Orphan Receptors Chicken Ovalbumin Upstream Promoter Transcription Factor II (COUP-TFII) and Retinoid X Receptor (RXR) Activate and Bind the Rat Cholesterol 7 (cid:97) -Hydroxylase Gene ( CYP7A )*

The cholesterol 7 (cid:97) -hydroxylase gene ( CYP7A ) is tran-scriptionally regulated by a number of factors, includ- ing hormones, bile acids, and diurnal rhythm. Previous studies have identified a region from nucleotides (nt) (cid:50) 74 to (cid:50) 55 of the rat CYP7A promoter that enhanced bile acid repression of the SV40 early promoter, as as- sayed with a luciferase reporter gene in transiently transfected HepG2 cells. The rat CYP7A promoter/re- porter activity was strongly stimulated by cotransfection with an expression plasmid encoding the nuclear hormone receptor chicken ovalbumin upstream promoter transcription factor II (COUP-TFII) in a dose-de- pendent manner. Site-directed mutagenesis in the region of nt (cid:50) 74 to (cid:50) 55 altered this stimulation. Recombinant COUP-TFII expressed in HepG2 or COS-1 cells were found to bind to nt (cid:50) 74 (cid:50) 55 and nt (cid:50) 149 (cid:50) 128 probes by electrophoretic mobility shift assay (EMSA) and by supershifting the corresponding band with COUP-TFII-specific antibodies. The region of nt (cid:50) 176 (cid:50) 117 was previously mapped as a retinoic acid response region and was found to bind retinoid X receptor (RXR). EMSA supershift assays of wild-type and mutant oligomers using antibody against RXR revealed that the

The CYP7A promoter structure is typical of a DNA-dependent RNA polymerase II promoter in that, upstream of the TATA box, there are several cis-acting elements that regulate the transcriptional activity of this promoter (7)(8)(9). Transient transfection assay of chimeric CYP7A promoter/luciferase constructs in HepG2 cells revealed that the region from nt Ϫ416 to ϩ32 of the rat CYP7A gene contained the promoter and regulatory domains conferring the activation of transcription by dexamethasone and retinoic acid and suppression by bile acids, phorbol esters, and insulin (2,3,9,10). We have mapped two footprints that are protected from DNase I digestion using rat liver nuclear extracts (11). Footprint (Fp) 1 A is located between nt Ϫ81 and Ϫ35, and FpB is located between nt Ϫ148 and Ϫ129. Nucleotide sequences in these two regions are highly conserved among homologous CYP7A genes of different species. These footprints contain liver-enriched transcription factors binding sites and hormone response elements. A putative bile acid response element (BARE) was mapped to the FpA region, which lead us to propose that a bile acid responsive nuclear receptor may be mediating bile acid response (11).
The molecular mechanism of bile acid repression is still unknown. Determining what factors interact with these BARE sequences is a first step in elucidating the nature of the bile acid response. A nuclear protein factor, which was found to bind a repeated sequence -65TCAAGTTCAAG-54 was named direct repeat binding protein (DRBP). Binding of DRBP to BARE was diminished when liver nuclear extract prepared from rats fed deoxycholate was used in EMSA (11). The sequence from nt Ϫ73 to Ϫ57 contains an imperfect direct repeat of the hormone response element (HRE, AGGTCA) separated by four nucleotides (DR4). According to the 3-4-5 rule for the binding specificity of the steroid/thyroid hormone family of transcription factors (12), a DR4 would be predicted to act as a thyroid hormone response element, hence "7␣TRE" as the tentative name for this DR4 sequence (11). However, CYP7A promoter constructs containing these sequences did not respond to thyroid hormones in HepG2 cells (2). A number of nuclear factors have affinity for DR4 sequences, including TR, XONR, and LXR that form heterodimers with RXR (13,14). COUP-TFII (or ARP-1) of orphan nuclear receptor family with promiscuous binding specificities also bind a DR4 sequence (15,16).
COUP-TFs are believed to influence developmental changes, which is supported by the report that the COUP-TFII promoter responds to differentiation signal, retinoic acid (17). Recently, we have mapped a complex retinoic acid response region located between Ϫ176 and Ϫ117 of the rat CYP7A promoter (10).
Using transient transfection assay of CYP7A promoter/reporter activity in HepG2 cells, we have screened a number of known steroid/thyroid hormone nuclear receptors for their effects on gene transcription. COUP-TFII produced one of the strongest effects on transcription. Here, we provide evidence that orphan receptor COUP-TFII activates transcription of rat CYP7A by binding to nt Ϫ74 to Ϫ54 and to nt Ϫ149 to Ϫ128 and that an RXR homodimer may bind to a DR1 motif between Ϫ146 to Ϫ134.
