Molecular characterization and role of bovine upstream stimulatory factor 1 and 2 in the regulation of the prostaglandin G/H synthase-2 promoter in granulosa cells.

The transcriptional activation of the prostaglandin G/H synthase-2 (PGHS-2) gene in granulosa cells is required for ovulation. To directly study the ability of upstream stimulatory factor 1 (USF1) and USF2 to trans-activate the bovine PGHS-2 promoter in granulosa cells, USF1 or USF2 expression vectors were cotransfected with the PGHS-2/luciferase (LUC) chimeric construct, -149/-2PGHS-2.LUC. Results revealed that overexpression of USF1 or USF2 caused a marked and significant increase in basal and forskolin-inducible promoter activities (p<0.05), and these effects were dependent on the presence of a consensus E-box cis-element within the promoter fragment. Co-transfections with different N- and C-terminal truncated USF mutants led to significant reductions in promoter activation, as compared with full-length constructs (p<0.05), thus allowing identification of putative bovine USF functional domains. Overexpression of a USF2 truncated mutant lacking the first 220 residues (U2Delta1-220) acted as a dominant negative mutant and blocked endogenous and USF-stimulated PGHS-2 promoter activation. Interestingly, transfections with U2Delta1-220 blocked the forskolin-dependent induction of PGHS-2 mRNA in granulosa cells, whereas transfections with full-length USF2 increased PGHS-2 transcript levels. Immunoblot analyses confirmed overexpression of full-length and truncated USF proteins, and electrophoretic mobility shift assays (EMSAs) and supershift EMSAs established that the observed effects were dependent on specific interactions between USF proteins and the consensus E-box cis-element. Stimulation of cells with forskolin increased, whereas treatment of extracts with phosphatase decreased USF binding activities to the E-box. Thus, this study presents for the first time direct evidence for a role of USF proteins in the regulation of the PGHS-2 promoter in preovulatory granulosa cells.

Prostaglandins are key mediators of inflammation, and the process of ovulation shares numerous signs of an acute inflammatory reaction, including hyperemia, edema, emigration of leukocytes, and induction of proteolytic and collagenolytic activities (1,2). Likewise, there is a marked increase in intrafollicular levels of prostaglandin E 2 (PGE 2 ) 1 and PGF 2␣ just prior to ovulation, and inhibitors of prostaglandin synthesis were shown to block ovulation in several species (3)(4)(5)(6). Prostaglandin G/H synthase (PGHS, also known as COX) is the first rate-limiting enzyme in the biosynthesis of all prostaglandins from arachidonic acid, and two PGHS isoforms derived from distinct genes located on separate chromosomes have been characterized, and are referred to as PGHS-1 and PGHS-2 (7)(8)(9). A third PGHS isoform produced as an alternate splice variant of the PGHS-1 gene was recently identified as PGHS-3 (10). Several studies have established that the increase in follicular prostaglandin synthesis prior to ovulation is caused by a selective, gonadotropin-dependent induction of PGHS-2 (11)(12)(13)(14). Within the preovulatory follicle, the enzyme was shown to be induced exclusively in granulosa cells, and not in theca interna (12,14,15). The obligatory role of PGHS-2 induction during ovulation was underscored in genetic studies in which ovulatory failure proved to be a common phenotype in PGHS-2-deficient female mice (16,17).
The upstream stimulatory factor (USF) proteins belong to a family of transcription factors characterized by a C terminuscontaining basic region (B), helix-loop-helix (HLH), and leucine zipper (LZ) motifs responsible for dimerization and DNA binding activities (18,19). Two distinct USF genes, referred to as USF1 and USF2, have been characterized and shown to encode full-length USF1 and USF2 proteins of 43 and 44 kDa, respectively (18 -22). Alternative USF1 and USF2 splicing variants have also been identified, but their biological significance remains incomplete (22,23). The C terminus B-HLH-LZ domains of USF1 and USF2 are highly conserved, and direct the formation of USF homodimers or heterodimers that interact with DNA with similar specificities (19,20,24). The cognate DNA binding cis-element for USF proteins, the E-box, contains the core nucleotide sequence CANNTG (18,24). Another very well conserved domain between USF1 and USF2 is a small domain termed USF-specific region (USR) that is located immediately upstream of the basic region and involved in nuclear localization and context-dependent transcriptional activation (25). In contrast, the N-terminal regions of USF1 and USF2 are much more divergent, but were shown to contain additional transactivation domains (25)(26)(27). USF proteins are ubiquitous transcription factors that play a key role during embryonic development (28) and that are known to activate or repress the expression of various genes (29 -35).
