/04/$15.00/0 Molecular Endocrinology 18(3): Copyright 2004 by The Endocrine Society doi: /me

Transcription

1 /04/$15.00/0 Molecular Endocrinology 18(3): Printed in U.S.A. Copyright 2004 by The Endocrine Society doi: /me Increased Cytochrome P Hydroxylase Promoter Function in Theca Cells Isolated from Patients with Polycystic Ovary Syndrome Involves Nuclear Factor-1 JESSICA K. WICKENHEISSER, VELEN L. NELSON-DEGRAVE, PATRICK G. QUINN, AND JAN M. MCALLISTER Department of Cellular and Molecular Physiology, Pennsylvania State University College of Medicine, Hershey, Pennsylvania Cytochrome P hydroxylase (CYP17) gene expression and androgen biosynthesis are persistently elevated in theca cells isolated from ovaries of women with polycystic ovary syndrome (PCOS). We previously reported that 235 to 109 bp of the CYP17 promoter confers increased CYP17 promoter function in PCOS theca cells. In this report, additional deletion and mutational analyses of the CYP17 promoter were performed to identify the sequences that contribute to increased CYP17 promoter function in PCOS theca cells. Results of these analyses established that augmented promoter function in PCOS theca cells results from preferentially increased basal regulation conferred by sequences between 188 and 147 bp of the CYP17 promoter. Scanning mutant analysis demonstrated that mutations within a 16-bp sequence, spanning 174 to 158 bp of the promoter, ablated increased basal CYP17 promoter function in PCOS theca cells. EMSA analysis demonstrated that the NF-1 family member, NF-1C, bound this sequence. Cotransfection of several NF-1C isoforms expressed in normal and PCOS cells repressed CYP17 promoter function. NF-1C protein and DNA binding were reduced in PCOS theca cell nuclear extracts, as compared with normal. Another NF-1C site between 102 and 90 bp of the promoter was also identified. However, mutation of this site had no effect on differential promoter function in PCOS theca cells. These studies demonstrate that 1) augmented CYP17 promoter function in PCOS theca cells results from increased basal regulation, and 2) diminished NF-1C-dependent repression may be one mechanism underlying increased basal CYP17 promoter activity and altered gene expression in PCOS theca cells. (Molecular Endocrinology 18: , 2004) POLYCYSTIC OVARY SYNDROME (PCOS) is a reproductive endocrine disorder characterized by hyperandrogenism, chronic anovulation, and polycystic ovaries (1). PCOS affects approximately 5 10% of premenopausal women (2 4) and is associated with elevated levels of circulating free testosterone, which is produced primarily in the ovary (5). Other clinical features of PCOS patients may include obesity, hirsutism, hyperinsulinism, and a predisposition for noninsulin-dependent diabetes mellitus. Atherosclerosis, hypertension, dyslipidemia, coronary artery disease, and endometrial carcinoma have also been reported Abbreviations: CYP17, Cytochrome P hydroxylase; CYP11A, cytochrome P450 cholesterol side chain cleavage; DBD, DNA-binding domain; FBS, fetal bovine serum; 3 -HSDII, 3 -hydroxysteroid dehydrogenase type II; hsf-1, human SF-1; NE, nuclear extract; NF-1, nuclear factor-1; PCOS, polycystic ovary syndrome; PRD, proline-rich domain; SF-1, steroidogenic factor-1; StAR, steroidogenic acute regulatory protein; TAD, transactivation domain; WT, wild-type. Molecular Endocrinology is published monthly by The Endocrine Society ( the foremost professional society serving the endocrine community. to be associated with PCOS. Familial clustering of PCOS and PCOS-associated phenotypes suggests that genetic factors are involved in the etiology of the disorder (1, 6). Unlike the ovary of normal cycling women, the PCOS ovary is characterized by multiple small follicles 4 7 mm in diameter, with a theca cell compartment that is often hypertrophied. Ovarian theca cells are recognized as one of the primary sources of excess androgen biosynthesis in women with PCOS (7 10). In response to LH-dependent increases in adenylate cyclase, theca cells express a variety of genes encoding components of the steroidogenic pathway that are necessary for androgen and progestin biosynthesis (11). The synthesis of thecal androgens is contingent on the expression of the cytochrome P hydroxylase (CYP17) gene, which encodes a single cytochrome P450 (P450c17) with both 17 -hydroxylase and C17, 20 lyase activities. We previously reported that androgen production is persistently elevated in theca cells isolated from the ovaries of women with PCOS and propagated for successive population doublings in vitro (12, 13). Augmented expression of several steroidogenic enzymes 588