Transfection Assays-Confluent cultures of HepG2 cells grown in 12-well tissue culture plates were transfected with DNA by calcium phosphate method as described previously (3,11). Luciferase activities were determined with the luciferase assay kit (Promega) according to manufacturer instructions using a Lumat LB9501 luminometer (Berthold Systems, Inc., Pittsburgh, PA). Luciferase activities were normalized for transfection efficiencies by dividing relative light units by ␤-galactosidase activity expressed from cotransfected pCMV␤ plasmid (Clontech). ␤-Galactosidase activities were determined using O-ni-trophenyl-␤-D-galactopyranoside as a substrate (18). The average of the corrected luciferase activity from cell extracts is given, and error bars indicate the standard deviation of activity from triplicate samples. All transfections were repeated at least two times. Statistical significances were analyzed by Student's t test using Sigma Plot software (Jandel Scientific, San Rafael, CA).
Preparation of Nuclear Extracts-Nuclei from HepG2 and COS-1 cells were isolated essentially as described (19,20). Recombinant COUP-TFII was prepared by transfecting confluent HepG2 or subconfluent COS-1 cells in 100-mm tissue culture plates with 20 g of pTF3A and 1 g of pCMV␤. The monolayers were overlaid with DMEM/ F12 containing 5 mM butyric acid and 10% fetal calf serum after glycerol shock, and nuclei were harvested 40 h later. Transfection efficiency was monitored by ␤-galactosidase activity.
Electrophoretic Mobility Shift Assay (EMSA)-Double-stranded synthetic probes for EMSA were prepared by heating equal molar amounts of complementary synthetic oligomers to 95°C in 2 ϫ SSC (0.3 M NaCl, 0.03 M Na 3 citrate, pH 7.0) and allowing them to cool to room temperature. The resulting double-stranded fragments were designed with single-stranded 5Ј overhangs for end-labeling by incorporating [␣-32 P]dCTP (3000 Ci/mol) with the Klenow fragment of DNA polymerase I. Oligonucleotides blunted with non-labeled dNTPs were used as cold competitors in EMSA. Labeled fragments were isolated from a 15% polyacrylamide gel or purified through two G-50 spin columns. Binding reactions were initiated with the addition of 2 g of nuclear extracts to 100,000 cpm of oligomer probe dissolved in 20 l of buffer containing 12 mM HEPES, pH 7.9, 50 mM KCl, 1 mM EDTA, 1 mM dithiothreitol, and 15% glycerol, and 2 g of poly(dI-dC)⅐poly(dI-dC). After incubation for the time indicated in the legends to Figs. 2-5, samples were run on 4% polyacrylamide gels, dried, and autoradiographed. EMSAs were quantitated with IP Lab Gel software (Signal Analytics, Corp., Vienna, VA) in conjunction with a PhosphorImager 445Si (Molecular Dynamics, Sunnyvale, CA).

Activation of CYP7A Promoter Activity by Recombinant COUP-TFII and Identification of a Negative Element-Co-
transfection of the rat CYP7A promoter fragment from nt Ϫ416 to ϩ32 fused to the luciferase gene (p416/ϩ32) (Fig. 1A) with a plasmid encoding recombinant COUP-TFII (pTF3A) was found to strongly stimulate reporter activity (Fig. 1B). The doseresponse curve resulting from cotransfecting with increasing amounts of pTF3A with p-200/ϩ32 was indistinguishable from the stimulation curve of p-416/ϩ32 (Fig. 1B) despite lower overall luciferase reporter activity (Fig. 1C), indicating the COUP-TF activation elements were downstream of position nt Ϫ200. Deletion of nt Ϫ74 to Ϫ54 (Fig. 1A) reduced the stimulatory effect of COUP-TFII (Fig. 1B).
