Enhancer-specific Modulation of E Protein Activity
2002; Elsevier BV; Volume: 277; Issue: 8 Linguagem: Inglês
10.1074/jbc.m110659200
ISSN1083-351X
AutoresMaurice Markus, Zhimei Du, Robert Benezra,
Tópico(s)Endoplasmic Reticulum Stress and Disease
ResumoHomodimeric complexes of members of the E protein family of basic helix-loop-helix (bHLH) transcription factors are important for tissue-specific activation of genes in B lymphocytes (Bain, G., Gruenwald, S., and Murre, C. (1993) Mol. Cell Biol. 13, 3522–3529; Shen, C. P., and Kadesch, T. (1995)Mol. Cell Biol. 15, 4518–4524; Jacobs, Y., et al. (1994)Mol. Cell Biol. 14, 4087–4096; Wilson, R. B., et al. (1991) Mol. Cell Biol. 11, 6185–6191). These homodimers, however, have little activity on myogenic enhancers (Weintraub, H., Genetta, T., and Kadesch, T. (1994) Genes Dev. 8, 2203–2211). We report here the identification of a novelcis-acting transcriptional repression domain in the E protein family of bHLH transcription factors. This domain, the Rep domain, is present in each of the known vertebrate E proteins. Extensive mapping analysis demonstrates that this domain is an acidic region of 30 amino acids with a predicted loop structure. Fusion studies indicate that the Rep domain can repress both of the E protein transactivation domains (AD1 and AD2). Physiologically, the Rep domain plays a key role in maintaining E protein homodimers in an inactive state on myogenic enhancers. In addition, we demonstrate that Rep domain mediated repression of AD1 is a necessary for the function of MyoD-E protein heterodimeric complexes. These studies demonstrate that the Rep domain is important for modulating the transcriptional activity of E proteins and provide key insights into both the selectivity and mechanism of action of E protein containing bHLH protein complexes. Homodimeric complexes of members of the E protein family of basic helix-loop-helix (bHLH) transcription factors are important for tissue-specific activation of genes in B lymphocytes (Bain, G., Gruenwald, S., and Murre, C. (1993) Mol. Cell Biol. 13, 3522–3529; Shen, C. P., and Kadesch, T. (1995)Mol. Cell Biol. 15, 4518–4524; Jacobs, Y., et al. (1994)Mol. Cell Biol. 14, 4087–4096; Wilson, R. B., et al. (1991) Mol. Cell Biol. 11, 6185–6191). These homodimers, however, have little activity on myogenic enhancers (Weintraub, H., Genetta, T., and Kadesch, T. (1994) Genes Dev. 8, 2203–2211). We report here the identification of a novelcis-acting transcriptional repression domain in the E protein family of bHLH transcription factors. This domain, the Rep domain, is present in each of the known vertebrate E proteins. Extensive mapping analysis demonstrates that this domain is an acidic region of 30 amino acids with a predicted loop structure. Fusion studies indicate that the Rep domain can repress both of the E protein transactivation domains (AD1 and AD2). Physiologically, the Rep domain plays a key role in maintaining E protein homodimers in an inactive state on myogenic enhancers. In addition, we demonstrate that Rep domain mediated repression of AD1 is a necessary for the function of MyoD-E protein heterodimeric complexes. These studies demonstrate that the Rep domain is important for modulating the transcriptional activity of E proteins and provide key insights into both the selectivity and mechanism of action of E protein containing bHLH protein complexes. The basic helix-loop-helix (bHLH) 1bHLHbasic helix-loop-helixAD1 and AD2E protein transactivation domains 1 and 2CMVcytomegalovirusCATchloramphenicol acetyltransferaseDMEMDulbecco's modified Eagle's mediumFBSfetal bovine serumDBDDNA binding domainPBSphosphate-buffered salineMCKmuscle creatine kinase family of transcription factors plays an important role in embryonic patterning, cell fate determination, cellular differentiation, and proliferation decisions (reviewed in Ref. 1Sadowski I. Bell B. Broad P. Hollis M. Gene. 1992; 118: 137-141Crossref PubMed Scopus (201) Google Scholar). Structurally, the bHLH domain is a 60-amino acid region containing two helices separated by a loop segment preceded by a region rich in basic residues. The basic region is responsible for DNA binding and the HLH region is important for dimerization with other members of the family (2Murre C. McCaw P.S. Baltimore D. Cell. 1989; 56: 