Prostaglandin F2α-induced Expression of 20α-Hydroxysteroid Dehydrogenase Involves the Transcription Factor NUR77*

Prostaglandin F2α (PGF2α) binding to its receptor on the rat corpus luteum triggers various signal transduction pathways that lead to the activation of a steroidogenic enzyme, 20α-hydroxysteroid dehydrogenase (20α-HSD), which in turn catabolizes progesterone. The molecular mechanism underlying PGF2α-induced 20α-HSD enzyme activity has not yet been explored. In this report we show, using mice lacking PGF2α receptor and pregnant rats, that PGF2α is responsible for the rapid and massive expression of the 20α-HSD gene at the end of pregnancy leading to a decrease in progesterone secretion. We also present evidence that PGF2α enhances 20α-HSD promoter activity. We have determined a region upstream of the −1590 position in the 20α-HSD promoter that confers regulation by PGF2α in ovarian primary cells. This region encompasses a unique transcription factor-binding site with a sequence of a NUR77 response element. Deletion of this motif or overexpression of a NUR77 dominant negative protein caused a complete loss of 20α-HSD promoter activation by PGF2α. NUR77 also transactivated the 20α-HSD promoter in transient transfection experiments in corpus luteum-derived cells (GG-CL). This induction required the NUR77-transactivating domain. We also show that PGF2α induces a very rapid expression of NUR77 that binds to a distal response element located at −1599/−1606 but does not interact with another proximal putative NUR77 response element located downstream in the promoter. A rapid increase in NUR77 mRNA was observed in mice corpora lutea just before parturition at a time when 20α-HSD becomes expressed. This increase in the expression of both genes was not seen in PGF2α receptor knockout mice. By using cyclosporin A and PGF2α treatment, we established that inhibition of NUR77 DNA binding in vivo prevents PGF2α induction of the 20α-HSD gene in the corpus luteum. Taken together, our results demonstrate, for the first time, that PGF2α induces in the corpus luteum the expression of the nuclear orphan receptor and transcription factor, NUR77, which in turn leads to the transcriptional stimulation of 20α-HSD, triggering the decrease in serum progesterone essential for parturition.

In all mammalian species, progesterone plays an essential role in reproduction. The precise timing of both the synthesis and degradation of this steroid hormone is crucial for reproductive success. The expression of enzymes implicated in the synthesis and catabolism of progesterone, therefore, needs to be accurately regulated during the different reproductive states of the animal. In rodents, the corpus luteum, which is the only source of progesterone throughout pregnancy (1), is also able to express the enzyme 20␣-hydroxysteroid dehydrogenase (20␣-HSD) 1 that converts progesterone into a biologically inactive steroid, thus playing a key role in the termination of pregnancy and allowing parturition to occur (2). Due to the detrimental effect of 20␣-HSD on luteal progesterone secretion, the corpus luteum of pregnancy does not express 20␣-HSD until 24 h before parturition (3). Our laboratory has previously demonstrated that the rapid appearance of luteal 20␣-HSD activity in the corpus luteum is due to the massive increase in 20␣-HSD gene expression and not to activation of an already existent enzyme (4).
Prostaglandin F 2 ␣ (PGF 2 ␣) administration to pregnant (5-8) and pseudopregnant (9) rats increases luteal 20␣-HSD activity. However, the molecular mechanism involved in 20␣-HSD activation is not known. PGF 2 ␣ is also known to induce abortion in many species including rodents (5,10). This last effect seems to be primarily a luteolytic one since progesterone administration prevents abortion. In rodents the participation of PGF 2 ␣ in luteolysis and the induction of labor was further demonstrated recently by the finding that mice deficient in the gene for the PGF 2 ␣ receptor (PGF 2 ␣-R) do not show the normal pre-partum drop in progesterone and do not exhibit parturition (11,12). However, whether the high levels of progesterone secreted at the end of pregnancy in knockout mice are due to the absence of 20␣-HSD activity and whether PGF 2 ␣ stimulates the activity of an already present enzyme or enhances/induces the expression of 20␣-HSD gene are still unknown.
To understand the molecular basis of PGF 2 ␣ regulation of 20␣-HSD enzyme activity, we first examined the effect of PGF 2 ␣ administration in vivo on 20␣-HSD gene expression and the expression of 20␣-HSD mRNA in PGF 2 ␣-R knockout mice. To determine whether PGF 2 ␣ stimulation involves transcrip-tional activation, we used both primary rat ovarian cells and a corpus luteum-derived cell line (GG-CL) to perform a series of gene transfer studies to examine the mechanisms mediating the effect of PGF 2 ␣ on 20␣-HSD promoter activity. We have shown that PGF 2 ␣ is responsible for the abrupt luteal expression of 20␣-HSD at the end of pregnancy. We have mapped a region located between position Ϫ1599 and Ϫ1606 containing a NUR77 response element that contributes to PGF 2 ␣-mediated activation. We have also provided evidence that PGF 2 ␣ stimulates the expression of the NUR77 transcription factor, which transactivates the 20␣-HSD promoter. In addition, by using an in vivo approach, we have shown that inhibition of NUR77 binding to DNA prevents PGF 2 ␣-mediated stimulation of 20␣-HSD gene expression.

EXPERIMENTAL PROCEDURES
Chemicals-Acrylamide and bisacrylamide were obtained from Accurate Chemical Inc. (Westbury, NY); Kodak NTB-2 liquid emulsion and Bio-Max AR Film were from Eastman Kodak Co.; [␣-32 P]dCTP was purchased from Amersham Pharmacia Biotech; Advantage RT-for-PCR kit was purchased from CLONTECH (Palo Alto, CA); dNTP, ExTaq DNA polymerase and ExTaq buffer were purchased from Takara Biomedicals (Shiga, Japan); the nucleotides used as primers in the RT-PCR analysis were obtained from Life Technologies, Inc.; Western blotting Luminol Reagent was obtained from Santa Cruz Biotechnology (Santa Cruz, CA); RPMI 1640 medium, nonessential amino acids, sodium pyruvate, trypsin-EDTA, antibiotics, and antimycotics were purchased from Mediatech (Herndon, VA). PGF 2 ␣, D-glucose, Tri-Reagent, aprotinin, leupeptin, PMSF, and all other reagent-grade chemicals were purchased from Sigma.