The induction of PGHS-2 in granulosa cells prior to ovulation was shown to be dependent on transcriptional events (36). Initial studies on the regulation of the rat and bovine PGHS-2 promoter in granulosa cells revealed that the proximal 150 -200 bp immediately upstream of the transcriptional start site were sufficient to confer basal and forskolin/ gonadotropin inducible activities (37,38). Different ciselements were identified in this region, including a C/EBP, an ATF/CRE and an E-box element. However, site-directed mutagenesis studies showed that only the E-box played a central role in the regulation of the PGHS-2 promoter in granulosa cells (38 -40). Endogenous USF proteins were detected in rat and bovine granulosa cells, and potential interactions with the E-box element were suggested from electrophoretic mobility shift assays (EMSAs) (38 -40). However, there has been no direct evidence demonstrating the capacity of USFs to trans-activate the PGHS-2 promoter in granulosa cells in any species. Therefore, the specific objectives of the present study were to clone and characterize the primary structure of bovine USF1 and USF2, and to provide direct evidence that USF proteins regulate PGHS-2 expression in preovulatory granulosa cells.

EXPERIMENTAL PROCEDURES
Materials-The QuikHyb hybridization solution and the ExAssist/ SOLR system were purchased from Stratagene Cloning Systems (La Jolla, CA). The bovine genomic library was obtained from Clontech (Bio/Can Scientific, Ontario, Canada). [␣-32 P]dCTP was purchased from PerkinElmer Canada Inc. (Woodbridge, Ontario, Canada). Lipo-fectAMINE PLUS reagent, TRIzol total RNA isolation reagent, avian myeloblastosis virus-modified reverse transcriptase, 1-kb DNA ladder, synthetic oligonucleotides, pcDNA 3.1(ϩ), 5Ј RACE system version 2.0, and culture media were obtained from Invitrogen. Fetal bovine serum was purchased from HyClone Laboratories (Logan, Utah). Prime-a-Gene labeling system, Dual-Luciferase Reporter Assay, RT-PCR Access kit, and plasmids pGEM-Teasy, pGL3-Basic, and pRL-SV40 were obtained from Promega (Madison, WI). Polyclonal antibodies against USF1 and USF2 were purchased from Santa Cruz Biotechnology Inc. (Santa Cruz, CA). Hybond-P polyvinylidene difluoride (PVDF) membranes, Rainbow molecular weight markers, ECL plus, horseradish peroxidase linked donkey anti-rabbit secondary antibody, and restriction enzymes were obtained from Amersham Biosciences. Expand High Fidelity DNA Polymerase was obtained from Roche Diagnostics (Laval, Québec, Canada). Phosphatase type VII-S was purchased from Sigma.
Isolation of Bovine USF1 and USF2 cDNAs-The cloning and characterization of full-length bovine USF cDNAs required four strategies, including cDNA library screening, 5Ј-rapid amplification of cDNA ends (5Ј-RACE), a three-step nested PCR approach and genomic screening (Fig. 1). A bovine cDNA library prepared from mRNA isolated from a preovulatory follicle (41) was initially screened with human USF1 (18) and mouse USF2 (25) cDNA probes labeled with [␣-32 P]dCTP using the Prime-a-Gene labeling system (Promega). Positive clones were purified, and pBluescript phagemids containing cloned DNA inserts were excised in vivo with the Ex-Assist/SOLR system (Stratagene). DNA sequencing was performed commercially (Université Laval, Québec, Canada), and revealed that the longest USF1 clone was lacking sequences corresponding to the first 31 amino acid residues (93 bp; Fig. 1A, a) of human USF1, whereas the longest USF2 clone was missing the region corresponding to the first 175 amino acid residues (525 bp; Fig. 1B, a) of human USF2.
Attempts to obtain the missing 5Ј-coding region of bovine USF1 by 5Ј-RACE were unsuccessful. The 93-bp region was characterized through partial sequencing of an overlapping genomic fragment iso-lated by genomic library screening with a bovine USF1 cDNA probe, as described (38) (Fig. 1A, b). The full-length USF1 coding region was isolated by RT-PCR using 0.5 g of preovulatory follicle RNA and sense primer 1 and antisense primer 2 (Fig. 1A, c).