777-783Abstract Full Text PDF PubMed Scopus (1863) Google Scholar, 3Davis R.L. Cheng P.F. Lassar A.B. Weintraub H. Cell. 1990; 60: 733-746Abstract Full Text PDF PubMed Scopus (528) Google Scholar, 4Voronova A. Baltimore D. Proc. Natl. Acad. Sci. U. S. A. 1990; 87: 4722-4726Crossref PubMed Scopus (233) Google Scholar). The bHLH family has been divided into three major classes. Class I consists of the ubiquitously expressed E proteins. There are three E protein family members in mammals, E2A (with three major splice products: E12, E47, and E2-5), E2-2 (also called ITF2) and HEB (2Murre C. McCaw P.S. Baltimore D. Cell. 1989; 56: 777-783Abstract Full Text PDF PubMed Scopus (1863) Google Scholar,5Corneliussen B. Thornell A. Hallberg B. Grundstrom T. J. Virol. 1991; 65: 6084-6093Crossref PubMed Google Scholar, 6Henthorn P. McCarrick-Walmsley R. Kadesch T. Nucleic Acids Res. 1990; 18: 678Crossref PubMed Scopus (21) Google Scholar, 7Henthorn P. Kiledjian M. Kadesch T. Science. 1990; 247: 467-470Crossref PubMed Scopus (336) Google Scholar, 8Hu J.S. Olson E.N. Kingston R.E. Mol. Cell Biol. 1992; 12: 1031-1042Crossref PubMed Scopus (255) Google Scholar, 9Javaux F. Donda A. Vassart G. Christophe D. Nucleic Acids Res. 1991; 19: 1121-1127Crossref PubMed Scopus (17) Google Scholar). Class II bHLH proteins, such as MyoD, are expressed in a tissue-specific manner (10Weintraub H. Davis R. Tapscott S. Thayer M. Krause M. Benezra R. Blackwell T.K. Turner D. Rupp R. Hollenberg S. Zhuang Y. Lassar A. Science. 1991; 251: 761-766Crossref PubMed Scopus (1238) Google Scholar) and require E proteins as obligate heterodimeric partners for DNA binding and transcriptional activation (11Lassar A.B. Davis R.L. Wright W.E. Kadesch T. Murre C. Voronova A. Baltimore D. Weintraub H. Cell. 1991; 66: 305-315Abstract Full Text PDF PubMed Scopus (688) Google Scholar). Cell fate decisions are often regulated by alterations in the expression of these tissue-specific proteins. Finally, class III members include the dominant negative Id proteins that heterodimerize with the E proteins but, because they lack a basic domain, form heterodimeric complexes incapable of binding DNA (12Benezra R. Davis R.L. Lockshon D. Turner D.L. Weintraub H. Cell. 1990; 61: 49-59Abstract Full Text PDF PubMed Scopus (1804) Google Scholar, 13Sun X.H. Baltimore D. Cell. 1991; 64: 459-470Abstract Full Text PDF PubMed Scopus (317) Google Scholar). basic helix-loop-helix E protein transactivation domains 1 and 2 cytomegalovirus chloramphenicol acetyltransferase Dulbecco's modified Eagle's medium fetal bovine serum DNA binding domain phosphate-buffered saline muscle creatine kinase In addition to their role as heterodimeric partners for tissue-specific bHLH proteins, E proteins are also capable of forming homodimeric complexes, which can activate a variety of B cell-specific genes and are necessary for B cell differentiation (14Bain G. Gruenwald S. Murre C. Mol. Cell Biol. 1993; 13: 3522-3529Crossref PubMed Scopus (140) Google Scholar, 15Shen C.P. Kadesch T. Mol. Cell Biol. 1995; 15: 4518-4524Crossref PubMed Scopus (160) Google Scholar, 16Bain G. Maandag E.C. Izon D.J. Amsen D. Kruisbeek A.M. Weintraub B.C. Krop I. Schlissel M.S. Feeney A.J. van Roon M. Cell. 1994; 79: 885-892Abstract Full Text PDF PubMed Scopus (650) Google Scholar, 17Choi J.K. Shen C.P. Radomska H.S. Eckhardt L.A. Kadesch T. EMBO J. 1996; 15: 5014-5021Crossref PubMed Scopus (90) Google Scholar, 18Kee B.L. Murre C. J. Exp. Med. 1998; 188: 699-713Crossref PubMed Scopus (212) Google Scholar, 19Sigvardsson M. O'Riordan M. Grosschedl R. Immunity. 1997; 7: 25-36Abstract Full Text Full Text PDF PubMed Scopus (226) Google Scholar, 20Zhuang Y. Soriano P. Weintraub H. Cell. 1994; 79: 875-884Abstract Full Text PDF PubMed Scopus (582) Google Scholar). The E proteins have two transcriptional activation domains, activation domain1 (AD1) and 2 (AD2) (21Aronheim A. Shiran R. Rosen A. Walker M.D. Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 8063-8067Crossref PubMed Scopus (106) Google Scholar, 22Massari M.E. Jennings P.A. Murre C. Mol. Cell Biol. 1996; 16: 121-129Crossref PubMed Scopus (87) Google Scholar, 23Quong M.W. Massari M.E. Zwart R. Murre C. Mol. Cell Biol. 1993; 13: 792-800Crossref PubMed Scopus (132) Google Scholar). AD1 is contained within the first 100 amino acids of the protein and contains a putative α-helix. AD2, also termed the loop-helix domain, is located midway between the AD1 and bHLH domains. The transcriptional activation domains of the E proteins play a critical role in cellular transformation in the context of chimeric oncoproteins such as E2A-HLF and E2A-Pbx1 (24Inaba T. Roberts W.M. Shapiro L.H. Jolly K.W. Raimondi S.C. Smith S.D. Look A.T. Science. 1992; 257: 531-534Crossref PubMed Scopus (249) Google Scholar, 25Yoshihara T. Inaba T. Shapiro L.H. Kato J.Y. Look A.T. Mol. Cell Biol. 1995; 1555: 3247-3255Crossref Scopus (47) Google Scholar, 26Lu Q. Wright D.D. Kamps M.P. Mol. Cell Biol. 1994; 14: 3938-3948Crossref PubMed Google Scholar). Site selection studies have suggested that enhancer selection by specific bHLH complexes is to some extent regulated by the sequences within and surrounding the E box (27Blackwell T.K. Weintraub H. Science. 1990; 250: 1104-1110Crossref PubMed Scopus (760) Google Scholar). Binding site selection is not sufficient, however, to explain the specificity of action of many bHLH proteins. It has been demonstrated that the myogenic bHLH protein MyoD can bind to the B cell-specific immunoglobulin enhancer, but is transcriptionally inactive as a result of a cis-acting repressive element in the enhancer (28Weintraub H. Genetta T. Kadesch T. Genes Dev. 1994; 8: 2203-2211Crossref PubMed Scopus (82) Google Scholar). Similarly, E protein homodimers can bind to E boxes within myogenic enhancers, but these complexes are almost entirely inactive (28Weintraub H. Genetta T. Kadesch T. Genes Dev. 1994; 8: 2203-2211Crossref PubMed Scopus (82) Google Scholar). The mode of repression of these E protein homodimers is unknown. In this study, we have identified a novel autoregulatory domain common to all of the vertebrate E proteins. This domain, which we term the Rep domain, is a potent inhibitor of the AD1 and AD2 transcriptional activation domains. We demonstrate that the Rep domain plays a key role in inactivating E protein homodimers bound to a myogenic enhancer. In addition, this domain is required to form active MyoD-E protein heterodimeric complexes by repressing AD1, which is inhibitory on theMCK enhancer. Thus, the Rep domain modulates the activity of E protein-containing bHLH protein complexes and may play a key role in regulating the activity of complexes bound to tissue-specific enhancers. Full-length E2-2 andE12 were amplified by PCR and cloned into a pcDNAIII vector (Invitrogen). All GAL4 DNA binding domain fusions were constructed by ligation of E protein DNA in-frame into a modified pM3 vector (1Sadowski I. Bell B. Broad P. Hollis M. Gene. 1992; 118: 137-141Crossref PubMed Scopus (201) Google Scholar). Deletions of E12 and E2-2 were generated by use of appropriate restriction enzyme sites and by PCR. E12ΔAD1 ΔAD2 and E12Δrep ΔAD1 ΔAD2 were cloned into pFLAG CMV2 vector (Kodak). All constructs were verified by sequencing. EMC11S is the expression vector for MyoD (29Davis R.L. Weintraub H. Lassar A.B. Cell. 1987; 51: 987-1000Abstract Full Text PDF PubMed Scopus (2499) Google Scholar). Reporter vectors used were: multimerized E box driving chloramphenicol acetyltransferase (CAT), 4RtkCAT (30Weintraub H. Davis R. Lockshon D. Lassar A. Proc. Natl. Acad. Sci. U. S. A. 1990; 87: 5623-5627Crossref PubMed Scopus (215) Google Scholar); multimerized GAL4 binding sites driving luciferase, GAL4-luc (31Forman B.M. Umesono K. Chen J. Evans R.M. Cell. 1995; 81: 541-550Abstract Full Text PDF PubMed Scopus (569) Google Scholar); muscle creatine kinase enhancer driving luciferase, MCK luciferase; multimerized immunoglobulin enhancer elements driving luciferase, (mE2mE5)6 luciferase (gift of T. Grundstrom). Site-directed mutagenesis of the Rep domain was performed with a Site-Directed Mutagenesis kit as per the manufacture's instructions (Stratagene). The following oligonucleotides were used for mutagenesis: 5′-GCGGCCGCGGCCGCTGCGGCCAAATCCGATGACGAG-3′ (mutant 1); 5′-GCGGCCGCTGAGATCGCAGCCGATGACGAGGGTG-3′ (mutant 2); 5′-GCTGAGATCAAATCCGCTGCCGAGGGTGATGAGAAC-3′ (mutant 3); 5′-GATCAAATCCGATGACGCGGCTGATGAGAACCTGCAAG-3′ (mutant 4); 5′-CCGATGACGAGGGTGCTGCGAACCTGCAAGACACG-3′ (mutant 5); 5′-GACGAGGGTGATGAGGCCGCGCAAGACACGAAATC-3′ (mutant 6); 5′-GGTGATGAGAACCTGGCAGCCACGAAATCTTCGGAG-3′ (mutant 7); 