Animals-Pregnant Harlan Sprague-Dawley rats (day 1 ϭ spermpositive) purchased from Sasco Animal Labs (Madison, WI) were housed at 24°C with a 14-h light, 10-h dark cycle (lights on 0500 -1900 h) and allowed free access to Purina Rat Chow and water. PGF 2 ␣ receptor knockout mice with a mixed genetic background of 129/Ola and C57BL/6 strains were used (11,12). Wild-type and PGF 2 ␣ receptor knockout mice were maintained at 23°C under a 12-h light cycle. Virgin females (9 -12 weeks of age) housed overnight with males were checked the following morning for vaginal plug. The day the plug was found was counted as day 1 of pregnancy. Animal care and handling conformed to the National Institutes of Health Guidelines for Animal Research. The experimental protocol was approved by the Institutional Animal Care and Use Committee.
Granulosa-luteinized Cells-27-28-day-old immature female Harlan Sprague-Dawley rats (Sasco, Madison, WI) were injected with 15 IU of pregnant mare serum gonadotropin intraperitoneally followed by 15 IU of human chorionic gonadotropin intraperitoneally 46 h afterward. Seven hours later ovaries were isolated and incubated sequentially in 6 mM EGTA in DMEM/F-12 and 0.5 M sucrose in DMEM/F-12. Cells were cultured at 37°C in Dulbecco's modified Eagle's medium/Ham's F-12 (DMEM/F-12, 1:1). After 3 days of culture, the cells were transfected by LipofectAMINE (Life Technologies, Inc.) with 0.5 g of 20␣-HSD-Luc constructs and 0.5 g of a control ␤-galactosidase expression vector (Life Technologies, Inc.) following the manufacturer's protocol. The cells were left overnight and then treated with PGF 2 ␣ for 12 h. To harvest cells, each well was washed twice with ice-cold PBS.
GG-CL-Cells derived from rat luteal cells were cultured in RPMI 1640 medium supplemented with L-glutamate, 1 mM nonessential amino acids, 1 mM sodium pyruvate, an additional 0.45% glucose, 1ϫ antibiotics and antimycotics, 10 g/ml nystatin, and 10% FBS at 33°C. Trypsinized cells were suspended in OPTI-MEM I (Life Technologies, Inc.) at 1 ϫ 10 6 cells/0.8 ml and transfected by electroporation with 0.5 g of 20␣-HSD-Luc construct and 0.5 g of a control ␤-galactosidase expression vector and a different amount of NUR77 expression plasmid using the gene pulser (Bio-Rad) at a capacitance setting of 975 microfarads and 280 V. The total amount of DNA per well was kept constant by the addition of an empty vector. After each electroporation, cells were pooled and resuspended in RPMI 1640 medium containing 10% FBS and plated into 6-well plates at a density of 1 ϫ 10 6 cells/well. Cells were cultured overnight and then the medium was changed to phenol red-free RPMI 1640 with 1% FBS. Cells were maintained for a maximum of 48 h after electroporation. To harvest cells, each well was washed twice with ice-cold PBS.
Luciferase Activity Measurement-Passive lysis buffer (100 l) (Promega) was added into each well, and 20 l of cell lysate was used to measure both firefly luciferase activity driven by the 20␣-HSD promoter. 20 l of cell lysate was also used to measure ␤-galactosidase activity using Promega's Luciferase Reporter or ␤-Galactosidase Assay System, respectively, in a Lumat LB 9507 Luminometer (EG & G Berthold). Relative light units were obtained by dividing the luciferase activity by the ␤-galactosidase activity.
Electrophoretic Mobility Shift Assay-To prepared whole cell extract, one 100-mm plate of GG-CL cells transfected with a NUR77 expression plasmid or empty vector were harvested in PBS and centrifuged for 5 min at 12,000 ϫ g to obtained the pellet. The pellet were resuspended in 150 l of 10 mM Tris-HCl buffer, pH 7.9, containing 1 mM EDTA, 1 mM dithiothreitol, 10% glycerol, 400 mM KCl, 1 mM vanadate, and leupeptin, aprotinin, and pepstatin (2 g/ml). Cells and nuclei were lysed by three rapid freeze-thaw cycles and centrifuged at 12,000 ϫ g, and protein concentration in the supernatant was measured (Bradford method, Pierce). Nuclear extract from corpus luteum of day 19 pregnant rats was obtained as described by Dignam et al. (13) with slight modifications. Corpora lutea were homogenized in solution A (10 mM Hepes, pH 7.9, 10 mM KCl, 1.5 mM MgCl 2 , 0.1 mM EGTA, 0.5 mM PMSF, 0.5 mM dithiothreitol), and the cells were broken by plunging the cells 10 -15 times in a Dounce homogenizer. The nuclear pellet was obtained by centrifugation for 30 min at 4°C in an Eppendorf centrifuge and resuspended in solution B, which was similar to solution A except that it contained 420 mM NaCl and 5% (v/v) glycerol and no KCl. The solution was rocked for 30 min at 4°C and then centrifuged at 14,000 ϫ g at 4°C for 20 min. The supernatant containing the nuclear extract was divided into portions and was stored at Ϫ80°C. In vitro translated NUR77 was produced with reticulocyte lysate kits purchased from Promega (TNT SP6/T7). The efficiency of protein synthesis was monitored by [ 35 S]methionine labeling of the reaction products, and about 10 ng of in vitro synthesized NUR77 was used. For electrophoretic mobility shift assays, 10 g of protein extract were incubated with the NUR77-binding sites found in the rat 20␣-HSD promoter (distal site, 5Ј-GCC ATG TGG GTA CTG GAA AAT GAA CAC AGA-3Ј, or the proximal site, 5Ј-TAG CCT CTT AAA TGG TCA TTA TAA TTCA CAA-3Ј) (50,000 cpm, 20 fmol) in 20 l of EMSA buffer (20 mM Hepes, pH 7.6, 1 mM dithiothreitol, 1.5 mM MgCl 2 , 10 M ZnCl 2 , 90 mM NaCl, 10% glycerol, NaF 20 mM), plus 2 g poly(dI-dC) as nonspecific competitor were incubated for 30 min at 22°C. Competition assays were carried out by including 10-, 50-, or 100-fold molar excess of unlabeled oligonucleotide in the reaction mixture. A 30-mer mutant oligonucleotide 5Ј-GCC ATG TGG GTA CTG GAcgcT GAA CAC AGA-3Ј (lowercase represents NUR77 RE mutations) was also used. For supershift assays, 0.5 or 1 l of NUR77 antiserum (Santa Cruz Biotechnology, Santa Cruz, CA) was added to the reaction mixture and incubated for 30 min at 22°C before the addition of the labeled probe. The DNA protein complexes were separated from the unbound DNA probe by non-denaturing PAGE (4% gel) at 4°C, in 0.5ϫ Tris borate EDTA buffer (44 mM Tris-HCl, 44 mM boric acid, 12.5 mM EDTA, pH 7.5).
