Title: The Human Medium Chain Acyl-CoA Dehydrogenase Gene Promoter Consists of a Complex Arrangement of Nuclear Receptor Response Elements and Sp1 Binding Sites
Abstract: Expression of the gene encoding the mitochondrial fatty acid β-oxidation enzyme, medium-chain acyl-CoA dehydrogenase (MCAD), is regulated among tissues during development and in response to alterations in substrate availability. To identify and characterize cis-acting MCAD gene promoter regulatory elements and corresponding transcription factors, DNA-protein binding studies and mammalian cell transfection analyses were performed with human MCAD gene promoter fragments. DNA:protein binding studies with nuclear protein extracts prepared from hepatoma G2 cells, 3T3 fibroblasts, or Y-1 adrenal tumor cells identified three sequences (nuclear receptor response element 1 or NRRE-1, NRRE-2, and NRRE-3) that bind orphan members of the steroid/thyroid nuclear receptor superfamily including chicken ovalbumin upstream promoter transcription factor and steroidogenic factor 1. Sp1 binding sites (A-C) were identified in close proximity to each of the NRREs. NRRE-3 conferred cell line-specific transcriptional repression by interacting with chicken ovalbumin upstream promoter transcription factor or activation via steroidogenic factor 1. In contrast, the Sp1 binding site A behaved as a transcriptional activator in all cell lines examined. We propose that multiple nuclear receptor transcription factors interact with MCAD gene promoter elements to differentially regulate transcription among a variety of cell types. Expression of the gene encoding the mitochondrial fatty acid β-oxidation enzyme, medium-chain acyl-CoA dehydrogenase (MCAD), is regulated among tissues during development and in response to alterations in substrate availability. To identify and characterize cis-acting MCAD gene promoter regulatory elements and corresponding transcription factors, DNA-protein binding studies and mammalian cell transfection analyses were performed with human MCAD gene promoter fragments. DNA:protein binding studies with nuclear protein extracts prepared from hepatoma G2 cells, 3T3 fibroblasts, or Y-1 adrenal tumor cells identified three sequences (nuclear receptor response element 1 or NRRE-1, NRRE-2, and NRRE-3) that bind orphan members of the steroid/thyroid nuclear receptor superfamily including chicken ovalbumin upstream promoter transcription factor and steroidogenic factor 1. Sp1 binding sites (A-C) were identified in close proximity to each of the NRREs. NRRE-3 conferred cell line-specific transcriptional repression by interacting with chicken ovalbumin upstream promoter transcription factor or activation via steroidogenic factor 1. In contrast, the Sp1 binding site A behaved as a transcriptional activator in all cell lines examined. We propose that multiple nuclear receptor transcription factors interact with MCAD gene promoter elements to differentially regulate transcription among a variety of cell types. The human medium chain acyl-CoA dehydrogenase gene promoter consists of a complex arrangement of nuclear receptor response elements and Sp1 binding sites.Journal of Biological ChemistryVol. 270Issue 41PreviewVol. 270, p. 16308 Full-Text PDF Open Access Medium chain acyl-CoA dehydrogenase (MCAD)1( 1The abbreviations used are: MCADmedium chain acyl-CoA dehydrogenaseCATchloramphenicol acetyltransferaseCOUP-TFchicken ovalbumin upstream promoter transcription factorEMSAelectrophoretic mobility shift assayFAOfatty acid oxidationFTZ-F1fushi tarazu factor 1HNF-4hepatocyte nuclear factor 4NRREnuclear receptor response elementSF-1steroidogenic factor 1TKthymidine kinasePCRpolymerase chain reaction.) (2,3-oxidoreductase, EC 1.3.99.3) is a nuclear-encoded mitochondrial enzyme that catalyzes the initial reaction in the fatty acid β-oxidation (FAO) pathway(1Beinert H. Boyer P.D. Lardy H.J. Myrback K. The Enzymes. 2nd Ed. Academic Press, New York1963: 447-476Google Scholar). The pivotal role of MCAD in cellular energy metabolism is underscored by the serious and often fatal clinical consequences of inherited MCAD deficiency including hypoglycemia, liver dysfunction, and sudden death (2, 3). MCAD expression is highly regulated in parallel with cellular FAO rates during development, among tissues, and by dietary factors and hormones(4Kelly D.P. Gordon J.I. Alpers R. Strauss A.W. J. Biol. Chem. 1989; 264: 18921-18925Abstract Full Text PDF PubMed Google Scholar, 5Nagao M. Parimoo B. Tanaka K. J. Biol. Chem. 1993; 268: 24114-24124Abstract Full Text PDF PubMed Google Scholar). medium chain acyl-CoA dehydrogenase chloramphenicol acetyltransferase chicken ovalbumin upstream promoter transcription factor electrophoretic mobility shift assay fatty acid oxidation fushi tarazu factor 1 hepatocyte nuclear factor 4 nuclear receptor response element steroidogenic factor 1 thymidine kinase polymerase chain reaction. To delineate transcriptional regulatory mechanisms involved in the control of nuclear genes encoding mitochondrial FAO enzymes, we have focused on the MCAD gene. As an initial step in the identification of MCAD gene transcriptional regulatory elements, genomic clones encoding the human MCAD gene and 5′-flanking region were isolated and characterized(6Zhang Z. Kelly D.P. Kim J.J. Zhou Y. Ogden M.L. Whelan A.J. Strauss A.W. Biochemistry. 