Characterization of SAP-1, a protein recruited by serum response factor to the c-fos serum response element

Attenuation of serum inducibility of immediate early genes by oncoproteins in tyrosine kinase signaling pathways

Immediate early genes involved in controlling cell proliferation are rapidly and transiently induced following stimulation of susceptible cells with serum. To study how oncoproteins regulate immediate early genes, we examined serum inducibility of these genes in cells transformed by various oncoproteins. We found that induction of the immediate early gene, c-fos, by serum stimulation was markedly attenuated in four independent cell lines stably transformed by the v-Src tyrosine kinase. Cells chronically transformed by other oncoproteins implicated in tyrosine kinase signaling pathways, including v-Sis, v-Ras, and v-Raf, showed the same pattern of attenuation. In contrast, serum inducibility of c-fos was not attenuated in cells transformed by simian virus 40, which is thought to transform cells through a different pathway. Cell cycle analyses showed that proliferation of these transformed cell lines could be arrested effectively in 0.1% serum, demonstrating that the attenuation was not simply due to continuous cycling of transformed cells after serum deprivation. Moreover, serum inducibility of other immediate early genes, including c-jun, junB, egr-1, and NGFI-B, also was strikingly attenuated by these same oncoproteins. Nuclear run-on transcription assays established that this attenuation of serum inducibility occurred at the transcriptional level. Finally, flow cytometric analysis demonstrated that serum-starved v-Src-transformed cells were viable and able to progress into S phase of the cell cycle after serum stimulation, even though the induction of immediate early genes was greatly attenuated in these cells. Our results suggest that activation of immediate early genes is repressed by chronic stimulation of tyrosine kinase signaling pathways in transformed cells.

Two factors that bind to highly conserved sequences in mammalian type C retroviral enhancers

The transcriptional enhancers of the Moloney and Friend murine leukemia viruses (MLV) are important determinants of viral pathogenicity. We used electrophoretic mobility shift and methylation interference assays to study nuclear factors which bind to a region of these enhancers whose sequence is identical between Moloney and Friend viruses and particularly highly conserved among 35 mammalian type C retroviruses whose enhancer sequences have been aligned (E. Golemis, N. A. Speck, and N. Hopkins, J. Virol. 64:534-542, 1990). Previous studies identified sites for the leukemia virus factor b (LVb) and core proteins in this region (N. A. Speck and D. Baltimore, Mol. Cell. Biol. 7:1101-1110, 1987) as well as a site, overlapping those for LVb and core, for a third factor (N. R. Manley, M. A. O'Connell, P. A. Sharp, and N. Hopkins, J. Virol. 63:4210-4223, 1989). Surprisingly, the latter factor appeared to also bind two sites identified in the Friend MLV enhancer, Friend virus factor a and b1 (FVa and FVb1) sites, although the sequence basis for the ability of the protein to bind these diverse sites was not apparent. Here we describe the further characterization of this binding activity, termed MCREF-1 (for mammalian type C retrovirus enhancer factor 1), and the identification of a consensus sequence for its binding, GGN8GG. We also identify a factor, abundant in mouse T-cell lines and designated LVt, which binds to two sites in the Moloney MLV enhancer, overlapping the previously identified LVb and LVc binding sites. These sites contain the consensus binding site for the Ets family of proteins. We speculate on how distinct arrays of these factors may influence the disease-inducing phenotype.

Attenuation of serum inducibility of immediate early genes by oncoproteins in tyrosine kinase signaling pathways

Immediate early genes involved in controlling cell proliferation are rapidly and transiently induced following stimulation of susceptible cells with serum. To study how oncoproteins regulate immediate early genes, we examined serum inducibility of these genes in cells transformed by various oncoproteins. We found that induction of the immediate early gene, c-fos, by serum stimulation was markedly attenuated in four independent cell lines stably transformed by the v-Src tyrosine kinase. Cells chronically transformed by other oncoproteins implicated in tyrosine kinase signaling pathways, including v-Sis, v-Ras, and v-Raf, showed the same pattern of attenuation. In contrast, serum inducibility of c-fos was not attenuated in cells transformed by simian virus 40, which is thought to transform cells through a different pathway. Cell cycle analyses showed that proliferation of these transformed cell lines could be arrested effectively in 0.1% serum, demonstrating that the attenuation was not simply due to continuous cycling of transformed cells after serum deprivation. Moreover, serum inducibility of other immediate early genes, including c-jun, junB, egr-1, and NGFI-B, also was strikingly attenuated by these same oncoproteins. Nuclear run-on transcription assays established that this attenuation of serum inducibility occurred at the transcriptional level. Finally, flow cytometric analysis demonstrated that serum-starved v-Src-transformed cells were viable and able to progress into S phase of the cell cycle after serum stimulation, even though the induction of immediate early genes was greatly attenuated in these cells. Our results suggest that activation of immediate early genes is repressed by chronic stimulation of tyrosine kinase signaling pathways in transformed cells.