RNA Isolation and RT-PCR Analysis-Total RNA from frozen rat corpus luteum or mouse ovary was isolated using Tri-Reagent following the manufacturer's instructions. For mRNA analysis by RT-PCR, 1 g of total RNA was reverse-transcribed at 42°C using Advantage RT-for-PCR kit (Promega, Madison). The PCR mixture containing specific oligonucleotide primers (20 pmol), [␣-32 P]dCTP (2 Ci of 3000 Ci/mmol), dNTP (150 M), ExTaq DNA polymerase (0.8 units), was added to each tube containing 5 l of reverse transcription product. Each PCR included primer for rat or mouse ribosomal protein L19 mRNA used as internal control. Before proceeding with the semi-quantitative PCR, the conditions were established such that the amplification of the products was in the exponential phase, and the assay was linear with respect to the amount of input RNA. After autoradiography, data were analyzed using a Molecular Dynamics PhosphorImager and ImageQuant version 3 software (Molecular Dynamics, Sunnyvale, CA).
Northern Blot Analysis-Total RNA was fractionated by electrophoresis in 1% agarose gels and blotted to nylon membranes. Ethidium bromide staining indicated whether ribosomal RNAs were intact and whether equal amounts of RNA were loaded in each lane and an equal rate of transfer. Full-length of rat 20␣-HSD cDNA (18) or rat NUR77 cDNA (19) was labeled with [␣-32 P]dCTP using the random hexamer primer and the Klenow fragment of Escherichia coli DNA polymerase. Blots were prehybridized overnight at 42°C in a solution containing 40% formamide, 6ϫ SSC, 5ϫ Denhardt's, 20 mM Na 2 HPO 4 , pH 7.0, 0.2% sodium dodecyl sulfate, and 100 g/ml heterologous DNA. Hybridization was completed in the same solution containing 32 P-labeled cDNA probe (1 ϫ 10 6 cpm/ml) at 42°C overnight. Blots were washed and then exposed to Kodak X-AR films (Kodak) with intensifying screen at Ϫ80°C.
Western Blotting Analysis-Ovaries from wild-type or PGF 2 ␣ receptor knockout mice were homogenized on ice by hand using a Potter-Elvejhem homogenizer in ice-cold lysis buffer (10 mM Tris-Cl, pH 8.0; 150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1% SDS, 40 M PMSF, 0.3 M aprotinin, and 1 M leupeptin). This was followed by a 30-min incubation on ice and centrifugation at 10,000 ϫ g for 20 min at 4°C. The supernatant was transferred to new tubes, aliquoted, and stored at Ϫ70°C until the time of electrophoresis. An aliquot of the supernatant was kept for protein measurement using bovine serum albumin as a standard. Samples were denatured by adding sample buffer (62.5 mM Tris-HCl, pH 6.8, 2% SDS, 10% glycerol, 0.01% bromphenol blue) followed by boiling for 10 min. 30 g of protein were separated on 10% SDS-PAGE gels in Tris glycine, 0.1% SDS buffer, and transferred to nitrocellulose paper in 25 mM Tris, 192 mM glycine, and 20% methanol buffer at 250 mA for 1.5 h. The blots were incubated for 2 h at room temperature in 5% non-fat dry milk to block unspecific binding. The blots were then washed and incubated with the polyclonal antibody against rat 20␣-HSD (20) (1:4000 dilution), overnight at 4°C, and then washed and incubated with a secondary antibody conjugated to horseradish peroxidase (1:6000 dilution) for 2 h at room temperature. Protein-antibody complexes were visualized using Western blotting Luminol Reagent following the manufacturer's protocol. The band densities were determined by digital analysis using a Kodak Digital Science DC 120 Zoom Digital Camera and Kodak Digital Science 1D 2.0.2 software (Kodak).
Radioimmunoassays-Serum progesterone concentrations were measured using a commercially obtained kit (Diagnostic Products Corp., Los Angeles, CA). The sensitivity of the assay was 0.02 ng/ml, and the inter-and intra-assay coefficients of variations were 5 and 6%, respectively. Serum 20␣-OH-progesterone was assayed after hexane extraction using a highly specific antiserum kindly provided by Dr. Quadri (Department of Anatomy/Physiology, Kansas State University, Manhattan, KS). The sensitivity of the assay was 0.01 ng/assay tube, and the inter-and intra-assay coefficients of variation were less than 10%.