2 Wickenheisser et al. Regulation of CYP17 Transcription in Human Theca Mol Endocrinol, March 2004, 18(3): is associated with increased androgen biosynthesis in PCOS theca cells. In addition to CYP17, these include cytochrome P450 cholesterol side chain cleavage (P450scc), encoded by the CYP11A gene, as well as 3 -hydroxysteroid dehydrogenase type II (3 -HSDII), encoded by the HSD3B2 gene (12, 14). A comparison of CYP17 promoter function in normal and PCOS theca cells has demonstrated that increased CYP17 expression in PCOS results, in part, from increased transactivation of the CYP17 gene (15). In contrast, the abundance of the steroidogenic acute regulatory protein (StAR) and 17 -hydroxysteroid dehydrogenase type V mrnas and the transcriptional activity of the StAR promoter are not different between normal and PCOS theca cells (13, 15). Therefore, the expression and activity of only a subset of proteins that are important for androgen biosynthesis are affected in PCOS theca cells. These observations are in agreement with recent microarray analysis of differential gene expression in normal and PCOS theca cells and demonstrate that dysregulation of androgen biosynthesis involves selective differences in networks of genes involved in steroid hormone biosynthesis as well as insulin and glucose homeostasis (1, 6, 16). In this report, we have focused and extended our studies on the regulation of CYP17 promoter function in normal and PCOS theca cells to examine the molecular mechanisms underlying dysregulated gene expression in the PCOS ovary. We previously reported that increased CYP17 gene expression in PCOS theca cells results from increased promoter function, within 235 bp to 109 bp of the start site of transcription of the CYP17 gene (15). Increased CYP17 gene expression in PCOS theca cells could result from alterations in basal promoter function, and/or changes in campdependent regulation. Here, further deletion and scanning mutational analysis of the human CYP17 promoter were performed to identify the elements involved in augmented CYP17 promoter function in PCOS theca cells. In this process, the DNA regulatory element(s) that are required for basal and campdependent CYP17 promoter regulation in normal and PCOS theca cells were also elucidated. The data presented in this report suggest that nuclear factor-1 (NF-1) is one factor involved in increased CYP17 gene expression in the PCOS ovary. The NF-1 family of transcription factors has been shown to regulate transcriptional initiation of a variety of genes by specifically binding to the bipartite recognition sequence, (C/T)TGGC(N) 6 CC(N) 3 (17, 18). The NF-1 family of transcription factors is comprised of four genes, NF-1A, B, C, and X, which are approximately 90% identical (19 23) and contain highly conserved amino terminal regions responsible for dimerization and DNA binding (24). NF-1C, also referred to as CAAT-box transcription factor (19), was the first NF-1 family member identified and has been reported to transactivate and repress the transcription of a wide variety of genes expressed in developmental and tissue-specific patterns (25 28). With respect to the CYP17 promoter, recent data in adrenal H295 cells have shown that NF-1C binds the CYP17 promoter (29); however, NF- 1C-dependent regulation of the CYP17 gene has not been characterized. This NF-1C binding site overlaps with putative binding sites for the orphan nuclear hormone receptor steroidogenic factor-1 (SF-1), which has also been reported to activate sequences within 235 bp of the CYP17 promoter in both ovarian and adrenal cell types (15, 30, 31). Examination of the cellular processes involved in the dysregulation of CYP17 gene expression in PCOS theca cells is crucial for understanding the underlying mechanisms involved in excessive ovarian androgen production associated with PCOS. In this report, we show that decreased expression of NF-1C contributes to differential regulation of CYP17 promoter function in normal and PCOS theca cells. These studies provide evidence to suggest that decreased NF-1C-dependent repression may play a role in increased CYP17 gene transcription in PCOS theca cells. RESULTS Deletion Analysis of the CYP17 Promoter To examine the regions within 235 bp of the CYP17 promoter involved in augmented CYP17 promoter regulation in PCOS theca cells, a series of promoter constructs containing successive deletions of the 5 flanking sequence of the human CYP17 gene were generated. Luciferase constructs containing 235, 188, 172, 147, 109, or 61 to 44 bp of the CYP17 promoter were transiently transfected into fourth passage theca cells isolated from normal cycling women and women with PCOS. To examine the regions of the CYP17 promoter involved in basal, as well as camp-dependent regulation, the cells were cultured in the absence (basal) or presence of 20 M forskolin for 72 h. The comparison of CYP17 promoter function in normal and PCOS demonstrated that both basal ( a P 0.01) and forskolin-stimulated ( b P 0.01) 235, 188, and 172 CYP17 promoter activities were increased in PCOS theca cells, as compared with normal theca cells (Fig. 1A). Deletion of sequences 5 of 147 bp of the CYP17 promoter resulted in a complete ablation of increased CYP17 promoter function in PCOS theca cells, suggesting that sequences within the general boundaries of 172 to 147 bp of the CYP17 promoter are involved in augmented promoter regulation in PCOS theca cells. Further trimming of promoter sequences to 109 bp did not affect CYP17 promoter function in normal or PCOS theca cells, as compared with the activity of 147 bp of the promoter (Fig. 1B). These data suggest that augmented promoter function in PCOS theca cells primarily results from preferentially increased regulation conferred by sequences between 172 and 147 bp of the CYP17 promoter.

