Portal de Periódicos da CAPES

Context-dependent Transcriptional Cooperation Mediated by Cardiac Transcription Factors Csx/Nkx-2.5 and GATA-4

1999; Elsevier BV; Volume: 274; Issue: 12 Linguagem: Inglês

10.1074/jbc.274.12.8231

1083-351X

Ichiro Shiojima, Issei Komuro, Toru Oka, Yukio Hiroi, Takehiko Mizuno, Eiki Takimoto, Koshiro Monzen, Ryuichi Aikawa, Hiroshi Akazawa, Tsutomu Yamazaki, Sumiyo Kudoh, Yoshio Yazaki,

RNA modifications and cancer

Although the cardiac homeobox geneCsx/Nkx-2.5 is essential for normal heart development, little is known about its regulatory mechanisms. In a search for the downstream target genes of Csx/Nkx-2.5, we found that the atrial natriuretic peptide (ANP) gene promoter was strongly transactivated by Csx/Nkx-2.5. Deletion and mutational analyses of the ANP promoter revealed that the Csx/Nkx-2.5-binding element (NKE2) located at −240 was required for high level transactivation by Csx/Nkx-2.5. We also found that Csx/Nkx-2.5 and GATA-4 displayed synergistic transcriptional activation of the ANP promoter, and in contrast to previous reports (Durocher, D., Charron, F., Warren, R., Schwartz, R. J., and Nemer, M. (1997) EMBO J. 16, 5687–5696; Lee, Y., Shioi, T., Kasahara, H., Jobe, S. M., Wiese, R. J., Markham, B., and Izumo, S (1998) Mol. Cell. Biol. 18, 3120–3129), this synergism was dependent on binding of Csx/Nkx-2.5 to NKE2, but not on GATA-4-DNA interactions. Although GATA-4 also potentiated the Csx/Nkx-2.5-induced transactivation of the artificial promoter that contains multimerized Csx/Nkx-2.5-binding sites, Csx/Nkx-2.5 reduced the GATA-4-induced transactivation of the GATA-4-dependent promoters. These findings indicate that the cooperative transcriptional regulation mediated by Csx/Nkx-2.5 and GATA-4 is promoter context-dependent and suggest that the complexcis-trans interactions may fine-tune gene expression in cardiac myocytes. Although the cardiac homeobox geneCsx/Nkx-2.5 is essential for normal heart development, little is known about its regulatory mechanisms. In a search for the downstream target genes of Csx/Nkx-2.5, we found that the atrial natriuretic peptide (ANP) gene promoter was strongly transactivated by Csx/Nkx-2.5. Deletion and mutational analyses of the ANP promoter revealed that the Csx/Nkx-2.5-binding element (NKE2) located at −240 was required for high level transactivation by Csx/Nkx-2.5. We also found that Csx/Nkx-2.5 and GATA-4 displayed synergistic transcriptional activation of the ANP promoter, and in contrast to previous reports (Durocher, D., Charron, F., Warren, R., Schwartz, R. J., and Nemer, M. (1997) EMBO J. 16, 5687–5696; Lee, Y., Shioi, T., Kasahara, H., Jobe, S. M., Wiese, R. J., Markham, B., and Izumo, S (1998) Mol. Cell. Biol. 18, 3120–3129), this synergism was dependent on binding of Csx/Nkx-2.5 to NKE2, but not on GATA-4-DNA interactions. Although GATA-4 also potentiated the Csx/Nkx-2.5-induced transactivation of the artificial promoter that contains multimerized Csx/Nkx-2.5-binding sites, Csx/Nkx-2.5 reduced the GATA-4-induced transactivation of the GATA-4-dependent promoters. These findings indicate that the cooperative transcriptional regulation mediated by Csx/Nkx-2.5 and GATA-4 is promoter context-dependent and suggest that the complexcis-trans interactions may fine-tune gene expression in cardiac myocytes. atrial natriuretic peptide base pair(s) interleukin-5 glutathione S-transferase polymerase chain reaction electrophoretic mobility shift assay hemagglutinin polyacrylamide gel electrophoresis serum response factor Murine Csx/Nkx-2.5 is a homeobox-containing gene originally identified as a potential vertebrate homolog of theDrosophila gene tinman (1Komuro I. Izumo S. Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 8145-8149Crossref PubMed Scopus (462) Google Scholar, 2Lints T.J. Parsons L.M. Hartley L. Lyons I. Harvey R.P. Development (Camb.). 1993; 119: 419-431Crossref PubMed Google Scholar). tinmanis expressed in all mesodermal cells during germ-band elongation, but its expression is subsequently restricted to the dorsal cardiac mesoderm and, in the later developmental stage, to the dorsal vessel, an insect equivalent of the vertebrate heart (3Bodmer R. Jan L.Y. Jan Y.N. Development (Camb.). 1990; 110: 661-669PubMed Google Scholar). A loss-of-function mutant of the Drosophila tinman gene exhibits complete loss of heart formation, indicating that tinman is essential for Drosophila heart formation (4Azpiazu N. Frasch M. Genes Dev. 1993; 7: 1325-1340Crossref PubMed Scopus (634) Google Scholar, 5Bodmer R. Development (Camb.). 1993; 118: 719-729PubMed Google Scholar). The expression ofCsx/Nkx-2.5 is also highly restricted to the heart and the heart progenitor cells from the very early developmental stage when the two heart primordia are symmetrically situated in the anterior lateral mesoderm (1Komuro I. Izumo S. Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 8145-8149Crossref PubMed Scopus (462) Google Scholar, 2Lints T.J. Parsons L.M. Hartley L. Lyons I. Harvey R.P. Development (Camb.). 1993; 119: 419-431Crossref PubMed Google Scholar), and targeted disruption of murineCsx/Nkx-2.5 results in embryonic lethality due to the abnormal looping morphogenesis of the primary heart tube (6Lyons I. Parsons L.M. Hartley L. Li R. Andrews J.E. Robb L. Harvey R.P Genes Dev. 1995; 9: 1654-1666Crossref PubMed Scopus (956) Google Scholar). Furthermore, Csx/Nkx-2.5 family genes are also identified in various vertebrate species from zebrafish to humans (7Tonissen K.F. Drysdale T.A. Lints T.J. Harvey R.P. Kreig P.A. Dev. Biol. 1993; 162: 325-328Crossref Scopus (195) Google Scholar, 8Schultheiss T.M. Xydas S. Lassar A.B. Development (Camb.). 