sites. For at least four loci, the potential initiation regions identified by sequence analysis aligned with previously mapped initiation events. A consensus DNA sequence, WAWTTDDWWWDHWGWHMAWTT, was found within the potential initiation regions in every case. An additional 35 kb of combined flanking sequences from the six loci were also analyzed, but no additional copies of this consensus sequence were found.

DNA amplifications and elevated expression of proto-oncogene in addition to altered DNA ploidy in metastatic brain tumors

The histopathological characteristics, proto-oncogene amplification, immunohistopathology of the c-erbB-2 product distribution, and the DNA content of nuclei were examined in metastatic brain tumors, which consisted of seven adenocarcinomas, a large cell carcinoma, a squamous cell carcinoma, a renal cell carcinoma and a mucoepidermoid carcinoma. A very high incidence of DNA changes was seen in these tumors. Proto-oncogene amplification and abnormal DNA content in the nuclear portion were found in 64% (7/11) and 67% (6/9) of cases, respectively. We also found double oncogene alteration in three cases metastasizing from lung, esophagus and kidney, and triple oncogene alteration in one case metastasizing from breast. We could not identify the common alterations in the group of metastatic brain tumor cells. These data suggest that the proto-oncogene amplifications and the alteration of DNA ploidy pattern may contribute to the metastatic process.

A sequence-specific, single-strand binding protein activates the far upstream element of c-myc and defines a new DNA-binding motif

The far upstream element (FUSE) of the human c-myc proto-oncogene stimulates expression in undifferentiated cells. A FUSE-binding protein (FBP) is present in undifferentiated but not differentiated cells. Peptide sequences from the purified protein allowed cloning of cDNAs encoding FBP. Expression of FBP mRNA declined upon differentiation, suggesting transcriptional regulation of FBP. Features in the FBP cDNA suggest that FBP is also regulated by RNA processing, translation, and post-translational mechanisms. Both cellular and recombinant FBP form sequence-specific complexes with a single strand of FUSE. Transfection of FBP into human leukemia cells stimulated c-myc-promoter-driven expression from a reporter plasmid in a FUSE-dependent manner. Deletion and insertion mutagenesis of FBP defined a novel single-strand DNA-binding domain. Analysis of the primary and predicted secondary structure of the amino acid sequence reveals four copies of a reiterated unit comprised of a 30-residue direct repeat and an amphipathic alpha-helix separated by an 18- to 21-residue spacer. The third and fourth copies of this repeat-helix unit constitute the minimum single-stranded DNA-binding domain. To determine whether the FUSE site, in vivo, possesses single-strand conformation, and therefore could be bound by FBP, cells were treated with potassium permanganate (KMnO4) to modify unpaired bases. Modification of genomic DNA in vivo revealed hyperreactivity associated with single-stranded DNA in the FUSE sequence and protection on the strand that binds FBP in vitro. The role of single-stranded DNA and single-strand binding proteins in c-myc regulation is discussed.

Intragenomic heterogeneity of DNA damage formation and repair: a review of cellular responses to covalent drug DNA interaction

Chemical DNA interaction and its processing can now be studied at the level of specific genomic regions. Such investigations have revealed important new information about the molecular biology of the cellular responses to genomic insult and especially of the repair processes. They also have demonstrated that both the formation and repair of DNA damage display patterns of intragenomic heterogeneity. Therefore, mechanistic studies should involve examination of DNA damage formation and repair in specific genomic sequences besides in the overall genome to provide clues to the way in which specific modifications of DNA or chromatin could have specific biological effects. This review primarily focuses on studies done to elucidate the nature of DNA damage induction and intragenomic processing provoked by covalent drug-DNA modification in mammalian cells. The involvement of DNA damage formation and cellular processing as critical factors for genomic injury is exemplified by studies of the novel alkylating morpholinyl anthracyclines and the bifunctional alkylating agent nitrogen mustard as a prototype agent for covalent drug DNA interaction.

Multiple changes in chromatin structure precede the transcriptional activation of the human growth hormone locus in placental cells

In addition to the growth hormone gene (hGH-N) itself, the human growth hormone (hGH) locus contains four related genes, namely hGH-V and hCS-L, -A and -B, which have appeared very recently in evolution and are specifically expressed in placenta. With the aim of identifying the regulatory elements responsible for this placental-specific expression, we have mapped the DNaseI hypersensitive sites present at the hGH gene cluster in a placental cell line (BeWo) that expresses the hGH-V and hCS genes. Our results reveal a complex pattern of hypersensitive sites distributed along the hGH locus, most of which appear to be cell type-specific. Thus, we have identified placental-specific hypersensitive sites within the first intron of the hGH-N and hGH-V genes, but not in the equivalent regions of the hCS genes. In addition, we have found several placental-specific hypersensitive sites downstream of the hCS-L and hCS-A genes, which might reflect the presence of enhancer elements similar to that located downstream of the hCS-B gene (Walker et al. (1990) J. Biol. Chem. 265, 12940). Comparison of BeWo cells with a placental cell line (JEG-3) which does not express the hGH-V and hCS genes revealed a very similar pattern of hypersensitive sites, suggesting that the sites detected are established before the onset of transcription. Our results indicate that the transition to an active hGH locus in placental cells requires multiple alterations in chromatin structure, and provide a framework for the molecular analysis of the regulatory elements and mechanisms mediating such processes.

