Nuclear Orphan Receptors Regulate Transcription of the Gene for the Human Luteinizing Hormone Receptor*

An imperfect estrogen receptor half-site response element direct-repeat, located within the TATA-less promoter of the human luteinizing hormone receptor (hLHR), was identified as an inhibitory site for Sp1/Sp3-driven basal transcription. Isolation of proteins recognizing this site by yeast one-hybrid screening of a human placenta cDNA library revealed three nuclear orphan receptors, EAR2, EAR3/COUP-TFI, and TR4. Electrophoresis mobility shift assays demonstrated that the in vitro translated nuclear orphan receptors specifically bound the direct-repeat motif of the hLHR promoter. Also, endogenous EAR2 and EAR3/COUP-TFI from JAR cell and human testis and TR4 from testes bound this motif in electrophoresis mobility shift assays. Functional analyses in CV-1 cells showed that EAR2 and EAR3/COUP-TFI repressed the hLHR promoter activity by up to 70% in a dose-dependent and sequence-specific manner. Conversely, TR4 activated the hLHR promoter activity up to 2.5-fold through binding to the same cis-element. The stimulation was reversed by coexpression of EAR2 or EAR3/COUP-TFI, indicating their competitive binding for this site. Such recognition of a common cognate site by the proteins with antagonistic functions implies that a net regulation of the hLHR gene may result from the relative availability of repressors and activator in a physiological state. This also may contribute to the differential expression of the hLHR gene in gonadal and non-gonadal tissues.

The luteinizing hormone receptor (LHR) 1 is an essential G-protein-coupled receptor located on the plasma membrane of gonadal cells. It mediates gonadotropin signals and triggers intracellular responses that participate in maturation and function of the gonads as well as the regulation of steroidogenesis and gametogenesis (1,2). The LHR has also been identified in several non-gonadal tissues, including human nonpregnant uterus, placenta, fallopian tubes, uterine vessels, umbilical cord, brain, and lymphocyte, and in rat prostate (for review, see Ref. 2).
Characterization of the human LHR gene has demonstrated genetic polymorphism with at least two forms of sequence heterogeneity by the presence or absence of a 6-bp insertion that encodes a leucine-glutamine dipeptide within the exon I coding region (3). The minimal promoter has been identified within the 5Ј 176-bp region (from ATG), which contains two Sp1 domains of central importance for the transcription of the TATA-less hLHR gene (4,5). Furthermore, an estrogen response element half-site located at the Ϫ171-bp position was indicated to be a repressive element on basal transcription (5). Mutation at the consensus AGGTCA sequence caused about 100% induction of the promoter activity in human placental choriocarcinoma JAR cells and SV40-transformed placental cells, suggesting release of a putative repressor(s) from this site (5). Electrophoresis mobility shift assays (EMSAs) and mutagenesis revealed that multiple specific DNA-protein complexes were formed between the hLHR EREhs and JAR cell nuclear extracts (5). This indicated that more than one protein could converge upon this common cognate site, although it does not exclude the possibility that the bands represent several truncated forms of a single protein. The estrogen receptor was not considered as a candidate since the hLHR promoter lacks an intact ERE composed of an inverted half-site repeat with a 3-base pair spacer. This is consistent with the lack of supershift of the complexes by estrogen receptor antibodies and of an estrogen effect on the hLHR promoter activity. Also, the hLHR EREhs did not bind monomeric orphan receptors SF-1 and ERR-1 (5). The hexa-nucleotide sequence has been reported to serve as a core motif for hormone response elements recognized by other members of the steroid/orphan nuclear receptor subfamily (for review, see Ref. 6). Therefore, we propose that one or more of these receptors may bind the EREhs domain and participate in the transcriptional regulation of the hLHR gene expression during differentiation and development cascades.
In this report, we have employed a yeast one-hybrid system to isolate regulatory proteins through binding to a target sequence containing the EREhs domain. Three human nuclear orphan receptors, EAR2, EAR3/COUP-TFI, and TR4, were cloned and shown to specifically bind the EREhs region in which an imperfect direct-repeat motif was identified as a cognate binding site. Functional analysis showed that EAR2 and EAR3/COUP-TFI repressed the hLHR promoter activity, whereas TR4 activated hLHR gene transcription. Furthermore, we demonstrated that the inhibition of the hLHR gene in JAR cells was caused by endogenous EAR2 and EAR3/COUP-TFI proteins binding to the direct-repeat motif.