1995; 121: 4203-4214PubMed Google Scholar, 9Chen J.N. Fishman M.C. Development (Camb.). 1996; 122: 3809-3816PubMed Google Scholar, 10Shiojima I. Komuro I. Mizuno T. Aikawa R. Akazawa H. Oka T. Yamazaki T. Yazaki Y. Circ. Res. 1996; 79: 920-929Crossref PubMed Scopus (64) Google Scholar). These results indicate that Csx/Nkx-2.5 is also essential for normal heart development and morphogenesis in vertebrates like thetinman gene in Drosophila and suggest that the regulatory mechanism of heart development controlled by Csx/Nkx-2.5 is highly conserved in evolution. To clarify the molecular framework of vertebrate cardiogenesis, the identification of upstream regulatory factors that control the expression of the Csx/Nkx-2.5 gene and the downstream targets of the Csx/Nkx-2.5 protein is necessary (11Bodmer R. Trends. Cardiovasc. Med. 1995; 5: 21-28Crossref PubMed Scopus (157) Google Scholar, 12Olson E.N. Srivastava D. Science. 1996; 272: 671-676Crossref PubMed Scopus (392) Google Scholar, 13Harvey R.P. Dev. Biol. 1996; 178: 203-216Crossref PubMed Scopus (494) Google Scholar). InDrosophila, tinman gene expression is controlled by several upstream factors including twist, decapentaplegic, and wingless (3Bodmer R. Jan L.Y. Jan Y.N. Development (Camb.). 1990; 110: 661-669PubMed Google Scholar, 14Wu X. Golden K. Bodmer R. Dev. Biol. 1995; 169: 619-628Crossref PubMed Scopus (189) Google Scholar, 15Staehling-Hampton K. Hoffmann F.M. Baylies M.K. Rushton E. Bate M. Nature. 1994; 372: 783-786Crossref PubMed Scopus (201) Google Scholar, 16Frasch M. Nature. 1995; 374: 464-467Crossref PubMed Scopus (374) Google Scholar), and tinman directly regulates the expression ofD-MEF2 via the cardiac enhancer in the D-MEF2gene (17Gajewski K. Kim Y. Lee Y.M. Olson E.N. Schulz R.A. EMBO J. 1997; 16: 515-522Crossref PubMed Scopus (128) Google Scholar). In vertebrates, however, there is only limited information regarding the upstream factors or the downstream targets of Csx/Nkx-2.5. Among the possible regulatory factors of cardiac development, bone morphogenetic proteins have recently been identified as potential inducing signals of Csx/Nkx-2.5 that are secreted from the endoderm (18Schultheiss T.M. Burch J.B.E. Lassar A.B. Genes Dev. 1997; 11: 451-462Crossref PubMed Scopus (583) Google Scholar). As for the downstream targets of Csx/Nkx-2.5, the expression of ventricular myosin light chain 2, CARP (cardiac ankyrin-repeatprotein), and a basic helix-loop-helix transcription factor (eHAND) is significantly reduced in Csx/Nkx-2.5 knockout mice (6Lyons I. Parsons L.M. Hartley L. Li R. Andrews J.E. Robb L. Harvey R.P Genes Dev. 1995; 9: 1654-1666Crossref PubMed Scopus (956) Google Scholar, 19Zou Y. Evans S. Chen J. Kuo H.C. Harvey R.P. Chien K.R. Development (Camb.). 1997; 124: 793-804PubMed Google Scholar, 20Biben C. Harvey R.P. Genes Dev. 1997; 11: 1357-1369Crossref PubMed Scopus (271) Google Scholar), suggesting that Csx/Nkx-2.5 may, either directly or indirectly, regulate the expression of these genes. In addition, since CARP promoter activity is down-regulated by the overexpression of the dominant-negative form of Csx/Nkx-2.5 in cultured cardiac myocytes (19Zou Y. Evans S. Chen J. Kuo H.C. Harvey R.P. Chien K.R. Development (Camb.). 1997; 124: 793-804PubMed Google Scholar), CARP may be a direct target of Csx/Nkx-2.5, although the precisecis-regulatory sequences in the CARP promoter are not known. In addition to Csx/Nkx-2.5, several transcription factors have also been identified as potential regulators of cardiac development. Among these molecules, MEF2C and GATA-4 are thought to be involved in the early step of cardiac development because both of these factors appear in the precardiac mesoderm almost simultaneously with Csx/Nkx-2.5 and earlier than any other known transcription factors implicated in cardiogenesis (21Edmondson D.G. Lyons G.E. Martin J.M. Olson E.N. Development (Camb.). 1994; 120: 1251-1263Crossref PubMed Google Scholar, 22Heikinheimo M. Scandrett J.M. Wilson D.B. Dev. Biol. 1994; 164: 361-373Crossref PubMed Scopus (250) Google Scholar). MEF2C belongs to the MEF2 (myocyteenhancer factor-2) subfamily of MADS box transcription factors and binds to the AT-rich element in the regulatory regions of numerous muscle-specific genes (23Olson E.N. Perry M. Schulz R.A. Dev. Biol. 1995; 172: 2-14Crossref PubMed Scopus (315) Google Scholar). GATA-4 is a member of the cardiac GATA subfamily, which is composed of GATA-4/5/6 and binds to the WGATAR motif in the promoter regions of cardiac- or gut-specific genes (24Evans T. Trends Cardiovasc. Med. 1997; 7: 75-83Crossref PubMed Scopus (62) Google Scholar). Targeted disruption of MEF2Cresults in embryonic lethal phenotype due to the right ventricular dysplasia (25Lin Q. Schwarz J. Bucana C. Olson E.N. Science. 