To obtain additional sequences at the 5Ј-end of bovine USF2, the 5Ј-RACE system version 2.0 (Invitrogen) was used according to the manufacturer's instructions. Reverse transcription was performed as directed using antisense primer 13 (Fig. 1B, b and C) and 5 g of RNA from a preovulatory follicle. The first 5Ј-RACE/PCR reaction was performed with sense abridged anchor primer 14 (Invitrogen) and antisense primer 15, whereas the second 5Ј-RACE/PCR reaction employed the sense abridged universal amplification primer 16 (Invitrogen) and antisense primer 17 (Fig. 1B, b and C). PCR reactions consisted of 35 cycles of 94°C for 30 s, 56°C for 60 s, and 72°C for 1 min, and products were subcloned into pGEM-Teasy. This approach provided only 157 bp of new 5Ј-USF2 sequences (Fig. 1B, b). To provide more upstream USF2 cDNA sequences, a three-step nested PCR approach was performed under conditions previously described (42), using the bovine follicular cDNA library (41), two sense primers specific for sequences located near the multiple cloning site of the pBluescript vector (primers 18 and 19) and three antisense primers derived from bovine USF2 (primers 13, 20 and 21; Fig. 1B, c and C). The first PCR amplification employed sense primer 18 and antisense primer 13, whereas the second and third PCR used primer pairs 19 and 20, and 19 and 21, respectively. The product of the third PCR was subcloned into pGEM-Teasy, sequenced and shown to miss only the first 18 bp of the USF2 open reading frame (ORF) (Fig. 1B, c). The complete USF2 coding region was isolated by RT-PCR using 0.5 g of preovulatory follicle RNA and sense primer 7 (designed from the highly conserved 5Ј-end of human and mouse USF2 ORF) and antisense primer 8 (Fig. 1B, d).
Cell Culture, Transient Transfections, and Reporter Activity Assays-Primary cultures of bovine granulosa cells, transient transfections, and reporter activity assays were performed as previously described (38). Briefly, cells were seeded in 24-well plates at a density of 1-2 ϫ 10 5 cells/0.5 ml of minimal essential medium (MEM) supplemented with L-glutamine, non-essential amino acids, 2% fetal bovine serum, insulin (1 g/ml), transferrin (5 g/ml), and penicillin (100 units/ml)/streptomycin (100 g/ml), and incubated at 37°C in a humidified atmosphere of 5% CO 2 . Cultures were transiently transfected with 90 fmol/well of Ϫ149/Ϫ2PGHS.LUC or Ϫ149/Ϫ2PGHS-2⌬E-box.LUC in the absence or presence of various USF expression constructs (1-50 fmol/well) using 2 g of LipofectAMINE PLUS (Invitrogen) in 0.3 ml of MEM, in accordance with the manufacturer's protocol. Co-transfection with the SV40 Renilla luciferase control vector (pRL.SV40; Promega) was performed to normalize results. Three hours after transfection, cells were incubated in fresh culture media for 36 h in the absence or presence of forskolin (10 M). At the end of the culture period, cell lysates were prepared, and firefly and Renilla luciferase activities were determined using Promega's Dual Luciferase Assay System and a Lumat LB 9507 luminometer (Berthold Technologies).
Granulosa Cell Nuclear Extracts and Electrophoretic Mobility Shift Assays (EMSAs)-Bovine granulosa cells were cultured in the absence or presence of forskolin (10 M) for 36 h, and nuclear extracts were prepared as described (37,38). Protein concentration in each extract was determined by the method of Bradford (43). EMSAs were performed as described (37,38), with minor modifications. Briefly, extracts of nuclear proteins (0.5 g/reaction) were incubated with 40,000 cpm of end-labeled Ϫ149/Ϫ2 PGHS-2 promoter fragment and 1 g of poly(dI/ dC) (Amersham Biosciences) in a final volume of 20 l of binding buffer containing 15 mM Tris-HCl (pH 7.5), 1 mM EDTA, 100 mM KCl, 5 mM MgCl 2 , 5 mM dithiothreitol, and 12% (v/v) glycerol. In some experiments aimed at studying the effect of protein dephosphorylation on binding activity, extracts were preincubated for 10 min with phosphatase type VII-S (0.5-1 IU; Sigma) prior to the addition of other reagents. When antibodies were used in supershift EMSAs, the nuclear extract was first incubated for 20 min with the antiserum prior to the addition of other reagents. Binding complexes were resolved by 5% acrylamide, 0.5ϫ TBE gel electrophoresis.