Repeated CT elements bound by zinc finger proteins control the absolute and relative activities of the two principal human c-myc promoters

Transcription of the human proto-oncogene c-myc is governed by two tandem principal promoters, termed P1 and P2. In general, the downstream promoter, P2, is predominant, which is in contrast to the promoter occlusion phenomenon usually observed in genes containing tandem promoters. A shift in human c-myc promoter usage has been observed in some tumor cells and in certain physiological conditions. However, the mechanisms that regulate promoter usage are not well understood. The present studies identify regulators which are required to promote transcription from both human c-myc promoters, P1 and P2, and have a role in determining their relative activities in vivo. A novel regulatory region located 101 bp upstream of P1 was characterized and contains five tandem repeats of the consensus sequence CCCTCCCC (CT element). The integrity of the region containing all five elements is required to promote transcription from P1 and for maximal activity from P2 in vivo. A single copy of this same element, designated CT-I2, also appears in an inverted orientation 53 bp upstream of the P2 transcription start site. This element has an inhibitory effect on P1 transcription and is required for P2 transcription. The transcription factor Sp1 was identified as the factor that binds specifically to the tandem CT elements upstream of P1 and to the CT-I2 element upstream of P2. In addition, the recently cloned zinc finger protein ZF87, or MAZ, was also able to bind these same elements in vitro. The five tandem CT elements can be functionally replaced by a heterologous enhancer that only in the absence of CT-I2 reverses the promoter usage, similar to what is observed in the translocated c-myc allele of Burkitt's lymphoma cells.

Repeated CT elements bound by zinc finger proteins control the absolute and relative activities of the two principal human c-myc promoters

Transcription of the human proto-oncogene c-myc is governed by two tandem principal promoters, termed P1 and P2. In general, the downstream promoter, P2, is predominant, which is in contrast to the promoter occlusion phenomenon usually observed in genes containing tandem promoters. A shift in human c-myc promoter usage has been observed in some tumor cells and in certain physiological conditions. However, the mechanisms that regulate promoter usage are not well understood. The present studies identify regulators which are required to promote transcription from both human c-myc promoters, P1 and P2, and have a role in determining their relative activities in vivo. A novel regulatory region located 101 bp upstream of P1 was characterized and contains five tandem repeats of the consensus sequence CCCTCCCC (CT element). The integrity of the region containing all five elements is required to promote transcription from P1 and for maximal activity from P2 in vivo. A single copy of this same element, designated CT-I2, also appears in an inverted orientation 53 bp upstream of the P2 transcription start site. This element has an inhibitory effect on P1 transcription and is required for P2 transcription. The transcription factor Sp1 was identified as the factor that binds specifically to the tandem CT elements upstream of P1 and to the CT-I2 element upstream of P2. In addition, the recently cloned zinc finger protein ZF87, or MAZ, was also able to bind these same elements in vitro. The five tandem CT elements can be functionally replaced by a heterologous enhancer that only in the absence of CT-I2 reverses the promoter usage, similar to what is observed in the translocated c-myc allele of Burkitt's lymphoma cells.

Overexpression of Myc suppresses CCAAT transcription factor/nuclear factor 1-dependent promoters in vivo

Overexpression of Myc in cells can suppress the transcription of specific genes. Because several of these genes have common transcriptional regulatory elements, we investigated the possibility that this effect of Myc is mediated through a specific transcription factor. In vitro DNA-binding assays detect only one form of CCAAT transcription factor/nuclear factor 1 (CTF/NF-1) in quiescent 3T3-L1 cells. By contrast, quiescent 3T3-L1 cells that stably overexpress either c-Myc or N-Myc contain at least three forms of CTF/NF-1. Biochemical characterization of the various CTF/NF-1 forms showed that they have the same native molecular weight but differ in charge density. The more negatively charged CTF/NF-1 forms present in Myc-overexpressing cells are converted into that found in normal cells by treatment with acid phosphatase, suggesting that they represent a more phosphorylated form of the CTF/NF-1 protein. The various CTF/NF-1 forms have a similar DNA-binding affinity. Transfection experiments demonstrated that transcription from CTF/NF-1-dependent promoters is specifically suppressed in cells that stably overexpress c-Myc. This effect requires CTF/NF-1 binding. CTF/NF-1-dependent promoter activity is also suppressed in 3T3-L1 cells during active growth (relative to the quiescent state). Interestingly, actively growing 3T3-L1 cells contain forms of CTF/NF-1 similar to those in quiescent cells that stably overexpress c-Myc. Thus, the CTF/NF-1 forms present in cells that express high amounts of c-Myc correlate with a lower transcription rate of CTF/NF-1-dependent promoters in vivo. Our results provide a basis for the suppression of specific gene transcription by c-Myc.

Overexpression of Myc suppresses CCAAT transcription factor/nuclear factor 1-dependent promoters in vivo

Overexpression of Myc in cells can suppress the transcription of specific genes. Because several of these genes have common transcriptional regulatory elements, we investigated the possibility that this effect of Myc is mediated through a specific transcription factor. In vitro DNA-binding assays detect only one form of CCAAT transcription factor/nuclear factor 1 (CTF/NF-1) in quiescent 3T3-L1 cells. By contrast, quiescent 3T3-L1 cells that stably overexpress either c-Myc or N-Myc contain at least three forms of CTF/NF-1. Biochemical characterization of the various CTF/NF-1 forms showed that they have the same native molecular weight but differ in charge density. The more negatively charged CTF/NF-1 forms present in Myc-overexpressing cells are converted into that found in normal cells by treatment with acid phosphatase, suggesting that they represent a more phosphorylated form of the CTF/NF-1 protein. The various CTF/NF-1 forms have a similar DNA-binding affinity. Transfection experiments demonstrated that transcription from CTF/NF-1-dependent promoters is specifically suppressed in cells that stably overexpress c-Myc. This effect requires CTF/NF-1 binding. CTF/NF-1-dependent promoter activity is also suppressed in 3T3-L1 cells during active growth (relative to the quiescent state). Interestingly, actively growing 3T3-L1 cells contain forms of CTF/NF-1 similar to those in quiescent cells that stably overexpress c-Myc. Thus, the CTF/NF-1 forms present in cells that express high amounts of c-Myc correlate with a lower transcription rate of CTF/NF-1-dependent promoters in vivo. Our results provide a basis for the suppression of specific gene transcription by c-Myc.