1992; 31: 81-89Crossref PubMed Scopus (43) Google Scholar). Transfection of plasmids containing varying lengths of the MCAD gene 5′-flanking region fused to a reporter gene into human hepatoma G2 (HepG2) cells localized the functional promoter to a region between 362 base pairs upstream and 189 base pairs downstream of the transcription start site. Recently, we identified a novel nuclear hormone receptor response element (NRRE-1) within this promoter region located 343 base pairs upstream of the transcription start site(7Raisher B.D. Gulick T. Zhang Z. Strauss A.W. Moore D.D. Kelly D.P. J. Biol. Chem. 1992; 267: 20264-20269Abstract Full Text PDF PubMed Google Scholar). In this study, we sought to delineate the cis-acting regulatory elements within the human MCAD gene promoter and to identify transcription factors that interact with these sequences. To this end, DNA-protein binding studies and mammalian cell gene transfer studies were performed. The results of these studies revealed that the MCAD gene promoter contains a series of cis-acting regulatory modules composed of Sp1 binding sites juxtaposed to nuclear receptor response elements. The transcriptional regulatory properties of the elements suggest a mechanism for modulation of MCAD gene expression among distinct cell types in vivo. Using internal restriction sites, three overlapping MCAD gene promoter DNA fragments were gel-isolated for the DNase I protection assay (all numbers are relative to the transcription start site = +1(6Zhang Z. Kelly D.P. Kim J.J. Zhou Y. Ogden M.L. Whelan A.J. Strauss A.W. Biochemistry. 1992; 31: 81-89Crossref PubMed Scopus (43) Google Scholar)): −362 (HindIII) to −12 (SphI), −102 (PstI) to +134 (XbaI), and −11 (SphI) to +189 (XbaI). DNase I footprint analyses were performed by a modification of a standard protocol(8Ausubel F.M. Brent R. Kingston R.E. Moore D.D. Seidman J.G. Smith J.A. Struhl K. Current Protocols in Molecular Biology. Vol. 2. John Wiley & Sons, Inc, New York1994: 12.4.1-12.4.6Google Scholar). In brief, 5-25 μg of nuclear protein was incubated with 3-5 pmol of labeled DNA fragment and digested with DNase I. Samples were separated on an 8% denaturing polyacrylamide gel followed by autoradiography. The location of the footprinted sequences was determined by a Maxam-Gilbert sequence reaction of the identical labeled fragment. Nuclear protein extracts were prepared as previously described(9Heberlein U. Tjian R. Nature. 1988; 331: 410-415Crossref PubMed Scopus (87) Google Scholar). Electrophoretic mobility shift assays (EMSAs) were performed as described(10Revzin A. BioTechniques. 1989; 7: 346-355PubMed Google Scholar). The following double-stranded oligonucleotides were prepared (sense strand sequence only, lower case nucleotides indicate restriction site ends added to allow labeling): site A (MCAD gene promoter region −282 to −258), 5′-gatccGGCCCCAGCCACGCCCTCTAACCCAg-3′; NRRE-1 (−343 to −311), 5′-gatccGGGTTTGACCTTTCTCTCCGGGTAAAGGTGAAGg-3′; NRRE-2 (−39 to −18), 5′-gatccACGGCGCACGCAAGGGTCACGGg-3′; NRRE-3 (+134 to +159), 5′-gatccGAGTATGTCAAGGCCGTGACCCGTGTg-3′; site B (−59 to −36), 5′-gatccCCCCTCCCCAGGTCGCAGCGACGGg-3′; site C (+39 to +86), 5′-gatccCCCCGTCCTTCCGCAGCCCAACCGCCTCTTCCCGCCCCGCCCCATCCCg-3′; myocyte nuclear factor, 5′-ACCACCCCACCCCCTGTGGC-3′; AP-2, 5′-GATCGAACTGACCGCCCGCGGCCCGT-3′; nuclear respiratory factor-1, 5′-gatccTGCTAGCCCGCATGCGCGCGCACCTTg-3′; Sp1, 5′-ATTCGATCGGGGCGGGGCGAGC-3′. Human recombinant Sp1 protein (11Kadonaga J.T. Carner K.R. Masiarz F.R. Tjian R. Cell. 1987; 51: 1079-1090Abstract Full Text PDF PubMed Scopus (1335) Google Scholar) was obtained from Promega. Antibody supershift experiments were performed with a polyclonal antibody to human COUP-TF kindly supplied by Dr. Bert O'Malley (Baylor College of Medicine)(12Wang L. Tsai S.Y. Cook R.G. Beattie W.G. Tsai M. O'Malley B.W. Nature. 