1997; 276: 1404-1407Crossref PubMed Scopus (784) Google Scholar), and GATA-4 −/− mice are also embryonic lethal because of the failure of the fusion of cardiac primordia at the midline due to the ventral folding/fusion defect of the developing embryo (26Kuo C.T. Morrisey E.E. Anandappa R. Sigrist K. Lu M.M. Parmacek M.S. Soudais C. Leiden J.M. Genes Dev. 1997; 11: 1048-1060Crossref PubMed Scopus (872) Google Scholar, 27Molkentin J.D. Lin Q. Duncan S.A. Olson E.N. Genes Dev. 1997; 11: 1061-1072Crossref PubMed Scopus (959) Google Scholar). All these results indicate that Csx/Nkx-2.5, MEF2C, and GATA-4 are essential transcription factors required for normal heart development and suggest that multiple transcription factors coordinately regulate the process of cardiogenesis. In this study, to clarify the genetic pathways that control vertebrate heart development, we tried to find the direct downstream target genes of Csx/Nkx-2.5. Cotransfection of the expression plasmid of human Csx/Nkx-2.5 and the reporter genes with promoters of various cardiac-specific genes linked to the firefly luciferase gene revealed that the atrial natriuretic peptide (ANP)1 gene promoter is strongly transactivated by Csx/Nkx-2.5 via the high affinity Csx/Nkx-2.5-binding element termed NKE2 (Nkx-2.5 responseelement-2) located at −250 in the 5′-flanking region of the rat ANP gene. Because the GATA motif is located near NKE2, we cotransfected the expression plasmids of Csx/Nkx-2.5 and GATA-4 simultaneously with the 300-bp ANP promoter-containing reporter gene and found that the ANP promoter is synergistically transactivated by these two factors. Deletion and mutational analyses indicated that this synergism is dependent on NKE2, but not on the two GATA sites within this proximal promoter. The artificial promoter construct containing multimerized Csx/Nkx-2.5-binding sequences was also synergistically transactivated by Csx/Nkx-2.5 and GATA-4. However, Csx/Nkx-2.5 and GATA-4 negatively cooperated on the promoters that are dependent on GATA-4. Co-immunoprecipitation experiments and glutathioneS-transferase (GST) pull-down assay demonstrated that Csx/Nkx-2.5 and GATA-4 interact with each other both in vivoand in vitro, and electrophoretic mobility shift assay revealed that the DNA-binding affinity of GATA-4 was reduced in the presence of Csx/Nkx-2.5, whereas that of Csx/Nkx-2.5 was not affected by GATA-4. Our present results indicate that the Csx/Nkx-2.5-GATA-4 synergism is differentially regulated depending on the context of thecis-elements with which the two factors interact and suggest that the complex interactions among transcription factors and target DNAs fine-tune gene expression in cardiac myocytes and regulate development of the heart. The expression plasmid of human Csx/Nkx-2.5 (pEFSA-CSX1) and that of human GATA-4 (pSSRa-hGATA4) were previously described (10Shiojima I. Komuro I. Mizuno T. Aikawa R. Akazawa H. Oka T. Yamazaki T. Yazaki Y. Circ. Res. 1996; 79: 920-929Crossref PubMed Scopus (64) Google Scholar, 28Yamagata T. Nishida J. Sakai R. Tanaka T. Honda H. Hirano N. Mano H. Yazaki Y. Hirai H. Mol. Cell. Biol. 1995; 15: 3830-3839Crossref PubMed Scopus (80) Google Scholar). To construct luciferase reporter genes for deletion analysis, various length of ANP genomic fragments were amplified by polymerase chain reaction (PCR) with primers containing aSacI site in the forward primer and an XhoI site in the reverse primer. For mutational analyses, two GATA elements (located at −120 and −270) and one Csx/Nkx-2.5-binding element (NKE2, located at −250) within the ANP(300)-luc construct containing the 300-bp 5′-flanking region of the ANP gene were mutated either alone or in combination. To introduce mutations into NKE2 and/or the GATA element at −270, forward primers that contained corresponding mutations were used for PCR. A two-step PCR method was used to introduce mutations into the GATA element at −120. PCR products were digested with SacI and XhoI; subcloned into theSacI/XhoI site of PGV-B (TOYO INKI), a promoter-less vector containing the luciferase gene; and verified by sequencing. The 4x(TTF-1)-tk-luc reporter gene (which contains four tandem copies of Csx/Nkx-2.5-binding elements), the 6x(GATA)-tk-luc reporter gene (which contains six tandem copies of GATA-4-binding elements), and a reporter gene containing the proximal 400 bp of the human IL-5 gene promoter were previously described (10Shiojima I. Komuro I. Mizuno T. Aikawa R. Akazawa H. Oka T. Yamazaki T. Yazaki Y. Circ. Res. 1996; 79: 920-929Crossref PubMed Scopus (64) Google Scholar, 28Yamagata T. Nishida J. Sakai R. Tanaka T. Honda H. Hirano N. Mano H. Yazaki Y. Hirai H. Mol. Cell. Biol. 1995; 15: 3830-3839Crossref PubMed Scopus (80) Google Scholar). COS-7 cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum. Transient transfections were performed 24 h after plating with the standard calcium phosphate method. For each 35-mm dish, the reporter construct, expression vectors of CSX1 and/or GATA-4, SV40-βgal as an internal control, and the appropriate amounts of parental expression plasmid were transfected with the total amount of DNA kept constant at 2.5 μg. Luciferase activities were measured 48 h after transfection with a Berthold Lumat LB9501 luminometer, and differences in transfection efficiency were corrected relative to the level of β-galactosidase activity. Experiments were repeated at least three times in triplicate, and representative data are shown. PCR-amplified cDNA fragments corresponding to the coding region of CSX1 derivatives and GATA-4 derivatives were subcloned into pSPUTK vector (Promega) containing the SP6 RNA polymerase promoter and the sequence of the Xenopus β-globin 5′-untranslated region. Proteins were transcribed and translated in vitro using the TNT SP6 coupled reticulocyte lysate system (Promega) with or without the [35S]methionine. Double-stranded oligonucleotides corresponding to the TTF-1 (thyroidtranscription factor-1)-binding