MATERIALS AND METHODS
cDNA Isolation by Yeast One-hybrid Screening-The MATCH-MAKER One-hybrid System (CLONTECH, Palo Alto, CA) was used to isolate the cDNA encoding for the EREhs domain binding proteins. The preparation of the target reporter constructs, integration of these constructs into yeast Saccharomyces cerevisiae strain YM 4271, isolation of plasmid from each candidate clone, as well as the screening procedures were preformed following established protocols recommended by the manufacturer. Basically, 4 tandem copies of 31-bp wild-type DNA frag-ment from the hLHR promoter (Ϫ176 to Ϫ145, 5ЈGTCGCAAGGT-CAAGGCAGAGCAGACTCAGCGG-3Ј) containing the EREhs were inserted into pHis, pHis-1, and pLacZ yeast reporter vectors. The pHis/ pHis-1 plasmids were then linearized and integrated into YM4271 chromosome to obtain yeast reporter strains YM-wtERE/His and YM-wtERE/His-1, respectively. After assessment of leaky His ϩ expression on titrated 3-AT (aminotriazole) (Sigma) plates, the YM-wtERE/His-1 was selected as the candidate strain. Then the pLacZ construct was linearized and integrated into the YM-wtERE/His-1 strain to obtain YM-wtERE/His/LacZ dual reporter strain. Another yeast reporter strain, YM-mERE/His-1, was also created in which mutation at the ERE half-site (m1) was introduced into the target sequence (5Ј-GTCG-CAATTTCAAGGCAGAGCAGACTCAGCGG-3Ј), and it was used in binding selectivity assay. The YM-wtERE/His/LacZ strain was then transformed with the GAL4 activation domain fused human placenta cDNA library (CLONTECH) and selected on histidine-deficient plates containing 30 mM 3-AT. Large colonies from the ϪHis plates were transferred onto Hybond N filters (Amersham Pharmacia Biotech) and further screened for ␤-galactosidase activity. The plasmid DNA, recovered from His ϩ /lacZ ϩ yeast colonies, was transformed into DH5␣ cells. Plasmid DNA extracted from bacteria was transformed back into both wild type and EREhs mutant yeast reporter cells to retest for His ϩ phenotype and the binding specificity. The cDNA from reproducible positive clones was partially sequenced. The nucleotide sequences were compared with sequences in the GenBank/EMSL data bases using Fasta program.
Reporter Gene Constructs and Expression Vectors-All plasmids were constructed by standard recombinant DNA techniques. A SacI/BglII fragment harboring the hLHR gene promoter region (Ϫ176 to ϩ1) was cloned into pGL2 basic vector (Promega, Madison, WI) just upstream of the luciferase reporter gene. Mutant constructs generated by polymerase chain reaction were designed to mutate the DNA sequence with putative protein binding activity. All mutants were verified by sequence analysis.
For construction of expression plasmids, DNA fragments containing full-length cDNAs of EAR2 (1.21 kilobases), EAR3/COUP-TFI (1.27 kilobases), and TR4 (1.79 kilobases) were amplified by polymerase chain reaction from yeast clones 9, 5, and 2, respectively. These three orphan receptors were cloned into pcDNA3.1 vector (Invitrogen, Carlsbad, CA). Full-length RAR␣ and RXR␤ expression plasmids were obtained in two steps. A 660-bp EcoRI/SalI fragment corresponding to N-terminal 220 amino acids of RAR␣ and a 420-bp EcoRI/NcoI fragment for the N-terminal 140 amino acids of RXR␤ was first amplified by polymerase chain reaction from QUICK-Clone cDNA of human placenta (CLONTECH). The resultant fragment was ligated with the respective C-terminal part of the molecule into pcDNA 3.1 vector. The fidelity of the clones was verified by DNA sequencing.
In Vitro Transcription and Translation of Nuclear Orphan Receptors-T7 promoter just upstream of the cloned genes in pcDNA 3.1 constructs was used to synthesize the three nuclear receptor proteins in an in vitro coupled transcription and translation system (TNT, Promega). The reaction was carried out in presence or absence of [ 35 S]methionine (Amersham Pharmacia Biotech) and 1 g of template DNA (linearized plasmid DNA). The molecular weights of the in vitro translated products were confirmed by electrophoresis/autoradiography before use in EMSA assays.