1989; 340: 163-166Crossref PubMed Scopus (409) Google Scholar), a rabbit polyclonal antibody against the amino-terminal region of steroidogenic factor-1 (SF-1) kindly supplied by Dr. Keith Parker (Duke University)(13Ikeda Y. Lala D.S. Luo X. Kim E. Moisan M.P. Parker K.K. Mol. Endocrinol. 1993; 7: 852-860Crossref PubMed Google Scholar), and a rabbit polyclonal Sp1 antibody raised to an epitope corresponding to amino acid residues 520-538 of human Sp1 (Santa Cruz Biotechnology). MCAD genomic DNA fragments were inserted into a promoterless chloramphenicol acetyltransferase (CAT) reporter plasmid (pCAT-Basic, Promega). Construction of MCADCAT(−362/+189), MCADCAT(−312/+189), pTKCAT, and pTKCAT(NRRE-1) have been described(7Raisher B.D. Gulick T. Zhang Z. Strauss A.W. Moore D.D. Kelly D.P. J. Biol. Chem. 1992; 267: 20264-20269Abstract Full Text PDF PubMed Google Scholar). Promoter fragments for MCADCAT(−170/+189) and MCADCAT (−102/+189) were generated by PCR. The PCR products were ligated into the pCAT-Basic plasmid via HindIII (incorporated into the 5′-PCR primer) and the internal SphI site. DNA sequence analysis was performed to confirm that the PCR product contained no mutations. The pCATSV40 plasmid was obtained from Promega (pCAT-Control). pTKCAT(A) and pTKCAT(NRRE-3) were constructed by ligating double-stranded site A or NRRE-3 oligonucleotides into the BamHI site present upstream of the thymidine kinase promoter. Human HepG2 cells, Y-1 mouse adrenal cells, and NIH-3T3 mouse fibroblasts were maintained in an atmosphere containing 5% CO2 in minimal essential medium supplemented with 10% Nu-serum (Collaborative Biomedical Products) (HepG2), F-10 supplemented with 15% heat inactivated horse serum and 5% fetal calf serum (Y-1), and minimal essential medium supplemented with 10% fetal calf serum (3T3). Cell transfection protocol and CAT and β-galactosidase assays were performed as described(6Zhang Z. Kelly D.P. Kim J.J. Zhou Y. Ogden M.L. Whelan A.J. Strauss A.W. Biochemistry. 1992; 31: 81-89Crossref PubMed Scopus (43) Google Scholar, 7Raisher B.D. Gulick T. Zhang Z. Strauss A.W. Moore D.D. Kelly D.P. J. Biol. Chem. 1992; 267: 20264-20269Abstract Full Text PDF PubMed Google Scholar). In brief, 10 μg of MCADCAT plasmid were transfected by the calcium-phosphate technique along with 1-3 μg of a murine sarcoma virus promoter-β-galactosidase chimeric gene plasmid (pMSV-βgal) to control for transfection efficiency. In experiments shown in Fig. 2A, in addition to normalizing to pMSV-βgal activity, the transcriptional activity of pCATSV40 was determined in parallel transfections to allow normalization for differences in background CAT activity among different cell lines. Cells were harvested 48 h after transfection. As an initial step in the identification of cis-acting regulatory elements within the MCAD gene promoter region, DNase I protection experiments were performed with nuclear protein extracts prepared from cell lines derived from liver (human hepatoma G2) and adrenal cortex (mouse Y-1 adrenal tumor cells), two tissues with a high capacity for FAO and relatively abundant MCAD expression, and a third cell line with low MCAD expression (mouse NIH 3T3 fibroblasts)(4Kelly D.P. Gordon J.I. Alpers R. Strauss A.W. J. Biol. Chem. 1989; 264: 18921-18925Abstract Full Text PDF PubMed Google Scholar, 14Kelly D.P. Kim J.-J. Billadello J.J. Hainline B.E. Chu T.W. Strauss A.W. Proc. Natl. Acad. Sci. U. S. A. 1987; 84: 4068-4072Crossref PubMed Scopus (121) Google Scholar). Overlapping end-labeled DNA fragments spanning from −362 to +189 (relative to the transcription start site = +1) of the MCAD gene promoter region were employed in these studies. Compared to the published human MCAD promoter sequence(6Zhang Z. Kelly D.P. Kim J.J. Zhou Y. Ogden M.L. Whelan A.J. Strauss A.W. Biochemistry. 1992; 31: 81-89Crossref PubMed Scopus (43) Google Scholar), we identified two additional guanine nucleotides (at −281 and −298) by Maxam-Gilbert sequencing. Repeat dideoxy sequence reactions revealed that compression artifacts in the dideoxy sequence accounted for this disparity. Accordingly, the numbering system used here to describe the MCAD promoter constructs and regulatory elements was revised to correspond to this corrected sequence. Six distinct DNase I protected sites were identified between −343 and +160 (Fig. 1, A-F). The footprinted regions include three similar GC-rich sequences (designated A-C) and three sites (NRRE-1, NRRE-2, and NRRE-3) containing hexamer sequences that conform to the consensus (RG(G/T)TNA) for binding members of the steroid/thyroid nuclear hormone receptor superfamily of transcription factors (Fig. 1F). We have shown previously that one of the latter sites, NRRE-1, is a functional nuclear receptor response element(7Raisher B.D. Gulick T. Zhang Z. Strauss A.W. Moore D.D. Kelly D.P. J. Biol. Chem. 1992; 267: 20264-20269Abstract Full Text PDF PubMed Google Scholar, 15Carter M.E. Gulick T. Raisher B.D. Caira T. Ladias J.A. Moore D.D. Kelly D.P. J. Biol. Chem. 1993; 268: 13805-13810Abstract Full Text PDF PubMed Google Scholar, 16Carter M.E. Gulick T.G. Moore D.D. Kelly D.P. Mol. Cell. Biol. 1994; 14: 4360-4372Crossref PubMed Google Scholar, 17Gulick T. Cresci S. Caira T. Moore D.D. Kelly D.P. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 11012-11016Crossref PubMed Scopus (498) Google Scholar). Each of sites A-C and NRRE-1 exhibited similar footprinting patterns with nuclear proteins derived from all three cell lines (Fig. 1, A-D). In contrast, the DNase I protection patterns at NRRE-2 and NRRE-3 were cell specific. At NRRE-2, DNase I protection was strong with 3T3 nuclear extracts, weak with HepG2 extracts, and absent with Y-1 nuclear proteins (Fig. 1C). In contrast, both Y-1 and 3T3 nuclear proteins conferred a prominent footprint at NRRE-3, whereas HepG2 nuclear extracts did not interact with this site (Fig. 1E). To determine if the MCAD gene promoter regions delineated by the DNase I protection studies function as transcriptional regulatory elements, the full-length MCAD promoter and a series of mutant promoter fragments with various deletions fused to a CAT gene were transiently transfected into HepG2, Y-1, and 3T3 cells. The initial transfection experiments evaluated the activity of the full-length MCAD gene promoter construct (MCADCAT(−362/+189)) among the three cell lines. The transcriptional activity of MCADCAT(−362/+189) was normalized for transfection efficiency by cotransfecting pMSV-βgal. MCADCAT(−362/+189) activity was compared to that of an SV40 promoter in the identical CAT reporter plasmid backbone (pCATSV40) to control for cell line-specific differences in background CAT activity. MCADCAT(−362/+189) activity was high in HepG2 cells (44% pCATSV40 activity) and Y-1 cells (38% pCATSV40 activity) but more than 10-fold lower in 3T3 fibroblasts (3.4% pCATSV40 activity, Fig. 2A). The steady-state MCAD mRNA levels in each of the cell lines, as determined by Northern blot analysis (data not shown), paralleled the CAT activities. A similar cell line-specific pattern of MCADCAT(−362/+189) transcription was observed when the data were compared to the activity of a herpes simplex virus thymidine kinase (TK) promoter-CAT plasmid standard (data not shown). These data indicate that elements involved in cell-specific control of MCAD gene transcription are contained within the MCAD(−362/+189) promoter fragment. An MCAD promoter deletion series was employed in transfection experiments to examine the transcriptional regulatory properties of the DNase I protected regions in the MCAD gene promoter. Deletion of the 50-base pair region from −362 to −312, which contains NRRE-1, resulted in a modest but statistically significant increase in transcriptional activity in 3T3 cells (1.4-fold, p = 0.006 based on four independent experiments) but no significant changes in HepG2 or Y-1 cells (Fig. 2B). Removal of the 5′-promoter region containing site A (from −312 to −170) resulted in a significant decrease in transcriptional activity in HepG2 (81%), Y-1 (49%), and 3T3 cells (64%). As predicted by the DNase I protection studies, further deletion of the region from −170 to −102, which lacks footprinted regions, had no significant effect on transcriptional activity in any cell line. These data confirm that promoter module A functions as a cis-acting transcriptional activator and suggests that NRRE-1 confers transcriptional repression in 3T3 cells. The transfection experiments indicated that MCAD promoter module A is a positive regulatory element in all three cell lines. To evaluate the transcriptional regulatory properties of the site A element in the absence of the NRREs and in the context of a heterologous promoter, a single copy of a double-stranded oligonucleotide containing site A was placed upstream of the TK promoter in a CAT reporter plasmid (pTKCAT(A)) and transfected into the three cell lines. The transcriptional activity of pTKCAT(A) was 2.3-, 3.3-, and 2.5-fold higher than that of pTKCAT lacking the site A element (pTKCAT(-)) in HepG2, Y-1, and 3T3 cells, respectively (Fig. 2C). A DNA fragment containing site A conferred orientation-independent transcriptional activation (8.6 ± 1.3-fold induction) when inserted at a remote distance upstream of the SV40 promoter (data not shown). Thus, site A is a transcriptional activator with enhancer-like properties. EMSAs were performed to