The retinoblastoma protein associates with the protein phosphatase type 1 catalytic subunit

The retinoblastoma protein (p110RB) interacts with many cellular proteins in complexes potentially important for its growth-suppressing function. We have developed and used an improved version of the yeast two-hybrid system to isolate human cDNAs encoding proteins able to bind p110RB. One clone encodes a novel type 1 protein phosphatase catalytic subunit (PP-1 alpha 2), which differs from the originally defined PP-1 alpha by an amino-terminal 11-amino-acid insert. In vitro-binding assays demonstrated that PP-1 alpha isoforms preferentially bind the hypophosphorylated form of p110RB. Moreover, similar p110RB sequences are required for binding PP-1 alpha 2 and SV40 large T antigen. Cell cycle synchrony experiments revealed that this association occurs from mitosis to early G1. The implications of these findings on the regulation of both proteins are discussed.

hMEF2C gene encodes skeletal muscle- and brain-specific transcription factors

The myocyte enhancer-binding factor 2 (MEF2) site is an essential element of many muscle-specific enhancers and promoters that binds nuclear proteins from muscle and brain. Recently, we have cloned a family of MEF2 transcription factors produced by two genes that, at the mRNA level, are broadly expressed and produce tissue-specific isoforms by posttranscriptional processes (Y.-T. Yu, R. E. Breitbart, L. B. Smoot, Y. Lee, V. Mahdavi, and B. Nadal-Ginard, Genes Dev. 6:1783-1798, 1992). Here, we report the isolation and functional characterization of cDNA clones encoding four MEF2 factors derived from a separate gene that we have named hMEF2C. In contrast to those of the previously reported genes, the transcripts of the hMEF2C gene are restricted to skeletal muscle and brain. One of the alternate exons is exclusively present in brain transcripts. The products of this gene have DNA-binding and trans-activating activities indistinguishable from those of the previously reported MEF2 factors. The hMEF2C gene is induced late during myogenic differentiation, and its expression is limited to a subset of cortical neurons. The potential targets for this transcription factor in a subset of neurons are not known at this time. The strict tissue-specific pattern of expression of hMEF2C in comparison with the more ubiquitous expression of other MEF2 genes suggests a different mode of regulation and a potentially important role of hMEF2C factors in myogenesis and neurogenesis.

Strategies for the identification of interacting proteins

Many problems in modern biology involve complex arrays of interacting protein and, in some cases, RNA molecules. The initial challenge facing investigators is to identify the important players that drive the process under study. This difficult task is ameliorated somewhat by the development of methods designed to keep pace with the magnitude of this challenge. I have outlined a few of these approaches at the cutting edge of cloning interacting proteins. A perhaps more daunting prospect is to dissect the important molecules once they are in hand, to identify key interactions, and, ultimately, to move to an understanding of function in cells. For this, of course, all of the tools of genetics, biochemistry, and molecular biology, extant and yet to be developed, will have to be tapped.

Functional and defective components of avian endogenous virus long terminal repeat enhancer sequences

Oncogenic avian retroviruses, such as Rous sarcoma virus (RSV) and the avian leukosis viruses, contain a strong enhancer in the U3 portion of the proviral long terminal repeat (LTR). The LTRs of a second class of avian retroviruses, the endogenous viruses (ev) lack detectable enhancer activity. By creating ev-RSV hybrid LTRs, we previously demonstrated that, despite the lack of independent enhancer activity in the ev U3 region, ev LTRs contain sequences that are able to functionally replace essential enhancer domains from the RSV enhancer. A hypothesis proposed to explain these data was that ev LTRs contain a partial enhancer that includes sequences necessary but not sufficient for enhancer activity and that these sequences were complemented by RSV enhancer domains present in the original hybrid constructs to generate a functional enhancer. Studies described in this report were designed to define sequences from both the ev and RSV LTRs required to generate this composite enhancer. This was approached by generating additional ev-RSV hybrid LTRs that exchanged defined regions between ev and RSV and by directly testing the requirement for specific motifs by site-directed mutagenesis. Results obtained demonstrate that ev enhancer sequences are present in the same relative location as upstream enhancer sequences from RSV, with which they share limited sequence similarity. In addition, a 67-bp region from the internal portion of the RSV LTR that is required to complement ev enhancer sequences was identified. Finally, data showing that CArG motifs are essential for high-level activity, a finding that has not been previously demonstrated for retroviral LTRs, are presented.