RNA Extraction and RT-PCR Analysis-Total RNA was extracted from granulosa cells cultured under various experimental conditions using TRIzol (Invitrogen). The Access RT-PCR System (Promega) was used to characterize expression of PGHS-2 and GAPDH transcripts. Reactions were performed with sense and antisense primers specific for PGHS-2 (5Ј-CACAGTGCACTACATACTTACCC-3Ј and 5Ј-GTCTGGAA-CAACTGCTCATC GC-3Ј; generates a 735-bp fragment) and GAPDH (5Ј-TGTTCCAGTATGATCCACCC-3Ј and 5Ј-TCCACCACCCTGTTGC-TGTA-3Ј; 850-bp fragment). Each reaction was performed using 100 ng of total RNA, and cycling conditions were one cycle of 48°C for 45 min and 94°C for 2 min, followed by 35 cycles of 94°C for 30 s, 59°C for 1 min and 68°C for 2 min. Following PCR amplification, samples were electrophoresed on 1% Tris acetate/EDTA-agarose gels.
Statistical Analyses-One-way ANOVA was used to test the effect of different treatments and constructs on reporter gene activity. When , and nested PCR (c), as described under "Experimental Procedures," arrows and numbers indicate the orientation, relative position, and identity of oligonucleotides used in each cloning procedures. The complete coding region of bovine USF2, as well as a series of truncated constructs were produced by PCR (c) using specific primers identified by numbered arrows. C, list of oligonucleotides used for USF1 and USF2 cloning. The abridged anchor primer (14) and abridged universal amplification primer (16) are components of the 5Ј-RACE system.
ANOVAs indicated significant differences (p Ͻ 0.05), Dunnett's test was used for multiple comparisons of means. Statistical analyses were performed using JMP software (SAS Institute, Inc., Carry, NC).

RESULTS
Characterization of Bovine USF1 and USF2-The isolated bovine USF1 cDNA was composed of an ORF of 933 bp (including the stop codon) and a 3Ј-untranslated region of 301 bp ( Fig.  2A). Its coding region encodes a 310-amino acid protein, which is identical in length to other known mammalian USF1 proteins (18,(45)(46)(47). Overall comparisons between bovine USF1 and the human (18), mouse (45), rat (46), and rabbit (47) homolog revealed a 97-99% identity at the amino acid level. The bovine USF2 cDNA was composed of an ORF of 1041 bp encoding a 346-amino acid protein, and a 3Ј-untranslated region of 268 bp (Fig. 2B). Comparisons between bovine USF2 and the human (19), mouse (20), and rat (48) homolog revealed a 97% identity at the amino acid level.
USFs-and E-box-dependent Transactivation of the Bovine PGHS-2 Promoter in Granulosa Cells-To provide direct evidence that USF proteins transactivate the bovine PGHS-2 promoter in granulosa cells, USF1 or USF2 expression vectors were co-transfected with the Ϫ149/Ϫ2PGHS-2.LUC chimeric construct into primary cultures of granulosa cells. The Ϫ149/ Ϫ2PGHS-2.LUC construct, which contains the proximal PGHS-2 promoter region spanning from Ϫ149 to Ϫ2 (ϩ1 ϭ transcription start site), was previously shown to contain a consensus E-box cis-element (5Ј-CACGTG-3Ј, position Ϫ50 to Ϫ45) essential for promoter activity (38). Results from the present study revealed that transfections of the PGHS-2 promoter construct generated basal activity that was significantly increased by forskolin (3.1 Ϯ 0.4 versus 6.1 Ϯ 0.5, p Ͻ 0.05; Fig.   3A). Cotransfections with USF1 had a marked effect on promoter activities, with a 3.3-and 6.6-fold increase in basal and forskolin-inducible activities, respectively (Fig. 3A). Likewise, a 5.2-and 8.6-fold rise in basal and inducible activities, respectively, were observed after cotransfections with USF2 ( Fig. 3A), clearly establishing the capacity of USF proteins to activate the PGHS-2 promoter in granulosa cells. The stimulatory effect of each USF construct was dose-dependent, with the highest amount (50 fmol) of USF1 or USF2 causing a 6.7-and 14-fold increase in basal activity, and 21.9-and 38.4-fold increase in forskolin-inducible activity, respectively (Fig. 3, B and C). Cotransfections with USF1 and USF2 produced levels of luciferase activities similar to those observed with USF2 alone (data not shown).