To define the sequences important for COUP-TFII activation, a number of deletion (⌬) and linker scanning mutations (LSM) were introduced into the nt Ϫ74 to Ϫ54 region. The dose-dependent stimulation of the promoter activity of the reporter plasmids by COUP-TFII fell into roughly three groups. In the first group, the wild-type promoter fragments (p-416/ ϩ32 and p-200/ϩ32) showed the greatest stimulation of transcriptional activity, and the effect was positive at all points of the curve (Fig. 1B). The second group, which includes p-416LSMϪ74/Ϫ71, p-416LSMϪ70/Ϫ66, p-416⌬Ϫ74/Ϫ65, p-416LSMϪ59/Ϫ54, p-416⌬Ϫ74/Ϫ54, and p200⌬74/54 displayed considerably flatter dose-response curves (Fig. 1B). The mutations in this group destroyed one or both of the putative HRE half-sites, AGGTCA, either by completely deleting the sequence (p-416⌬Ϫ74/Ϫ65, p-416⌬Ϫ74/Ϫ54, and p200⌬74/54) or by replacing the sequence with an SpeI restriction site while preserving the relative spacing of sequences on either side of the mutation (p-416LSMϪ74/Ϫ71, p-416LSMϪ70/Ϫ66, and p-416LSMϪ59/Ϫ54) (Fig. 1A). Reporter plasmids with both HRE half-sites deleted (p-416⌬Ϫ74/Ϫ54 and p200⌬74/54) retained significant stimulation by COUP-TFII (approximately 3-4 fold), suggesting there may be additional COUP-TFII sites elsewhere in the promoter. Consistent with the hypothesis that the putative COUP-TFII activation elements were downstream FIG. 1. Effect of overexpressing recombinant COUP-TFII on transcriptional activity of the nt ؊416 to ؉32 fragment of the rat CYP7A promoter. A, schematic representation of mutations introduced into the region of nt Ϫ74 to nt Ϫ54 of the rat CYP7A promoter, as described under "Experimental Procedures." Wild-type sequence from nt Ϫ81 to Ϫ48 is diagrammed on top. All the clones have the promoter sequences from nt Ϫ416 to ϩ32 ligated into pGL2-Basic. Restriction sites are underlined, dashes in sequence represent deleted base pairs, and non-homologous sequences are indicated in lowercase letters. The DR4 is overlined, and the direct repeat described in (11) is boxed. B, increasing mass of the expression plasmid pTF3A encoding COUP-TFII was cotransfected with CYP7/luciferase reporter plasmids diagrammed in panel A. The total mass of plasmid was kept constant by the addition of vector without promoter. Confluent monolayers of HepG2 cells were transfected with 5 g of the indicated reporter plasmids, various amounts of pTF3A, and 0.5 g of pCMV-␤ as described under "Experimental Procedures" and then overlaid with DMEM/F12 medium without serum after glycerol shock. The cells were harvested 40 h after glycerol shock, and the luciferase activity was assayed and corrected for transfection efficiency with the ␤-galactosidase activity. The activation by COUP-TFII of the transcriptional activities were expressed as percentage of corrected luciferase activity without coexpressed COUP-TFII. The percentage activation differed from wild type p Ͻ 0.01 (p-200/ϩ32 versus p-200⌬-74/54; p-416/ϩ32 versus remainder of the mutants), as determined by Student's t test; error bars omitted for clarity. Data are representative of two individual experiments. C, basal transcriptional activities of CYP7A promoter/luciferase reporter plasmids with mutations in the region from nt Ϫ74 to Ϫ54. Error bars indicate standard deviation of the triplicate samples for each construct. *, indicates statistical significance for difference from wild type p Ͻ 0.01. To control for variation in measured activities due to differences in individual plasmid preparations, relative activities of p-416/ϩ32, p-416⌬Ϫ74/54Spe, p-416⌬Ϫ74/65, p-416⌬Ϫ59/54, p-416⌬Ϫ74/54 CYP7A promoter/luciferase gene constructs were further assayed with three separate plasmid preparations transfected in triplicate in two different experiments as described under "Experimental Procedures." The results obtained were essentially the same as those shown in panel C. of position nt Ϫ200, p-200⌬74/54 was stimulated to approximately the same extent as p-416⌬Ϫ74/Ϫ54, which has the same deletion but was 216 base pairs longer (Fig. 1B).
The third category was represented by an insertion mutant (IM), p-416IMϪ64/Ϫ63 that displayed an activation intermediate between the first and second groups. This plasmid, in which the HRE half-sites were intact but the spacing was increased from 4 to 5 base pairs (Fig. 1A), was strongly activated but to a somewhat lesser extent than the wild type, reaching a maximum of 1740 Ϯ 99% verses 2960 Ϯ 209%, respectively, at a transcription factor plasmid mass to reporter plasmid mass ratio of 0.2 (Fig. 1B).