Human FLI-1 localizes to chromosome 11Q24 and has an aberrant transcript in neuroepithelioma

The v-ets oncogene family shares a conserved motif, termed the ETS-domain, that mediates sequence-specific DNA binding. This motif is unique among transcription factor families. Using partially degenerate oligonucleotides to highly conserved amino acids in this motif as primers for the polymerase chain reaction, a novel ETS-domain cDNA fragment was generated. This fragment was subsequently used to clone both mouse and human full length cDNAs for this gene. The amino acid sequence of the longest open reading frame showed that this gene was homologous to the mouse FLI-I gene, an ETS family gene activated by Friend erythroleukemia virus insertion. The gene is normally expressed only in hematopoietic cells. The gene was localized to chromosome 11q24, a region of aberrations in Ewing's sarcoma and neuroepithelioma. In the neuroepithelioma cell line TC-32 the FLI-1 transcript is present but has an aberrant structure, indicating that it may be rearranged in neuroepithelioma.

Identification of amino acids essential for DNA binding and dimerization in p67SRF: implications for a novel DNA-binding motif

The serum response factor (p67SRF) binds to a palindromic sequence in the c-fos serum response element (SRE). A second protein, p62TCF binds in conjunction with p67SRF to form a ternary complex, and it is through this complex that growth factor-induced transcriptional activation of c-fos is thought to take place. A 90-amino-acid peptide, coreSRF, is capable for dimerizing, binding DNA, and recruiting p62TCF. By using extensive site-directed mutagenesis we have investigated the role of individual coreSRF amino acids in DNA binding. Mutant phenotypes were defined by gel retardation and cross-linking analyses. Our results have identified residues essential for either DNA binding or dimerization. Three essential basic amino acids whose conservative mutation severely reduced DNA binding were identified. Evidence which is consistent with these residues being on the face of a DNA binding alpha-helix is presented. A phenylalanine residue and a hexameric hydrophobic box are identified as essential for dimerization. The amino acid phasing is consistent with the dimerization interface being presented as a continuous region on a beta-strand. A putative second alpha-helix acts as a linker between these two regions. This study indicates that p67SRF is a member of a protein family which, in common with many DNA binding proteins, utilize an alpha-helix for DNA binding. However, this alpha-helix is contained within a novel domain structure.

Two domains of ISGF3 gamma that mediate protein-DNA and protein-protein interactions during transcription factor assembly contribute to DNA-binding specificity

Alpha interferon (IFN-alpha) induces the transcription of a large set of genes through activation of multimeric transcription factor ISGF3. This factor can be dissociated into two protein components, termed ISGF3 gamma and ISGF3 alpha. ISGF3 gamma is a 48-kDa protein related at the amino terminus to members of the IFN-regulatory factor (IRF) and Myb families of DNA-binding proteins; ISGF3 alpha consists of three polypeptides of 84, 91, and 113 kDa that self-assemble to form an activated component in response to IFN-alpha. DNA-binding studies indicated that ISGF3 gamma binds DNA alone, recognizing the IFN-stimulated response element, while the ISGF3 alpha polypeptides alone display no specific interactions with DNA. A complex between ISGF3 gamma and activated ISGF3 alpha binds the IFN-stimulated response element with much greater affinity than does the 48-kDa ISGF3 gamma protein alone. The DNA-binding domain of ISGF3 gamma and regions responsible for protein-protein interaction with ISGF3 alpha were identified by using deleted forms of ISGF3 gamma expressed in vitro. The amino-terminal region of ISGF3 gamma homologous to the IRF and Myb proteins was sufficient for interaction with DNA and displayed the binding specificity of the intact protein; phosphorylation of this region was necessary for activity. A second region of 160 amino acids separated from the DNA-binding domain by over 100 amino acids contained a domain capable of associating with ISGF3 alpha and was sufficient to confer specific ISGF3 alpha interaction to a heterologous protein. Interaction of the ISGF3 alpha component with the protein interaction domain of ISGF3 gamma altered the DNA-binding specificity of the resulting complex, suggesting that one or more of the ISGF3 alpha polypeptides make base-specific contacts with DNA. This interaction defines a mechanism through which IRF-like proteins complexed with regulatory components can display novel DNA-binding specificities.