C-myc gene chromatin of estrogen receptor positive and negative breast cancer cells

Expression of the c-myc protooncogene is estrogen regulated in estrogen receptor (ER) positive, hormone-dependent human breast cancer cells, but it is constitutively active in ER negative, hormone-independent breast cancer cells. To determine whether these differences are reflected in c-myc chromatin, DNase I hypersensitive sites (DHS) were mapped. Six DHS were detected in all cell lines studied, with DHS 3(2) being more prominent than DHS 3(1). The accessibility of DHS 2 was markedly greater in ER negative cells than in ER positive cells, and this relative accessibility remained unchanged when cells were grown in estrogen free medium. DHS 2, 3(1) and 3(2) map near the P0, P1 and P2 promoters, respectively. An analysis of promoter usage demonstrated that P2 was the preferred promoter. Thus, the differences in the accessibility of DHS 2 in c-myc chromatin of ER positive and negative cells likely reflects alterations in DNA-protein interactions in this region.

Reverse transformation, genome exposure, and cancer

The reverse transformation reaction whereby malignant cells are restored to a more normal phenotype has been reviewed. The primary causative action is ascribed to the genome exposure reaction in which a peripheral nuclear DNA region is restored to high sensitivity to DNase I, like that in normal cells. Various aspects of genome exposure around the nucleoli and the nuclear periphery are considered. The special role of the cytoskeleton in regulating exposure resulting in normal differentiation on the one hand and malignant transformation on the other is discussed. The action of the two-level system for regulation of the mammalian genome previously proposed is reviewed in relation to normal differentiation and malignancy with brief indication of roles played by various metabolites, transcription factors, protooncogenes, cell organelles, and processes like specific phosphorylation and dephosphorylation. Possible implications for cancer therapy and prevention and for the fields of genetic disease and toxicology are indicated.

Activation of the beta-globin locus control region precedes commitment to the erythroid lineage

The beta-globin locus control region (LCR) is characterized by erythroid-specific DNase I hypersensitive sites and is involved in the chromatin organization, transcriptional potentiation, developmental regulation, and replication timing of the entire beta-globin gene cluster. When and how the LCR is first activated during erythropoiesis is not known. Here we analyze the chromatin structure of the LCR during early hematopoietic differentiation using nontransformed, multipotential, growth factor-dependent, murine hematopoietic progenitor cells. We show that LCR hypersensitive sites characteristic of erythroid cells are present in three independent multilineage progenitors [FDCP (factor-dependent cell, Paterson)-mix A4, B6SUtA, and LyD9] under conditions of self-renewal. Induction of differentiation down a nonerythroid pathway causes a progressive loss of hypersensitivity in the LCR. These results show that the beta-globin LCR is in an active chromatin configuration prior to erythroid commitment and indicate a significant role for selective gene repression in lineage specification.

The mouse beta-globin locus control region: hypersensitive sites 3 and 4

The human beta-globin LCR plays a key role in the transcriptional regulation of the beta-globin locus and comprises four erythroid specific DNase I hypersensitive sites, designated 5'HS1-4. We have now isolated genomic clones containing 5'HS3 and 5'HS4 of the mouse beta-globin LCR. 5'HS3 and 5'HS4 are located 15 kb and 22 kb upstream of the mouse epsilon y-globin gene, respectively. Sequence analysis of murine 5'HS3 and 5'HS4 reveals a significant degree of sequence conservation with their human homologues, including the presence of recognition sites for functionally relevant transcription factors. 5'HS3 and 5'HS4 regions were found to form hypersensitive sites in nuclei from murine erythroid cells, but not in nuclei from a variety of nonerythroid haematopoietic cell lines. Analysis of different mouse strains revealed the existence of a polymorphism that alters the spacing between 5'HS3 and 5'HS4. Taken together, our results emphasize the extent of evolutionary conservation and complexity of mammalian beta-globin LCRs. Finally, the cloning of mouse 5'HS3 and 5'HS4 will facilitate the molecular analysis of LCR function in the mouse model.

Modulation of c-myc transcription by triple helix formation

The human c-myc oncogene promoter was used as a model with which to study the mechanism of action of oligodeoxyribonucleotides targeted to a gene regulatory region. The nuclease-hypersensitive element, NHE, lying -115 bp from the P1 promoter of the human c-myc gene, is known to be required in cis for transcription of the gene from both P1 and P2 promoters (Fig. 1). Inhibition of c-myc transcription by an oligonucleotide designed to bind to NHE by triplex formation has been observed in a cell-free transcription assay. Using a reconstituted transcription system with the semipurified PuF transcription factor whose site of interaction resides within the NHE, it is shown here that the oligonucleotide inhibits PuF-mediated transcription. These findings, together with data presented elsewhere showing that: (1) PU1 binds to cloned DNA fragments to form a colinear triplex; (2) PU1 inhibits transcription in nuclear extracts; (3) triple helix formation inhibits the binding of PuF to its target NHE element in an in vitro binding competition assay (E. Postel, R. Durland, and M. Hogan, submitted); (4) triplex formation at the NHE target site can occur in living HeLa cells treated with the triplex-forming PU1 oligomer, and (5) c-myc mRNA synthesis in these treated cells is repressed, clearly support the proposed model in which the oligonucleotide targeted against the c-myc NHE promoter region binds to form a triplex, thereby blocking access to the regulatory protein PuF. This results in promoter-sensitive repression of transcriptional activation of the c-myc gene. The potential for manipulation of gene expression by oligonucleotides targeted to a DNA sequence of the c-myc oncogene promoter and other gene promoters is clear.