EMSA-Nuclear proteins from JAR cells (American Type Culture Collection, Manassas, VA) and normal human testis tissues (Cooperative Human Network, Eastern Division, Philadelphia, PA) were prepared as described previously (7). Oligonucleotide probes for the EM-SAs were end-labeled with [␥-32 P]ATP (3000 or 6000 mCi/mmol, NEN life Science Products) by T4 polynucleotide kinase (Life Technologies, Inc.). 5 l of EAR2, EAR3/COUP-TFI, or TR4 cDNA-programmed rabbit reticulocyte lysate or 5 g of nuclear protein was added to 20 l of binding reaction containing 12 mM HEPES (pH 7.9), 60 mM KCl, 4 mM Tris-HCl, 5% glycerol, 1 mM EDTA, 1 mM dithiothreitol, 0.5 mM ZnSO 4 , 1 g of polydeoxyinosinic deoxycytidylic acid, 1 g of sheared salmon sperm DNA (only for TNT proteins), and a 5-10 ϫ 10 4 cpm probe on ice for 15 min. As a control, the probe was also incubated with the same amount of unprogrammed TNT lysate. In competition assays, unlabeled DNA sequences were added to the reaction 15 min before the addition of the probe. In supershift assays, specific antibodies or 2 l of normal rabbit IgG (Santa Cruz Biotechnology, Inc., Santa Cruz, CA) were preincubated with proteins for 30 -60 min at 4°C before the probe was added. DNA-protein complexes were resolved on 5% native polyacrylamide gel electrophoresis containing 0.5 ϫ TBE (0.89 M Tris, 0.89 M boric acid, 0.02 M disodium EDTA for 10 ϫ TBE).
Cell Culture and Transient Transfection-CV-1 cells (African green monkey kidney cells, American Type Culture Collection) were routinely maintained in minimal essential medium (Life Technologies, Inc.) supplemented with 10% fetal bovine serum, nonessential amino acids, sodium pyruvate, and L-glutamine. JAR cells were maintained in RPMI 1640 medium (Life Technologies, Inc.) with 10% fetal bovine serum. JAR cells were transfected with 0.5 g of different hLHR promoter/ luciferase constructs and 0.3 g of pRSV-␤gal plasmid DNA using LipofectAMINE Plus reagent (Life Technologies, Inc.) according to the protocols from the manufacturer. CV-1 cells (2 ϫ 10 5 ) were seeded on 6-well culture plates 24 h before transfection. The medium was replaced before transfection with 1 ml of medium without serum and antibiotics. Cells were transfected with 0.5 g of promoter/luciferase construct and different doses (0.2, 0.4, and 0.6 g) of expression vectors harboring EAR2, EAR3/COUP-TFI, or TR4. 0.3 g of pRSV-␤gal plasmid DNA was also cotransfected as an internal control to normalize the transfection efficiency. The final DNA concentration was adjusted using empty vector DNA. Cells were replaced 3 h later with normal culture medium and collected 40 h after transfection. The cell lysates were extracted, and the luciferase activity was measured by luminometry (8) and normalized by ␤-galactosidase activity (9). All experiments were performed at least three times in triplicate. The statistical significance was evaluated by analysis of variance test. Fig. 1 illustrates the 176-bp promoter region of the human LHR gene in which the functional cis-elements regulating the promoter activity have been indicated. These include two Sp1-activating domains and a consensus ERE half-site (AGGTCA) located at the Ϫ171-bp position (from ATG) that displayed marked inhi-FIG. 1. Promoter structure of the hLHR. The DNA sequence of the 176-bp human LHR promoter is shown, and the nucleotides are numbered relative to the translation initiation codon (ATG, ϩ1). In vitro transcription start sites are indicated with arrows. The functional domains are shown in bold italics, and the putative direct-repeat motif resembling a nuclear orphan receptor cognate site is indicated by horizontal arrows placed under the hs1 (EREhs) and hs2. An adjacent nonconsensus GREhs is underlined. bition on the hLHR gene transcription in JAR-and SV40transformed placental cells (5). Since putative repressor proteins from the nuclear receptor family may bind to this site and binding selectivity could depend not only on the core EREhs sequence but also on the adjacent nucleotides, the sequences surrounding the EREhs were also considered. The lack of ATrich nucleotides preceding the EREhs indicated that nuclear receptors that bind to the AGGTCA sequence as monomers were not likely candidate ligands. For these reasons, initial studies were centered on an imperfect direct-repeat motif that contains the consensus EREhs (termed as hs1) and a putative imperfect half-site (hs2) 3Ј adjacent to the EREhs. Luciferasereporter gene analyses were carried out to evaluate the function of these putative binding sites as well as to address their binding specificity. Wild type hLHR promoter (WT) and constructs with mutated hs1 (m1) and hs2 (m2) were transfected into JAR cells. The luciferase activity of the WT hLHR promoter is shown in Fig. 2. Mutation at either the hs1 (m1) or the hs2 (m2) caused a marked increase of promoter activity by 100%, indicating that both hs1 and hs2 participated in the repression of the hLHR gene promoter. In contrast, mutation of the 3Ј adjacent GREhs (see Fig. 1) had no effect. These results demonstrated that the inhibition of hLHR transcription was conferred by the imperfect direct-repeat motif. The required functionality of the two half-sites indicated that the repressor protein(s) likely binds to the cognate site as a dimer. Members of the RAR/thyroid receptor subfamily of the nuclear receptor superfamily including many orphan receptors can bind to a direct-repeat sequence as homodimers or heterodimers (for review, see Refs. 6 and 10). Thus, it was reasonable to assume that one or more of these may contribute to the observed inhibition of the hLHR promoter.