A short, disordered protein region mediates interactions between the homeodomain of the yeast alpha 2 protein and the MCM1 protein

Homeodomains are folded into a characteristic three-dimensional structure capable of recognizing DNA in a sequence-specific manner. We show that correct target site selection by the yeast alpha 2 protein requires, as well as its homeodomain, an adjacent short and apparently unstructured region of the protein. This flexible homeodomain extension is responsible for specifying an interaction with a second regulatory protein, MCM1, which permits the cooperative binding of the two proteins to an operator. Two additional experiments suggest that this extension-homeodomain arrangement is likely to have some generality. First, when the extension of alpha 2 is grafted onto the Drosophila engrailed homeodomain, it yields a protein with the DNA binding specificity of engrailed and the ability to bind cooperatively to DNA with MCM1. Second, the alpha 2 extension specifies interaction not only with the yeast MCM1 protein, but also with the related human protein SRF.

The Ets family of transcription factors

Interest in the Ets proteins has grown enormously over the last decade. The v-ets oncogene was originally discovered as part of a fusion protein expressed by a transforming retrovirus (avian E26), and later shown to be transduced from a cellular gene. About 30 related proteins have now been found in species ranging from flies to humans, that resemble the vEts protein in the so-called 'ets domain'. The ets domain has been shown to be a DNA-binding domain, that specifically interacts with sequences containing the common core trinucleotide GGA. Furthermore, it is involved in protein-protein interactions with co-factors that help determine its biological activity. Many of the Ets-related proteins have been shown to be transcription activators, like other nuclear oncoproteins and anti-oncoproteins (Jun, Fos, Myb, Myc, Rel, p53, etc.). However, Ets-like proteins may have other functions, such as in DNA replication and a general role in transcription activation. Ets proteins have been implicated in regulation of gene expression during a variety of biological processes, including growth control, transformation, T-cell activation, and developmental programs in many organisms. Signals regulating cell growth are transmitted from outside the cell to the nucleus by growth factors and their receptors. G-proteins, kinases and transcription factors. We will discuss how several Ets-related proteins fit into this scheme, and how their activity is regulated both post- and pre-translationally. Loss of normal control is often associated with conversion to an oncoprotein. vEts has been shown to have different properties from its progenitor, which might explain how it has become oncogenic. Oncogene-related products have been implicated in the control of various developmental processes. Evidence is accumulating for a role for Ets family members in Drosophila development, Xenopus oocyte maturation, lymphocyte differentiation, and viral infectious cycles. An ultimate hope in studying transformation by oncoproteins is to understand how cells become cancerous in humans, which would lead to more effective treatments. vEts induces erythroblastosis in chicken. Cellular Ets-family proteins can be activated by proviral insertion in mice and, most interestingly, by chromosome translocation in humans. We are at the beginning of understanding the multiple facets of regulation of Ets activity. Future work on the Ets family promises to provide important insights into both normal control of growth and differentiation, and deregulation in illness.

Identification of amino acids essential for DNA binding and dimerization in p67SRF: implications for a novel DNA-binding motif

The serum response factor (p67SRF) binds to a palindromic sequence in the c-fos serum response element (SRE). A second protein, p62TCF binds in conjunction with p67SRF to form a ternary complex, and it is through this complex that growth factor-induced transcriptional activation of c-fos is thought to take place. A 90-amino-acid peptide, coreSRF, is capable for dimerizing, binding DNA, and recruiting p62TCF. By using extensive site-directed mutagenesis we have investigated the role of individual coreSRF amino acids in DNA binding. Mutant phenotypes were defined by gel retardation and cross-linking analyses. Our results have identified residues essential for either DNA binding or dimerization. Three essential basic amino acids whose conservative mutation severely reduced DNA binding were identified. Evidence which is consistent with these residues being on the face of a DNA binding alpha-helix is presented. A phenylalanine residue and a hexameric hydrophobic box are identified as essential for dimerization. The amino acid phasing is consistent with the dimerization interface being presented as a continuous region on a beta-strand. A putative second alpha-helix acts as a linker between these two regions. This study indicates that p67SRF is a member of a protein family which, in common with many DNA binding proteins, utilize an alpha-helix for DNA binding. However, this alpha-helix is contained within a novel domain structure.