5′-GAGAACCTGCAAGACGCGGCATCTTCGGAGGACAAG-3′ (mutant 8); 5′-CTGCAAGACACGAAAGCTGCGGAGGACAAGAAATTAG-3′ (mutant 9); 5′-GACACGAAATCTTCGGCGGCCAAGAAATTAGATGAC-3′ (mutant 10); 5′-CGAAATCTTCGGAGGACGCGGCATTAGATGACGACAA-3′ (mutant 11); 5′-CGGAGGACAAGAAAGCAGCTGACGACAAGAAG-3′ (mutant 12); 5′-GACAAGAAATTAGATGCCGCCAAGAAGGATGAATTC-3′ (mutant 13); 5′-GAAATTAGATGACGACGCGGCGGCTGAATTCTAGATA-3′ (mutant 14). Mismatched nucleotides are indicated in boldface. All constructs described above were verified by DNA sequencing. HeLa, C3H10T1/2, and COS-7 cells were maintained in Dulbecco's modified Eagle medium (DMEM) with 10% fetal bovine serum (FBS) (HyClone) and plated at 50% confluence into six-well dishes (HeLa and C3H10T1/2) or 6-cm dishes (COS-7 cells) 12–24 h prior to transfection. HeLa cells were transfected with 0.5 mg of expression vector along with either 0.1 mg of 4RtkCAT or 0.1 mg of GAL4-luc reporter by the calcium phosphate method and maintained in DMEM with 10% FBS for 48 h before harvesting for chloramphenicol acetyltransferase (CAT) or luciferase assays. C3H10T1/2 cells were transfected with 0.5 mg of expression plasmid and 0.5 mg of MCK luciferase using LipofectAMINE reagent (Invitrogen) or FuGENE6 (Roche Molecular Biochemicals) according to the manufacturer's instructions and maintained in DMEM containing 10 mg/ml insulin 5 mg/ml transferrin without FBS for 48 h before harvesting for luciferase assays. The total amount of transfected DNA was kept constant. Cells were harvested by disruption with Reporter Lysis Buffer (for CAT assays) or Cell Culture Lysis Buffer (for luciferase assays) (Promega). Equal amounts of protein were assayed for each sample. Luciferase activity was measured with a luminometer (Berthold Lumat LB9501) using a luciferase assay kit (Promega). CAT activity was measured by a non-chromatographic CAT assay (32Sleigh M.J. Anal. Biochem. 1986; 156: 251-256Crossref PubMed Scopus (297) Google Scholar). Each transfection was repeated at least three times in triplicate. Data shown represent the average of triplicate values for a typical assay with standard deviations. COS-7 cells were plated at a 50% confluence and transfected with 3 mg of expression plasmid 12–24 h later by the calcium phosphate method. Nuclear extracts were prepared after 48 h according to the procedure of Schreiber et al. (33Schreiber E. Matthias P. Muller M.M. Schaffner W. Nucleic Acids Res. 1989; 17: 6419Crossref PubMed Scopus (3918) Google Scholar). 60 mg of protein was fractionated on a 10% discontinuous SDS gel and electrotransferred to a nitrocellulose filter. After transfer, the filter was incubated for 30 min at room temperature in 5% milk. Primary antibody incubation was carried out at room temperature for 1 h in PBS containing 0.05% Tween 20 and 200 ng/ml anti-GAL4 DBD antibody (sc-577, Santa Cruz Biotechnologies). After primary antibody binding, the filter was washed three times with PBS containing 0.05% Tween 20. Secondary antibody incubation was carried out at room temperature for 1 h in PBS containing 0.05% Tween 20 and donkey anti-rabbit immunoglobulin conjugated to horseradish peroxidase diluted 1:10,000. The filter was then washed five times with PBS containing 0.05% Tween 20, and proteins were detected by chemiluminescence (Amersham Biosciences, Inc.). All mobility shift assays were performed as described previously (12Benezra R. Davis R.L. Lockshon D. Turner D.L. Weintraub H. Cell. 1990; 61: 49-59Abstract Full Text PDF PubMed Scopus (1804) Google Scholar). 6 mg of soluble nuclear protein was incubated with probe DNA for 5 min at room temperature. Supershifts were performed by addition of 1 mg of anti-FLAG M2 monoclonal antibody (Sigma Chemical Co.) to the binding reaction prior to gel loading. The DNA binding buffer contained 20 mm HEPES (pH 7.6), 50 mm KCl, 1 mm EDTA, 6% glycerol, and 100 mg/ml poly(dI-dC). The upper strand of the oligonucleotides used as probes were as follows: E box (sense), 5′-CCCCAACACCTGCCTGCCTGA-3′; Gal binding site (sense), 5′-GATCCGGAGTACTGTCCTCCG-3′. C3H10T1/2 cells transfected with expression plasmids were fixed in 2% paraformaldehyde for 30 min at room temperature and washed three times in PBS. Slides were blocked with MOM™ Kit (Vector Laboratories) blocking reagent (Zymed Laboratories Inc.), incubated with anti-myocin-heavy-chain antibody (Zymed Laboratories