To investigate the role of the E-box element in PGHS-2 promoter activities, transfection experiments were performed with the Ϫ149/Ϫ2PGHS-2⌬E-box.LUC construct, which contains a mutated E-box element in the context of the Ϫ149/Ϫ2 PGHS-2 promoter fragment. Results showed that transfections with Ϫ149/Ϫ2PGHS-2⌬E-box. LUC alone abolished or severely reduced basal and forskolin-inducible activities (Fig. 3A), as previously reported (38). Basal and forskolin-dependent activities were also ablated when granulosa cells were cotransfected with the mutated promoter construct and with USF1 or USF2 expression vectors (Fig. 3A), suggesting a role for the E-box in both normal and USF overexpressing cells.
To determine whether the inhibitory effect of U2⌬1-220 on PGHS-2 promoter activation could translate into the regulation of PGHS-2 transcripts in granulosa cells, cells were transfected with U2⌬1-220 24 h prior to forskolin stimulation, and RT-PCR was performed to study changes in mRNA levels. Results showed that U2⌬1-220 blocked the forskolin-dependent induction of PGHS-2 mRNA in granulosa cells (Fig. 5C), whereas transfections with the empty expression vector had no effect (data not shown). In contrast, cells transfected with fulllength U2wt led to higher basal and forskolin-dependent induction of PGHS-2 transcripts (Fig. 5C), in keeping with effects observed on promoter activities (Fig. 5A).
Binding Activity of USF Proteins to the E-box cis-Element within the PGHS-2 Promoter-To relate the ability of USF proteins to trans-activate the PGHS-2 promoter to specific protein-DNA interactions, EMSAs were performed with nuclear extracts prepared from granulosa cells transfected with USF constructs and the Ϫ149/Ϫ2 PGHS-2 promoter fragment. Results showed that a major protein/DNA complex was formed with cells transfected with U1wt or U2wt (Fig. 6A, lanes 3 and  4, respectively), with the U1wt band migrating slightly faster than U2wt in keeping with minor differences in their molecular weight. A larger complex encompassing both USF bands was observed when cells were co-transfected with U1wt and U2wt (Fig. 6A, lane 5). Stimulation with forskolin did not change the nature of the bands, but an apparent increase in the intensity of the complexes was observed (compare lanes 7, 8, and 9 with  lanes 3, 4, and 5, respectively; Fig. 6A).
To characterize the specificity of protein/DNA interactions, and to establish the role of the E-box element present in the Ϫ149/Ϫ2 PGHS-2 promoter fragment, EMSAs were performed using molar excess of various competitor DNA. Results showed that U1wt and U2wt complexes were competed with excess  Fig. 6B). In a complementary experiment, no protein/DNA complex was observed when the same extracts were incubated with labeled Ϫ149/Ϫ2PGHS-2 DNA fragment containing a mutated E-box element (lanes 2 and 3, Fig. 6C). Collectively, these results illustrate the central role of the E-box in producing USFs/DNA binding activities.
To confirm the presence of USF proteins in binding complexes, supershift EMSAs were performed using antibodies against USF1 and USF2. The major band produced with extracts prepared from cells transfected with U1wt was markedly shifted with the anti-USF1 antibody but was not affected by the anti-USF2 antibody (lanes 1, 2, and 3, Fig. 7). Conversely, the band produced with the U2wt extract was shifted with anti-USF2, but not with anti-USF1 antibody (lanes 7, 8, and 9;  Fig. 7). Interestingly, extracts prepared from cells transfected with N-terminal truncated U1⌬1-129 and U2⌬1-220 generated bands with higher mobility that were also shifted by their respective antibody (lanes 4 and 5 and lanes 10 and 11, Fig. 7). In contrast, no complex was formed with extracts from cells expressing C-terminal truncated U1⌬249 -310 and U2⌬299 -346, which lack portion of the HLH and all the LZ domain ( lanes 6 and 12, Fig. 7).