Two cellular single-strand-specific DNA-binding proteins interact with two regions of the bovine papillomavirus type 1 genome, including the origin of DNA replication

We have identified and purified to near homogeneity two specific single-stranded DNA-binding factors (SPSF I and II) with molecular masses of 42 and 39 kDa, respectively, from calf thymus. Gel retention analysis and competition experiments demonstrate that the ubiquitous proteins SPSF I and II specifically interact with single-stranded DNA derived from the minimal in vitro origin of replication of bovine papillomavirus type 1 and a region of the viral genome proposed to be involved in plasmid maintenance. Bovine papillomavirus type 1 proteins do not interfere with DNA binding of SPSF I and II. The exact location of the binding domains of SPSF I and II on the DNA has been determined by methylation interference and T4 DNA polymerase footprinting. A potential cellular binding site for SPSF I and II is the major promoter (P2) of the human c-myc gene.

Direct demonstration of transcriptional activation of collagen gene expression in systemic sclerosis fibroblasts: insensitivity to TGF beta 1 stimulation

Lesional fibroblasts propagated from the skin of patients with scleroderma, when compared to normal fibroblasts, show increased synthesis of several collagens and increased levels of their corresponding mRNAs. Using constructs (COL1A2/CAT) containing the promoter for the alpha 2 (I) collagen gene in transient transfection assays with matched pairs of scleroderma and normal skin fibroblasts, we observed higher transcriptional activity of the COL1A2 gene in scleroderma fibroblasts and, in contrast to normal fibroblasts, no further expression was observed in the presence of TGF beta 1. Analysis of the expression of COL1A2 promoter deletion constructs indicates that the TGF beta responsive element functional in normal fibroblasts and the sequence involved in intrinsic upregulation of COL1A2 gene expression in scleroderma fibroblasts are both located between bp-376 (Bgl II) and bp-108 (Sma I) sites. These data may indicate that intrinsic upregulation of extracellular matrix genes in scleroderma fibroblasts utilizes a TGF beta dependent pathway.

Chromatin structure changes suggest a compensatory response to c-myc gene amplification in malignant fibrous histiocytoma

Changes in chromatin structure as determined from DNAse I hypersensitive site analysis are associated with c-myc amplification and increased transcript/protein levels in malignant fibrous histiocytoma (MFH) cell lines. A DNAse I hypersensitive site near the PO promoter region was observed in one MFH cell line (UR HCL 1), and in normal fibroblasts (HFF), but not in an MFH cell line with an amplified c-myc gene (P3C). A DNAse I hypersensitive site exclusive to P3C amplified c-myc was identified slightly 3' of exon one. No alterations in c-myc DNAse I hypersensitive site patterns were observed in HFF fibroblasts following serum release, when peak levels of c-myc transcript were induced. DNAse I hypersensitive site patterns associated with gene amplification may reflect a compensatory response by P3C cells to an abundance of c-myc transcript. Furthermore, elevated levels of protein in P3C cells provide additional evidence that amplified c-myc is an oncogene in MFHs.

Transcriptional down-regulation of c-myc in human prostate carcinoma cells by the synthetic androgen mibolerone

The mechanism of down-regulation of c-myc RNA associated with androgen-induced suppression of the transformed phenotype in the human prostate carcinoma cell line LNCaP was investigated. The synthetic androgen mibolerone (7 alpha-17 alpha-Dimethyl-19-nortestosterone) reversibly inhibits the proliferation of LNCaP cells and, from 12-72 h after hormone addition reduces the level of c-myc transcripts to a few per cent of controls. P1, P2, and P0 c-myc transcripts decline at the same rate, whereas P3 transcripts are much less hormone sensitive. Nuclear run-on analysis revealed that c-myc is down-regulated at the level of transcription initiation in LNCaP cells. The level of c-myc transcripts prevailing in untreated control cells can be restored in androgen-induced cells by excess antiandrogen, indicating the involvement of the androgen receptor in c-myc down-regulation.

Chromatin structure, DNA methylation, and gene expression at sites of viral integration in human fibroblasts. Implications for chromosomal fragility

We analyzed the structural and functional properties of a chromosomal region in which a recombinant hybrid virus adenovirus 5/SV40 preferentially integrates. Our results demonstrated that the structure of the cellular targets for DNA and RNA viruses is very similar and that the cellular sequence flanking the integrated virus possesses, simultaneously, all the features postulated to be the molecular basis for chromosomal fragility.

Functional analysis of the regulatory region of a zein gene in transiently transformed protoplasts

The transcription of zein genes in maize is tissue-specific and developmentally regulated. The 5' regulatory region of many zein genes contains two promoters, P1 and P2, lying approximately 1000 bases apart. The promoter/enhancer activity of various fragments of the two promoter regions of the zein gene E19 have been analysed by means of transient expression experiments. The results indicate that the various regions differentially affect the expression of the GUS reporter gene activity in protoplasts from tobacco leaves, maize immature endosperms and in vitro endosperm cell cultures. In tobacco protoplasts only the proximal promoter region, P2, activates GUS expression, while in endosperm culture cells only the distant promoter, P1, gives significant activity. The P1 region, both in direct and opposite orientation, stimulates a low level of GUS expression in protoplasts from immature endosperms.

The gene for erythropoietin receptor is expressed in multipotential hematopoietic and embryonal stem cells: evidence for differentiation stage-specific regulation

The principal regulator of erythropoiesis is the glycoprotein erythropoietin, which interacts with a specific cell surface receptor (EpoR). A study aimed at analyzing EpoR gene regulation has shown that both pluripotent embryonal stem cells and early multipotent hematopoietic cells express EpoR transcripts. Commitment to nonerythroid lineages (e.g., macrophage or lymphocytic) results in the shutdown of EpoR gene expression, whereas commitment to the erythroid lineage is concurrent with or followed by dramatic increases in EpoR transcription. To determine whether gene activity could be correlated with chromatin alterations, DNase-hypersensitive sites (HSS) were mapped. Two major HSS located in the promoter region and within the first intron of the EpoR gene are present in all embryonal stem and hematopoietic cells tested, the intensities of which correlate well with EpoR expression levels. In addition, a third major HSS also located within the first intron of the EpoR gene is uniquely present in erythroid cells that express high levels of EpoR. Transfection assays show that sequences surrounding this major HSS impart erythroid cell-specific enhancer activity to a heterologous promoter and that this activity is at least in part mediated by GATA-1. These data, together with concordant expression levels of GATA-1 and EpoR in both early multipotent hematopoietic and committed erythroid cells, support a regulatory role of the erythroid cell-specific transcription factor GATA-1 in EpoR transcription in these cells. However, the lack of significant levels of GATA-1 expression in embryonal stem cells implies an alternative regulatory mechanism of EpoR transcription in cells not committed to the hematopoietic lineage.