Two domains of ISGF3 gamma that mediate protein-DNA and protein-protein interactions during transcription factor assembly contribute to DNA-binding specificity

Alpha interferon (IFN-alpha) induces the transcription of a large set of genes through activation of multimeric transcription factor ISGF3. This factor can be dissociated into two protein components, termed ISGF3 gamma and ISGF3 alpha. ISGF3 gamma is a 48-kDa protein related at the amino terminus to members of the IFN-regulatory factor (IRF) and Myb families of DNA-binding proteins; ISGF3 alpha consists of three polypeptides of 84, 91, and 113 kDa that self-assemble to form an activated component in response to IFN-alpha. DNA-binding studies indicated that ISGF3 gamma binds DNA alone, recognizing the IFN-stimulated response element, while the ISGF3 alpha polypeptides alone display no specific interactions with DNA. A complex between ISGF3 gamma and activated ISGF3 alpha binds the IFN-stimulated response element with much greater affinity than does the 48-kDa ISGF3 gamma protein alone. The DNA-binding domain of ISGF3 gamma and regions responsible for protein-protein interaction with ISGF3 alpha were identified by using deleted forms of ISGF3 gamma expressed in vitro. The amino-terminal region of ISGF3 gamma homologous to the IRF and Myb proteins was sufficient for interaction with DNA and displayed the binding specificity of the intact protein; phosphorylation of this region was necessary for activity. A second region of 160 amino acids separated from the DNA-binding domain by over 100 amino acids contained a domain capable of associating with ISGF3 alpha and was sufficient to confer specific ISGF3 alpha interaction to a heterologous protein. Interaction of the ISGF3 alpha component with the protein interaction domain of ISGF3 gamma altered the DNA-binding specificity of the resulting complex, suggesting that one or more of the ISGF3 alpha polypeptides make base-specific contacts with DNA. This interaction defines a mechanism through which IRF-like proteins complexed with regulatory components can display novel DNA-binding specificities.

Different binding site requirements for binding and activation for the bipartite enhancer factor EF-1A

The human transcription factor EF-1A binds to the purine-rich E1A core enhancer sequence in the adenovirus E1A and E4 and polyomavirus enhancer regions. The consensus binding site for EF-1A resembles that of members of the ets domain protein family. EF-1A activation of transcription requires a dimeric binding site. Analysis of binding sites containing point mutations revealed that EF-1A binding is determined by the core nucleotides of the binding site, while transcriptional activation is determined both by the core and some peripheral nucleotides that do not affect binding. We have purified EF-1A and analyzed its two constituent subunits, EF-1A alpha and EF-1A beta. EF-1A alpha (MW approximately 60kD) makes the primary DNA contacts. EF-1A beta (MW approximately 50 kD) forms a heteromultimeric complex with EF-1A alpha both in solution and on a dimeric binding site. Binding of both EF-1A subunits is necessary, but not sufficient, for transcriptional activation. We present immunochemical and functional evidence that EF-1A alpha is related to the murine ets-related protein GABP alpha and that EF-1A beta is related to the murine protein GABP beta.

The isolation of transcription factors from lambda gt11 cDNA expression libraries: human steroid 5 alpha-reductase 1 has sequence-specific DNA binding activity

The Surf-1/Surf-2 bi-directional promoter contains binding sites for at least three transcription factors (Su1, Su2, and Su3). By screening a lambda gt11 HeLa cell cDNA expression library with a concatenated Su2 factor binding site, we isolated a cDNA which encodes a protein with sequence-specific DNA binding activity. Gel retardation assays showed that the cloned factor binds specifically to the Su2 factor binding site present in the human Surf-1/Surf-2 promoter but not to an Su2 site containing mutated base pairs. Co-transfection experiments demonstrated that the cloned cDNA had little or no effect on the expression of a reporter gene under the control of multiple Su2 factor binding sites. Similarly a fusion protein in which the acidic activation domain of HSV VP16 was linked to the cloned factor had no effect, implying that the factor does not function as a DNA binding protein in vivo. DNA sequence analysis revealed that the cloned cDNA is identical to that of human steroid 5 alpha-reductase 1, an enzyme which converts testosterone to dihydrotestosterone. These results are discussed with respect to other putative transcription factors which have been isolated from cDNA expression libraries on the basis of their sequence-specific DNA binding activity.