Phosphatase Treatment Decreases USFs/DNA Binding Activities-To study the effect of protein dephosphorylation on USF1 and USF2/DNA binding activity, nuclear extracts from cells transfected with U1wt or U2wt cultured in the absence or presence of forskolin were treated with phosphatase prior to EMSAs. Results showed that phosphatase treatment decreased the intensity of U1wt and U2wt bands (Fig. 8). However, binding complexes produced with extracts from forskolin-treated cells appeared more resistant to phosphatase treatment than extracts from untreated cells, especially with the U2wt cell extract (lanes 4 -6 versus lanes 1-3, Fig. 8B).

DISCUSSION
The process of follicular rupture is essential to all mammalian reproductive cycles, and numerous investigations have established that the induction of prostaglandin synthesis is required for ovulation (1-4, 16, 17). A number of studies revealed that the obligatory preovulatory rise in follicular prostaglandin synthesis results from a selective, gonadotropin-dependent induction of PGHS-2 transcript and protein in granulosa cells (11)(12)(13)(14). The previous characterization of the rat and bovine PGHS-2 promoter showed that a conserved E-box element located in the proximal promoter region plays a central role in basal and agonist inducible activities in granulosa cells (37,38). The present study provides for the first time direct evidence that USF proteins trans-activate the PGHS-2 promoter in granulosa cells. Using the bovine model, our results demonstrate that overexpression of full-length USF1 and USF2 proteins resulted in a marked increase in PGHS-2 promoter activities. In contrast, overexpression of a USF2 dominant negative mutant repressed promoter activation, and ultimately blocked forskolin-dependent induction of PGHS-2 mRNA in granulosa cells. These effects involved specific interactions between USF proteins and the consensus E-box element located between Ϫ45 and Ϫ50 bp upstream the bovine PGHS-2 transcription start site.
The present study also establishes some of the functional domains of bovine USF proteins. Deletion of USF1 and USF2 regions upstream of the USR-B-HLH-LZ domain resulted in a marked decrease in PGHS-2 promoter activation but had no effect on DNA binding activity, in keeping with the presence of FIG. 5. Dominant negative effect of U2⌬1-220 on PGHS-2 promoter activity and mRNA expression. A, cells were co-transfected simultaneously with Ϫ149/Ϫ2PGHS-2.LUC (Ϫ149/Ϫ2PGHS) in the absence or presence of constructs expressing U1wt, U2wt, U1⌬1-129, and U2⌬1-220, as described under "Experimental Procedures." All cultures were also co-transfected with the SV40 Renilla luciferase vector (pRL.SV40) as an internal control to normalize experimental reporter activity. After transfection, cells were cultured for 36 h in the absence (control) and presence of forskolin (10 M; FSK). Results are presented as relative luciferase activity (firefly/Renilla; mean Ϯ S.E. of triplicate cultures from three experiments). B, cells were first co-transfected with U2⌬1-220 in the absence or presence of U2wt, and cultured for 24 h prior to a second co-transfection with Ϫ149/Ϫ2 PGHS and pRL.SV40 (internal control). Cells were then cultured for 24 h in the absence (control) and presence of forskolin (10 M; FSK). Results are presented as relative luciferase activity (firefly/Renilla; mean Ϯ S.E. of triplicate cultures from three experiments). C, cells were transfected with U2⌬1-220 or U2wt and cultured for 24 h prior to a second period of culture of 24 h in the absence (ϪFSK) or presence of forskolin (ϩFSK). RNA samples were prepared and changes in PGHS-2 and GAPDH (internal control) transcripts were analyzed by RT-PCR, as described under "Experimental Procedures." FIG. 6. USFs-and E-box-dependent protein/DNA binding activities in granulosa cell nuclear extracts and the bovine PGHS-2 promoter. A, nuclear protein extracts were prepared from mock-transfected granulosa cells (control) and from cells transfected with fulllength wild type USF1 (U1wt) and USF2 (U2wt) and cultured in the absence or presence of forskolin (FSK) for 36 h, as described under "Experimental Procedures." Extracts were incubated with 32 P-labeled PGHS-2 promoter fragment Ϫ149/Ϫ2, and protein/DNA interactions were studied by EMSAs. Binding reactions were resolved by 5% polyacrylamide, 0.5ϫ TBE gel electrophoresis. For reference purposes, the major USF1 and USF2 protein/DNA complexes are designated as bands a and b, respectively. B, competitive EMSAs were performed in the presence of nuclear extracts prepared from granulosa cells cultured in the absence of forskolin and transfected with U1wt and U2wt, 32 Plabeled DNA fragment Ϫ149/Ϫ2 and various unlabeled competitor DNA. They included 75-molar excess of wild type Ϫ149/Ϫ2, as well as of an E-box site-directed Ϫ149/Ϫ2 mutant (⌬E-box) and a set of complementary oligonucleotides containing a consensus E-box element (E-box) (38). Identical results were obtained with competitive EMSAs using nuclear extracts prepared from cells treated with forskolin (data not shown). C, EMSAs were performed with nuclear extracts prepared from cells transfected with U1wt and U2wt, and a 32 P-labeled E-box sitedirected Ϫ149/Ϫ2 DNA as probe (Ϫ149/Ϫ2⌬Ebox).