sequence, NKE, NKE2, and the GATA site with or without mutations were synthesized with GATC overhanging at the 5′ terminus of each oligonucleotide. The two complementary oligonucleotides were annealed and labeled with [α-32P]dCTP using Klenow enzyme. Labeled probes were incubated with 5 μl of programmed or unprogrammed reticulocyte lysate and 2 μg of poly(dI-dC) in 20 μl of binding buffer (10 mm Tris-HCl (pH 7.5), 50 mm NaCl, 10% glycerol, 0.5 mm dithiothreitol, and 0.05% Nonidet P-40) for 30 min at room temperature. The protein/DNA mixture was resolved on 5% polyacrylamide gel in 0.5× Tris borate/EDTA buffer at 4 °C for 2 h at 150 V. The expression plasmid of HA-tagged CSX1 (pEFSA-HA-CSX1) was described previously (10Shiojima I. Komuro I. Mizuno T. Aikawa R. Akazawa H. Oka T. Yamazaki T. Yazaki Y. Circ. Res. 1996; 79: 920-929Crossref PubMed Scopus (64) Google Scholar). To generate the expression plasmid of Myc-tagged GATA-4 (pBK-CMVMyc-GATA4), the corresponding cDNA fragment was amplified by PCR and subcloned into the XbaI/ClaI site of modified pBK-CMV vector (Stratagene), which contains the Myc tag sequence at the NheI/XbaI site. These constructs were transiently transfected into COS-7 cells by the calcium phosphate method, and 48 h after transfection, nuclear extracts were prepared as described previously (10Shiojima I. Komuro I. Mizuno T. Aikawa R. Akazawa H. Oka T. Yamazaki T. Yazaki Y. Circ. Res. 1996; 79: 920-929Crossref PubMed Scopus (64) Google Scholar). Nuclear extracts were then incubated with anti-HA monoclonal antibody 12CA5 in NTEN binding buffer (150 mm NaCl, 50 mm Tris (pH 7.5), 0.5 mm EDTA, 0.5% Nonidet P-40, 1 mmdithiothreitol, 0.5 mm phenylmethylsulfonyl fluoride, and 0.25% bovine serum albumin) for 2 h at 4 °C with gentle rotation, and the immune complex was precipitated with protein A-Sepharose beads, washed three times with NTEN binding buffer, resuspended in sample buffer, and subjected to SDS-PAGE. Resolved proteins were transferred onto Hybond ECL nitrocellulose membrane (Amersham Pharmacia Biotech) and immunoblotted with anti-Myc polyclonal antibody. Horseradish peroxidase-conjugated anti-rabbit IgG was used as the secondary antibody, and the immune complex was visualized using the ECL detection kit (Amersham Pharmacia Biotech). PCR-amplified cDNA fragments corresponding to full-length CSX1 and CSX1 derivatives were subcloned in frame into the BamHI/EcoRI site of pGEX-2T (Amersham Pharmacia Biotech). To express GST fusion proteins in bacteria, JM109 cells were transformed with pGEX-2T-CSX1 constructs, and fusion protein expression was induced by 0.1 mmisopropyl-β-d-thiogalactopyranoside for 5 h at 25 °C. Cells were harvested, resuspended in phosphate-buffered saline containing 1% Triton X-100, sonicated, and centrifuged. GST fusion proteins were purified by binding to glutathione-Sepharose 4B beads (Amersham Pharmacia Biotech) according to the manufacturer's directions. The concentrations and purity of the glutathione-bound GST-CSX1 proteins were estimated by SDS-PAGE and Coomassie Blue staining. In vitro interaction assays were performed with GST-CSX1 fusion proteins and in vitrotranslated GATA-4 derivatives. GST or GST-CSX1 fusion proteins immobilized on glutathione-Sepharose 4B beads and in vitrotranslated GATA-4 derivatives labeled with [35S]methionine were mixed in NTEN binding buffer, incubated for 2 h at 4 °C with gentle rotation, and washed three times with binding buffer. Beads were resuspended in SDS-PAGE sample buffer, and GATA-4 derivatives bound to GST-CSX1 fusion proteins were resolved by SDS-PAGE and visualized by fluorography. To search for the direct downstream target genes of the Csx/Nkx-2.5 protein, we cotransfected luciferase reporter constructs containing various cardiac-specific gene promoters and the Csx/Nkx-2.5 expression vector into COS-7 cells and examined whether the transcriptional activation of native promoters was observed by the overexpression of Csx/Nkx-2.5. Among the reporter constructs tested, Csx/Nkx-2.5 most strongly transactivated the ANP-luc reporter containing the 600-bp 5′-flanking region of the rat ANP gene (Fig. 1), suggesting that the ANP gene is a downstream target of Csx/Nkx-2.5. To identify the cis-regulatory elements in the ANP promoter that mediate the transcriptional activation by Csx/Nkx-2.5, luciferase reporter constructs containing various lengths of the ANP promoter region were cotransfected with the Csx/Nkx-2.5 expression vector. Deletion of the ANP 5′-flanking region from −600 to −390 did not significantly change the -fold activation of the ANP promoter by Csx/Nkx-2.5. However, deletion between −390 and −160 markedly reduced the -fold induction of the promoter activity, although further deletion up to −125 had little effect (Fig.2 A), indicating that the Csx/Nkx-2.5-responsive element is situated between −390 and −160 of the ANP 5′-flanking region. Because the consensus Csx/Nkx-2.5-binding element (TGAAGTG) was located at −250 in the ANP promoter, we speculated that this Csx/Nkx-2.5-binding element might mediate the transcriptional activation of the ANP promoter by the Csx/Nkx-2.5 protein. To examine this possibility, the deletion and mutant constructs shown in Fig.2 B were cotransfected with the expression vector of Csx/Nkx-2.5. Deletion or mutation of the Csx/Nkx-2.5-binding element located at −250 in the ANP promoter markedly reduced the -fold induction of transactivation by Csx/Nkx-2.5, whereas deletion up to −270 of the ANP