Inc.) overnight at 4 °C, then incubated 1 h with anti-mouse IgG-Texas Red-conjugated antibody (Jackson ImmunoResearch) at room temperature. After three final washes in PBS containing 3% bovine serum albumin and 0.5% IGPEL, slides were mounted with SlowFade mounting reagent (Molecular Probe). In a screen to map functional domains in E2-2, sequences in and around activation domain 2 (AD2) (226–450) of E2-2 were fused to the DNA binding domain of GAL4. These expression plasmids were cotransfected into HeLa cells with a luciferase reporter plasmid containing five GAL4 DNA binding sites, and luciferase activity was quantitated after 24 h. A GAL-E2-2 fusion that spans residues 226–540 is a very poor transcriptional activator (Fig.1) whereas a shorter GAL4-E2-2 fusion containing amino acids 226–493 is 82.3-fold more active. This suggested that E2-2 contains a domain capable of partially inactivating AD2. To test whether E12 (a product of the E2A gene) also contains this domain, the GAL4 fusion domain studies were repeated with E12 (Fig. 1). A fusion of amino acids 230–524, containing AD2 and significant C-terminal residues, was also a poor stimulator of the GAL4 reporter. Deletion of 33 amino acids (residues 492–524) activated luciferase expression by 18.3-fold. These data led us to hypothesize that E proteins contain a transcriptional repression domain downstream of AD2. To further map the repression domain of E2-2, amino acids between residues 450 and 578 were added incrementally to a transcriptionally active GAL4-E2-2 fusion containing the minimal AD2 domain (amino acids 226–450). As shown in Fig.2 A, addition of amino acids 450–493 had no significant effect on transcriptional activity. Inclusion of an additional 28 residues (amino acids 493–521) produced a 2-fold repression. Addition of amino acids 521–530 led to a further 1.5-fold repression and amino acids 530–540 resulted in an additional 4.1-fold repression. Overall, 12.2-fold repression of AD2 was observed by inclusion of residues 493–540. Larger additions, including a previously characterized dimerization repression domain (amino acids 540–556) (13Sun X.H. Baltimore D. Cell. 1991; 64: 459-470Abstract Full Text PDF PubMed Scopus (317) Google Scholar) and the basic domain (amino acids 556–578), led to no further effects on transcriptional activity. A refinement of the location of the repression domain was achieved by removing residues C-terminal to amino acid 450 incrementally in a fully repressed GAL4-E2-2 (226–540) construct (Fig. 2 A). Deletion of residues 493–502, 493–511, and 450–511 had no effect on the transcriptional activity of the Gal4 E2-2 construct. Deletion of residues C-terminal to 511 (Gal E2-2 (226–540 Δ450–521) or Gal E2-2 (226–540 Δ450–528)) or N-terminal to 540 (Gal E2-2 (226–533 Δ450–521)), however, resulted in significant derepression. From these data, we conclude that the repression domain (which we term the Rep domain) is located between amino acids 511 and 540 (highlighted in Fig. 2 A) and is capable of inactivating AD2 of E2-2. The Rep domain functions in multiple cell types as identical results were observed in COS-7, 293T, Namalwa, and 2017 cells (data not shown). To rule out the possibility that the alterations in transcriptional activities described in Fig. 2 A were the result of alterations in GAL4 fusion protein levels, Western blot analysis using an anti-GAL4 antibody was performed on soluble nuclear protein extracted from transfected cells. As shown in Fig. 2 B, each of the constructs yields a comparable amount of Gal4 fusion protein recovered from the transfected cells. To determine if the GAL4 fusions bound DNA with equal efficiency, these extracts were incubated with a radiolabeled oligonucleotide containing the minimal 21-bp GAL4 DNA element, and protein-DNA complexes were resolved on non-denaturing gels. As shown in Fig. 2 C, each fusion protein was equally efficient in forming a complex with DNA. These results suggest that the Rep domain affects transcriptional activity of E2-2 without significantly altering protein steady-state levels or DNA binding efficiency. We sought to determine whether the repression domain between positions 511 and 540 defined above was also capable of altering transcriptional activity in the context of non-GAL4 linked E2-2. Transfection of