Immunoglobulin heavy-chain and CD3 delta-chain gene enhancers are DNase I-hypersensitive in hemopoietic progenitor cells

Multipotential interleukin 3-dependent non-immortalized murine hemopoietic progenitor cells have DNase I-hypersensitive sites in the immunoglobulin heavy-chain and CD3 delta enhancers and transcribe germ-line T-cell antigen receptor gamma-chain (TCR gamma), but not IgM or TCR beta, genes. Induction of myeloid differentiation in these cells clones down expression and/or transcriptional accessibility of the immunoglobulin heavy-chain and TCR gamma genes. The CD3 delta enhancer region remains DNase I-hypersensitive but closes down in B cells. In embryonic stem cells and pan-mesodermal cells, these genes or enhancer regions are neither expressed nor DNase I-hypersensitive. These data suggest that lineage potential may be programmed, at least in part, by alterations in the accessibility or conformation of regulatory regions of genes and that some promiscuity of gene expression and/or accessibility can precede lineage commitment and maturation in progenitor cells induced to self-renew by interleukin 3.

The HeLa Pur factor binds single-stranded DNA at a specific element conserved in gene flanking regions and origins of DNA replication

A major site of DNA bending is located 1.6 kb upstream of the P1 transcription start site of the human c-myc gene, near the center of a reported zone of initiation of DNA replication. A repeated, purine-rich element, termed PUR, at the bend site is specifically bound by a protein in HeLa cell nuclear extracts. This protein has specific affinity for the purine-rich single strand of the element. Methylation interference maps a pattern of specific contact points with guanosine bases in a 24-mer oligonucleotide containing the element. UV cross-linking reveals that contact is made by a polypeptide of approximately 28 kDa. The PUR element is present at origins of replication and in gene flanking regions in a variety of eukaryotes from yeasts through humans. The consensus sequence GGNNGAGGGAGARRRR has been derived. This element is present near centers of regions of two mammalian loci (human c-myc and hamster dhfr) recently reported as initiation zones for DNA replication. A 24-mer oligonucleotide representing the hamster dhfr version of the PUR element effectively competes with the human c-myc version for binding to Pur.

The gene for erythropoietin receptor is expressed in multipotential hematopoietic and embryonal stem cells: evidence for differentiation stage-specific regulation

The principal regulator of erythropoiesis is the glycoprotein erythropoietin, which interacts with a specific cell surface receptor (EpoR). A study aimed at analyzing EpoR gene regulation has shown that both pluripotent embryonal stem cells and early multipotent hematopoietic cells express EpoR transcripts. Commitment to nonerythroid lineages (e.g., macrophage or lymphocytic) results in the shutdown of EpoR gene expression, whereas commitment to the erythroid lineage is concurrent with or followed by dramatic increases in EpoR transcription. To determine whether gene activity could be correlated with chromatin alterations, DNase-hypersensitive sites (HSS) were mapped. Two major HSS located in the promoter region and within the first intron of the EpoR gene are present in all embryonal stem and hematopoietic cells tested, the intensities of which correlate well with EpoR expression levels. In addition, a third major HSS also located within the first intron of the EpoR gene is uniquely present in erythroid cells that express high levels of EpoR. Transfection assays show that sequences surrounding this major HSS impart erythroid cell-specific enhancer activity to a heterologous promoter and that this activity is at least in part mediated by GATA-1. These data, together with concordant expression levels of GATA-1 and EpoR in both early multipotent hematopoietic and committed erythroid cells, support a regulatory role of the erythroid cell-specific transcription factor GATA-1 in EpoR transcription in these cells. However, the lack of significant levels of GATA-1 expression in embryonal stem cells implies an alternative regulatory mechanism of EpoR transcription in cells not committed to the hematopoietic lineage.

The HeLa Pur factor binds single-stranded DNA at a specific element conserved in gene flanking regions and origins of DNA replication

A major site of DNA bending is located 1.6 kb upstream of the P1 transcription start site of the human c-myc gene, near the center of a reported zone of initiation of DNA replication. A repeated, purine-rich element, termed PUR, at the bend site is specifically bound by a protein in HeLa cell nuclear extracts. This protein has specific affinity for the purine-rich single strand of the element. Methylation interference maps a pattern of specific contact points with guanosine bases in a 24-mer oligonucleotide containing the element. UV cross-linking reveals that contact is made by a polypeptide of approximately 28 kDa. The PUR element is present at origins of replication and in gene flanking regions in a variety of eukaryotes from yeasts through humans. The consensus sequence GGNNGAGGGAGARRRR has been derived. This element is present near centers of regions of two mammalian loci (human c-myc and hamster dhfr) recently reported as initiation zones for DNA replication. A 24-mer oligonucleotide representing the hamster dhfr version of the PUR element effectively competes with the human c-myc version for binding to Pur.