identify the protein(s) that bind MCAD promoter elements A-C. A double-stranded oligonucleotide containing the site A element was employed as a probe in EMSA performed with nuclear protein extracts prepared from Y-1, HepG2, and 3T3 cells. Two DNA:protein complexes were observed with each nuclear extract (Fig. 3A, lanes 2-4, I and II). Complex I was the most prominent in each case. Competition with 100-fold molar excess unlabeled site A oligonucleotide or a size-matched, unrelated DNA fragment confirmed that complexes I and II represented a specific protein-DNA interaction (Fig. 3A, lanes 5 and 6). Parallel experiments revealed that an identical mobility shift pattern was observed with oligonucleotide probes containing site B or C sequences and that a molar excess of unlabeled site B diminished or site C abolished the formation of complexes I and II (data not shown). These data suggest that identical or similar transcription factors bind MCAD promoter sites A-C in all three cell lines. Analysis of the DNA sequence of sites A-C revealed that each contained several homologous GC-rich sequences, including the known core consensus for binding members of the Sp1 transcription factor family (CCCGCCC) or a related motif, C(C/G)C(A/T)(C/G)(C/G)(C/G) (Fig. 1F). Competition experiments were performed with unlabeled oligonucleotides containing consensus binding sites for factors known to bind GC-rich sequences including Sp1(11Kadonaga J.T. Carner K.R. Masiarz F.R. Tjian R. Cell. 1987; 51: 1079-1090Abstract Full Text PDF PubMed Scopus (1335) Google Scholar), AP-2(18Mitchell P.J. Timmons P.M. Hebert J.M. Rigby P.W. Tjian R. Genes & Dev. 1991; 5: 105-119Crossref PubMed Scopus (510) Google Scholar, 19Williams T. Admon A. Luscher B. Tjian R. Genes & Dev. 1988; 2: 1557-1569Crossref PubMed Scopus (487) Google Scholar), nuclear respiratory factor-1(20Chau C.A. Evans M.J. Scarpulla R.C. J. Biol. Chem. 1992; 267: 6999-7006Abstract Full Text PDF PubMed Google Scholar), and myocyte nuclear factor (21Bassel-Duby R. Hernandez M.D. Yang Q. Rochelle J.M. Seldin M.F. Williams R.S. Mol. Cell. Biol. 1994; 14: 4596-4605Crossref PubMed Scopus (81) Google Scholar) in an attempt to identify the transcription factor that binds sites A-C. Competition was observed only with the oligonucleotide containing an Sp1 binding consensus sequence (Fig. 3B and data not shown). To confirm that Sp1 or a related protein was contained in complexes I and II, antibody "supershift" experiments were performed with anti-Sp1 antibody and the Y-1 cell nuclear protein extract. Addition of anti-Sp1 antibody, but not pre-immune serum, resulted in a mobility supershift of complex I, confirming that Sp1 was present in this DNA:protein complex (Fig. 3C, lanes 1-3). Identical results were obtained with an Sp1 consensus oligonucleotide probe and the Y-1 cell extract (Fig. 3C, lanes 4-6). Incubation of the site A probe with purified recombinant Sp1 formed a single complex that comigrated with complex I. In contrast to the partial supershift observed with the Y-1 nuclear extract, all of the DNA:recombinant Sp1 complex was supershifted by the anti-Sp1 antibody (Fig. 3C, lane 8). These data suggest either that the anti-Sp1 antibody does not recognize all forms of Sp1 present in complexes I and II or that other proteins participate in this DNA-protein interaction. Similar EMSA results were obtained using HepG2 and 3T3 cell-derived nuclear protein extracts with site A, B, and C probes (data not shown). Taken together, the transfection and DNA-protein binding studies indicate that Sp1 or a highly related transcription factor plays a major role in MCAD gene promoter function. Analysis of the DNA sequences of sites A-C reveals the presence of both typical and atypical Sp1 binding sites. Site C contains three sequences that match the known core Sp1 binding consensus (CCCGCCC, Fig. 1F). In contrast, sites A and B lack the Sp1 binding consensus but do contain several related sequences (CCCAGCC, CGCAGCG, and CCCTCCC) (Fig. 1F). Atypical Sp1 binding sites have been identified in promoters upstream of a variety of other genes, including those encoding the mouse secretory protease inhibitor P12(22Robidoux S. Gosselin P. Harvey M. Leclerc S. Guerin S.L. Mol. Cell. Biol. 1992; 12: 3796-3806Crossref PubMed Scopus (37) Google Scholar), the T-cell receptor α(23Kingsley C. Winoto A. Mol. Cell. Biol. 1992; 12: 4251-4261Crossref PubMed Scopus (491) Google Scholar), and alcohol dehydrogenases 2 and 3(24Brown C.J. Baltz K.A. Edenberg H.J. Gene (Amst.). 