N-terminal trans-activation domains identified in human and mouse USFs (25)(26)(27). The bovine USF2 region spanning from Met 1 to Pro 149 proved to contain a key element(s) involved in regulating basal PGHS-2 promoter activity, whereas bovine USF1 region from Met 1 to Gly 129 and USF2 region from Met 1 to Pro 220 appeared central for forskolin-inducible promoter activation. The other extremity of the bovine USF proteins was equally important for their overall activity, as removal of the last 62 and 48 amino acid residues of USF1 (U1⌬249 -310) and USF2 (U2⌬299 -346), respectively, resulted in a loss of their capacity to stimulate transcription. However, in these latter cases, the inability of C-terminal-truncated USF mutants to bind to the E-box element present in the proximal PGHS-2 promoter fragment appeared as the underlying mechanism responsible for the loss of activity, in keeping with the dimerization/DNA binding functions of the HLH-LZ domains identified in this region (18 -20, 24).
The production of a bovine USF dominant negative mutant was sought as an important tool to demonstrate the functional role of the transcription factor in the regulation of the PGHS-2 promoter in granulosa cells. This approach has been used previously to establish the function of USF proteins in the regulation of other promoters (49 -53). However, although N-terminal and C-terminal truncations of USF1 and USF2 reduced or blocked their trans-activating capacity, these mutants were unable to inhibit endogenous PGHS-2 promoter activities when transfected simultaneously with the promoter construct. As a potential explanation for this outcome, it was hypothesized that a rapid stimulation of basal and forskolin-inducible PGHS-2 promoter activities by endogenous USF proteins preceded expression of sufficient USF mutants when both constructs were co-transfected contemporaneously. The hypothesis appeared supported by results from experiments in which endogenous activities were abrogated when cells were allowed to accumulate a USF mutant (U2⌬1-220) for 24 h prior to transfection with the PGHS-2 promoter construct. EMSAs and supershift EMSAs confirmed the ability of overexpressed U2⌬1-220 to bind the E-box element present in the bovine PGHS-2 promoter, thus establishing the saturation of a critical cis-element by non-functional truncated transcription factors as the putative mechanism of action for this dominant negative mutant.
While this study establishes that binding of USF proteins to the E-box is a prerequisite for the regulation of the PGHS-2 promoter in granulosa cells, the complete molecular control of PGHS-2 induction during ovulation remains to be characterized. Importantly, previous studies have shown that USF proteins are expressed at constant levels in rat and bovine granulosa cells during the ovulatory process, indicating that PGHS-2 induction is not caused by changes in levels of these transcription factors (38,39). Alternatively, the involvement of coactivator(s) regulated by the LH preovulatory surge may control the transcriptional activity of USFs in granulosa cells. The presence of a specialized unidentified coactivator has been implicated in the transcriptional activity of USF proteins in HeLa cells (54). The coactivator p300, which has intrinsic acetyltransferase activity, was shown to interact with USF2 to potentiate transactivation of the mouse cytochrome c oxidase subunit Vb promoter (51) and the human telomerase reverse transcriptase promoter (55), and was thus considered as a potential candidate for the PGHS-2 promoter in granulosa cells. However, co-transfections with a p300-expressing construct had no effect on USF1 or USF2 regulated PGHS-2 promoter activity in granulosa cells (data not shown). This result is in keeping with another report that revealed that p300 was unable to affect USF1-dependent transactivation of the TGF-␤2 promoter in F9 differentiated embryonal carcinoma cells (56), suggesting that the involvement of p300 in USF action may be promoter-and cell-specific.