Identification of a protein that interacts with the nuclear factor-1 (NF-1) binding site in cells that do not express NF-1: comparison to NF-1, cellular distribution, and effect on transcription

We examined expression of nuclear factor-1 (NF-1) in different cell lines. Expression was low or undetectable in T and B lymphocyte cell lines, whereas fibroblasts and other adherent cell lines generally had a relatively high level of NF-1 mRNA. In cell lines that did not express NF-1, gel retardation assays, nevertheless, indicated complexes between a protein or proteins and the NF-1 site. These complexes were less abundant than those formed with NF-1, they migrated more slowly, and they appeared as single species instead of the multiple species observed with NF-1. NF-1 site-binding proteins were compared in the fibrosarcoma cell line HT-1080 (expressed the highest level of NF-1 in our study) and the B cell line Raji (does not express NF-1). UV-crosslinking studies indicated that the NF-1 site-binding proteins in both cell lines were similar in size. Proteolytic clipping band shift assays suggested that the Raji protein and NF-1 share structural similarity in their DNA binding domains, but are distinct proteins. The NF-1 site mediated transcriptional stimulation in cell lines where NF-1 is expressed; however, this element did not affect transcription in cell lines that do not express NF-1, suggesting that the NF-1 site-binding protein in these cells is functionally distinct from NF-1.

Conserved chromatin structure in c-myc 5'flanking DNA after viral transduction

The role of local sequence information in establishing the chromatin structure of the human c-myc upstream region (MUR) was investigated. Adeno-associated virus (AAV)-mediated gene transduction was used to introduce an additional unrearranged copy of the 2.4 kb HindIII-XhoI fragment of the MUR into a novel location in the genome in each of two cloned HeLa cell lines. The AAV-based rep- cap- viral vector SKMA used to transduce the MUR retained only 1.4 kb (24%) of the AAV genome and could accommodate inserts as large as 2.4 kb. SKMA was capable of infecting HeLa cells and integrating into the host genome at single copy number. Integration may have occurred at a preferred site in the HeLa genome, but this site was apparently distinct from the previously identified preferred AAV integration site on human chromosome 19. Indirect end-labelling was used to map DNase I and micrococcal nuclease (MNase) cleavage sites over the transduced c-myc sequences and the endogenous c-myc loci in infected HeLa cells. A similarly ordered chromatin domain, extending 5' from c-myc promoter P0, was found to exist at the transduced c-myc locus in each clone. The position and relative sensitivity of 13 MNase cleavage sites and five DNase I hypersensitive sites, originally identified at the endogenous MUR in non-transduced cells, were shown to be conserved when this DNA was moved to a new chromosome site. A conserved DNase I hypersensitive site also was mapped to the region between the left AAV terminal repeat and AAV promoter P5. These results suggest that the information required to establish the particular chromatin structure of the MUR resides within the local DNA sequence of that region.

Differential DNase I hypersensitivity of ras oncogenes in B6C3F1, C3H/He, and C57BL/6 mouse liver

The male hybrid B6C3F1 mouse exhibits a 30% spontaneous hepatoma incidence, whereas the paternal C3H/He strain and the maternal C57BL/6 strain exhibit a 60% and a negligible incidence, respectively. In addition, both male and female B6C3F1 mice are extremely sensitive to chemical induction of hepatocarcinogenesis. The Ha-ras, Ki-ras, and myc oncogenes have been implicated in a variety of solid tumors. Specifically, Ha- and, less frequently, Ki-ras have been reported to be activated in B6C3F1 mouse liver tumors. The objective of this study was to examine a possible point of transcriptional control of Ha-ras, Ki-ras, and myc in all three mouse strains, our hypothesis being that these oncogenes may be primed for expression in the nascent liver of those strains exhibiting a high spontaneous hepatoma incidence. A positive correlation has been established between gene expression and the presence of DNase I hypersensitive sites. DNase I hypersensitive sites were observed in the Ha-ras and myc oncogenes in the three mouse strains. However, Ha-ras appears to possess an additional site in B6C3F1 and C3H/He as compared to C57BL/6. Similarly, the Ki-ras oncogene exhibited a DNase I hypersensitive site only in B6C3F1 and C3H/He mouse liver. These results indicate that the hepatoma-prone strains (B6C3F1 and C3H/He) may have a greater potential for Ha- and Ki-ras expression than does the non-hepatoma-prone strain (C57BL/6).

Evidence that a triplex-forming oligodeoxyribonucleotide binds to the c-myc promoter in HeLa cells, thereby reducing c-myc mRNA levels

A synthetic 27-base-long oligodeoxyribonucleotide, termed PU1, has been shown to bind to duplex DNA to form a triplex at a single site within the human c-myc P1 promoter. PU1 has been administered to HeLa cells in culture to examine the feasibility of influencing transcription of the c-myc gene in vivo. It is shown that uptake of PU1 into the nucleus of HeLa cells is efficient and that the compound remains intact for at least 4 hr. In nuclei extracted from PU1-treated cells, inhibition of DNase I cleavage is detected within the c-myc P1 promoter at the target site for triplex formation. The inhibition is shown to be both site and oligodeoxyribonucleotide specific. After cellular uptake of PU1, it is shown that steady-state mRNA arising from the c-myc P1 initiation site is selectively reduced relative to total mRNA, relative to mRNA from the alternative c-myc P2 initiation site, and relative to mRNA derived from the beta-actin promoter. Significant mRNA repression is not seen upon treating cells with oligodeoxyribonucleotides that fail to bind to the P1 promoter target. Taken together, these data suggest that triplex formation can occur between an exogenous oligodeoxyribonucleotide and duplex DNA in the nucleus of treated cells.