Identification of an Imperfect Direct-repeat Motif That Mediates Repression on the hLHR Gene Transcription-
Isolation of Nuclear Receptors through Binding to the hLHR Promoter in Yeast-The yeast one-hybrid system was employed to identify regulatory proteins that specifically bind to the hLHR promoter via the direct-repeat motif. A target sequence containing tandem hs1, hs2, and adjacent sequences (designated as HS) was used to screen a human placenta cDNA library. Nine His ϩ and LacZ ϩ dual positive clones were selected from 3.5 million yeast transformants and then verified by restored His ϩ /LacZ ϩ phenotypes in the subsequent several rounds of transformation into wild-type reporter strain (see "Materials and Methods"). Sequence analysis of the nine positive clones followed by a search in the GenBank identified three human nuclear orphan receptors and two human retinoid acid receptors. Two clones with different lengths of 5Ј-and 3Ј-flanking regions encoded the full-length EAR2 receptor. Three clones corresponded to the EAR3/COUP-TFI orphan receptor; one of these was full-length, and the others were identical clones lacking 11 amino acids at the N terminus. The third gene was identified as the full-length orphan receptor TR4 in two clones. The other two clones that corresponded to RAR␣ and RXR␤ receptors were N-terminus-truncated forms lacking 110 amino acids and 38 amino acids, respectively.
These results suggested that multiple proteins recognized the same HS region, and this was confirmed in binding selectivity assays. A mutant yeast reporter strain in which the direct-repeat motif was disrupted through mutation of the hs1 site (m1) was established. The cDNAs corresponding to the five receptors were transformed into this mutant strain. The complete lack of growth (data not shown) of the transformants in histidine-deficient plates confirmed that we had cloned EAR2, EAR3/COUP-TFI, TR4, and RAR␣ and RXR␤ through their specific binding to the direct-repeat motif of the hLHR gene promoter.
Characterization of the Binding Specificity of the Nuclear Orphan Receptors for the hLHR Promoter in Vitro-The nuclear orphan receptors EAR2, EAR3/COUP-TFI, and TR4 belong to the same RAR/TR/orphan receptor subfamily within the nuclear receptor superfamily. These three receptors have been reported to recognize a cis-element that is composed of a directrepeat of the AGGTCA sequence, with elasticity regarding interspacing and nucleotide differences from the consensus core sequences (11). The initial identification of an imperfect directrepeat motif as an inhibitory domain of hLHR gene promoter indicated that the hLHR could be transcriptionally regulated by these orphan receptors. EMSAs were therefore carried out to characterize their binding to the hLHR promoter. The receptors were first in vitro translated, and [ 35 S]methionine was incorporated for detection of nascent protein synthesis. Expression of EAR2, EAR3/COUP-TFI, and TR4 cDNAs yielded proteins with the expected molecular mass of 42.9, 46.1, and 65 kDa, respectively, whereas no product was detected in mock translation control (not shown). EMSAs were performed using an oligonucleotide probe corresponding the HS region of the hLHR promoter (Ϫ176 to Ϫ145). Incubation of the probe with in vitro translated EAR3/ COUP-TFI resulted in formation of a single DNA-protein complex, whereas no band was observed when unprogrammed lysate (NP) was used (Fig. 3A, lanes 1 and 2). The complex was eliminated upon the addition of a 100-fold excess of unlabeled wild-type competitor (lane 3), demonstrating its specific binding to the HS region. The identity of the complex was verified by the evidence that it was completely supershifted by an EAR3/COUP-TFI antibody, but it was not affected by normal rabbit IgG (lanes 7 and 8). Furthermore, the EAR3/COUP-TFI complex was not competed by mutated oligomers m1 or m2 (lanes 4 and 5), indicating that both half-sites (hs1 and hs2) were required for the binding. The results were compatible with the notion that EAR3/COUP-TFI is able to recognize a broad range of direct-repeat motifs to which it binds as a stable homodimer (18). In addition, the GREhs, shown not to participate in the repression of hLHR gene (Fig. 2), when mutated, competed for the EAR3/COUP-TFI binding as the wild-type DNA (Fig. 3A, lane 6). This further confirmed that EAR3/ COUP-TFI specifically bound to the imperfect direct-repeat motif of the hLHR promoter.