3 590 Mol Endocrinol, March 2004, 18(3): Wickenheisser et al. Regulation of CYP17 Transcription in Human Theca Table 1. Fold Induction of CYP17 Promoter Function by Forskolin in Normal and PCOS Theca Cells Construct Normal PCOS a b a b a b a b a b / a b ( 188/ 140)/ a b The data represent the mean fold induction observed following transfection of five independent normal or PCOS theca cell cultures. Significant fold induction by forskolin (P 0.05), are indicated for normal ( a ) and PCOS ( b ) theca cells. Fig. 1. Deletion Analysis of the CYP17 Promoter in Normal and PCOS Theca Cells Fourth passage theca cells, isolated from either normal or PCOS patients, were transiently transfected with pgl3 luciferase constructs containing 235, 188, 172, 147, 109, or 61 to 44 bp of the CYP17 promoter (20 g/dish) and an expression vector encoding hsf-1 (0.1 g/dish). All constructs contain the endogenous TATA box and transcriptional start site. Following transfection, the cells were cultured in the presence or absence of forskolin (20 M) for 72 h. Data are presented as relative luciferase (LUC) activity that has been normalized with -galactosidase. Data represent the mean SEM of independent experiments performed with theca cells isolated from five independent normal and five independent PCOS patients. CYP17 promoter activity was increased in PCOS theca cells as compared with normal theca cells, under basal ( a P 0.01) and forskolin-stimulated ( b P 0.01) conditions. B, Expanded scale of the relative LUC activities of the 172, 147, 109, or 61 CYP17 promoter constructs. Successive 5 deletions of the CYP17 promoter similarly affected forskolin-dependent regulation in both normal and PCOS theca cells (Table 1 and Fig. 1). Significant forskolin-stimulated promoter function was observed with all of the CYP17 promoter constructs (i.e. 235, 188, 172, 147, and 109), with the exception of the 61 CYP17 promoter construct, in both normal and PCOS theca cells. The activity of the 61 CYP17 construct was similar to that observed with the promoterless pgl3-basic construct (data not shown). The observed reduction in forskolin-stimulated CYP17 promoter activity after removal of sequences upstream of 147 bp, as well as 61 bp, suggests the presence of two putative camp-regulatory elements within the general boundaries of 172 to 147 bp, and 109 to 61 bp, of the CYP17 promoter. A comparison of the fold-induction with forskolin in normal vs. PCOS theca cells (Table 1) indicates that forskolin-dependent induction is less robust in PCOS theca cells. However, the forskolin induced promoter activity, like basal activity, is significantly higher is PCOS cells. These data are consistent with quantitative real-time PCR data that also demonstrate that the fold increase in forskolin-stimulated CYP17 mrna accumulation is blunted in PCOS theca cells, although the absolute amount is higher than in normal theca cells (data not shown). Examination of the 188- to 140-bp Region of the CYP17 Promoter To determine whether the region between 188 and 140 bp of the CYP17 promoter is necessary and sufficient for increased CYP17 promoter activity in PCOS theca cells, this sequence was: 1) deleted from the 235 CYP17 construct, generating 235( -188/ 140) CYP17; or 2) ligated to the minimal 61 CYP17 promoter, generating ( 188/ 140)/ 61 CYP17. Basal and forskolin-stimulated promoter function was compared with the activity of the wild-type (WT) 235 CYP17 or minimal 61 CYP17 constructs after transient transfection of normal and PCOS theca cells. As shown in Fig. 2B, both basal ( a P 0.01) and forskolinstimulated ( b P 0.01) 235 CYP17 (WT) promoter activities were increased more than 2-fold in PCOS theca cells compared with normal theca cells. In contrast, 235 ( 188/ 140) promoter activities were comparable in normal and PCOS theca cells under basal or forskolin-stimulated conditions. Forskolin treatment increased 235 ( 188/ 140) CYP17 promoter activity approximately 2-fold over basal promoter activity in both normal and PCOS theca cells (Table 1). Similar results were observed after deletion of the 188- to 140-bp sequence from a promoter construct containing 750 bp of the CYP17 promoter (data not shown). These data suggest that one or more elements between 188 and 140 of the CYP17 pro-

4 Wickenheisser et al. Regulation of CYP17 Transcription in Human Theca Mol Endocrinol, March 2004, 18(3): contribute to camp-dependent regulation in both normal and PCOS theca cells. The data presented in Figs. 1 and 2 indicate that increased promoter function in PCOS results from an increase in basal regulation, rather than altered camp-dependent regulation. Scanning Mutant Analysis of the 188- to 140- bp Region of the CYP17 Promoter Fig. 2. Sequences between 188 and 140 bp of the CYP17 Promoter Function Confer Increased CYP17 Promoter Function in PCOS Theca Cells A, Schematic representation of the CYP17 promoter constructs used in these studies. 235 Contains 235/ 44 bp of the 5 -flanking sequence of the CYP17 gene / 140 Contains 235/ 44 bp of the CYP17 promoter with a deletion of sequences from 188 to 140 bp (PCOS) suspected to be involved in differential regulation in PCOS theca cells. (188/ 140)/ 61 Contains the 188/ 140-bp element (PCOS) fused to the minimal 61/ 44 bp promoter. B, Fourth passage theca cells isolated from either normal or PCOS patients were transfected and cultured in the absence or presence of forskolin (20 M) for 72 h. Data represent the mean SEM of independent experiments performed with theca cells isolated from five independent normal and five independent PCOS patients. 