Statistical Analysis-One-way analysis of variance (ANOVA I) followed by the Tukey test was used for the statistical analysis of plasma steroid concentrations, relative mRNA expression, and luciferase activity data using the Prism software (Graph Pad Software, Inc., San Diego, CA). Values were considered statistically significant at p Ͻ 0.05.

PGF 2 ␣ Induces 20␣-HSD Gene Expression in Rats and
Mice-To determine whether PGF 2 ␣ can induce 20␣-HSD gene expression, either PGF 2 ␣ (400 g/rat intraperitoneal) or vehicle (saline solution) was administered to rats on day 19 of pregnancy 2 days before this gene becomes expressed in the corpus luteum. Twenty four hours later (day 20 of pregnancy), corpora lutea were dissected from each rat and pooled independently. Total RNA was isolated and subjected to RT-PCR using L19 as an internal control. As expected from previous results (4), no 20␣-HSD expression could be detected in the corpus luteum of all vehicle-treated rats (Fig. 1A, left panel,  lanes 1, 3, and 5). In contrast, high expression of 20␣-HSD mRNA was found in corpora lutea of PGF 2 ␣-treated rats (Fig.  1A, left panel; lanes 2, 4 and 6) accompanied by a significant reduction in the circulating levels of progesterone (Fig. 1A, middle panel) and a rise in the levels of 20␣-OH-progesterone (Fig. 1A, right panel). To examine the time course of PGF 2 ␣ stimulation, day 19 pregnant rats were injected with PGF 2 ␣ (400 g/rat intraperitoneal) and sacrificed 0.5-24 h thereafter. Total RNA was isolated and subjected to Northern blot analysis. The results (Fig. 1B) show that PGF 2 ␣ treatment induces 20␣-HSD gene expression within 2 h, and mRNA levels increased progressively for up to 24 h.
To confirm further the stimulatory effect of PGF 2 ␣ on 20␣-HSD expression at the end of pregnancy in rodents, we exam-FIG. 1. Effect of PGF 2 ␣ treatment on rat luteal 20␣-HSD mRNA levels and serum progesterone and 20␣-OHprogesterone levels. A, rats were injected with either 400 g of PGF 2 ␣ intraperitoneally (ϩ) or vehicle (Ϫ) on day 19 of pregnancy. Luteal 20␣-HSD mRNA levels by RT-PCR and serum steroid levels by radioimmunoassay were determined 24 h after administration. Bars represent the mean Ϯ S.E. (n ϭ five animals). ***, p Ͻ 0.001 versus vehicle (Ϫ). B, rats on day 19 of pregnancy were treated with 400 g of PGF 2 ␣ intraperitoneally, and corpora lutea were isolated at 0.5-24 h thereafter. Total RNA was prepared, and Northern blot analyses were performed by using 20 g of each sample and the rat 20␣-HSD cDNA as probe. Photography of the blot membrane after transfer is shown as control of loading and transfer.
ined by RT-PCR and Western blot analysis the expression of this gene on days 18 -20 of pregnancy in wild-type and PGF 2 ␣ receptor knockout mice. 20␣-HSD mRNA ( Fig. 2A) and protein (Fig. 2B) were undetectable on day 18 in corpora lutea of wild-type mice but became abruptly expressed on day 19 and remained elevated on day 20, the day of parturition. In sharp contrast, no 20␣-HSD expression could be detected in corpora lutea of PGF 2 ␣ receptor-deficient mice on any day examined, establishing further the importance of PGF 2 ␣ in the induction of 20␣-HSD at the end of pregnancy. This also provides an explanation as to why progesterone secretion remains elevated and why parturition is prevented in PGF 2 ␣ receptor knockout mice. Indeed, along with the appearance of 20␣-HSD protein in wild-type mice corpora lutea was a rise in the circulating levels of 20␣OH-progesterone and a drop in progesterone (Fig. 2C). PGF 2 ␣-R knockout mice did not show these changes. Serum progesterone levels remained elevated (11) (Fig. 2C) and 20␣-OH-progesterone level remained low.
PGF 2 ␣ Enhances 20␣-HSD Promoter Activity in Granulosaluteinized Cells-To investigate further the effect of PGF 2 ␣ on the expression of 20␣-HSD gene, we transfected luteinized granulosa cells with a 2.5-kb 20␣-HSD promoter, previously isolated in our laboratory, linked to a luciferase reporter gene (20␣-HSD-Luc) (21). The cells were also co-transfected with an internal reference plasmid expressing ␤-galactosidase, allowing the transfection efficiencies to be normalized. PGF 2 ␣ treatment induced a 4 -5-fold increase in the activity of 2.5-kb 20␣-HSD-Luc construct (Fig. 3A). Deletion of the Ϫ2467 to Ϫ1590 5Ј-fragment completely blunted responsiveness to PGF 2 ␣. PGF 2 ␣ had no effect on the activity of constructs carrying further downstream deletions of the 20␣-HSD promoter.
These results indicate that a region upstream of Ϫ1590 in the promoter confers regulation by PGF 2 ␣.
Analysis of the Ϫ2467/Ϫ1590 20␣-HSD promoter region revealed the presence of one perfect binding site (ACTGGAAAA) for the transcription factor NUR77 at position Ϫ1599 to Ϫ1606. Another proximal, less perfect, putative binding site, with a T insertion (Ϫ1539) AAATGGTCA (Ϫ1531) was found (Fig. 3B).