full-length E2-2 results in a 3.1-fold activation of a multimerized E box reporter above background (Fig. 2 D). Deletion of amino acids 450–540 resulted in a 4.4-fold increase in activity above full-length E2-2. Deletion of residues 493–540 or 521–540 had a similar stimulatory effect indicating that critical elements of the repression domain are present between amino acids 521 and 540 in the context of full-length E2-2. Deletion of residues 493–502, 493–511, and 226–280 resulted in no stimulation of E2-2. This analysis demonstrates that the Rep domain functions in full-length E2-2 homodimers and maps to the same region (residues 511–540) defined in the Gal4 fusion experiments (highlighted in Fig. 2 A). To confirm that repression is not due to an alteration in the DNA binding capacity of the E2-2 constructs, nuclear extracts from transfected cells were incubated with a radiolabeled E-box containing oligonucleotide, and protein-DNA complexes were resolved by non-denaturing gel electrophoresis (Fig. 2 E). An equivalent concentration of E box binding activity was present in cells transfected with the various E2-2 constructs. These data demonstrate that the transcriptional repression mechanism is unrelated to alterations in the dimerization or DNA binding capacity of E2-2. The minimal E2-2 Rep domain was fused to AD1 and AD2 of E2-2 and to the activation domain of the viral transcription activator VP16. Each of these chimeric proteins was expressed as a fusion with the GAL4 DNA binding domain. As shown in Fig. 3 A, fusion of the Rep domain onto AD1 (top panel) and AD2 (middle panel) repressed transactivation to nearly background levels indicating that the Rep domain is sufficient for transcriptional inhibition. Fusion of the Rep domain onto the VP16 activation domain, however, had no significant effect on transcription, indicating at least some degree of specificity for E protein activation domains. We observed, however, that the Rep domain cannot inhibit E protein activation domains when bound separately to an adjacent DNA element (data not shown), suggesting that the Rep domain may function exclusively as an intramolecular transcriptional repressor. Examination of the Rep domain sequence reveals a region rich in polar and charged amino acids (particularly lysine, aspartic, and glutamic acid) that is remarkably conserved within the E protein family in all characterized vertebrate E proteins (Fig. 3 B). Overall, the Rep domain of human E2-2 and Zebrafish E12 are 43% identical. PHD (Profile network predictionHeiDelberg) analysis predicted this stretch of amino acids to adopt a flexible loop structure. To assess the contribution of each amino acid residue to the Rep domain function, we performed alanine scan mutagenesis. Each residue of the Rep domain was altered to alanine (two at a time), and mutants were assayed for the ability to repress a minimal AD2 domain. Two mutants, m4 and m6 (corresponding to resides 518–519 and 522–523, respectively) clearly affect the inhibitory function of Rep, which results in a dramatic restoration of AD2 transcriptional activity (Fig.3 C). Mutation of individual residues of the pairs had little effect by themselves (data not shown). These data suggest that residues Glu-518 and Glu-519, as well as Asn-522 and Leu-523 are crucial for Rep domain inhibitory function. Given that the Rep domain is highly conserved in E proteins (Figs. 1 and 3 B), we utilized the E12 molecule exclusively for the remainder of the experiments because of its high expression level and transactivation potential in a variety of lymphocytic and myogenic genes. We sought to determine if the Rep domain was responsible for maintaining E protein homodimers in an inactive state on the MCK enhancer. As shown in Fig.4 A, MyoD strongly activates transcription from the MCK enhancer, whereas E12 is completely inactive. Deletion of the Rep domain alone (E12Δrep) had no effect on E12 activity. However, when activation domain 1 (AD1) was also deleted (E12Δrep ΔAD1), E12 became a potent transcriptional activator on the MCK enhancer. A molecule missing only AD1 (E12ΔAD1), however, lacked detectable transactivation potential. These data suggested that both AD1 and Rep interfere with homodimer activation of the MCK enhancer. We also demonstrated that E12Δrep ΔAD1 drives the MCK enhancer via AD2, because deletion of AD2 (E12Δrep ΔAD1 ΔAD2) results in complete inactivation of the molecule. These results demonstrate that E proteins lacking a Rep domain can inappropriately stimulate a myogenic enhancer via AD2 but AD1 interferes with this activation. To evaluate this phenomenon on endogenous muscle gene expression, we performed immunofluorescence analysis on 10T1/2 cells transiently transfected with our E protein mutants. 