promoter, which did not affect the Csx/Nkx-2.5-binding element, had little effect on the transactivation by Csx/Nkx-2.5 (Fig.2 B). Because previous reports have shown that Csx/Nkx-2.5 binds to another Nkx-2.5-responsive element called NKE located at −80 of the ANP promoter, we designated the Csx/Nkx-2.5-binding element located at −250 as NKE2. Our present data suggest that NKE2 is required for high level transactivation of the ANP promoter by Csx/Nkx-2.5. To examine whether Csx/Nkx-2.5 binds to NKE2, EMSA was performed with in vitro translated Csx/Nkx-2.5 protein and oligonucleotide probes corresponding to the TTF-1-binding element, wild-type NKE2, mutant NKE2, and wild-type NKE. Under our experimental conditions, the Csx/Nkx-2.5 protein bound to NKE2 with almost equal binding affinity compared with the TTF-1-binding element, although it bound to NKE with much lower binding affinity and did not bind to mutant NKE2 (Fig. 2, B, right; and C, right). Comparison of the sequences of the human, mouse, and rat ANP promoters revealed that NKE2 is highly conserved in these species (Fig. 2 C, left). Together, these results strongly suggest that NKE2, as well as proximal NKE, plays a critical role in the transcriptional regulation of the ANP gene by Csx/Nkx-2.5. Because the GATA site was located at −280 just distal from NKE2, we examined whether Csx/Nkx-2.5 and GATA-4 cooperatively regulate the transcription of the ANP gene. When the ANP(300)-luc construct containing the 300-bp 5′-flanking region of the ANP promoter was cotransfected with the Csx/Nkx-2.5 expression plasmid alone, modest transactivation of the ANP promoter was observed. When ANP(300)-luc was cotransfected with the GATA-4 expression plasmid alone, no significant increase in the promoter activity was observed, suggesting that GATA-4 is a weak transactivator of the ANP promoter under our assay conditions. However, overexpression of both GATA-4 and Csx/Nkx-2.5 induced much stronger activity of the ANP promoter than that induced by the expression of Csx/Nkx-2.5 alone (Fig.3 A, left), suggesting that Csx/Nkx-2.5 and GATA-4 synergistically transactivate the ANP promoter. The apparent cooperativity of Csx/Nkx-2.5 and GATA-4 was further investigated using deletion and mutant constructs of the ANP promoter. Removal of the region between −300 and −270 of the ANP promoter, which resulted in the deletion of the GATA site at −280, had little effect on the Csx/Nkx-2.5-GATA-4 synergism as well as Csx/Nkx-2.5-induced activation (Fig. 3 A, middle). However, deletion of the promoter region between −270 and −240, in which NKE2 is located, abolished not only Csx/Nkx-2.5-induced activation of the ANP promoter, but also the cooperativity between Csx/Nkx-2.5 and GATA-4 (Fig.3 A, right), suggesting that NKE2, but not the GATA site at −280, is required for the Csx/Nkx-2.5-GATA-4 transcriptional synergism. To test whether NKE2 (but not GATA elements) is important for the Csx/Nkx-2.5-GATA-4 synergism, mutations were introduced into NKE2 and/or two GATA sites within the ANP(300)-luc construct either alone or in combination, and the effects of coexpression of Csx/Nkx-2.5 and GATA-4 on these reporter genes were examined. Cotransfection of the expression plasmids of Csx/Nkx-2.5 and GATA-4 with wild-type or mutant ANP(300)-luc constructs revealed that NKE2, but not the two GATA sites, was required for the synergistic transactivation of the ANP promoter by the two factors (Fig.3 B). For example, even when the two GATA sites were mutated, Csx/Nkx-2.5-GATA-4 synergism was still observed if NKE2 remained intact (Fig. 3 B, ANP(300-m5-)-luc), whereas Csx/Nkx-2.5-GATA-4 synergism was abolished by the mutation of NKE2 irrespective of the presence or absence of the intact GATA sites (ANP(300-m2-)-luc, ANP(300-m3-)-luc,ANP(300-m6-)-luc, and ANP(300-m7-)-luc). The dependence of Csx/Nkx-2.5-GATA-4 transcriptional synergism on the Csx/Nkx-2.5-binding site was also investigated with the artificial reporter construct containing multimerized Csx/Nkx-2.5-binding sites (4x(TTF-1)-tk-luc). Cotransfection experiments revealed that 4x(TTF-1)-tk-luc is synergistically transactivated by Csx/Nkx-2.5 and GATA-4 (Fig. 4). All these results collectively suggest that the Csx/Nkx-2.5-binding site is essential, but that the GATA-4-binding site is dispensable, for the transcriptional synergy between Csx/Nkx-2.5 and GATA-4. To further investigate the Csx/Nkx-2.5-GATA-4 transcriptional cooperation, we examined the effects of the coexpression of Csx/Nkx-2.5 and GATA-4 on GATA-4-dependent promoters. For this purpose, we utilized the artificial promoter that contains multimerized GATA sites (6x(GATA)-tk-luc) and the IL-5 promoter that has been shown to be positively regulated by GATA-4 and does not contain the putative Csx/Nkx-2.5-binding sequence in the proximal promoter region. As expected, when GATA-4 was cotransfected with the 6x(GATA)-tk-luc reporter gene, modest transcriptional activation was observed, whereas Csx/Nkx-2.5 did not affect the activity of the reporter gene. Unexpectedly, however, cotransfection of Csx/Nkx-2.5 with GATA-4 partially inhibited the GATA-4-induced transactivation of the reporter gene (Fig. 5, left). Essentially the same results were obtained when the IL-5 promoter-containing reporter gene was examined (Fig. 5,right). These results suggest that the transcriptional cooperation between Csx/Nkx-2.5 and GATA-4 is promoter context-dependent; these two factors exhibit positive cooperation on