NUR77-induced Expression of 20␣-HSD-nur77 is an immediate-early gene which is known to play a major role in hor-  mone-induced expression of other steroidogenic enzymes (22)(23)(24). Since the region in the 20␣-HSD promoter that confers responsiveness to PGF 2 ␣ contains a perfect NUR77 RE, we examined whether this transcription factor may mediate the stimulatory effect of PGF 2 ␣. To test this possibility we examined first whether the overexpression of NUR77 protein affects the 20␣-hsd gene expression in a cell line derived from the rat corpus luteum, named GG-CL cells (25). We have previously demonstrated that the rat 20␣-HSD promoter is active in this cell line (21). We have also observed that this cell line does not express NUR77 under basal conditions (25). To assess whether NUR77 could affect the endogenous expression of 20␣-HSD, we transfected GG-CL cells with either a NUR77 expression plasmid or with a dominant negative NUR77 (DN NUR77), which encodes a NUR77 protein lacking the N-terminal transactivation domain (26). Controls were transfected with an empty plasmid. As a member of the steroid/thyroid hormone receptor superfamily, NUR77 is composed of a non-conserved N-terminal domain that plays a role in the transcriptional regulation, a highly conserved central zinc finger DNA-binding domain, and a moderately conserved C-terminal "ligand-binding" domain (27,28). The mutant NUR77 protein forms specific DNAprotein complexes (26).
Transfection of NUR77 into GG-CL cells enhanced the endogenous expression of 20␣-HSD (Fig. 4A, compare lanes 1 and  2). No stimulation of 20␣-HSD mRNA levels was observed with the dominant negative NUR77 (Fig. 4A, lane 3). Transfection of equal amounts of either NUR77 or DN NUR77 constructs results in approximately equivalent levels of protein expression (26).
To determine the effect of NUR77 on 20␣-HSD promoter activity, we co-transfected the 2.5-kb 20␣-HSD promoter luciferase construct with different concentrations of NUR77 expression vector or with DN NUR77 expression vector into GG-CL cells. The total amount of DNA per well was kept constant by the addition of empty plasmid. The cells were also co-transfected with an internal reference plasmid expressing ␤-galactosidase, allowing the transfection efficiencies to be normalized. The results (Fig. 4B) showed a dose-related stimulation of the 2.5-kb 20␣-HSD-Luc activity by NUR77, whereas no stimulation was observed with DN NUR77.
Deletion or Mutation of the Distal NUR77-binding Site Abolishes the NUR77 Stimulation of 20␣-HSD Promoter Activity-As mentioned before, the 2.5-kb 20␣-HSD promoter contains two putative NUR77-binding sites. The 5Ј distal putative NUR77 RE is a perfect consensus NUR77-binding site, 5Ј-ACTGGAAAA-3Ј, whereas the proximal putative NUR77-binding site is imperfect with an insertion of a T, 5Ј-AAATGGTCA-3Ј. As shown in Fig. 4C, the Ϫ1590/ϩ49 construct lacks the distal site but retains the proximal NUR77-like binding-like sequence at Ϫ1531, whereas both sites were deleted in the Ϫ289/ϩ49 construct. Deletion of the distal NUR77 consensusbinding site abolished the NUR77-induced stimulation of 20␣-HSD promoter activity. NUR77 did not affect the activity of the construct lacking either the distal or both putative NUR77binding sites (Fig. 4C). In order to confirm the participation of the distal NUR77-binding site in the regulation of 20␣-HSD, we mutated the ACTGGAAAA sequence to ATCGGAcgc. This mutation has been shown to prevent NUR77 DNA binding (27). The activity of the mutated promoter was evaluated in a tran- sient co-transfection assay and compared with that of the wildtype promoter. Mutation of the distal NUR77 RE prevented the stimulation of the 20␣-HSD promoter activity induced by NUR77 (Fig. 4C).

NUR77 Binds to the Distal NUR77 RE Found in the 20␣-HSD Promoter
Regions-To test NUR77 interaction with the 20␣-HSD promoter, we examined whether whole-cell extracts from GG-CL cells transfected with the NUR77 expression vector binds the putative NUR77-binding sites found in this promoter. Gel shift analysis demonstrated that the distal site, which is a consensus NUR77-binding sequence, forms a major protein-DNA complex with extracts from GG-CL cells transfected with the NUR77 expression vector (Fig. 5, lane 2) and with an in vitro translated NUR77 protein (lane 4). This complex was not detected when labeled probe was incubated with extracts from cells transfected with an empty plasmid (lane 1). The addition of unlabeled probe to the reaction mixture inhibited the formation of the DNA-protein complex (lane 3). In contrast, no protein-DNA complex was detected when the proximal site was used (Fig. 5, lanes 6 and 7). This experiment demonstrates clearly that the distal NUR77 RE binds NUR77, whereas the imperfect proximal NUR77 RE does not.
nur77 Expression Is Increased in Pregnant Rats by Treatment with a Luteolytic Dose of PGF 2 ␣-Since our present results have demonstrated that both PGF 2 ␣ and NUR77 stimulated 20␣-HSD mRNA expression and promoter activity, we thought that the PGF 2 ␣ stimulation of 20␣-HSD may be mediated by NUR77. To examine this possibility, we first analyzed the effect of PGF 2 ␣ on luteal NUR77 expression. Northern blot analysis (Fig. 6A) shows that administration of PGF 2 ␣ to rats on day 19 of pregnancy increased NUR77 mRNA expression within 0.5 h. Like many immediate-early response genes, NUR77 mRNA reached maximal levels after 30 min of treatment and diminished with time. This pattern of nur77 expression has been also observed in Y1 adrenocortical tumor cells where NUR77 mediated the increase of P450 C21 gene expression induced by ACTH (22).
To compare the temporal expression of NUR77 and 20␣-HSD mRNA levels after PGF 2 ␣ administration, we determined their levels by RT-PCR using primer that allows the co-amplification of these mRNAs. As shown in Fig. 6B, the rapid induction of nur77 gene expression by PGF 2 ␣ in the rat corpus luteum precedes that of the 20␣-HSD supporting the possible participation of this transcription factor in the PGF 2 ␣-induced 20␣-HSD expression.