235 CYP17 and ( 188/ 140)/ 61 promoter activities were increased in PCOS theca cells as compared with normal theca cells, under basal ( a P 0.01) and forskolin-stimulated ( b P 0.01) conditions. Forskolin treatment increased 235, / 140, and ( 188/ 140)/ 61 promoter activity, in both normal and PCOS theca cells (P 0.01, see Table 1). moter is required for increased basal regulation in PCOS theca cells. As shown in Fig. 2B, 61 CYP17 promoter activity was not different in normal and PCOS theca cells, and was not induced by forskolin. Both basal ( a P 0.01) and forskolin-stimulated ( 188/ 140)/ 61 CYP17 promoter function ( b P 0.01) was increased 2- to 3-fold in PCOS theca cells, as compared with normal theca cells. In both normal and PCOS theca cells, ( 188/ 140)/ 61 CYP17 promoter activity was coordinately increased by 3- to 4-fold in response to forskolin treatment (Table 1). These studies suggest that sequences between 188 and 140 bp confer increased basal CYP17 promoter function in PCOS, and Linker-scanning mutational analysis was performed to identify the nucleotide sequences within 188 and 140 bp of the CYP17 promoter responsible for increased promoter activity in PCOS theca cells. Mutant CYP17 promoter constructs were generated within the context of the 188 CYP17/LUC construct by introducing overlapping 5-bp mutations spanning sequences between 188 and 140 bp (Table 2 and Fig. 3). The relative basal promoter activities of the mutant CYP17 promoter constructs (i.e. M1 11) were compared with the WT 188 CYP17 promoter construct (WT) by transient transfection analysis in independent theca cell cultures isolated from five normal and five PCOS patients. In comparison to the 188 WT construct, mutation of any 5-bp sequence within 188/ 140, with the exception of M5, resulted in a decrease in basal promoter activity, in both normal and PCOS theca cells. This suggests that several factors with overlapping sites contribute to regulation. The strongest reduction (90%) in promoter activity was observed with the M3 and M7 mutants. M1, M2, M3, M5, M8, M9, M10, and M11 CYP17 promoter function was significantly increased in PCOS theca cells as compared with normal theca cells. In contrast, the promoter function of mutants M4, M6, or M7 was equivalent in normal and PCOS theca cells, suggesting that one or more elements mediating differential regulation were disrupted. These data implicate nucleotide sequences from 174 to 158 bp of the CYP17 promoter (M4, M6, and M7) in differential regulation of CYP17 promoter activity in normal and PCOS theca cells. This is in agreement with results from deletion analysis, suggesting that sequences within approximately 172 bp of the CYP17 promoter are sufficient for increased CYP17 promoter function in PCOS theca cells (Fig. 1). The significant reduction in basal activity after mutation of several regions of the 188- to 140-bp element, with the exception of M5, further confirms that this region plays an important role in the basal regulation of the CYP17 promoter. Consistent with our data demonstrating that increased CYP17 promoter function in PCOS theca cells results from altered basal regulation, all of the CYP17 scanning mutants (M1-M11) retained forskolin inducibility (data not shown). Increased CYP17 Promoter Function in PCOS Maps to a Putative NF-1 Site The sequences altered in M4, M6, and M7 correspond to a NF-1 consensus element (Fig. 4A). Lin et al. (29) have recently reported that NF-1 binds to the CYP17

5 592 Mol Endocrinol, March 2004, 18(3): Wickenheisser et al. Regulation of CYP17 Transcription in Human Theca Table 2. Oligonucleotides Used to Generate CYP17 Promoter Mutant Constructs Construct/Oligonucleotide Nucleotide Sequence WT 188/ CCGGAGGTTTGCCCTGGAGTTGAGCCAGCCCTTGAGGAGGCCTTCACTC3 Scanning mutations a Mut1( 186/ 182) CCAAGAATTTGCCCTGGAGTTGAGCCAGCCCTTGAGGAGGCCTTCACTC Mut2( 182/ 178) CCGGAGACCCACCCTGGAGTTGAGCCAGCCCTTGAGGAGGCCTTCACTC Mut3( 178/ 174) CCGGAGGTTTATTTCGGAGTTGAGCCAGCCCTTGAGGAGGCCTTCACTC Mut4( 174/ 170) CCGGAGGTTTGCCCCAAGATTGAGCCAGCCCTTGAGGAGGCCTTCACTC Mut5( 170/ 166) CCGGAGGTTTGCCCTGGAACCACGCCAGCCCTTGAGGAGGCCTTCACTC Mut6( 166/ 162) CCGGAGGTTTGCCCTGGAGTTGGATTGGCCCTTGAGGAGGCCTTCACTC Mut7( 162/ 158) CCGGAGGTTTGCCCTGGAGTTGAGCCGATTTTTGAGGAGGCCTTCACTC Mut8( 158/ 154) CCGGAGGTTTGCCCTGGAGTTGAGCCAGCCTCCAGGGAGGCCTTCACTC Mut9( 154/ 150) CCGGAGGTTTGCCCTGGAGTTGAGCCAGCCCTTGGAAGAGCCTTCACTC Mut 10 ( 150/ 146) CCGGAGGTTTGCCCTGGAGTTGAGCCAGCCCTTGAGGAAATTCTCACTC Mut 11 ( 146/ 142) CCGGAGGTTTGCCCTGGAGTTGAGCCAGCCCTTGAGGAGGCCCCTGTTC Mut 4/6 ( 174/ 170, 166/ 162) CCGGAGGTTTGCCCCAAGATTGGATTGGCCCTTGAGGAGGCCTTCACTC a The sense strand of the CYP17 promoter from 188 bp to 140 bp is shown for scanning mutation sequences. Transition mutations (A N G, C N T) are indicated in boldface. Fig. 3. Scanning Mutational Analysis of Sequences between 188 and 140 bp of the CYP17 Promoter in Normal and PCOS Theca Cells Schematic of the CYP17 promoter constructs used in these studies. Overlapping 5-bp transition mutations spanning sequences from 188 to 140 bp of the CYP17 promoter were generated in the context of the full-length 188 CYP17 construct (WT), generating eleven individual mutant 188 constructs (M1-M11). Fourth passage theca cells isolated from either normal or PCOS patients, were transfected with CYP17 promoter constructs containing 188 bp (WT) of the CYP17 promoter, or 188 bp with scanning mutations within sequences from 188 to 140 bp (M1-M11) and cultured under basal conditions for 72 h. Data represent the mean SEM of independent experiments performed with theca cells isolated from five independent normal and five independent PCOS patients. 