Ionizing radiation activates transcription of the EGR1 gene via CArG elements

The present studies have examined the effects of ionizing radiation on control of the early growth response 1 (EGR1) gene. Exposure of human HL-525 cells to x-rays was associated with increases in EGR1 mRNA levels. Nuclear run-on assays showed that this effect is related at least in part to activation of EGR1 gene transcription. Sequences responsive to ionizing radiation-induced signals were determined by deletion analysis of the EGR1 promoter. The results demonstrate that x-ray inducibility of the EGR1 gene is conferred by a region containing six serum response or CC(A+T-rich)6GG (CArG) motifs. Further analysis confirmed that the region encompassing the three distal or upstream CArG elements is functional in the x-ray response. Moreover, this region conferred x-ray inducibility to a minimal thymidine kinase gene promoter. Taken together, these results indicate that ionizing radiation induces EGR1 transcription through CArG elements.

A phorbol ester response element within the human T-cell receptor beta-chain enhancer

The activity of the T-cell receptor beta-chain gene enhancer is increased by activators of the protein kinase C pathway during T-cell activation. Analysis of mutant enhancer constructs identified two elements, beta E2 and beta E3, conferring phorbol ester inducibility. Multimerized beta E2 acted in isolation as a phorbol ester-responsive element. Both beta E2 and beta E3, which contain a consensus Ets-binding site, were shown to bind directly to the product of the c-ets-1 protooncogene. Both regions also bound a second factor, core-binding factor. Mutation of the beta E2 Ets site abolished the inducibility of the beta E2 multimer. beta E2 and beta E3 Ets site mutations also profoundly affected activity and inducibility of the enhancer. In contrast, enhancer activity but not its inducibility was affected by mutation of the beta E2 core-binding factor site. Cotransfection studies showed that Ets-1 specifically repressed activity of the multimerized beta E2 element and the complete T-cell receptor beta-chain enhancer. These data show that the T-cell receptor beta-chain enhancer responds to protein kinase C-mediated activation signals via a functional domain, composed of two elements, which contains binding sites for Ets transcription factors and which is negatively regulated by Ets-1.

Transcriptional activation by the human c-Myc oncoprotein in yeast requires interaction with Max

The c-myc protein (Myc) contains an amino-terminal transcriptional activation domain and a carboxy-terminal basic helix-loop-helix-leucine zipper (bHLH-Z) domain that directs dimerization of Myc with its partner, the max protein (Max), and promotes DNA binding to sites containing a CACGTG core consensus sequence. Despite these characteristics and the observation that Myc can modulate gene expression, a direct role for Myc or Max as transcription factors has never been demonstrated. Here we use Saccharomyces cerevisiae as an in vivo model system to show that the Myc protein is a sequence-specific transcriptional activator whose DNA binding is strictly dependent on dimerization with Max. Transactivation is mediated by the amino-terminal domain of Myc. We find that Max homodimers bind to the same DNA sequence as Myc+Max but that they fail to transactivate and thus can antagonize Myc+Max function. We also show that the Max HLH-Z domain has a higher affinity for the Myc HLH-Z domain than for itself, and suggest that the heterodimeric Myc+Max activator forms preferentially at equilibrium.

An inhibitory carboxyl-terminal domain in Ets-1 and Ets-2 mediates differential binding of ETS family factors to promoter sequences of the mb-1 gene

The mb-1 gene is expressed only during the early stages of B-lymphocyte differentiation. Here we show that the mb-1 proximal promoter region contains a functionally important binding site for members of the ETS family of DNA-binding proteins. We found that both the E26 virus-encoded v-ets and the myeloid/B-cell-specific factor PU.1 bind efficiently to this site in vitro. By contrast, Ets-1, the lymphocyte-specific cellular homologue of v-ets, and the related, more ubiquitously expressed Ets-2 protein interacted weakly with this binding site. DNA binding by both Ets-1 and Ets-2, however, could be increased 20- to 50-fold by deleting as few as 16 carboxyl-terminal amino acids. The inhibitory carboxyl-terminal amino acid sequence is highly conserved between Ets-1 and Ets-2 but is not present in either v-ets or PU.1. Replacement of the carboxyl-terminal amino acids of v-ets with those of Ets-1 decreased DNA binding by v-ets drastically. Cotranslation of Ets-1 transcripts encoding proteins of different lengths suggested that Ets-1 binds DNA as a monomer. Therefore, the carboxyl-terminal inhibitory domain appears to interfere directly with DNA binding and not with homodimerization. Finally, the functional relevance of ETS factor binding to the mb-1 promoter site was evidenced by the stimulation of transcription through this site by a v-myb-v-ets fusion protein. Together, these data suggest that one or more ETS family factors are involved in the regulation of mb-1 gene expression.