Sequence-specific binding of a c-myc nuclear-matrix-associated region shows increased nuclear matrix retention after leukemic cell (HL-60) differentiation

HL-60 cells, a human promyelocytic leukemia cell line, contain amplified c-myc DNA sequences and mRNA transcripts. These cells can be induced to undergo macrophage differentiation by phorbol esters, which results in suppression of c-myc expression and cessation of cell proliferation. The nuclear matrix (NM), a nuclear skeleton resistant to DNase I digestion and high salt extraction, is proposed to be involved in DNA replication, gene regulation, and the correct distribution of DNA at mitosis. We have previously identified a nuclear-matrix-associated region (MAR) of the c-myc protooncogene to reside in a 1.4-kb region between Cla I and Eco RI restriction sites at the 3'-end of the gene. A 172-bp Dra I/Dra I subfragment of the 1.4-kb region was shown to be a major component of the MAR (myc-MAR), and this subfragment was demonstrated to be recognized by a nuclear protein (p25). In this report we demonstrate that phi X174 DNA, or the synthetic copolymers poly[d(G.C)] and poly[d(A.T)], are not effective suppressors of the binding of the myc-MAR to isolated NM, indicating that the binding sequence(s) are unique. We find that the addition of partially purified protein p25 increases the relative affinity of the myc-MAR for HL-60 NM in an in vitro assay system. NM isolated from HL-60 macrophages induced by phorbol esters retains significantly more myc-MAR DNA fragment in the presence of an excess amount of competitor DNA than does NM from untreated HL-60 cells. These data suggest that a change of the myc-MAR association with the NM occurs after monocytic differentiation of HL-60 cells.

Cyclosporin A regulates the expression of HLA-DR on human monocytes by two different mechanisms

Cyclosporin A (CsA), but not its nonimmunosuppressive analog cyclosporin H (CsH), inhibited the expression of HLA-DR in human monocytes. Induction of HLA-DR by interferon (IFN)-gamma in fresh monocytes was also inhibited by CsA and not by CsH. However, when monocytes were pretreated with either CsA or CsH for 16 hr prior to the addition of IFN-gamma, HLA-DR expression was increased, probably because of a cyclosporin-induced increase in the number of IFN-gamma receptors. Down-regulation of the HLA-DR mRNA by CsA was found to be dependent on continuous protein synthesis. IFN-alpha also inhibited the IFN-gamma-induced HLA-DR mRNA expression and showed synergy with CsA at low concentrations but not at high concentrations of the drugs. A common mechanistic element in the pathways of CsA and IFN-alpha is proposed.

Cloning and characterization of a human c-myc promoter-binding protein

A human cDNA clone encoding a c-myc promoter-binding protein was detected by screening a HeLa cell lambda phage expression cDNA library. The library was screened by using an XhoI-NaeI human c-myc P2 promoter fragment as a probe. The recombinant phage encoded a fusion protein, myc-binding protein 1 (MBP-1), which had an apparent molecular size of 40 kDa. A corresponding protein with a molecular size of 35 kDa was present in a HeLa cell extract. Sequence analysis of the cloned gene reveals an open reading frame of 1,038 bp with a 3' untranslated region of 378 bp. The predicted protein sequence contains a proline-rich region in the amino terminus but does not demonstrate a known DNA-binding domain. DNase I footprint analysis demonstrates that MBP-1 binds to the sequence just 5' of the TATA box sequence of the human c-myc P2 promoter. MBP-1 cDNA hybridizes to a 1.4-kb mRNA from HeLa and HL-60 cells, indicating that the cDNA insert (1,416 bp) is a full-length clone. Coexpression of the MBP-1 protein repress transcription from the human c-myc promoter, suggesting that MBP-1 may act as a negative regulatory factor for the human c-myc gene.

Cloning and characterization of a human c-myc promoter-binding protein

A human cDNA clone encoding a c-myc promoter-binding protein was detected by screening a HeLa cell lambda phage expression cDNA library. The library was screened by using an XhoI-NaeI human c-myc P2 promoter fragment as a probe. The recombinant phage encoded a fusion protein, myc-binding protein 1 (MBP-1), which had an apparent molecular size of 40 kDa. A corresponding protein with a molecular size of 35 kDa was present in a HeLa cell extract. Sequence analysis of the cloned gene reveals an open reading frame of 1,038 bp with a 3' untranslated region of 378 bp. The predicted protein sequence contains a proline-rich region in the amino terminus but does not demonstrate a known DNA-binding domain. DNase I footprint analysis demonstrates that MBP-1 binds to the sequence just 5' of the TATA box sequence of the human c-myc P2 promoter. MBP-1 cDNA hybridizes to a 1.4-kb mRNA from HeLa and HL-60 cells, indicating that the cDNA insert (1,416 bp) is a full-length clone. Coexpression of the MBP-1 protein repress transcription from the human c-myc promoter, suggesting that MBP-1 may act as a negative regulatory factor for the human c-myc gene.

Evolutionary conservation of target sequences for cis-acting regulation in c-myc exon 1 and its upstream region

Transcriptional control of the c-myc proto-oncogene, an important factor in cellular growth, differentiation and in the genesis of various neoplasms, is mediated by multiple positive and negative regulators in the 5' end region of the gene. Here, we report the nucleotide sequence of the first c-myc exon and its upstream region from woodchuck, a rodent which can develop liver tumors associated with c-myc activation [Möröy et al., Nature 324 (1986) 276-279]. Alignment of these sequences with the corresponding human and murine regions shows a surprisingly high homology between woodchuck and human, and suggests the absence of species-specificity in the fundamental regulatory elements which govern c-myc expression.

Characterization of the human cyclophilin gene and of related processed pseudogenes

The human cyclophilin gene was isolated from a genomic library derived from leucocyte DNA and sequenced. The gene contains five exons and four introns. The amino acid sequence deduced from the exons matches perfectly the one previously determined from the T-cell cyclophilin cDNA. A TATA box is visible in the promoter region and putative Sp1 binding sites are also found there as well as in the first intron. Six members of the middle repetitive Alu gene family are present in one or other orientation in the non-coding regions of the cyclophilin gene. Hybridisation of genomic DNA to probes derived from the promoter region or the first intron indicates that the cyclophilin gene is present as a single copy in the human haploid genome. Seven other cyclophilin-related DNA clones isolated from the same library were also characterized. They show a high degree of similarity to the cyclophilin cDNA and are colinear to it. However, multiple genetic lesions, often including deletion and/or insertion events which modify the reading frame, are found in these clones which are therefore likely to represent processed pseudogenes.