188 (WT), M1, M2, M3, M5, M8, M9, M10, and M11 CYP17 promoter activities were significantly increased in PCOS theca cells as compared with normal theca cells, under basal ( a P 0.05) conditions. M4, M6, and M7 CYP17 promoter activities were not significantly different in PCOS theca cells as compared with normal theca cells. promoter and contributes to basal regulation in adrenocortical H295 cells. To examine whether NF-1 binds to the region of the CYP17 promoter that confers increased regulation in PCOS theca cells, EMSA was performed using nuclear extracts (NEs) isolated from normal theca cells maintained under basal conditions. Incubation of theca cell NEs (3 g) with a radioactively labeled oligonucleotide probe corresponding to sequences between 188 and 109 bp of the CYP17 promoter resulted in the formation of multiple protein- DNA complexes (Fig. 4B). Addition of NF-1 antisera that recognizes an amino-terminal region common to all NF-1 family members, resulted in a decrease in intensity of two specific protein/dna complexes. These data demonstrate that NF-1 proteins in theca NEs form specific complexes with the region of the CYP17 promoter required for differential regulation. To identify the sequences of the CYP17 promoter that specifically bind NF-1, competition analysis was performed by addition of 50-fold excess unlabeled WT ( 188/ 109) oligonucleotide or mutant oligonucleotides (Mut 4, 5, 6) corresponding to the scanning mutants. Competition with cold WT 188/ 109 oligonucleotide resulted in a dimunition in the intensity of all of the complexes (Fig. 4B). Addition of an NF-1 consensus oligonucleotide prevented the formation of NF-1 complexes, further demonstrating that proteins present in these complexes interact with a consensus NF-1 binding site. Incubation of the Mut 4 or Mut 6 oligonucleotides did not compete for the NF-1 complexes, indicating that 174/ 170 bp and 166/ 162 bp of the CYP17 promoter are required for binding. In contrast, the Mut 5 oligonucleotide competed for NF-1 binding. These results are in agreement with the bipartite nature of the NF-1 binding site [i.e. (C/ T)TGGC(N) 6 CC(N) 3 ] because the sequences altered in mutants 4 and 6 correspond with the two halves of the NF-1 site required for NF-1 binding (Fig. 4A). The slower migrating, more intense complex was identified

6 Wickenheisser et al. Regulation of CYP17 Transcription in Human Theca Mol Endocrinol, March 2004, 18(3): Fig. 4. NF-1 Specifically Interacts with Sequences within 174 and 162 bp of the CYP17 Promoter A, Schematic of the 188- to 140-bp region of the CYP17 promoter used in these studies. The NF-1 site within 174 and 162 bp of the CYP17 promoter is boxed. The nucleotides mutated in Mut 4, Mut 5, and Mut 6 are underlined. The sequence of the consensus NF-1 oligonucleotide used in these studies is also presented. B, EMSA utilizing a 32 P-labeled probe corresponding to 188/ 109 bp of the CYP17 promoter and 3 g of NEs isolated from theca cells cultured under basal conditions. Antisera to the highly conserved N-terminal DNA-binding domain of all NF-1 proteins (NF-1 Ab) was incubated with NEs before the addition of labeled probe. For competition analysis, 50-fold excess of unlabeled oligonucleotide corresponding to the WT 188/ 109 (WT), Mut 4, Mut 5, Mut 6, or a consensus NF-1 site, were included in the reaction. The NF-1/DNA complexes are indicated with arrows. C, Comparison of the promoter activities of the 188 WT, Mut 4, Mut 6, and double-mutant Mut 4/6 CYP17 constructs after transient transfection of normal or PCOS theca cells under basal conditions. Data represent the mean SEM of independent experiments performed with theca cells isolated from five independent normal and five independent PCOS patients. 188 WT promoter activity was significantly increased in PCOS theca cells as compared with normal theca cells ( a P 0.01). Mut 4, Mut 6, and Mut 4/6 CYP17 promoter activities were not significantly different in PCOS theca cells as compared with normal theca cells. as the regulatory subunits of DNA-dependent protein kinase, Ku 86 and Ku70, which interact with double stranded DNA primarily in a nonspecific fashion (32, 33) because they can be competed by the addition of unlabeled probe (Fig. 4B). To investigate whether the sequences mutated in Mut 4 (between 174/ 170 bp) and Mut 6 (between 166/ 162 bp) comprise a single bipartite NF-1- regulatory element, as discussed above, we generated a double mutant (Mut 4/6). The promoter activity of the Mut 4/6 construct was compared with the WT 188 construct, Mut 4, and Mut 6 constructs under basal conditions, in normal and PCOS theca cells. As shown in Fig. 4C, mutation of either half of the putative NF-1 site (Mut 4 or Mut 6) or both half sites of the bipartite recognition sequence (Mut 4/6) was observed to ablate increased promoter regulation in PCOS theca cells. Furthermore, mutation of both half sites did not further affect promoter function in PCOS theca cells. Examination of NF-1C-Dependent Regulation in Normal and PCOS Theca Cells The genes of NF-1 family members are subject to differential splicing, resulting in the formation of multiple isoforms of each gene (34). To examine whether the NF-1 family member, NF-1C, contributes to NF-1 binding to the CYP17 promoter, supershift analysis was performed with antisera specific to the C terminus of NF-1C. This antisera does not cross react with other NF-1 family members (35). Addition of NF-1C (8199)- specific antibody resulted in the formation of two supershifted complexes, whereas incubation with the NF-1 antibody (common to all NF-1 family members) competed for NF-1 binding (Fig. 5A). Along with the data in Fig. 4, these data demonstrate that an element within 174/ 162 bp of the CYP17 promoter, that is required for increased promoter function in PCOS theca cells, specifically binds NF-1C.