FIG. 6. PGF 2 ␣ induce a rapid increase in nur77 gene expression in rat corpus luteum at the end of pregnancy. A, rats on day 19 of pregnancy were treated with 400 g of PGF 2 ␣ intraperitoneally and sacrificed at 0.5-24 h thereafter. Total RNA was prepared, and Northern blotting analyses were performed by using 20 g of each RNA sample and a rat NUR77 cDNA as a probe. Photography of the blot membrane after transfer is shown as control of loading and transfer. B, RT-PCR analysis was performed using plasmids that allow the coamplification of the NUR77 and 20␣-HSD messages. Normalized mRNA levels are graphically represented in the bottom panel. Bars represent means Ϯ S.E., n ϭ 3.
FIG. 5. NUR77 binds to the distal NUR77 RE found in the 20␣-HSD promoter region. EMSA were carried out using 32 P-labeled double-stranded oligonucleotide probes representing the distal (lanes 1-4) or the proximal (lanes 5-7) putative NUR77-binding site. Lanes 1 and 5, whole-cell extracts of GG-CL cells transfected with empty plasmid. Lanes 2 and 6, whole-cell extracts of GG-CL cells transfected with a NUR77 expression plasmid. Lanes 3 and 7, whole-cell extracts of GG-CL cells transfected with a NUR77 expression plasmid plus 50 molar excess of the homologous unlabeled probe. Lane 4, EMSA was performed using in vitro translated NUR77 protein.
To confirm the stimulatory effect of PGF 2 ␣ on NUR77, we examined the expression of this transcription factor in corpora lutea of wild-type and PGF 2 ␣-R knockout mice. An increase in nur77 expression was seen at the end of pregnancy in the wild-type mouse (Fig. 7). Similar to 20␣-HSD (see Fig. 2A), an increase in nur77 expression did not take place in corpora lutea of mice lacking PGF 2 ␣ receptor (Fig. 7).
PGF 2 ␣ Induces Binding of NUR77 to the 20␣-HSD Promoter-We next examined whether PGF 2 ␣-induced NUR77 binds to the functional NUR77 RE found in the 20␣-HSD promoter. EMSA analysis showed that nuclear extracts from corpora lutea of PGF 2 ␣ treated rats formed two prominent complexes, I and II (Fig. 8, lane 2), as compared with the control rat, where only one major shifted complex is formed (complex I) (Fig. 8, lane 1). Complexes I and II were inhibited by 10, 50, and 100 molar excess of unlabeled probe (lanes 3-5). Oligonucleotides containing the mutated distal NUR77 RE ACTGGAAAA motif to ACTGGAcgc did not competitively inhibit protein binding (lane 6), indicating that a perfect NUR77 RE sequence is essential for the protein/DNA interaction. Furthermore, formation of complex II was not inhibited by an oligonucleotide containing the proximal imperfect NUR77 response (lane 7). The formation of a broad shift band by NUR77, observed also by Okabe et al. (29), might be due to post-transcriptional modification (phosphorylation in multiple sites), which has been identified as the cause of the heterogeneity in NUR77 protein size (26,30,31).
To determine whether NUR77 is present in either of these complexes, we preincubated luteal nuclear extracts with a monoclonal antibody that specifically recognizes NUR77 and inhibits DNA-NUR77 complex formation (29). The addition of 0.5 l of this antibody (Fig. 8, lane 9) prevented the formation of the PGF 2 ␣-induced complex II but had no effect on complex I, indicating the participation of NUR77 protein in the formation of this complex. When 1 l of NUR77 antibody was added, complex I was also decreased, indicating that NUR77 may be part of this complex found in corpus luteum of untreated rats, which agrees with basal labels of NUR77 expression found during gestation in rats (data not shown). These results demonstrated that the PGF 2 ␣-stimulated NUR77 binds to its response elements in the 20␣-HSD promoter.
To confirm, in vivo, that the binding of NUR77 to the 20␣-HSD promoter induced by PGF 2 ␣ is responsible for the increase in the expression of this enzyme, we used cyclosporin A (CsA), which has been shown to prevent NUR77 binding to its response element (30). CsA was locally administered into the FIG. 7. NUR77 expression at the end of pregnancy in wild-type or PGF 2 ␣ receptor knockout mice. Total RNA was subjected to RT-PCR analysis using specific primers for mouse NUR77 and L19 as internal control. Data were quantified by densitometry and corrected using L19 as an internal standard. Normalized mRNA levels are graphically represented in the right panel as the means Ϯ S.E., n ϭ 3. ***, p Ͻ 0.01 versus the previous day of gestation.

FIG. 8.
PGF 2 ␣-induced NUR77 binds to the distal NUR77-binding site found in the 20␣-HSD promoter. Rats on day 19 of pregnancy were treated with 400 g of PGF 2 ␣ intraperitoneally, and luteal nuclear proteins were isolated 3 h after. An oligonucleotide containing the distal NUR77-binding site was used to perform EMSA analysis. Lane 1, luteal nuclear extract from vehicle-treated rats; lanes 2-9, luteal nuclear extract from PGF 2 ␣-treated rats; lanes 3-5, unlabeled distal site probe in 10-, 50-, or 100-fold molar excess; lane 6, 50-fold molar excess of unlabeled mutated distal NUR77 RE; lane 7, 50-fold molar excess of unlabeled proximal NUR77 RE; lanes 8 and 9, 1 or 0.5 g of a NUR77 antibody was added to the reaction mixture prior to the addition of labeled probe. ovarian bursa 30 min prior to intraperitoneal injection of PGF 2 ␣ or vehicle. Corpora lutea isolated 2 h after treatment were used to obtain nuclear protein for gel mobility shift assay, whereas total RNA was isolated 24 h after treatment to measure 20␣-HSD mRNA levels. As shown in Fig. 9A (lane 1), nuclear extract from vehicle-treated rats resulted in a retarded radiolabeled complex I. PGF 2 ␣ treatment caused a second prominent shift in complex II, containing NUR77 as show in Fig. 8, to form (Fig. 9A, lane 2) and induced high levels of 20␣-HSD mRNA expression (Fig. 9, B and C, compare lanes 1  and 2). CsA pretreatment prevented the formation of the NUR77-DNA complex II induced by PGF 2 ␣ (Fig. 9A, lane 3) and reduced markedly the PGF 2 ␣-induced 20␣-HSD mRNA expression (Fig. 9, B and C, lane 3). Administration of CsA to vehicle treated rat did not affect 20␣-HSD expression (Fig. 9, B  and C, lane 4).