The serum response element

The promoters of many genes whose transcription is rapidly and transiently induced following growth factor or mitogen stimulation of susceptible cells contain a common regulatory element, the serum response element (SRE). As the transcription factors that bind the SRE and the signalling molecules that affect its activity are characterized in more detail, the major challenge is to elucidate the signalling pathways that link cell-surface receptors to the SRE, and to determine the mechanism by which signalling events modulate transcription factor activity.

Functional antagonism between YY1 and the serum response factor

The rapid, transient induction of the c-fos proto-oncogene by serum growth factors is mediated by the serum response element (SRE). The SRE shares homology with the muscle regulatory element (MRE) of the skeletal alpha-actin promoter. It is not known how these elements respond to proliferative and cell-type-specific signals, but the response appears to involve the binding of the serum response factor (SRF) and other proteins. Here, we report that YY1, a multifunctional transcription factor, binds to SRE and MRE sequences in vitro. The methylation interference footprint of YY1 overlaps with that of the SRF, and YY1 competes with the SRF for binding to these DNA elements. Overexpression of YY1 repressed serum-inducible and basal expression from the c-fos promoter and repressed basal expression from the skeletal alpha-actin promoter. YY1 also repressed expression from the individual SRE and MRE sequences upstream from a TATA element. Unlike that of YY1, SRF overexpression alone did not influence the transcriptional activity of the target sequence, but SRF overexpression could reverse YY1-mediated trans repression. These data suggest that YY1 and the SRF have antagonistic functions in vivo.

Functional antagonism between YY1 and the serum response factor

The rapid, transient induction of the c-fos proto-oncogene by serum growth factors is mediated by the serum response element (SRE). The SRE shares homology with the muscle regulatory element (MRE) of the skeletal alpha-actin promoter. It is not known how these elements respond to proliferative and cell-type-specific signals, but the response appears to involve the binding of the serum response factor (SRF) and other proteins. Here, we report that YY1, a multifunctional transcription factor, binds to SRE and MRE sequences in vitro. The methylation interference footprint of YY1 overlaps with that of the SRF, and YY1 competes with the SRF for binding to these DNA elements. Overexpression of YY1 repressed serum-inducible and basal expression from the c-fos promoter and repressed basal expression from the skeletal alpha-actin promoter. YY1 also repressed expression from the individual SRE and MRE sequences upstream from a TATA element. Unlike that of YY1, SRF overexpression alone did not influence the transcriptional activity of the target sequence, but SRF overexpression could reverse YY1-mediated trans repression. These data suggest that YY1 and the SRF have antagonistic functions in vivo.

Human and Drosophila homeodomain proteins that enhance the DNA-binding activity of serum response factor

Cells with distinct developmental histories can respond differentially to identical signals, suggesting that signals are interpreted in a fashion that reflects a cell's identity. How this might occur is suggested by the observation that proteins of the homeodomain family, including a newly identified human protein, enhance the DNA-binding activity of serum response factor, a protein required for the induction of genes by growth and differentiation factors. Interaction with proteins of the serum response factor family may allow homeodomain proteins to specify the transcriptional response to inductive signals. Moreover, because the ability to enhance the binding of serum response factor to DNA residues within the homeodomain but is independent of homeodomain DNA-binding activity, this additional activity of the homeodomain may account for some of specificity of action of homeodomain proteins in development.

Phosphorylation of transcription factor p62TCF by MAP kinase stimulates ternary complex formation at c-fos promoter

Transcription of the proto-oncogene c-fos is stimulated rapidly and transiently by serum growth factors and mitogens. Critical for this response is the serum-response element which is bound in vivo in a ternary complex containing the transcription factors p67SRF and p62TCF (ref. 2). Disruption of the ternary complex correlates with impaired induction by serum and phorbol ester. Mitogen-activated protein (MAP) kinase is a serine/threonine kinase which is activated 1-5 minutes after treatment of cells with mitogens and growth factors that induce re-entry into the cell cycle, making MAP kinase a candidate for the transmission of proliferative signals. Here we show that p62TCF is phosphorylated by MAP kinase in vitro and that phosphorylation results in enhanced ternary complex formation. Serum-starved Swiss 3T3 cells treated with epidermal growth factor, which induces MAP kinase in these cells, are induced to express c-fos and yield p62TCF active in ternary complex formation. In contrast, treatment of Swiss 3T3 cells with insulin, which does not activate MAP kinase under these conditions, does not lead to enhanced ternary complex formation nor does it induce c-fos transcription. Our results link the expression of the human c-fos proto-oncogene to signal transduction pathways known to be activated before its own induction.