7 594 Mol Endocrinol, March 2004, 18(3): Wickenheisser et al. Regulation of CYP17 Transcription in Human Theca Fig. 5. Characterization of NF-1C Expression and NF-1C-Dependent Regulation of the CYP17 Promoter in Normal and PCOS Theca Cells A, Supershift analysis was performed by incubation of theca cell NEs in the absence or presence of NF-1 antibodies before addition of radiolabeled oligonucleotide corresponding to sequences between 188 and 140 bp of the CYP17 promoter. The positions of NF-1 complexes are indicated by arrows. In the presence of antisera specific for NF-1C (8199), the NF-1 complexes are supershifted (*), whereas addition of the antibody to the amino terminus of all NF-1 family members competes for NF-1 complex formation. B, Schematic of the domains of NF-1C and NF-1C mrna species generated by alternative splicing. The DBD and dimerization domain, encoded by exon 2, is localized to the first 220 residues of NF-1C. The C-terminal region encodes domains required for transactivation/repression, including the PRD and an internal TAD, within exons 5 and 6 (37). The splicing of exon 9 in NF-1C2 results in an altered reading frame (indicated by a dashed line) and translational stop in exon 10 (19). C, NF-1C mrna species were amplified from total RNA isolated from normal or PCOS theca cells, or adrenal H295 cells, using primers specific for exon 3 and exon 11 of NF-1C and separated on a 1.5% agarose gel. The predicted sizes of NF-1C1, NF-1C2, and NF-1C5 products using these primers are 946, 792, and 706 bp, respectively. For comparison, bona fide NF-1C1, 2, and 5, were also amplified. These primers amplify all identified NF-1C isoforms, with the exception of NF-1C3, which does not contain exon 3. NF-1C3, NF-1C4, and NF-1C6 were not detected in human theca cells using alternative primer pairs. D, Normal and PCOS theca cells were transiently transfected with the 188 CYP17 construct in the presence of pcdna3.1 expression vector and/or increasing amounts of pcdna expression vectors for NF-1C1, NF-1C2, or NF-1C5 (0.1 g, 0.3 g, or 1.0 g/dish). Data represent the mean SEM of independent experiments performed with theca cells isolated from five independent normal and five independent PCOS patients. In normal theca cells, CYP17 promoter function was reduced in the presence of 1 g/dish of all three NF-1C isoforms (P 0.05). In PCOS theca cells, CYP17 promoter activity was reduced by all amounts of cotransfected NF-1C5, or g/dish of cotransfected NF-1C1 or NF-1C2 (P 0.05). Seven different isoforms of NF-1C have been reported in various species (34, 36). As presented in Fig. 5B, these NF-1C isoforms result primarily from alternative splicing within the C-terminal region of NF-1C, which contains two domains involved in transactivation, the proline-rich domain (PRD) and a weaker transactivation domain (TAD) within the central region of the NF-1 protein (37). The C terminus of NF-1C has also been reported to function as a transcriptional repression domain, although, the specific domains involved in repression have not been identified (34). The DNA-binding domain (DBD) of NF-1 proteins is highly conserved and resides within the first 220 residues (34). To investigate the NF-1C isoforms expressed in normal and PCOS theca cells, RT-PCR was performed with primers specific for NF-1C isoforms. The primers used in these experiments amplify exon 3 through exon 11 of NF-1C (Fig. 5B), and therefore have the potential to amplify all known isoforms of NF-1C, with the exception of NF-1C3, which lacks exon 3. As shown in Fig. 5C, three NF-1C isoforms, of approximately 950, 800, and 700 bp in length were similarly

8 Wickenheisser et al. Regulation of CYP17 Transcription in Human Theca Mol Endocrinol, March 2004, 18(3): expressed in normal and PCOS theca cells. Based on the predicted size of known NF-1C isoforms, these mrna species were identified as NF-1C1 (946 bp), NF-1C2 (792 bp), and NF-1C5 (706 bp), respectively. RT-PCR of the bona fide NF-1C1, NF-1C2, and NF- 1C5 cdnas (see Materials and Methods) resulted in PCR products with sizes identical to those observed after RT-PCR of theca cell RNA. As a positive control, NF-1C isoforms were also amplified from RNA isolated from adrenocortical H295 cells, where the expression of NF-1C2 and NF-1C5 protein has been described (29). RT-PCR using alternative NF-1C specific primer pairs did not detect NF-1C3, NF-1C4, NF-1C6, or NF-1C7 in human theca cells (data not shown). Together, these data confirm that similar NF-1C mrna splice variants are expressed in normal and PCOS theca cells. To investigate whether NF-1C isoforms affect CYP17 gene expression, normal and PCOS theca cells were transfected with increasing concentrations of NF-1C1, NF-1C2, or NF-1C5 expression vectors, and 188 CYP17 promoter function was examined under basal conditions. In agreement with previous observations, 188 CYP17 promoter activity was 2- to 3-fold higher in PCOS theca cells as compared with normal theca cells. Cotransfection of each of the three NF-1C expression vectors resulted in a dose-dependent decrease in 188 CYP17 promoter function in both normal and PCOS theca cells (Fig. 5D). Cotransfection of 1.0 g/dish of NF-1C1, NF-1C2, or NF-1C5 resulted in approximately 60 70% reduction in 188 CYP17 promoter activity in normal theca cells, and a 75 85% reduction in PCOS theca cells. Similar reductions in CYP17 promoter function by NF-1C isoforms were observed after forskolin treatment (data not shown). These studies demonstrate that each of the three NF-1C isoforms expressed in theca cells have the potential to inhibit CYP17 promoter function in normal and PCOS theca cells. Comparison of NF-1 Binding in Normal and PCOS Theca Cells To compare the relative levels of NF-1 binding in normal and PCOS theca cells, EMSA was performed