Csx/Nkx-2.5-dependent promoters and negative cooperation on GATA-4-dependent promoters. To examine whether Csx/Nkx-2.5 and GATA-4 directly interact with each other in vivo, co-immunoprecipitation experiments were performed. COS-7 cells were transiently transfected with HA-tagged Csx/Nkx-2.5 and Myc-tagged GATA-4, and nuclear extracts were immunoprecipitated with anti-HA antibody, followed by immunoblotting with anti-Myc antibody. Western blot analysis revealed that Myc-GATA-4 protein was co-immunoprecipitated with HA-Csx/Nkx-2.5 (Fig.6 A), indicating the interaction of HA-Csx/Nkx-2.5 and Myc-GATA-4 in vivo, either direct or indirect. To further confirm the direct interaction of these two factors, GST pull-down assay was performed with GST-Csx/Nkx-2.5 derivatives andin vitro translated GATA-4 derivatives. Full-length GST-Csx/Nkx-2.5 immobilized on glutathione-Sepharose beads could retainin vitro translated GATA-4, whereas GST protein could not bind to GATA-4, indicating that Csx/Nkx-2.5 directly interacts with GATA-4 in vitro (Fig. 6 B). Examination of the binding of GATA-4 derivatives to full-length GST-Csx/Nkx-2.5 revealed that the zinc-finger domain of GATA-4 (GATA-4(ZF)) is sufficient for interacting with Csx/Nkx-2.5. Since GATA-4(ZF) bound to GST-Csx/Nkx-2.5(HD), the homeodomain of Csx/Nkx-2.5 is sufficient for interacting with GATA-4 (Fig. 6 C). To determine the mechanism of positive and negative cooperation of Csx/Nkx-2.5 and GATA-4 on Csx/Nkx-2.5- and GATA-4-dependent promoters, respectively, the DNA-binding activities of Csx/Nkx-2.5 and GATA-4 were examined in the presence or absence of each other. EMSA revealed that the DNA-binding activity of Csx/Nkx-2.5 was not affected by GATA-4 (Fig. 7, left), whereas that of GATA-4 was reduced when the Csx/Nkx-2.5 protein coexisted (right). These results suggest that the negative cooperation of Csx/Nkx-2.5 and GATA-4 on GATA-4-dependent promoters is at least in part due to the reduced DNA-binding affinity of GATA-4 in the presence of Csx/Nkx-2.5, whereas the positive cooperation of Csx/Nkx-2.5 and GATA-4 on Csx/Nkx-2.5-dependent promoters is not due to the cooperative binding of these factors to DNA. In this study, we have obtained the following results. (i) The ANP promoter is strongly transactivated by Csx/Nkx-2.5. (ii) High level transactivation of the ANP promoter by Csx/Nkx-2.5 is dependent on NKE2, a Csx/Nkx-2.5-binding element located at −250. (iii) NKE2 is a high affinity binding site for Csx/Nkx-2.5 and is highly conserved in evolution. (iv) Csx/Nkx-2.5 and GATA-4 synergistically transactivate the ANP promoter. (v) Cooperative transcriptional activation of the ANP promoter by Csx/Nkx-2.5 and GATA-4 is mediated by NKE2, but is independent of GATA-4-DNA interaction. (vi) Cooperative transactivation by Csx/Nkx-2.5 and GATA-4 is also observed on the artificial promoter containing multimerized Csx/Nkx-2.5-binding sites. (vii) Csx/Nkx-2.5 and GATA-4 exhibit negative cooperation on the GATA-4-dependent promoters such as the human IL-5 gene or the artificial promoter containing multimerized GATA sites. (viii) Csx/Nkx-2.5 and GATA-4 directly interact with each other both in vivo and in vitro, and the homeodomain of Csx/Nkx-2.5 and the zinc-finger domain of GATA-4 are required for the interaction. (viii) Csx/Nkx-2.5 reduces the DNA-binding affinity of GATA-4. Our initial objective in this study was to identify the direct downstream targets of Csx/Nkx-2.5 in the genetic pathway of heart formation. Among the various cardiac-specific promoters tested, ANP was strongly transactivated by Csx/Nkx-2.5, and high level transactivation of the ANP promoter by Csx/Nkx-2.5 was mediated by NKE2, the Csx/Nkx-2.5-binding site located at −250 in the rat ANP promoter. Recently, other investigators have shown that Csx/Nkx-2.5 transactivates the ANP promoter via NKE, the proximal Csx/Nkx-2.5-binding element located at −80 (29Durocher D. Charron F. Warren R. Schwartz R.J. Nemer M. EMBO J. 1997; 16: 5687-5696Crossref PubMed Scopus (550) Google Scholar). Consistent with their results, modest transcriptional activation of the ANP promoter by Csx/Nkx-2.5 was observed when NKE2 was deleted or mutated. Together, these results suggest that the activity of the ANP promoter is regulated by Csx/Nkx-2.5 via multiple cis-elements. ANP is one of the few genes whose promoter has been identified to have Csx/Nkx-2.5-binding elements and to be transactivated by Csx/Nkx-2.5. In addition, endogenous ANP mRNA is increased in the hearts of human Csx/Nkx-2.5-overexpressing mice recently generated in our laboratory, 2E. Takimoto, I. Komuro, I. Shiojima, and T. Mizuno, unpublished observations. strongly suggesting that ANP is a direct target gene of Csx/Nkx-2.5 in vivo. In Drosophila, D-MEF2 is the only known direct target of tinman (17Gajewski K. Kim Y. Lee Y.M. Olson E.N. Schulz R.A. EMBO J. 1997; 16: 515-522Crossref PubMed Scopus (128) Google Scholar).D-MEF2 expression in cardiac cells is positively regulated by tinman, and the transactivation of D-MEF2 by tinman is mediated via the two tinman-binding elements in the cardiac enhancer located in the 5′-flanking region of the D-MEF2gene. The tinman-binding elements in the D-MEF2 enhancer and NKE2 completely match the consensus high affinity binding sequence for Csx/Nkx-2.5, TNAAGTG (30Chen C.Y. Schwartz R.J. J. Biol. Chem. 1995; 270: 15628-15633Crossref PubMed Scopus (271) Google Scholar), whereas two Csx/Nkx-2.5-binding sites in NKE do not (29Durocher D. Charron