Identification of tissue specific nuclear proteins: DNA sequence and protein binding regions in the T cell receptor beta J-C intron

We have determined the DNA sequences in the J2-C2 intron of the T cell receptor (TCR) beta gene and analyzed nuclear proteins binding to this region. Previously, we identified two tissue-specific DNase I hypersensitive regions, potential regulatory regions, in the J-C intron. The DNA sequence of the J2-C2 intron revealed that both DNase I hypersensitive regions have similar DNA sequences, suggesting that these regions are evolutionarily conserved. We have also identified tissue-specific nuclear-protein binding regions downstream of the DNase hypersensitive regions. Although transcriptional enhancer activity was not observed in the hypersensitive regions or the adjacent protein binding regions in the J-C intron, our findings suggest that the TCR-beta J-C intron may contain some other type of regulatory element.

Transcriptional activation of a conserved sequence element by ras requires a nuclear factor distinct from c-fos or c-jun

The expression of transforming growth factor beta type 1 mRNA was increased by conditional expression of ras. A 31-base-pair sequence found approximately 420 base pairs upstream of the gene encoding human transforming growth factor beta 1 acted as a ras-responsive enhancer element in transient transfection assays. The human sequence contains the element TGACTCT that also is found in a murine ras-responsive enhancer. Analysis of nuclear factors present in cells stably transformed by ras indicated that both human and murine sequences were recognized by the same nuclear factor. The role of fos and jun in ras transcriptional activation was analyzed in transfection assays using murine elements that contained either TGACTCT or TGAGTAA. These experiments showed that while both elements are activated by fos/jun expression to nearly the same event, only the former element responded to ras. In addition, activation of reporters containing TGACTCT is 6-fold higher by ras than by fos/jun. Gel retention experiments revealed that the nuclear factor present in cells transformed by ras exhibited the same sequence preference as demonstrated in the transient transfection assays. UV-crosslinking experiments identify a protein of apparent molecular mass 120 kDa that recognizes the ras-responsive element. This work identifies a persistent signal transduction pathway that links ras to nuclear transcription and indicates that a 120-kDa protein is a target of this pathway.

Transcriptional down-regulation of the rearranged C-myc expression in murine cell hybrids between a plasmacytoma and a T-cell lymphoma

Regulation of the rearranged and non-rearranged c-myc expression was studied in murine cell hybrids (SBWI and SBWII) between plasmacytoma (S194) and T-cell lymphoma (BW5147) cells. Expression of the rearranged c-myc of heterogeneous mRNA sizes (1.8 approximately 2.4 kb) was markedly down-regulated in these hybrids regardless of retention of the gene. On the other hand, expression of the non-rearranged c-myc (2.4 kb) was not significantly affected in these hybrids. Treatment of SBWI hybrid cells with cycloheximide enhanced the non-rearranged c-myc 2- to 4-fold but did not release the down-regulation of the rearranged c-myc at all, suggesting that the down-regulation of the rearranged c-myc in the hybrid cells was mainly at a transcriptional rather than a post-transcriptional level. This was supported by the results of nuclear run-on assay: the high level of run-on transcripts in S194 cells declined in SBWI hybrid cells comparable to the level in BW5147 cells. The rearranged c-myc was hemi-methylated in S194 cells and the pattern was the same in SBWI hybrid cells. Furthermore, down-regulation of the rearranged c-myc in the hybrid was also not restored by treatment with 5-azacytidine (5-AzaC), 12-O-tetradecanoylphorbol-13-acetate (TPA) or forskolin, suggesting no causative involvement of DNA methylation or protein phosphorylation in down-regulation. Higher DNase I sensitivity of the rearranged c-myc in S194 cells decreased to a similar extent to that of the non-rearranged c-myc after cell fusion with BW5147 cells. These results suggest that expression of the rearranged c-myc is down-regulated at the level of transcription in murine cell hybrids between a plasmacytoma and a T-cell lymphoma, probably by changing chromatin configuration around the gene from the open to the closed state.

Autonomous replication of a DNA fragment containing the chromosomal replication origin of the human c-myc gene

The c-myc genes of HeLa cells are preferentially replicated in the transcriptional direction, from chromosomal origin sequences which display cell type-specific activity. Using a run-off replication assay involving in vitro extension of replication forks initiated in intact HeLa cells, bidirectional replication was observed to begin within a 3.5 kb domain 5' to the c-myc gene. To characterize the replication origin further a 2.4 HindIII-Xhol subfragment of the c-myc 5' flanking DNA was cloned in a selectable vector and transfected into HeLa cells. The resulting pNeo.Myc-2.4 construct persisted as a circular extrachromosomal element for more than 300 cell generations under selection, with recovery of approximately 500-1000 times the mass of plasmid initially introduced into the cells. Extrachromosomal circular pNeo.Myc-2.4 monomer was reisolated in supercoiled form, along with oligomeric and miniplasmid variants which had been generated in vivo; however, chromosomally integrated copies of the plasmid were not detectable in cultures containing extrachromosomal pNeo.Myc-2.4. The recovered pNeo.Myc-2.4 plasmid was resistant to Dpnl digestion and sensitive to Mbol digestion. After transfection with pNeo.Myc-2.4 BrUdR pulse labeling of long-term or short-term cultures demonstrated that the plasmid replicated semiconservatively, under controls similar to those imposed on chromosome replication. Bisection of the pNeo.Myc-2.4 insert suggested that c-myc 5' flanking DNA within 1.2 kb 5' to promoter P1 was sufficient to confer autonomously replicating sequence activity on the plasmid vector in transient replication assays.