Similar results were obtained when in vitro translated EAR2 and TR4 were investigated in EMSAs (Fig. 3, B and C). Both   FIG. 2. Identification of an imperfect direct-repeat element that mediates repression on the hLHR promoter activity. The wild-type 176-bp hLHR promoter/luciferase reporter gene (WT) and its mutant constructs with mutated hs1 (m1), hs2 (m2), or GREhs (m3) were transiently expressed in JAR cells. The promoter-less vector (Basic) was used as a negative control. Relative promoter activities are indicated as the percentage of the luciferase activity of the wild-type promoter (100%). The results were normalized with ␤-galactosidase activity and expressed as the mean Ϯ S.E. of three independent experiments of triplicate wells for each construct.
EAR2 and TR4 formed specific complexes with the HS probe (lanes 10 and 18), and these were absent in the controls lanes (lanes 9 and 17). The complexes were abolished by the wildtype competitor but not by oligomers with mutation at either hs1 or hs2 (lanes 11-13 and 19 -21). The mutant GREhs inhibited the binding as the wild-type oligomer (lanes 14 and 22). The supershifted bands caused by an EAR2 antibody (lane 15) and a TR4 antibody (lane 23) verified the binding attributed to the expressed receptors. No supershift was observed with normal rabbit IgG (lanes 16 and 24). Thus, the EMSAs illustrate that the nuclear orphan receptors EAR2, EAR3/COUP-TFI, and TR4 specifically bound the hLHR promoter region through the direct-repeat sequence.
In contrast, EMSAs failed to detect RAR␣ or RXR␤ binding to the hLHR HS probe. Furthermore, functional analyses in CV-1 cells did not show retinoid acid or 9-cis-retinoid acidmediated effect on hLHR transcription when the promoter was cotransfected with full-length RAR␣ or RXR␤ or both together (data not shown). These results are consistent with the notion that the RAR and RXR retinoid receptors, unlike EAR2, EAR3/ COUP-TFI, and TR4, require a restricted spacer for high affinity binding (10,(12)(13)(14). It is therefore evident that RAR␣ and RXR␤ were detected in yeast by the highly sensitive system employed in this study through binding to tandem copies of the HS sequence containing the direct-repeat motif with zero spacing (DR0). However, such binding is not present or is too weak (not detected in the EMSA) to mediate a change in the hLHR promoter activity.
Transcriptional Repression of the hLHR Gene by EAR2 and EAR3/COUP-TFI-The specific recognition of the hLHR promoter by the nuclear orphan receptors made it necessary to examine their potential functions in the regulation of the hLHR gene transcription. For these studies, cotransfection assays of the nuclear receptors with the hLHR promoter/luciferase reporter constructs (wild type or mutated hs1 or hs2) were carried out in CV-1 cells. In Fig. 4A, it is shown that EAR2 repressed the hLHR wild-type promoter activity but had no effect on the basic promoter-less construct. Increasing the dose of EAR2 lowered the promoter activity by up to 70%. In contrast, the hs1 mutation (m1) abolished the inhibitory effect of EAR2, indicating that this site was essential for EAR2 to exert its action on the hLHR gene. Furthermore, coexpression of EAR2 did not repress the hs2-mutated construct (m2) but inhibited the activity of the GREhs mutant promoter (m3) to the same extent as the wild-type promoter (Fig. 4B). The results demonstrated that the negative regulation of the hLHR gene by EAR2 depended on the presence of an intact direct-repeat motif but not on the adjacent GREhs. These findings were consistent with those derived from the EMSA binding analysis (Fig. 3). Taken together, it was shown that EAR2 was able to repress the hLHR promoter activity in a dose-dependent and sequencespecific manner.