F. Warren R. Schwartz R.J. Nemer M. EMBO J. 1997; 16: 5687-5696Crossref PubMed Scopus (550) Google Scholar). Our EMSA analysis also showed that the binding affinity of NKE2 is much higher than that of NKE. These results suggest that NKE2 is a critical cis-regulatory element in the induction and activation of ANP gene. The existence of the GATA site near NKE2 prompted us to examine the possible cooperative action of Csx/Nkx-2.5 and GATA-4. Although GATA-4 was a relatively weak transactivator of the ANP promoter under our experimental conditions, the simultaneous overexpression of GATA-4 with Csx/Nkx-2.5 markedly potentiated the Csx/Nkx-2.5-induced transactivation of the ANP promoter, indicating that there exists a positive cooperation between these two factors. While this manuscript was in preparation, Durocher et al.(29Durocher D. Charron F. Warren R. Schwartz R.J. Nemer M. EMBO J. 1997; 16: 5687-5696Crossref PubMed Scopus (550) Google Scholar) reported that Csx/Nkx-2.5 could cooperate with GATA-4 to activate transcription of the ANP promoter. Our results are essentially similar to theirs, but they differ in two respects. (i) They used a reporter construct containing the 130-bp 5′-flanking region of the rat ANP promoter and observed the cooperative transcriptional activation by Csx/Nkx-2.5 and GATA-4, whereas we could not detect such synergism when NKE2 at −250 was deleted or mutated; and (ii) they showed that DNA binding of the two factors is required for synergistic transcriptional activation, whereas we could detect the cooperative action even when GATA-4 sites were deleted or mutated. Although the reasons for such discrepancies are unclear at present, they may be due to the differences in the cell types or the reporter and effector constructs used in the transfection experiments. The finding that the positive cooperation of Csx/Nkx-2.5 and GATA-4 on the ANP promoter was independent of GATA-4-DNA interaction prompted us to examine the effects of coexpression of these two factors on the Csx/Nkx-2.5-dependent artificial promoter, 4x(TTF-1)-tk-luc. Cotransfection experiments revealed that Csx/Nkx-2.5 and GATA-4 also exhibited positive cooperation on this promoter, although GATA-4 alone induced no significant increase in the promoter activity. Essentially the same results were recently reported by Sepulveda et al. (31Sepulveda J.L. Belaguli N. Nigam V. Chen C.Y. Nemer M. Schwartz R.J. Mol. Cell. Biol. 1998; 18: 3405-3415Crossref PubMed Scopus (277) Google Scholar) regarding the cardiac α-actin promoter, which has no functional GATA elements and is not significantly activated by overexpression of GATA-4. These results suggest that the Csx/Nkx-2.5-GATA-4 synergism depends on Csx/Nkx-2.5-DNA interaction and Csx/Nkx-2.5-GATA-4 association, but not on GATA-4-DNA interaction. If the direct protein-protein interaction of Csx/Nkx-2.5 and GATA-4 mediates the transcriptional synergy, one might expect that these two factors also exhibit positive transcriptional cooperation on GATA-4-dependent promoters. Unexpectedly, however, Csx/Nkx-2.5 and GATA-4 displayed negative cooperation on both the artificial and native promoters that are positively regulated by GATA-4, but not by Csx/Nkx-2.5. Together with the findings on the ANP promoter and the 4x(TTF-1)-tk-luc reporter gene, these results suggest that the Csx/Nkx-2.5-GATA-4 transcriptional cooperation depends on the context of the promoter; their cooperation is positive when the promoter contains Csx/Nkx-2.5-binding sites irrespective of the presence or absence of GATA sites and negative when the promoter contains only GATA sites. GST pull-down assay and co-immunoprecipitation experiments demonstrated that Csx/Nkx-2.5 and GATA-4 directly interact with each other bothin vivo and in vitro via the homeodomain of Csx/Nkx-2.5 and the zinc-finger domain of GATA-4, which is consistent with recent reports (31Sepulveda J.L. Belaguli N. Nigam V. Chen C.Y. Nemer M. Schwartz R.J. Mol. Cell. Biol. 1998; 18: 3405-3415Crossref PubMed Scopus (277) Google Scholar, 32Lee Y. Shioi T. Kasahara H. Jobe S.M. Wiese R.J. Markham B. Izumo S. Mol. Cell. Biol. 1998; 18: 3120-3129Crossref PubMed Scopus (245) Google Scholar). Several transcription factors have been shown to associate with Csx/Nkx-2.5, and the domains of the Csx/Nkx-2.5 protein required for the association are different depending on the interacting partner. For example, the serum response factor (SRF) associates with Csx/Nkx-2.5 via the MADS box of SRF and the homeodomain of Csx/Nkx-2.5 and synergistically transactivates the cardiac α-actin gene promoter, and this cooperative action of SRF and Csx/Nkx-2.5 requires SRF-DNA interaction, but is independent of Csx/Nkx-2.5-DNA interaction (33Chen C.Y. Schwartz R.J. Mol. Cell. Biol. 1996; 16: 6372-6384Crossref PubMed Google Scholar). Our previous study (10Shiojima I. Komuro I. Mizuno T. Aikawa R. Akazawa H. Oka T. Yamazaki T. Yazaki Y. Circ. Res. 1996; 79: 920-929Crossref PubMed Scopus (64) Google Scholar) suggested that the N-terminal domain of human Csx/Nkx-2.5 interacts with other nuclear factor(s) and mediates the transactivation activity without directly binding to DNA. Furthermore, we have recently isolated a multiple zinc finger-containing putative transcription factor that interacts with the N-terminal domain and the homeodomain of Csx/Nkx-2.5 using a yeast two-hybrid interaction screen. 