10T1/2 cells were cultured on chamber slides and transfected with plasmids, and immunofluorescence was performed using anti-Myocin-heavy-chain antibody. As shown in Fig. 4 B, pcDNA (vector control)-, E12-, E12Δrep-, and E12ΔAD1-transfected cells showed little staining, whereas MyoD- and E12Δrep ΔAD1-transfected cells showed intense cytoplasmic staining. These results indicate that E12Δrep ΔAD1 alone has the ability to activate endogenous MCK gene expression, whereas wild type E12, E12Δrep, and E12ΔAD1 cannot. Taken together, these results imply that the AD2 domain can induce a full range of active myogenic responses, and that both the Rep and AD1 domains play a negative regulatory role by direct or indirect inactivation of the AD2 domain, thereby preventing E12 homodimer transactivation ofMCK. During muscle development, E proteins activate the MCK enhancer as heterodimeric complexes with MyoD. To determine if the Rep domain plays a role in heterodimer activity, E12 and E12Δrep were cotransfected with MyoD and anMCK reporter. Surprisingly, removal of the Rep domain of E12 consistently repressed MyoD-dependent transcription in a dose-dependent fashion (Fig.5 A). Electrophoretic mobility shift analysis demonstrates that this repression event is not due to an altered efficiency in forming MyoD-E12 heterodimers or in binding DNA, because both E12 and E12Δrep form strong DNA binding complexes with MyoD (Fig. 5 B, lanes 5 and 6). Because the Rep domain behaves as an intramolecular transcriptional repressor, and AD1 also acts as a repressor that inhibits E protein homodimer transactivation function (Fig. 4 A), we hypothesized that activity of MyoD-E12 complexes is dependent on masking of the AD1 repressive activity by the Rep domain. MyoD-E12Δrep complexes would be unable to perform this masking event but should regain wild type activity if the relevant activation domain was inactivated. To test this, we made use of the E12 mutant proteins with targeted deletions of AD1 and AD2. As shown in Fig. 5 C, the E12 mutant lacking both Rep and AD1 domains (E12ΔRepΔAD1) are fully capable of acting synergistically with MyoD on the MCK enhancer, confirming our hypothesis. AD1's repressive function in the MyoD-E12 complex was further supported by the observation that the E12 mutant lacking AD1 alone (E12ΔAD1) showed higher transactivation potential compared with wild type (Fig. 5 C). In contrast, mutants with targeted deletions of AD2 are inactive in MyoD-dependent transcription, both in the presence and absence of the Rep domain. Electrophoretic mobility shift analysis using extracts from transfected cells demonstrates that these activity differences are not the result of altered expression or DNA binding efficiency (data not shown). These results demonstrate that the MyoD-E heterodimer stimulates a myogenic enhancer via AD2 and requires Rep-domain-mediated repression of AD1 (see model, Fig. 6 B). Thus the AD1 domain of E protein is repressive on the MCK enhancer, both in the context of E protein homodimer or MyoD-E protein heterodimer, and its repression activity is suppressed by the Rep domain only as a heterodimer (see model, Fig. 6). E protein complexes are potent transcriptional activators and are highly specific for the enhancers they stimulate. As homodimers, E proteins directly stimulate transcription from the B lymphocyte-specific IgH and Igk enhancers and thereby facilitate immunoglobulin gene rearrangement (14Bain G. Gruenwald S. Murre C. Mol. Cell Biol. 1993; 13: 3522-3529Crossref PubMed Scopus (140) Google Scholar, 15Shen C.P. Kadesch T. Mol. Cell Biol. 1995; 15: 4518-4524Crossref PubMed Scopus (160) Google Scholar, 34Jacobs Y. Xin X.Q. Dorshkind K. Nelson C. Mol. Cell Biol. 1994; 14: 4087-4096Crossref PubMed Scopus (19) Google Scholar, 35Wilson R.B. Kiledjian M. Shen C.P. Benezra R. Zwollo P. Dymecki S.M. Desiderio S.V. Kadesch T. Mol. Cell Biol. 1991; 11: 6185-6191Crossref PubMed Scopus (120) Google Scholar, 36Kadesch T. Immunol. Today. 