Similar results were obtained with EAR3/COUP-TFI, which repressed the hLHR promoter activity by up to 55% in a dosedependent manner. The inhibition was also sequence-specific since the repression was eliminated by mutation of either hs1 or hs2 site but was not affected by the GREhs mutation (Fig. 4,  C and D). In the absence of EAR2 or EAR3/COUP-TFI in CV-1 cells, the wild-type hLHR promoter was released from the inhibition, and therefore, the luciferase activity of the hs1-or hs2-mutated construct was not different from the wild-type promoter. The effect of the orphan receptor TR4 on the hLHR promoter activity was next examined in the cotransfection experiments in CV-1 cells. Unexpectedly, cotransfection of TR4 with the wild-type hLHR 176-bp promoter/luciferase construct activated the transcriptional activity up to 2.5-fold in a dose-  1, 2, 9, 10, 17, and 18) or presence of unlabeled 100-fold excess wild-type (lanes 3, 11, and 19) or mutated oligomers (lanes 4 -6, 12-14, and 20 -22) or in the presence of normal rabbit IgG (N-IgG, lanes 8, 16, and 24)  dependent manner (Fig. 5A). The activation was abolished by mutation of either half-site (hs1 or hs2), but it was present in cells transfected with the GREhs mutant construct (Fig. 5, A  and B). The results show that TR4 is a transcriptional activator of the hLHR gene, and it exerts its function through the imperfect direct-repeat element. Occupancy of the same cognate site by EAR2, EAR3/COUP-TFI, and TR4 with apparent opposite functions suggested that they may antagonize each other in the regulation of the hLHR gene. This possibility was tested by coexpression of TR4 with equal amounts of expression plasmid for EAR2 or EAR3/COUP-TFI in the presence of the wildtype hLHR promoter. Upon the addition of EAR2 or EAR3/ COUP-TFI, the TR4-mediated induction was completely abolished, and it was replaced by a repression comparable with that exhibited by EAR2 or EAR3/COUP-TFI alone (Fig. 5C). This suggested that the three proteins competitively bound to the same site. Taken together, these results demonstrate that the hLHR gene was subject to transcriptional repression and activation by nuclear orphan receptors.
Transcriptional Repression of the hLHR Gene in JAR Cells Resulted from Endogenous Nuclear Orphan Receptors-The results obtained from cotransfection assays in CV-1 cells demonstrate that nuclear orphan receptors EAR2 and EAR3/ COUP-TFI potently repressed hLHR gene promoter activity in a sequence-specific manner. The subsequent studies were directed to determine whether the inhibition of hLHR gene in JAR cells was caused by the corresponding endogenous nuclear orphan receptors. EMSAs performed (Fig. 6A) by incubation of JAR cell nuclear extracts with the HS probe revealed the formation of three major DNA-protein complexes (a, b, and c as indicated by arrows, lane 1). The complexes a and b were competed by the 100-fold unlabeled wild-type competitor (WT, lane 2) but remained unchanged in the presence hs1 (m1)-or hs2 (m2)-mutated oligonucleotides (lanes 3 and 4). However, complex c was significantly but not completely abolished by the wild-type oligomer. It was also not fully retained upon the addition of mutated competitors when compared with the complexes a and b. This indicated that the complex c was composed of more than one protein, some of which bound the HS probe nonspecifically. In addition, the GREhs mutant oligomers competed binding of the complexes (lane 5) as the wild-type DNA. The results illustrated two major complexes (a and b), and part of the complex (c) bound strictly to the direct-repeat motif of the hLHR promoter. These findings were similar to the binding specificity results from the EMSAs using in vitro translated orphan receptors. Supershift assays using antibodies directed to the three receptors were carried out to elucidate the identities of the proteins binding to this site. The complex b was supershifted by the EAR2 antibody, and the complex a was supershifted by the EAR3/COUP-TFI antibody. The TR4 antibody had no influence on the observed complexes as the normal rabbit IgG (lanes 8 and 10), indicating TR4, did not participate in the regulation of the hLHR gene in JAR cells. This is consistent with the observed marked repression of the hLHR gene through the direct-repeat motif in these cells. A human SF-1 (steroidogenic factor 1) antibody against its DNA binding domain was also included in the supershift assays as a negative control (lane 9). SF-1 recognizes the AGGTCA as its cognate site to which it binds as a monomer, and the binding also strictly depends on the presence of AT nucleotides just preced- FIG. 4. Transcriptional repression of the hLHR gene by EAR2 and EAR3/ COUP-TFI. Left, the WT or hs1-mutated (m1) hLHR promoter/luciferase constructs were cotransfected into CV-1 cells with increasing dose (0.2, 0.4, and 0.6 g) of EAR2 cDNA (A) or EAR3/COUP-TFI cDNA (C). The promoter-less vector (Basic) was used as a negative control. Right, the WT or hs1 (m1)-, hs2 (m2)-, or GREhs (m3)-mutated hLHR promoter/luciferase constructs were contransfected into CV-1 cells with 0.6 g of EAR2 cDNA (B) or 0.6 g of EAR3/COUP-TFI cDNA (D). Relative promoter activities are indicated as the percentage of luciferase activity (100%) from the wild-type promoter in the absence of coexpression of nuclear orphan receptors. Results were normalized with ␤-galactosidase activity and expressed as mean Ϯ S.E. of three independent experiments in triplicate wells for each cotransfection.