The various mutations affected the basal activity. Mutant reporter plasmids p-416⌬Ϫ74/54, p-416LSMϪ74/Ϫ71, p-416LSMϪ59/Ϫ54, p-416⌬Ϫ74/65, and p200⌬74/54 had increased basal promoter activity relative to their parental plasmids (p-416/ϩ32 and p-200/ϩ32), as would be observed if a repressor binding site was mutated. With one exception, the basal promoter activity also fell into three groups, corresponding with the COUP-TFII responsiveness of the reporter plasmid (Fig. 1C). The promoter/reporter constructs with the lowest response to COUP-TFII had the highest basal activities (p-416⌬Ϫ74/54 and p-416LSMϪ74/Ϫ71). p-416IMϪ64/Ϫ63 basal transcriptional activity was not significantly higher than the wild type. The exception was p-416LSMϪ70/Ϫ66, which was activated by COUP-TFII to the same extent as p-416LSMϪ59/ Ϫ54 and p-416⌬Ϫ74/65 but did not display the same transcriptional activation over p-416/ϩ32. The activation of p-416⌬Ϫ74/ Ϫ54 over p-416/ϩ32 was not observed in the kidney cell line, COS-1.
Interaction of COUP-TFII with the CYP7A Promoter at nt Ϫ74 to Ϫ54 -To determine if COUP-TF activates transcriptional activity by directly binding to the sequence between nt Ϫ74 and Ϫ54, the ability of COUP-TF to interact with the CYP7A sequences in vitro, was measured with the EMSA. Also, other promoter fragments from nt Ϫ200 to Ϫ74 were screened to determine if there were additional COUP-TFII binding sites.
Nuclear extracts were prepared from HepG2 cells transfected with pTF3A, an expression plasmid specifically encoding COUP-TFII. The gel shift patterns of the nt Ϫ72 to Ϫ59 and nt Ϫ65 to 54 oligomers are similar in HepG2 cells, regardless of COUP-TFII overexpression (Fig. 2B). Two extra bands were shifted with nt Ϫ74 to Ϫ53 probe that were enhanced by COUP-TFII overexpression (Fig. 2). The nt Ϫ74 to Ϫ53 included the putative DR4, which is overlined in Fig. 2A; the nt Ϫ72 to Ϫ59 or nt Ϫ65 to 54 oligomers do not. The diffuse nature of these bands was attributed to the low abundance of the nuclear factors.
Recombinant COUP-TFII expressed in the kidney cell line COS-1, bound the nt Ϫ74 to Ϫ53 (Fig. 2C), indicating the binding of COUP-TF was not dependent on liver-specific factors. The increased band shift with COS-1 extracts was attributed to the higher activity of the MT2 promoter of pTF3A, which drives transcription of the recombinant gene encoding COUP-TFII, in COS-1 cells relative to HepG2 cells. This presumably results in higher levels of expression of recombinant COUP-TFII in this cell line. Overexpression of COUP-TFII was associated with a reduction in DRBP binding, even in probes unable to be shifted by COUP-TFII (nt Ϫ72/Ϫ59 and Ϫ65/Ϫ54). The reduction in DRBP binding may be attributed to the reduced proportion of DRBP in the given mass of protein in COS-1 nuclear extracts overexpressed with recombinant COUP-TFII. Alternatively, if COUP-TFII interacts with DRBP in solution, the overexpressed COUP-TFII may be sequestering the DRBP and forms larger complexes with the probe nt Ϫ74/ Ϫ53. COUP-TFII did not shift the probes lacking the COUP- TFII binding site (nt Ϫ72/Ϫ59 and Ϫ65/Ϫ54) (Fig. 2C). It is apparent that COUP-TFII interacts with the DR4 of nt Ϫ74 to Ϫ53 and DRBP interacts with AGTTCAAG sequence (Fig. 2A).
To further confirm that COUP-TFII binds to this sequence, antiserum raised against COUP-TF was added to an EMSA using the nt Ϫ74 to Ϫ55 oligomer as a probe of nuclear extract from HepG2 cells. The band that increased in intensity when shifted by nuclear extracts from HepG2 overexpressing COUP-TFII was supershifted by the anti-COUP-TF antibody (Fig. 3), indicating that COUP-TFII was interacting directly with the region. Extracts prepared from cells overexpressing recombinant C/EBP␣, LAP, HNF4, or HNF3␣ did not produce any changes in the band-shift pattern (data not shown), and antiserum directed against C/EBP␣ (potential binding site at nt Ϫ52 to Ϫ41) and c-erbA (a homologue of the thyroid receptor that is regulated by phosphorylation events) did not change the gel shift pattern (Fig. 3). Antibody against RAR and RXR did not supershift the nt Ϫ74 to Ϫ53 probe. However, when performed with extracts prepared from HepG2 cells treated with all trans-retinoic acid, the ligand for RAR, the nt Ϫ74 to Ϫ54 probe was weakly supershifted with anti-RXR (data not shown). DRBP, the major nuclear factor that binds to the sequence was not affected by the antibody, indicating that the proteins responsible for that band are distinct from COUP-TFII and RXR. These results confirm that the COUP-TFII binds to DR4 motif, which does not bind thyroid hormone receptor. C/EBP␣, LAP, HNF4, and HNF3␣ do not bind to this region of CYP7A promoter.