1992; 121: 313-320Crossref PubMed Scopus (19) Google Scholar). The wide variance in Sp1 binding site sequences may define diverse binding affinities necessary for competitive or synergistic interactions of Sp1 with other transcription factors. Alternatively, sequence-specific binding by related but distinct members of the Sp1 family may occur at sites A-C in vivo. In support of the latter possibility, several new Sp1 isoforms have recently been identified(23Kingsley C. Winoto A. Mol. Cell. Biol. 1992; 12: 4251-4261Crossref PubMed Scopus (491) Google Scholar, 25Hagen G. Muller S. Beato M. Suske G. Nucleic Acids Res. 1992; 20: 5519-5525Crossref PubMed Scopus (527) Google Scholar). We have shown previously that multiple nuclear receptor-mediated regulatory pathways converge on the complex MCAD gene promoter element NRRE-1 to bidirectionally modulate MCAD gene promoter activity(7Raisher B.D. Gulick T. Zhang Z. Strauss A.W. Moore D.D. Kelly D.P. J. Biol. Chem. 1992; 267: 20264-20269Abstract Full Text PDF PubMed Google Scholar, 15Carter M.E. Gulick T. Raisher B.D. Caira T. Ladias J.A. Moore D.D. Kelly D.P. J. Biol. Chem. 1993; 268: 13805-13810Abstract Full Text PDF PubMed Google Scholar, 16Carter M.E. Gulick T.G. Moore D.D. Kelly D.P. Mol. Cell. Biol. 1994; 14: 4360-4372Crossref PubMed Google Scholar, 17Gulick T. Cresci S. Caira T. Moore D.D. Kelly D.P. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 11012-11016Crossref PubMed Scopus (498) Google Scholar). Our cell culture cotransfection experiments revealed that NRRE-1 is activated by 9-cis retinoic acid and all-trans retinoic acid receptors in the presence of ligand(7Raisher B.D. Gulick T. Zhang Z. Strauss A.W. Moore D.D. Kelly D.P. J. Biol. Chem. 1992; 267: 20264-20269Abstract Full Text PDF PubMed Google Scholar). NRRE-1 also confers transcriptional activation by the orphan receptor, hepatocyte nuclear factor 4 (HNF-4)(15Carter M.E. Gulick T. Raisher B.D. Caira T. Ladias J.A. Moore D.D. Kelly D.P. J. Biol. Chem. 1993; 268: 13805-13810Abstract Full Text PDF PubMed Google Scholar), and repression by COUP-TF and its isoform, apolipoprotein regulatory protein 1(15Carter M.E. Gulick T. Raisher B.D. Caira T. Ladias J.A. Moore D.D. Kelly D.P. J. Biol. Chem. 1993; 268: 13805-13810Abstract Full Text PDF PubMed Google Scholar, 16Carter M.E. Gulick T.G. Moore D.D. Kelly D.P. Mol. Cell. Biol. 1994; 14: 4360-4372Crossref PubMed Google Scholar). These studies also demonstrated that additional unidentified COUP-TF-responsive element(s) are present in the MCAD gene promoter region downstream of NRRE-1. The DNase I protection studies shown here (Fig. 1) identified two additional MCAD promoter sites (NRRE-2 and NRRE-3) that contain hexamer sequences conforming to the known consensus (RG(G/T)TGNA) for binding to members of the thyroid/retinoid subgroup of the nuclear receptor superfamily (Fig. 1F). The DNA footprinting patterns suggested that these NRREs interact with multiple nuclear receptors in a cell line-specific and element-specific manner. We have shown previously that the orphan receptor HNF-4 present in nuclear extracts derived from HepG2 cells binds NRRE-1(15Carter M.E. Gulick T. Raisher B.D. Caira T. Ladias J.A. Moore D.D. Kelly D.P. J. Biol. Chem. 1993; 268: 13805-13810Abstract Full Text PDF PubMed Google Scholar). To identify the endogenous nuclear receptors that interact with the NRREs in 3T3 and Y-1 cells, double-stranded oligonucleotides containing the NRRE-1, NRRE-2, or NRRE-3 sequences were employed in EMSA with nuclear protein extracts prepared from these cells. With 3T3 cell nuclear extracts, NRRE-1 and NRRE-3 probes both formed single prominent complexes with identical mobilities (Fig. 4A). Formation of the complexes was inhibited by addition of 100-fold molar excess specific unlabeled oligonucleotide but not by an unrelated double-stranded DNA, confirming that the NRRE-1-protein and NRRE-3-protein interactions were specific. Cross-competition experiments revealed that the NRRE-1:protein complex could be specifically inhibited by unlabeled NRRE-3, and conversely the NRRE-3:protein complex was specifically inhibited by unlabeled NRRE-1 (Fig. 4A). Formation of the complexes formed with NRRE-1 or NRRE-3 probes was also inhibited by unlabeled NRRE-2, but to a lesser extent, suggesting that the NRRE-2 binds the nuclear protein(s) with lower affinity (Fig. 4A). Antibody supershift assays were performed with antisera raised to a variety of nuclear receptors to identify the specific factor that bound NRRE-1, NRRE-2, and NRRE-3 in 3T3 cells. Anti-COUP-TF antibody supershifted the entire DNA:protein complex formed with each of the probes (Fig. 4B). In contrast, pre-immune serum, anti-HNF-4, or anti-9-cis retinoic acid receptor α had no effect (Fig. 4B and data not