FIG. 8. Effect of phosphatase treatment on USFs/DNA binding activities. Nuclear extracts were prepared from granulosa cells transfected with U1wt (panel A) or U2wt (panel B), and cultured in the absence or presence of forskolin (FSK) for 36 h. Extracts were preincubated with graded doses (0, 0.5, and 1 unit) of phosphatase for 10 min prior to addition of 32 P-labeled PGHS-2 promoter fragment Ϫ149/Ϫ2 and other reagents, and EMSAs were performed as described under "Experimental Procedures." Binding reactions were resolved by 5% acrylamide, 0.5ϫ TBE gel electrophoresis. Major USF1 and USF2 protein/DNA complexes are designated as bands a and b, respectively. and ovulation (1,2). The LH receptor (LHR) is a member of the rhodopsin/␤2-adrenergic receptor subfamily of G protein-coupled receptors, and LHR-mediated effects in preovulatory granulosa cells are directed mostly, but not exclusively, through the activation of the G protein/adenylyl cyclase/cAMP/cAMP-dependent protein kinase (PKA) pathway (57). The phosphorylation of transcription factors by activated PKA represents an important mechanism by which gene expression is controlled (58). The stimulatory effect of forskolin, a potent activator of adenylyl cyclase, on endogenous and USF-driven PGHS-2 promoter activities in the present study raises the possibility that USF phosphorylation may be involved in the regulation of the promoter in granulosa cells. A number of reports have documented that USF1 can be phosphorylated by various kinases, including cyclin A2-and B1-p34 cdc2 , stress-responsive p38 kinase, PKA and proteine kinase C, and that USF1 phosphorylation led to increases in DNA binding and transactivation of different promoters (59 -61). Interestingly, results from the present study showing apparent forskolin-dependent increase, and phosphatase-dependent decrease, in USFs binding activities to the E-box also argue for a potential role of USF phosphorylation in PGHS-2 transcriptional regulation.
Lastly, this study reports for the first time the cloning and characterization of bovine USF1 and USF2, with results underscoring the remarkable similarities among the primary structure of USFs across species. The length of bovine USF1, 310 amino acids, is identical to that of the four other mammalian orthologs characterized thus far (18,(45)(46)(47). More importantly, comparative analyses revealed that the amino acid sequence of bovine USF1 is almost identical to that of other species, being 99% identical to human (18) and 97% identical to mouse (45), rat (46), and rabbit USF1 (47). Comparisons at the nucleic acid level also revealed that the corresponding coding region of the bovine USF1 cDNA was highly conserved, being 89 -94% identical to these mammalian orthologs (18,(45)(46)(47). As for bovine USF2, the size of the full-length protein (346 amino acids) is identical to that of the three other characterized mammalian counterparts (19,20,48), and levels of similarities among amino acid sequences (97% identity) and corresponding nucleic acid sequences (90 -92% identity) were very high when the bovine factor was compared with human (19), mouse (20), and rat USF2 (48). Consequently, all putative structural and functional domains are thought to be conserved in bovine USF proteins, including the characteristic C terminus B-HLH-LZ domains responsible for dimerization and DNA binding activities (18,19,24). It should also be noted that, as for USF2 of other species, the bovine USF2 N terminus was particularly rich in guanine and cytosine residues. The high GC content of the first 150 nucleotides (72% GC) and 400 nucleotides (69% GC) located immediately after the translation initiation codon were likely responsible for difficulties encountered in cloning the 5Ј-end of bovine USF2.
In summary, this is the first study to characterize the primary structure of bovine USF1 and USF2, and to provide direct evidence that USF proteins are involved in the trans-activation of the PGHS-2 promoter in granulosa cells. Thus, considering the obligatory role of the gonadotropin-dependent induction of PGHS-2 in ovarian follicles during the ovulatory process, this study clearly contributes to our understanding of the molecular mechanisms responsible for this key physiological process. However, more efforts will be needed to unravel the precise molecular control of USF action on PGHS-2 induction in granulosa cells, with future studies focusing on the potential role of USF phosphorylation and involvement of transcriptional coactivators.