Interaction of COUP-TFII and RXR with the CYP7A Promoter at nt Ϫ149 to Ϫ128 -The probe nt Ϫ149 to Ϫ128 shifted one band in HepG2 nuclear extracts. Antibody against COUP-TFII supershifted the oligomer and changed the pattern of bands as if a complex was disrupted (Fig. 4). The amount of probe supershifted was not increased with increasing amounts of antibody, indicating that antibody was in excess. Antibody against RXR supershifted a band in HepG2 nuclear extract treated with retinoic acid. The antibody directed against RAR did not supershift the probe even with extract prepared from HepG2 cells treated with all trans-retinoic acid. Retinoic acid treatment may have increased the expression of endogenous RXR and showed the supershift with anti-RXR.
Once it had been determined that the antibody specific for RXR supershifted the nt Ϫ149 to Ϫ128 probe, nucleotides important for the binding of RXR were determined by EMSA with mutant oligonucleotides and the antibody against RXR. Double-stranded oligonucleotides were synthesized in which three bases were changed by transversions. Trinucleotide mutations in nt Ϫ145 to Ϫ134 completely abolished or reduced the supershift by antibody against RXR (Fig. 5). Mutations in nt Ϫ133/ Ϫ131 did not have any effect on supershift by antibody against RXR. These results indicate that RXR was interacting with the DR1 motif located from nt Ϫ146 to Ϫ134.

DISCUSSION
Cotransfection assays of COUP-TFII expression plasmid with wild-type and mutant CYP7A proximal promoter constructs and EMSA results indicate that COUP-TFII interacts with nt Ϫ74 to Ϫ53 and nt Ϫ149 to Ϫ128 of the rat CYP7A promoter. The region of nt Ϫ146 to Ϫ134 was shown by EMSA to bind RXR and contains HRE half-sites separated by one nucleotide (DR1), the preferred binding site for RXR ho- FIG. 4. EMSA in the presence of RAR-, RXR-, and COUP-TFspecific antibodies using the nt ؊149 to ؊128 probe. EMSA was performed using nt Ϫ149 to Ϫ128 as a probe, generated by digesting the double-stranded synthetic oligomer nt Ϫ149 to Ϫ118 with MspI (diagrammed, top; imperfect HRE half-sites underlined), end-labeling as described under "Experimental Procedures," and gel isolating the nt Ϫ149 to Ϫ128 fragment. The probe was incubated with 5 g of nuclear extracts from untreated HepG2 cells (HepG2) or cells treated with 10 M all-trans retinoic acid (HepG2 ϩ retinoic acid) for 24 h before harvesting. modimer (14). The pattern of response of the mutants in nt Ϫ74 to Ϫ54 suggested that the nt Ϫ74 to Ϫ54 was the element with the strongest potential for COUP-TF activation of transcription. The deletion of the sequences from nt Ϫ74 to Ϫ54 results in the increase of transcriptional activity of CYP7A promoter/ reporter plasmids. This sequence was previously found to mediate the bile acid response and to interact with DRBP, whose binding was sensitive to deoxycholate feeding in rats. Fig. 6 is a model which shows the binding sites for these transcription factors on the rat CYP7A proximal promoter. Interactions of transactivators COUP-TFII and RXR with DRBP and their competition for binding to these bile acid response elements may regulate the CYP7A gene transcription by bile acids and hormones. The CCAAT enhancer binding proteins (C/EBP␣ and C/EBP␤ or LAP) (21) consensus binding site flank the COUP-TFII binding site at nt Ϫ74 to Ϫ54. The C/EBP family of transcription factors are temporally regulated and are thus important developmental regulators of gene expression. COUP-TFII may recruit C/EBP on to the apolipoprotein A1 promoter (22). The DBP B-site, which also is the LAP binding site, was located in the nt Ϫ149 to Ϫ118 region (23). However, antibody directed against C/EBP␣ and LAP did not supershift the nt Ϫ149/Ϫ118 probe. This is consistent with the lack of effect of these transcription factors on CYP7A promoter activity we observed. The juxtaposition of the COUP-TF and C/EBP binding sites at nt Ϫ74/Ϫ54 and nt Ϫ149/Ϫ118 may prevent the weak binding of these factors.