using NEs isolated from normal and PCOS theca cells maintained under basal conditions, and a radiolabeled oligonucleotide corresponding to 180 to 150 bp of the CYP17 promoter. To better resolve DNA/protein binding to this element we used 8% polyacrylamide gels in these experiments. The observed intensities of NF-1-containing complexes were less in PCOS NEs than in normal NEs (Fig. 6A). A comparable reduction in NF-1 complexes in PCOS theca cell NEs was also observed under forskolin-stimulated conditions (data not shown). Quantitation of the relative NF-1 binding obtained from EMSA of five individual normal and five individual PCOS NEs demonstrates that NF-1 binding is reduced by approximately 70% in PCOS NEs (Fig. 6A). In view of the observation that the sequences flanking the NF-1 site, within 180/ 150 bp of the promoter, appear to contain several putative SF-1 consensus sequences (PyAGGPyC or PuPuAGGTCA), supershift analysis was performed to examine whether SF-1 also binds to this region. Addition of SF-1 antisera competed for a specific complex, indicating that SF-1 binds this sequence. A comparison of SF-1 binding in normal and PCOS NEs demonstrated that the intensity of this SF-1 complex was variable in both cell types (Fig. 6A), with a trend toward increased SF-1 binding in PCOS NEs. In both normal and PCOS, SF-1 binding in individual NEs appeared to have a reciprocal relationship with respect to NF-1C binding (compare normal NEs, lanes 2 and 4, or PCOS NEs, lanes 5 and 6). Together, these data demonstrate that both NF-1 and SF-1 bind sequences within 180/ 150 bp of the CYP17 promoter and suggest that NF-1 and SF-1 may compete for binding to overlapping elements. Additional EMSA was performed utilizing normal and PCOS NEs and a probe corresponding to a shorter sequence of the promoter, from 178 to 158 bp, which retains the NF-1 site, but lacks the SF-1 site. As shown in Fig. 6B, the level of NF-1C binding to this sequence was also decreased approximately 70% in PCOS, similar to those presented for the 180/ 150 bp probe (Fig. 6A). As expected, no SF-1 binding was observed. These data suggest that in the absence of SF-1 binding, NF-1 binding to sequences between 178 and 158 bp of the CYP17 promoter is reduced in PCOS theca cells and does not result from competition with SF-1. Subsequent Western analysis using NF-1C antisera detected several NF-1C proteins ( 50 kda) in both normal and PCOS theca cell NEs, suggesting the presence of several NF-1C isoforms (Fig. 6C). The amount of immunoreactive NF-1C isoforms was reduced, to approximately the same extent as binding was reduced, in NEs isolated from PCOS theca cells as compared with normal theca cells. These data demonstrate that the observed decrease in NF-1C binding in PCOS theca cells may result from a reduction in nuclear NF-1C. Together with the finding that NF-1C represses promoter function, these studies suggest that increased CYP17 promoter activity in PCOS may result, in part, from a relief of NF-1C-dependent repression. Comparison of SF-1 Binding in Normal and PCOS Theca Cells To examine SF-1 binding, EMSA was performed using a probe corresponding to 163/ 140 bp of the CYP17 promoter, which contains a consensus SF-1 site but lacks the NF-1 binding site, and NE isolated from normal and PCOS theca cells. As shown in Fig. 7A, addition of SF-1 antisera resulted in a reduction in binding of a specific complex, indicating that SF-1 binds this sequence. A comparison of SF-1 binding in

9 596 Mol Endocrinol, March 2004, 18(3): Wickenheisser et al. Regulation of CYP17 Transcription in Human Theca Fig. 6. Comparison of NF-1C Binding and NF-1C Proteins in Normal and PCOS Theca Cell NEs A, EMSA utilizing a 32 P-labeled probe corresponding to 180/ 150 bp of the CYP17 promoter and 3 g of NEs isolated from four independent normal and PCOS theca cell cultures under basal conditions. The positions of the NF-1C and SF-1 complexes are indicated with arrows. Addition of antisera specific for SF-1 (SF-1 Ab) competes for SF-1 binding. Quantitation of NF-1C binding after EMSA of five independent normal and five independent PCOS NEs is graphed on the right. Relative NF-1C binding for each NE was obtained by normalization of total NF-1C binding by unlabeled probe. Relative NF-1C binding in PCOS theca NEs was reduced as compared with normal theca NEs (*, P 0.01). B, EMSA utilizing a probe corresponding to 178/ 158 bp of the CYP17 promoter, in which the NF-1 site is retained but the SF-1 sites are deleted, and 3 g of NEs isolated from three independent normal and PCOS theca cell cultures under basal conditions. The positions of the NF-1 complexes are indicated with arrows. Relative NF-1C binding in PCOS theca NEs was similarly reduced as compared with normal theca NEs utilizing this probe. C, Western analysis was performed with 15 g/lane NE isolated from theca cells propagated from three individual normal and PCOS women using NF-1C and Ku-70 antibody. Quantitation of immunoreactive NF-1C protein after immunoblot analysis of five independent normal and five independent PCOS NEs. The nuclear level of NF-1C protein was reduced in PCOS NEs, as compared with normal theca NEs (*, P 0.01). normal and PCOS NEs indicated that the intensity of SF-1 binding was similar in normal and PCOS NEs. Because SF-1 cotransfection was required to investigate differential regulation of this element in normal and PCOS theca cells, experiments were also performed to examine whether cotransfected SF-1 modulates SF-1 and NF-1C binding. To investigate SF-1 binding alone, EMSA was performed using the 163- to 140-bp SF-1 site and NE isolated from cells transiently transfected with either empty pcdna expres-