Finally, to examine further whether the activation of 20␣-HSD promoter by PGF 2 ␣ is mediated by NUR77, we determined whether dominant negative NUR77 can prevent the induction of 20␣-HSD promoter activity by PGF 2 ␣. Luteinized granulosas cells were co-transfected with the 2.5-kb 20␣-HSD-Luc promoter together with 1 or 2 g of the mutant NUR77 (DN NUR77) or empty plasmid as control. Cells were then treated with either PGF 2 ␣ or vehicle. As shown before in Fig. 3, PGF 2 ␣ caused a severalfold increase in the 20␣-HSD promoter activity (Fig. 10). Increasing amounts of DN NUR77 inhibited this stimulation. 2 g of the mutant NUR77 prevented totally the stimulation on 20␣-HSD promoter activity induced by PGF 2 ␣ confirming that the stimulation of promoter activity by PGF 2 ␣ is mediated by the transcription factor, NUR77.

DISCUSSION
The stimulatory effect of PGF 2 ␣ on the activity of the 20␣-HSD enzyme in the rat corpus luteum is well recognized (5)(6)(7)(8). The corpus luteum expresses PGF 2 ␣ receptor (32)(33)(34) and is able to produce the ligand (35,36). However, whether PGF 2 ␣ activates an already pre-existing enzyme or whether it induces/ increases the expression of the 20␣-HSD gene is unknown. In addition, the cellular signals linking the PGF 2 ␣ receptor to the 20␣-HSD gene remains uninvestigated. In this report, we have identified the 20␣-HSD gene as a novel target for the transcription factor NUR77, and we have demonstrated that PGF 2 ␣ stimulates rapidly the expression of NUR77, which binds to a specific response element in the 20␣-HSD promoter and stim-ulates its transcriptional activity. Moreover, we have shown that NUR77 and 20␣-HSD become expressed in the corpus luteum only at the end of pregnancy and that these expressions are totally obliterated in PGF 2 ␣ receptor knockout mice. These results together with our ability to prevent (i) the PGF 2 ␣induced expression of 20␣-HSD in vivo by inhibiting NUR77 DNA binding activity and (ii) the PGF 2 ␣-induced increase in 20␣-HSD promoter activity in vitro by overexpressing a dominant negative NUR77 protein give evidence for the involvement of NUR77 in the PGF 2 ␣-mediated functional transcriptional activation of 20␣-HSD gene at the end of pregnancy.
In contrast to transcription factors that are constitutively present and ready to be activated by post-translation modification or by interaction with allosteric effectors, NUR77 is encoded by an immediate-early gene, whose expression is tightly regulated by extracellular signals. Classified as an orphan nuclear receptor, NUR77 displays the tripartite domain structure of members of the steroid receptor family but binds no known ligand. However, when synthesized, NUR77 is constitutively active under all conditions examined (19) and binds as a monomer to a specific response element (27,37) which consists of the classical estrogen receptor half-site preceded by two adenines (5Ј-AAAGGTCA-3Ј). In contrast to other orphan receptors that bind rather promiscuously to their response elements and other response elements (38), NUR77 is highly selective for its response element. Indeed our results show clearly that despite the presence of two putative binding sites in the 20␣-HSD promoter, NUR77 associates only with the perfect response element at Ϫ1606/Ϫ1599. NUR77 was shown previously to regulate the transcriptional activity of two hydroxylase genes, the rat P450 C21 (22,23) and the mouse P450 C17 (24), leading to increased synthesis of cortisol and androgen, respectively. Results of the present investigation have revealed a role for NUR77 in the transcriptional activity of another steroidogenic gene, which encodes an enzyme with dehydrogenase activity and causes the catabolism of progesterone synthesized in the corpus luteum. NUR77 is a homologue to the steroidogenic factor 1 that is constitutively active in many steroidogenic tissues and binds to an element similar to but distinct from NUR77 (38). In contrast to steroidogenic factor 1, NUR77 is present at low levels under basal conditions and becomes rapidly and highly expressed in response to external stimuli (39,40). In addition to its effect on steroidogenic FIG. 9. Cyclosporin A, an inhibitor of NUR77 action, prevents the PGF 2 ␣-induced NUR77-DNA complex formation and the increase in 20␣-HSD expression. Rats on day 19 of pregnancy were injected with cyclosporin A (3 g/ovary) or vehicle (metil cellulose) locally into the ovarian bursa 30 min prior to the administration of 400 g of PGF 2 ␣ intraperitoneally. A, luteal nuclear extract, from either control rats (lane 1), PGF 2 ␣-treated rats (lane 2), or CsA plus PGF 2 ␣-treated rats (lane 3) were isolated 2 h after PGF 2 ␣ administration. An oligonucleotide containing the distal NUR77binding site was used to performed EMSA analysis. B, total RNA was prepared 24 h after the administration of PGF 2 ␣, and RT-PCR analysis was performed using primers for rat 20␣-HSD and L19 as internal control. C, densitometry analysis of 20␣-HSD expression normalized against ribosomal L19 mRNA expression. Bars represent means Ϯ S.E., n ϭ 3. Columns with different letters differ significantly; a-b, p Ͻ 0.01 (ANOVA I).
genes, NUR77 has been implicated as modulator of the retinoic acid signaling pathway (41,42) and in the apoptotic process in T-lymphocytes. Indeed, the induction of NUR77 is required for the negative selection of thymocytes by apoptosis (30,(43)(44)(45)(46). Interestingly, PGF 2 ␣ was also shown to initiate programmed cell death in the corpus luteum (47,48). The possibility that the induction of nur77 expression by PGF 2 ␣ leads to the initiation of an apoptotic process in the corpus luteum remains to be investigated.