1992; 13: 31-36Abstract Full Text PDF PubMed Scopus (119) Google Scholar, 37Park C.W. Walker M.D. J. Biol. Chem. 1992; 267: 15642-15649Abstract Full Text PDF PubMed Google Scholar). E protein homodimers also directly or indirectly stimulate the terminal deoxynucleotidyl transferase (TdT), l5, EBF, and RAG genes in vivo (18Kee B.L. Murre C. J. Exp. Med. 1998; 188: 699-713Crossref PubMed Scopus (212) Google Scholar, 38Schlissel M. Voronova A. Baltimore D. Genes Dev. 1991; 5: 1367-1376Crossref PubMed Scopus (196) Google Scholar). As heterodimers with MyoD, E protein complexes stimulate transcription of muscle-specific genes such as muscle creatine kinase (MCK) and myosin light chain (39Lassar A.B. Buskin J.N. Lockshon D. Davis R.L. Apone S. Hauschka S.D. Weintraub H. Cell. 1989; 58: 823-831Abstract Full Text PDF PubMed Scopus (570) Google Scholar). MyoD-E protein heterodimers bind E boxes in the immunoglobulin enhancer but are transcriptionally inactive as a result of a cis-acting repression region on the enhancer (28Weintraub H. Genetta T. Kadesch T. Genes Dev. 1994; 8: 2203-2211Crossref PubMed Scopus (82) Google Scholar). Conversely, E protein homodimers bind E boxes in the MCK enhancer, but their activities are completely repressed. The mechanisms the cell uses for recognition and inactivation of E protein homodimers bound to non-lymphocyte enhancers are largely unknown. In this study, we have identified a novel cis-acting transcriptional repression domain in the E proteins that plays a role in modulating the activity of bHLH complexes in an enhancer-specific manner. The Rep domain is a charged 30-amino acid region, highly conserved among the vertebrate E proteins, which is capable of repressing E protein activation domains 1 and 2. In E protein homodimers bound inappropriately to a myogenic enhancer, the Rep domain and AD1 cooperate in preventing transcriptional activation. In MyoD-E protein heterodimers bound to the MCK enhancer, the Rep domain is critical for transcriptional activation by masking negative signals derived from AD1. The Rep domain, then, is an intramolecular transcriptional repressor that plays a role in modulating the activity of enhancer-bound bHLH protein complexes. The Rep domain is a potent transcriptional repressor that inhibits the activities of the AD1 and AD2 domains of the E proteins. This repression is somewhat specific to these domains, because VP16 activity is unaffected by the addition of this domain. Work to determine Rep domain-interacting partners is ongoing. We have demonstrated that AD1 is repressive for theMCK enhancer both in the context of E protein homodimeric and MyoD-E heterodimeric complexes. This domain is normally active in E protein homodimers and is indeed required for activation of the immunoglobulin enhancer (data not shown). In MyoD-E heterodimers, AD1 is masked by the Rep domain, which allows AD2 to stimulate myogenic enhancers. The mechanism by which AD1 represses myogenic enhancers is unknown. We propose that either AD1 recruits a transcriptional repressor when bound to the MCK enhancer or AD1-recruited transcriptional coactivators interfere with myogenic enhancer function. These possibilities are currently being evaluated. The Rep domain may inhibit AD1's repression function in E protein heterodimers by preventing the recruitment of RNAPII or coactivators, such as histone acetyltransferase containing transcriptional cofactor, SWI/SNF complexes or mediators (40Eckner R. Yao T.P. Oldread E. Livingston D.M. Genes Dev. 1996; 10: 2478-2490Crossref PubMed Scopus (320) Google Scholar, 41Qiu Y. Sharma A. Stein R. Mol. Cell Biol. 1998; 18: 2957-2964Crossref PubMed Scopus (123) Google Scholar). We suspect, therefore, that the Rep domain is capable of preventing the assembly of several different transcriptional cofactor complexes. The simplest model we proposed in Fig. 6 with multiple functional interactions observed between AD1, AD2, and Rep domain is that multiple, sequential intramolecular interactions contribute to transactivation specificity. We presume that this kind of complex intramolecular regulation in E protein exists in part to permit a broad range of regulatory capabilities.
Referência(s)