shown). Thus, the orphan nuclear receptor COUP-TF, or a closely related protein, binds each of the NRRE sites in 3T3 cells. The low MCAD gene expression in 3T3 fibroblasts is consistent with the known transcriptional repressive effect of COUP-TF via the majority of known COUP-TF-responsive elements, including NRRE-1 (16, 26-28). Mobility shift experiments were also performed with the NRRE probes and Y-1 cell nuclear extracts. The NRRE-3 probe, but not NRRE-1 or NRRE-2 probes, formed a prominent complex with Y-1 cell nuclear extract (Fig. 4C, complex III, and data not shown). In addition, two faint complexes of lower mobility (I and II) formed with all three NRREs (Fig. 4C and data not shown). Competition experiments confirmed that complexes I, II, and III represent specific NRRE-3-protein interactions. Cross-competition with unlabeled NRRE-1 and NRRE-2 abolished complexes I and II but not complex III, indicating that only NRRE-3 contained a recognition sequence for the Y-1 cell nuclear protein present in this complex (Fig. 4C, lanes 4 and 5). The mobility shift pattern observed with the NRRE-3 probe and 3T3 cell nuclear extract was distinct from that with Y-1 cell extract (Fig. 4C, lanes 6 and 7). Antibody supershift and competition experiments were employed to identify the specific nuclear receptors present in the complexes obtained with the NRRE-3 probe and Y-1 cell nuclear proteins (Fig. 4D). The same panel of antibodies employed for the experiments shown in Fig. 4B were used. In addition, because complex III was obtained exclusively with nuclear extracts derived from Y-1 adrenal cells, a steroid-producing cell line derived from a mouse adrenal tumor, an antibody to SF-1 (the rabbit homolog of Drosophila nuclear receptor FTZ-F1) was used in these experiments(13Ikeda Y. Lala D.S. Luo X. Kim E. Moisan M.P. Parker K.K. Mol. Endocrinol. 1993; 7: 852-860Crossref PubMed Google Scholar). This anti-SF-1 antiserum has been shown previously to specifically abolish SF-1:DNA complexes (13Ikeda Y. Lala D.S. Luo X. Kim E. Moisan M.P. Parker K.K. Mol. Endocrinol. 1993; 7: 852-860Crossref PubMed Google Scholar) and virtually abolished the formation of complexes I-III (Fig. 4D). Anti-COUP-TF supershifted complexes I and II but had no effect on complex III, indicating that COUP-TF was also present in the two minor complexes. In contrast to the results obtained with 3T3 cell nuclear extract, the amount of COUP-TF:NRRE-3 complex formed with Y-1 cell-derived extract was minimal. Neither anti-9-cis retinoic acid receptor α nor anti-HNF-4 affected the mobility of any of the complexes (data not shown). These results demonstrate that the adrenal-specific orphan receptor, SF-1, binds NRRE-3. Comparison of the NRRE-3 sequence with the known SF-1/FTZ-F1 binding consensus sequence (PyCAAGGPyCPu) reveals a potential SF-1 binding site at +141 to +149 (TCAAGGCCG, Fig. 1F). The interaction of both SF-1 and COUP-TF at a single site (complexes I and II, Fig. 4D) has also recently been described for a transcriptional regulatory element in the oxytocin promoter(29Wehrenberg U. Ivell R. Jansen M. von Goedecke S. Walther N. Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 1440-1444Crossref PubMed Scopus (113) Google Scholar), suggesting that this may represent a general transcriptional regulatory mechanism. The biological role of SF-1 in regulating MCAD gene expression is not clear. SF-1 is an adrenal cortex-enriched orphan receptor that plays a crucial role in steroidogenesis and adrenal gland development(30Luo X. Ikeda Y. Parker K.L. Cell. 1994; 77: 481-490Abstract Full Text PDF PubMed Scopus (1418) Google Scholar). Acetyl-CoA, the major product of fatty acid oxidation, is necessary for the biosynthesis of steroid compounds in the adrenal cortex. Accordingly, it is possible that SF-1 coordinately regulates the expression of genes encoding enzymes involved in fatty acid β-oxidation and steroidogenesis. Our previously published work with NRRE-1 and the results of the transfection and DNA-protein binding studies shown here suggested that the MCAD gene promoter NRREs function as cell-specific transcriptional regulatory elements by interacting with multiple nuclear receptor transcription factors. To determine whether NRRE-3 mediates transcriptional activation or repression in accordance with cell-specific nuclear receptor-DNA interactions, a single copy of a double-stranded oligonucleotide containing NRRE-3 was inserted upstream of a TK promoter in a CAT reporter plasmid (pTKCAT(NRRE-3)) and transfected into Y-1 and 3T3 cells. The transcriptional activity of pTKCAT(NRRE-3) was compared to that of a pTKCAT lacking NRRE-3 (pTKCAT (-)). As predicted by the DNA binding studies, NRRE-3 conferred a 2.7-fold transcriptional activation in Y-1 adrenal cells and a 70% repression in 3T3 cells (Fig. 5). These data are consistent with the binding studies and indicate that, as with NRRE-1 (16Carter M.E. Gulick T.G. Moore D.D. Kelly D.P. Mol. Cell. Biol. 1994; 14: 4360-4372Crossref PubMed Google Scholar), either activator (SF-1) or repressor (COUP-TF) orphan receptors interact with NRRE-3 in a cell-specific manner to modulate transcriptional activity differentially in a cell-specific fashion. In contrast, NRRE-2 binds COUP-TF in all three cell lines. Evidence is emerging that Sp1 enhances cooperative interactions among multiple transcription factors to juxtapose the transcriptional regulatory domains of the proteins with the transcription initiation complex(31Seto E. Lewis B. Shenk T. Nature. 1993; 365: 462-464Crossref PubMed Scopus (253) Google Scholar, 32Lee J.S. Galvin K.M. Shi Y. Proc. Natl. Acad. Sci. U. S. A. 1993; 90: 6145-6149Crossref PubMed Scopus (280) Google Scholar, 33Su W. Jackson S. Tjian R. Echols H. Genes & Dev. 1991; 5: 820-826Crossref PubMed Scopus (265) Google Scholar, 34Dittmer J. Gegonne A. Gitlin S.D. Ghysdael J. Brady J. J. Biol. Chem. 1994; 269: 21428-21434Abstract Full Text PDF PubMed Google Scholar, 35Fischer K.D. Haese A. Nowick J. J. Biol. Chem. 1993; 268: 23915-23923Abstract Full Text PDF PubMed Google Scholar, 36Bessereau J.L. Mendelzon D. LePoupon C. Fiszman M. Changeux J.P. Piette J. EMBO J. 1993; 12: 443-449Crossref PubMed Scopus (52) Google Scholar). The interposition of Sp1 binding sites between multiple nuclear receptor response elements shown here suggests that Sp1-nuclear receptor cooperative interactions may play a role in MCAD gene promoter function and regulation. The lactoferrin (37Teng C.T. Liu Y. Yang N. Walmer D. Panella T. Mol. Endocrinol. 1992; 6: 1969-1981Crossref PubMed Scopus (51) Google Scholar) and Oct-4 (38Sylvester I. Scholer H.R. Nucleic Acids Res. 1994; 22: 901-911Crossref PubMed Scopus (73) Google Scholar) gene promoters also contain multiple Sp1 binding sites in close proximity to nuclear receptor response elements. In other gene regulatory regions, such as the human immunodeficiency virus long terminal repeat (39Desai-Yajnik V. Samuels H.H. Mol. Cell. Biol. 1993; 13: 5057-5069Crossref PubMed Google Scholar) and promoters of the genes encoding growth hormone (40Tansey W.P. Catanzaro D.F. J. Biol. Chem. 1991; 266: 9805-9813Abstract Full Text PDF PubMed Google Scholar), chorionic somatomammotropin(40Tansey W.P. Catanzaro D.F. J. Biol. Chem. 1991; 266: 9805-9813Abstract Full Text PDF PubMed Google Scholar), and the epidermal growth factor receptor(41Xu J. Thompson K.L. Shephard L.B. Hudson L.G. Gill G.N. J. Biol. Chem. 1993; 268: 16065-16073Abstract Full Text PDF PubMed Google Scholar), Sp1 binding sites overlap thyroid hormone receptor binding sites, indicating that, in some instances, Sp1 may compete with nuclear receptors in binding cognate DNA elements. Thus, competitive and cooperative interactions among Sp1 and other transcription factors, including members of the nuclear hormone receptor superfamily, are a potential mechanism, whereby promoter activity is modulated via many diverse signaling pathways. In summary, we have demonstrated that the MCAD gene promoter consists of a complex series of nuclear receptor response elements and Sp1 binding sites. We propose that multiple nuclear receptor transcription factors interact with the MCAD gene promoter in a cell-specific and element-specific manner to influence transcriptional activity in vivo. Our work strongly suggests that in cells and tissues with high capacity for fatty acid β-oxidation, MCAD promoter NRREs interact with tissue-specific expressed activator orphan nuclear receptors. In contrast, in low MCAD expressing cells and tissues, transcriptional repression is mediated by members of the COUP-TF family of receptors. Moreover, as we have recently shown, fatty acids regulate MCAD gene transcription via the orphan receptor PPAR through NRRE-1. Thus, this complex promoter structure allows multiple regulatory pathways to differentially regulate MCAD gene expression in a variety of cell types and in response to diverse metabolic and physiologic conditions. We thank Kelly Hall for expert secretarial assistance. Note Added in Proof-During review of this manuscript, Krey et al. (42Krey G. Mahfoudi A. Wahli W. Mol. Endocrinol. 1995; 9: 219-231Crossref PubMed Scopus (63) Google Scholar) reported that the acyl-coenzyme-A oxidase gene promoter contains multiple Sp1 sites and nuclear receptor response elements.