PGF 2 ␣-induced luteolysis is believed to be initiated through ligand receptor activation of the phospholipase C that in turn induces production of inositol 1,4,5-trisphosphate and diacylglycerol, causing an increase in intracellular calcium mobilization and activation of Ca 2ϩ /calmodulin kinase and protein kinase C (for review, see Ref. 49). Interestingly, either activation of PKC by phorbol ester or increase in intracellular levels of calcium with a calcium ionophore leads to the induction of nur77 expression (50,51) suggesting that PGF 2 ␣-induced expression of nur77 in the corpus luteum may result from the activation of either one of these signaling pathways.
Despite the well established role of NUR77 in many processes, mice deficient in NUR77 have no apparent phenotype suggesting that biological alternative pathways complement the functional defects (52). Indeed NURR1 may compensate for the loss of NUR77. NURR1 is also an immediate-early gene whose DNA and ligand-binding domain are similar to NUR77 (53). NUR77 was shown to mediate ACTH stimulation of P450 C21 in adrenal cells (22), yet the expression of this gene is normal in mice deficient in NUR77 (52); however, NURR1 becomes highly expressed after ACTH treatment in the adrenals of those animals and appears to compensate for NUR77.
The development of mice lacking the PGF 2 ␣ receptor revealed that this molecule plays a key role in the induction of parturition and that the lack of delivery in these knockout mice is due to the persistent production of progesterone. Indeed progesterone levels did not decline around day 20, the day of parturition in mice, whereas ovariectomy of PGF 2 ␣ receptordeficient mice led to parturition within 24 h (11, 12). Our results indicate that these high levels of progesterone are due to the lack of luteal 20␣-HSD expression in the PGF 2 ␣ receptor knockout mice. This together with the results showing that administration of PGF 2 ␣ to pregnant rats induces premature expression of 20␣-HSD indicates that PGF 2 ␣ is crucial for the massive increase in the luteal expression of this gene at the end of pregnancy, and that it is the expression of 20␣-HSD that gives the signal for parturition in rodent.
Because PRL and PRL-related hormones silence the expression of 20␣-HSD throughout pregnancy (4, 54 -56) and because a marked decline in the level of PRL receptor takes place at a time when 20␣-HSD levels rise (57), we previously suggested that this reduction in the PRL-R levels renders the corpus luteum less responsive to circulating PRL and PRL-related hormones allowing 20␣-HSD expression before parturition. However, the results presented herein do not support such a conclusion. Indeed, PGF 2 ␣ can induce the expression of 20␣-HSD when administered on day 19 of pregnancy, a time when circulating levels of rat placental lactogen-2 are elevated (58,59) and when luteal PRL receptor is still highly expressed (57).
Although it has been suggested that PGF 2 ␣ can affect the synthesis of progesterone by reducing cholesterol transport to the inner mitochondrial membrane (60,61), it appears clear that at least in rodent, the main effect of PGF 2 ␣ is to enhance the catabolism rather than inhibit the synthesis of progesterone. Results of this investigation indicate that levels of the progesterone metabolites, 20␣-OH-progesterone secreted by the ovaries of rats treated with PGF 2 ␣, exceed the levels of progesterone produced by non-treated rats. In addition to its ability to decrease luteal progesterone production, PGF 2 ␣ was also shown to affect its own expression in ovine large luteal cells (34) and that of its own receptor in rat ovary (32).
Our studies have focused on the role of NUR77 in the control of 20␣-HSD promoter activity. However, the 20␣-HSD promoter contains two AP-1 sites (21), which associate with two subunits encoded by the nuclear oncogenes c-fos and c-jun (62,63). PGF 2 ␣ treatment in vivo increases the expression of C-JUN mRNA in bovine corpus luteum (64) and that of c-JUN and c-FOS in cultured bovine luteal cells (65) and is also shown to increase the level of DAX-I expression (66). However, our FIG. 10. Overexpression of a dominant negative NUR77 protein prevents the increase in 20␣-HSD promoter activity induced by PGF 2 ␣. Luteinized granulosas cells were transfected with 0.5 g/well of the 2.5-kb 20␣-HSD-Luc construct together with two concentrations of the N-terminal truncated NUR77 expression vector (DN-Nur77). The amount of DNA per well was kept constant by addition of an empty plasmid (pSG5). 24 h after transfection the cells were treated with either PGF 2 ␣ (10 -6 M) or vehicle for 12 h. Transient expression of the reporter gene was quantified by a standard luciferase bioluminescence assay and normalized against ␤-galactosidase. Bars represent means Ϯ S.E. of three replicates. Columns with different letters differ significantly; a-c and a-b, p Ͻ 0.01; b-c, p Ͻ 0.05 (ANOVA I).
findings that PGF 2 ␣ stimulation of 20␣-HSD mRNA expression and promoter activity are respectively obliterated by cyclosporin A, which prevents NUR77 DNA binding, and by a dominant negative NUR77, suggesting that NUR77 is the sole transcription factor that mediates PGF 2 ␣ stimulation of 20␣-HSD.