Elk-1 protein domains required for direct and SRF-assisted DNA-binding

The Ets-related Elk-1 protein can bind to purine-rich DNA target sites in a sequence specific fashion and, in addition, can form a ternary complex with the c-fos serum response element (SRE) and the serum response factor (SRF). We demonstrate that Elk-1 can readily interchange between its different interaction partners. The amino terminal ETS-domain of Elk-1 was shown to be necessary and sufficient for direct DNA-binding activity. For ternary complex formation with the SRE and SRF, both the Elk-1 ETS-domain as well as flanking sequences up to amino acid 169 were required. Removal of sequences between the ETS-domain and amino acids 137-169 did not abolish ternary complex formation. This suggests the Elk-1 region spanning amino acids 137-169 to contain a protein-protein interaction domain. Furthermore, we have shown that a single amino acid exchange introduced into the ETS-domain can drastically alter the direct DNA-binding affinity of Elk-1 without severely affecting SRF-assisted binding to the SRE. Thus, Elk-1 requires different propensities of the ETS-domain to exert its different modes of DNA sequence recognition.

SRF and MCM1 have related but distinct DNA binding specificities

The mammalian transcription factor SRF and the yeast regulatory protein MCM1 contain DNA binding domains that are 70% identical; moreover, both proteins can bind the serum response element in the human c-fos promoter. Here we present an analysis of MCM1 sequence specificity by selection of sites from random sequence oligonucleotides. In this assay the MCM1 DNA binding domain selects binding sites containing the consensus (NotC)CCY(A/T)(A/T)(T/A)NN(A/G)G, distinct from the SRF binding consensus CC(A/T)6GG. Carboxylethylation interference analysis of a set of selected sites suggests that MCM1 contacts DNA in its major groove throughout one helical turn. These differences in specificity are largely due to sequence differences between the N terminal basic parts of the SRF and MCM1 DNA binding domains. Comparison of the relative binding affinities of MCM1 and SRF for a panel of representative binding sites showed that many high affinity MCM1 sites have negligible affinity for SRF and vice versa. Thus MCM1 and SRF have significantly different sequence specificities.

The ets gene family

Ets proteins have a conserved DNA-binding domain and regulate transcriptional initiation from a variety of cellular and viral gene promoter and enhancer elements. Some members of the Ets family, Ets-1 and Ets-2, cooperate in transcription with the AP-1 transcription factor, the product of the proto-oncogene families, fos and jun, while others, Elk-1 and SAP-1, form ternary complexes with the serum response factor (SRF). Certain ets gene family members possess transforming activity while others are activated by proviral integration in erythroleukaemias.

Interaction of murine ets-1 with GGA-binding sites establishes the ETS domain as a new DNA-binding motif

The proto-oncogene ets-1 is the founding member of a new family of eukaryotic transcriptional regulators. Using deletion mutants of murine ets-1 cDNA expressed in Escherichia coli, we show that the DNA-binding domain corresponds closely to the ETS domain, an 85-amino-acid region that is conserved among ets family members. To investigate the specificity of DNA binding of the ETS domain, we mapped the DNA contacts of a monomeric Ets-1 fragment by chemical protection and interference assays. DNA backbone interactions span a 20-nucleotide region and are localized on one face of the helix. Close phosphate and base contacts are restricted to 10 central nucleotides. Contacts map to the major groove in the center of the site. Flanking minor groove interactions also are predicted. To determine the sequence preference in the close contact zone, we selected a pool of high-affinity binding sites using a purified Ets-1 carboxy-terminal fragment. Our Ets-1-selected consensus, 5'-A/GCCGGAA/TGT/C-3', differs from the binding consensus for the Drosophila ETS domain protein E74A, suggesting that specificity of action of ets family members is mediated by the ETS domain. Compared to other well-characterized classes of DNA-binding proteins, Ets-1 produces a unique pattern of DNA contacts. These studies demonstrate that the ETS domain proteins bind DNA in a novel manner.

The SRF and MCM1 transcription factors

The mammalian transcription factor SRF (serum-response factor) and the related Saccharomyces cerevisiae transcription factor MCM1 are the prototypes of a new class of dimeric DNA-binding proteins. Their function is regulated in part by the interactions of their DNA-binding domains with accessory proteins. Recent work has advanced the functional characterization of the contributions of SRF and MCM1, and their accessory proteins to transcriptional activation.