ing the consensus half-site, which are not present in the hLHR sequences. Furthermore, RAR␣ and RXR␤ did not participate in the regulation of the hLHR promoter activity, since no supershift bands were formed upon the addition of the antibodies of these two receptors in the EMSAs. Also, retinoid acid hormone treatment of JAR cells in reporter gene analyses did not show any effect on the hLHR promoter activity (data not shown). In addition, the observation that complex c was not supershifted by the antibodies tested suggested that the specific component of this complex, yet to be determined, may also participate in hLHR gene regulation. Therefore, transcriptional repression of the hLHR gene in JAR cells resulted from endogenous EAR2 and EAR3/COUP-TFI proteins, if not exclusively. Furthermore, EMSAs using human testes nuclear extract were performed to examine the binding to the HS probe of the hLHR promoter by the nuclear proteins from normal gonadal tissues (Fig. 6B). Two major DNA-protein complexes (1 and 2) were formed (indicated by arrows, lane 11). Upon the addition of wild-type competitor, complex 2 was completely abolished, whereas complex 1 was significantly reduced (lane 12). Both complexes were retained in the presence of mutated hs1 (m1) and hs2 (m2) oligomers (lanes 13 and 14); however, complex 2 but not complex 1 was completely abolished by the mutated GREhs (lane 15, m3). The results indicated that complex 2 was a specific band whose formation was solely dependent on the direct-repeat motif. Antibody supershift assays demonstrated that this complex could be supershifted by the EAR2, EAR3/COUP-TFI, and TR4 antibodies (lanes 16 -18) but not recognized by the SF-1 antibody or normal rabbit IgG (lanes 19  and 20). The results demonstrated that the direct-repeat motif of the hLHR promoter was specifically bound by the orphan receptors EAR2 and EAR3/COUP-TFI in nuclear extracts from both JAR cells and gonadal tissues. In addition, TR4 present in the testes nuclear extracts also bound specifically to the same motif. Taken together, this is the first line of evidence showing that the hLHR gene is transcriptionally regulated by nuclear orphan receptors. DISCUSSION An imperfect EREhs direct-repeat motif, compatible with the recognition site bound by members of RAR/TR/orphan nuclear receptor subfamily, was identified in this study as an inhibitory element for hLHR gene transcription in JAR cells. This domain specifically bound the nuclear proteins in JAR cells and human gonadal tissues in the EMSA analyses. The isolation of three nuclear orphan receptors, EAR2, EAR3/COUP-TFI, and TR4, from a one-hybrid screening of a human placenta library in yeast indicated that multiple proteins could be involved in the hLHR promoter regulation through this element. EAR2 is a subtype of EAR3/COUP-TFI, and TR4 shares 70% and 50% sequence identity with EAR3/COUP-TFI at DNA binding domain and putative ligand binding domain, respectively (15). The conserved DNA binding domain of these orphan receptors confers their ability to recognize an identical/similar cognate element that is composed of a direct-repeat (DR) of AGGTCA core sequence with a spacer of variable length. The binding of these orphan receptors to the imperfect DR0 motif of the hLHR promoter was shown to be strictly mediated through the two half-sites (hs1 and hs2).
Members of the orphan receptor subfamily have been found to play an important role in gonad and brain development (16 -18). Therefore, it is of major significance that they are FIG. 5. Transcriptional activation of the hLHR gene by TR4. A, wild-type hLHR promoter/luciferase construct (WT) was cotransfected into CV-1 cells with increasing doses (0.2, 0.4, and 0.6 g) of TR4 cDNA. The basic promoter-less vector was used as negative control. B, the WT or hs1 (m1)-, hs2 (m2)-, or GREhs (m3)-mutated hLHR promoter/luciferase construct were cotransfected into CV-1 cells with 0.6 g of TR4 cDNA. C, the wild-type hLHR promoter/Luciferase construct was cotransfected into CV-1 cells with 0.6 g of TR4 cDNA in the absence or presence of coexpression of equal concentrations of EAR2 cDNA (0.6 g) or EAR3/ COUP-TFI cDNA (0.6 g). The relative luciferase activities were indicated as the percentage of the wild-type promoter activity (100%) in the absence of nuclear receptors. Results were normalized with ␤-galactosidase activity and expressed as the mean Ϯ S.E. of three independent experiments in triplicate wells for each cotransfection.
recognized as participants in the transcriptional regulation of the hLHR gene. Functional analysis demonstrated that EAR2 and EAR3/COUP-TFI repressed the hLHR promoter activity in a dose-dependent and sequence-specific manner. EAR2 and, particularly, EAR3/COUP-TFI are generally recognized as repressor proteins regulating an array of different target genes (for review, see Ref. 19). The mechanism of their repression includes active inhibition of basal or activated transcription, quenching a transactivator-regulated transcription, and transrepression (20). The inhibition of the hLHR gene appears to result from the active silencing function of EAR2 and EAR3/ COUP-TFI rather than competition with, quenching, or titrating out a hormonal partner (14,21,22).