3Y. Hiroi, S. Kudoh, I. Shiojima, and I. Komuro, unpublished observations. Taken together, these results suggest that the transcriptional regulation by Csx/Nkx-2.5 may be controlled by multiple protein-protein interactions mediated by distinct interaction domains of Csx/Nkx-2.5. In our EMSA analysis with in vitro translated GATA-4 and Csx/Nkx-2.5, we detected a reduced DNA-binding affinity of GATA-4 in the presence of Csx/Nkx-2.5. Sepulveda et al. (31Sepulveda J.L. Belaguli N. Nigam V. Chen C.Y. Nemer M. Schwartz R.J. Mol. Cell. Biol. 1998; 18: 3405-3415Crossref PubMed Scopus (277) Google Scholar) recently reported that GATA-4 slightly increased the DNA-binding affinity of Csx/Nkx-2.5. However, we could not detect such an increase in the DNA-binding affinity of Csx/Nkx-2.5 in the presence of GATA-4, although Csx/Nkx-2.5 and GATA-4 might affect each other's DNA-binding affinity in an in vivo situation. Our results suggest that the negative cooperation of Csx/Nkx-2.5 and GATA-4 on GATA-4-dependent promoters is at least in part due to the decrease in the DNA-binding affinity of GATA-4 in the presence of Csx/Nkx-2.5. One general mechanism for positive transcriptional synergy is the cooperative binding of activators to DNA. For example, the homeodomain protein Phox directly binds to SRF and increases the DNA-binding affinity of SRF (34Grueneberg D.A. Natesan S. Alexandre C. Gilman M.Z. Science. 1992; 257: 1089-1095Crossref PubMed Scopus (256) Google Scholar). Likewise, Drosophila extradenticle associates with Ultrabithorax and raises the DNA-binding affinity of the extradenticle-Ultrabithorax protein complex (35Chan S.K. Jaffe L. Capovilla M. Botas J. Mann R.S. Cell. 1994; 78: 603-615Abstract Full Text PDF PubMed Scopus (333) Google Scholar, 36van Dijk M.A. Murre C. Cell. 1994; 78: 617-624Abstract Full Text PDF PubMed Scopus (235) Google Scholar, 37Rauskolb C. Wieschaus E. EMBO J. 1994; 13: 3561-3569Crossref PubMed Scopus (134) Google Scholar), and so is the case with the vertebrate Pbx and Hox products (38Lu Q. Knoephler P.S. Scheele J. Wright D.D. Kamps M.P. Mol. Cell. Biol. 1995; 15: 3786-3795Crossref PubMed Scopus (136) Google Scholar). In these cases, direct contact between two factors increases their binding affinity for DNA, which can be detected by EMSA analysis. However, in our experiments, the degree of the cooperative binding of Csx/Nkx-2.5 to its target site in the presence of GATA-4 seems to be relatively small and does not seem to fully explain the positive cooperation observed in this study. In another model of transcriptional synergy, the interaction between two factors induces a conformational change in one of the factors and enables it to efficiently activate transcription. Two recent reports suggest that GATA-4 relieves the autorepression of Csx/Nkx-2.5 by the C-terminal inhibitory domain (29Durocher D. Charron F. Warren R. Schwartz R.J. Nemer M. EMBO J. 1997; 16: 5687-5696Crossref PubMed Scopus (550) Google Scholar, 31Sepulveda J.L. Belaguli N. Nigam V. Chen C.Y. Nemer M. Schwartz R.J. Mol. Cell. Biol. 1998; 18: 3405-3415Crossref PubMed Scopus (277) Google Scholar). The interaction of GATA-4 with Csx/Nkx-2.5 enhances the transcriptional activity of Csx/Nkx-2.5 by displacing the C-terminal inhibitory domain, and therefore, C-terminally truncated Csx/Nkx-2.5 does not exhibit synergy with GATA-4. However, in our hands, the Csx/Nkx-2.5-GATA-4 transcriptional synergy on NKE2 or on multimerized Csx/Nkx-2.5 sites was still observed when the C-terminally deleted form of Csx/Nkx-2.5 was used instead of wild-type Csx/Nkx-2.5 (data not shown), suggesting that there exist multiple modes of cooperative transactivation mediated by Csx/Nkx-2.5 and GATA-4. An alternative model is that Csx/Nkx-2.5 and GATA-4 create a specific protein complex that interacts with the basal transcriptional machinery more efficiently than the individual factors. In this case, the "efficient" interaction with the basal transcription complex may result from simultaneous interactions with different binding sites between the Csx/Nkx-2.5-GATA-4 protein complex and the basal transcription complex and/or from the stabilization of the basal transcription complex by the specific assembly of the multiprotein complex. Further studies are necessary to determine the mechanism of positive synergistic action mediated by Csx/Nkx-2.5 and GATA-4. Accumulating evidence suggests that multiple cis-trans interactions between sequence-specific transcription factors and factor-binding sites in the enhancer as well as the protein-protein interactions among transcription factors and cofactors regulate gene transcription during development or in response to external stimuli. Furthermore, in some natural enhancers, changes in the relative positions or orientations of protein-binding sites within the enhancer lead to the inactivation of the enhancer activity (39Giese K. Cox J. Grosscheld R. Cell. 1992; 69: 185-196Abstract Full Text PDF PubMed Scopus (558) Google Scholar), and some transcription factors act as activators in one promoter and as repressors in other promoters (40Natesan S. Gilman M.Z. Genes Dev. 1993; 7: 2497-2509Crossref PubMed Scopus (211) Google Scholar). Our present results suggest that synergistic action between two transcription factors also depends on the promoter context (Fig.8) and that such transcriptional regulatory mechanisms fine-tune gene expression in cardiac myocytes and regulate the process of heart development. We thank Chika Masuo and Kaoru Kuwabara for technical assistance and Hisamaru Hirai and Tetsuya Yamagata for plasmids.