In light of the well-characterized negative effect of COUP-TFII on gene promoters, removing a COUP-TFII binding site in CYP7A promoter may explain the much higher activity of the deletion mutants (Fig. 1). However, several recent reports support our finding that COUP-TFII also can function as a transactivator (24 -27). COUP-TFII binds as a stable homodimer to many variations of GGTCA sequences in direct repeats and interferes with vitamin D 3 , thyroid hormone, and retinoic acid receptors by competing for the AGGTCA sites, the cognate sequences for the steroid/thyroid hormone receptors (15). However, if the negative effect of COUP-TFII is due to interfering with a positive factor, over-expression of COUP-TFII should repress the transcriptional activity of the CYP7A promoter. COUP-TFII activated transcription of the CYP7A promoter at all COUP-TFII expression plasmid to reporter plasmid ratios tried. The deletion of the strong proximal COUP-TFII site at nt Ϫ74 to Ϫ55 may allow enhancers in the promoter to interact more readily with the transcriptional machinery and resulting in higher transcriptional activity. COUP-TFII has been reported to interact with TFIIB (28) and to affect the transcriptional activity by interacting with other transcription factors. For example, COUP-TFII is required for activation of the apolipoprotein A1 promoter by RXR (22). As the COUP-TFII and RXR both recognized the nt Ϫ149 to Ϫ128 probe, it is possible that COUP-TFII may interact with RXR. The heterodimer of COUP-TFII and RXR is known to bind a DR1 motif and transrepress gene transcription (13).
Despite the finding that COUP-TF binds to a region responsive to bile acid repression, the role of COUP-TFII in the bile acid regulation of CYP7A is not clear and requires further study. The region from nt Ϫ65 to Ϫ54 can confer the bile acid response to SV40 promoter, but the probe does not bind COUP-TFII in vitro (Fig. 2, B and C). However, the DRBP and COUP-TFII apparently share binding sequence, the 3Ј COUP-TFII half-site overlaps with the DRBP binding site ( Fig. 2A). COUP-TF clearly influences cholesterol metabolism, in that this transcription factor has previously been shown to affect the transcription of the promoters for apolipoprotein AI (16), apolipoprotein AII, apolipoprotein B (29), apolipoprotein CIII (29,30) and cholesteryl ester transfer protein gene (27).
The physiological role of the sequences from nt Ϫ74 to Ϫ54 may be to restrain CYP7A expression. This down-regulation may be necessary because of the cytotoxicity of bile acids and the need to preserve cholesterol for synthesis of steroid hormones and membrane components. The exact interactions of the factors involved has yet to be defined; however, the data are consistent with a model in which COUP-TFII, RXR, and DRBP interact with and bind to a negative element in the region between nt Ϫ74 and Ϫ54 and a positive element from nt Ϫ149 to Ϫ118 (10) and lead to the regulation of the CYP7A gene transcription in response to signals from bile acids and hormones. This work contributes to the increasing body of evidence indicating COUP-TF has a prominent role in transcriptional regulation of lipid metabolism and that factor binding to the nt Ϫ74 to Ϫ54 and nt Ϫ149 to Ϫ128 regions are important determinants of CYP7A transcription. FIG. 5. EMSA supershift in the presence of RXR-specific antibodies using probes based on the sequence from nt ؊149 to ؊128. EMSA was performed as in Fig. 4. Shown is the wild-type (WT) oligomer probe sequence on top with trinucleotide transversion mutations indicated below. Numbers over lanes correspond to the nucleotides mutated according to the following rule: A to C, C to A, G to T, and T to G. For example, 145/143 corresponds to the nt Ϫ149 to Ϫ128 probe with nt Ϫ145 G changed to T, nt Ϫ144 G changed to T, and nt Ϫ143 A changed to C. Mutation from wild-type is indicated with lowercase letters. Supershifted probes are indicated with arrow.
FIG. 6. Diagram of CYP7A proximal promoter with DRBP, COUP-TFII, and RXR recognition sites indicated in relation to the transcription start site, D-site binding protein (DBP) and liver activating protein (LAP) binding sites. DBP and LAP (C/ EBP␤) binding sites were determined previously (4,23).