EMSAs with JAR cell nuclear extracts revealed that EAR2 and EAR3/COUP-TFI antibodies caused supershift of two DNA-protein complexes of distinct migrating properties, indicating that recruitment of other cofactors via protein-protein interactions may participate in the regulation of hLHR gene transcription. EAR2 and EAR3/COUP-TFI were reported to interact directly with TFIIB, a component of basal transcriptional machinery (23). Their repressive functions were found enhanced by interaction with two common nuclear receptor corepressors, NcoR (nuclear receptor corepressor) and SMRT (silencing mediator of retinoic acid and thyroid hormone receptor (24)). Furthermore, EAR3/COUP-TFI has been shown to interact directly with Sp1 protein, independent of the EAR3/ COUP-TFI binding site, to enhance the activation of target gene transcription (25,26). It still remains to be determined whether EAR3/COUP-TFI-mediated repression of the hLHR promoter activity requires a direct interaction between EAR3/ COUP-TFI and Sp1 protein. However, if such interaction occurs, it would involve EAR3/COUP-TFI bound to its cognate site, since our cotransfection studies with the hs1-or hs2mutated constructs did not reveal the additional increase on the promoter activity over what was observed by exclusion of the inhibitory effect of the EAR3/COUP-TFI (Fig. 4B). The repression of the TATA-less hLHR gene by EAR2 and EAR3/ COUP-TFI is probably achieved by direct or corepressorbridged interaction with one or more components of the basal transcriptional machinery or, alternatively, by direct interaction with Sp1 to perturb the Sp1 driven hLHR gene transcription.
In contrast to the actions of EAR2 and EAR3/COUP-TFI, the nuclear orphan receptor TR4 was demonstrated to be a transcriptional activator of the hLHR promoter through the same direct-repeat motif (DR0). To our knowledge, this is first demonstration that TR4 regulates target gene transcription through a DR0 direct-repeat cis-element. It has been reported that TR4 up-regulated the rat ␣-myosin heavy chain and S14 genes through a DR4 motif, whereas it inhibited the SV40 gene through a DR2 and squelched RAR/RXR-mediated stimulation of several genes through the DR1/DR5 element (14,27,28). In addition, it has been recently proposed that TR4 may adopt different conformations once bound to a DR3 vitamin D response element (DR3VDRE) or a DR4 thyroid hormone response element (DR4T 3 RE). Such change would allow TR4 to recruit different coregulators and, hence, to function as a repressor through DR3VDRE or as an activator through DR4T 3 RE (21). It is proposed that the induction of the hLHR gene by TR4 through the DR0 element will be dependent on the context of its full promoter region, taking into account the possible participation of coactivators. Coactivator proteins as transcriptional mediator/intermediary factor 2, steroid receptor coactivator 1 (SRC-1), and human receptor-interacting protein 140 (RIP140) have been found to interact with several nuclear receptors (for review, see Ref. 29). However, it remains to be determined whether such interaction is present in TATAless genes and involved in the TR4-mediated induction of the hLHR gene. In JAR cells, the hLHR promoter appeared to be regulated solely by EAR2 and EAR3/COUP-TFI, since no DNA-TR4 complexes were observed in the EMSA. However, such complexes, including TR4, are readily observed in gonadal nuclear proteins, indicating that dual regulation of the hLHR gene by orphan receptors with opposite functions could occur in these tissues. Cotransfection of TR4 with EAR2 or EAR3/ COUP-TFI converted the marked activation of the hLHR promoter induced by TR4 into an inhibitory effect conferred by EAR2 or EAR3/COUP-TFI through their competitive occupancy of the same binding site. Therefore, binding of the hLHR promoter by these nuclear orphan receptors with apparent opposite functions implies that the net outcome of the hLHR gene transcription could be determined by the relative availability of the repressors (EAR2, EAR3/COUP-TFI) and the activator (TR4) at a physiological or pathological stage. Such a mechanism provides a mode for regulation of the Sp1-driven hLHR gene expression in various states of differentiation and development.