10 Wickenheisser et al. Regulation of CYP17 Transcription in Human Theca Mol Endocrinol, March 2004, 18(3): Fig. 7. Comparison of SF-1 Binding in Normal and PCOS Theca Cell NEs A, To examine SF-1 binding EMSA was performed utilizing a 32 P-labeled probe corresponding to 163/ 140 bp of the CYP17 promoter, which contains a consensus SF-1 site (but lacks the NF-1 site), and 3 g of NEs isolated from three independent normal and PCOS theca cell cultures under basal conditions. Where indicated, antisera to SF-1 ( ) was incubated with NE before the addition of labeled probe, which competes for SF-1 binding. The positions of the SF-1 complexes are indicated with arrows. SF-1 binding to this sequence was not significantly different in NEs isolated from normal and PCOS cells. B, To examine SF-1 binding in transfected cells, EMSA was performed utilizing the 163- to 140-bp probe and 3 g of NEs isolated from normal theca cells transfected with either empty pcdna expression vector or pcdna containing full-length hsf-1 cdna (SF-1, 0.1 g/dish). The positions of the SF-1 complexes are indicated with arrows. C, To examine both NF-1 and SF-1 binding in transfected cells, EMSA was performed utilizing a 180/ 150 bp probe, which contains both the NF-1 and SF-1 binding sites, and 3 g of NEs from cells transfected with pcdna or SF-1 pcdna. Significant changes in NF-1 and SF-1 binding were not observed after SF-1 cotransfection. sion vector or pcdna encoding human SF-1 (hsf-1) (0.1 g/dish). As shown in Fig. 7B, SF-1 cotransfection did not appear to affect SF-1 binding. To investigate both NF-1 and SF-1 binding in transfected cells, these experiments were also performed with the 180 to 150-bp probe. As shown in Fig. 7C, DNA/protein binding was similar in NEs isolated from both transfected and nontransfected cells and both SF-1 and NF-1-containing complexes were observed. Furthermore, SF-1 cotransfection did not alter overall SF-1 and/or NF-1C binding. Together, these data demonstrate that SF-1 interacts with sequences within 163 / 140 bp of the CYP17 promoter and suggests that cotransfection of SF-1 does not appear to dramatically modify binding to this region. NF-1C-Dependent Repression of CYP17 Promoter Function The observation of NF-1C-dependent repression was unexpected given our previous data demonstrating that mutations that ablate NF-1C binding decrease

11 598 Mol Endocrinol, March 2004, 18(3): Wickenheisser et al. Regulation of CYP17 Transcription in Human Theca CYP17 promoter function. This suggested more complex regulation of the CYP17 promoter. NF-1C-dependent repression of gene expression has been reported to involve the C-terminal repression domain, which can actively repress transcription through the recruitment of corepressors, as well as by the DNA-binding domain (DBD), through direct competition with other transcription factors for DNA binding. Therefore, to begin to investigate the mechanism of NF-1C-dependent repression, we examined whether the DBD of NF-1C could compete with endogenous NF-1 to relieve repression of CYP17 promoter function in normal and PCOS theca cells. As shown in Fig. 8, CYP17 promoter activity was examined after cotransfection of full-length NF-1C2, or an expression vector encoding the DBD of NF-1C (NF-1C-DBD). As expected, inhibition was observed with full-length NF-1C2, whereas cotransfection with NF-1C-DBD resulted in an approximately 4-fold increase in CYP17 promoter activity (P 0.001) in normal theca cells and an approximately 2-fold increase in CYP17 promoter activity in PCOS theca cells (P 0.01), as compared with the empty vector (pcdna) (Fig. 8). The observation of increased CYP17 promoter activity in the presence of cotransfected NF-1C-DBD also suggests that the DBD may compete with endogenous NF-1C proteins for DNA binding, and interfere with NF-1C-dependent repression. The reduced ability of the DBD to augment CYP17 promoter activity in PCOS theca cells may result from decreased endogenous NF-1 binding to Fig. 8. Differential Regulation of the CYP17 Promoter by Full-Length NF-1C2 and the NF-1C DNA-Binding Domain Theca cells isolated from normal and PCOS patients were transfected with the WT 188 CYP17 promoter construct in the presence of empty pcdna3.1, NF-1C2, or NF-1C-DBD expression vectors (1.0 g/dish). The NF-1C-DBD expression vector encodes the DNA-binding domain of NF-1C (1 220AA). Data represent the mean SEM of independent experiments performed with theca cells isolated from five independent normal and five independent PCOS patients. In both normal and PCOS theca cells, 188 CYP17 promoter function was decreased in the presence of NF-1C2 (P 0.05). In both normal and PCOS theca cells, CYP17 promoter activity was increased by NF-1C-DBD (P 0.05). the CYP17 promoter in PCOS theca cells. This suggests that a relief of repression model may explain the differential expression because CYP17 promoter function was similar in both normal and PCOS theca cells after cotransfection with NF-1C or NF-1C-DBD. Sequences within 109/ 75 bp of the CYP17 Promoter Also Bind NF-1C To determine whether the NF-1 binding site between 174 and 162 bp of the CYP17 promoter confers NF-1C-dependent repression, the promoter activity of the Mut 4/6 CYP17 construct was examined after cotransfection with the NF-1C2 in normal theca cells under basal conditions. As shown in Fig. 9A, mutation of the NF-1 site did not ablate NF-1C2-dependent inhibition of CYP17 promoter activity. Cotransfection of NF-1C2 inhibited Mut 4/6 promoter activity by approximately 60% (*, P 0.05). These data suggest that an additional NF-1 binding site may be involved in NF-1C-dependent repression of the CYP17 promoter. Recently, Lin et al. (29) identified a more proximal NF-1 binding site within 107 to 85 bp of the CYP17 promoter involved in basal regulation in adrenal H295 cells. To determine if NF-1C binds to this more proximal sequence in the CYP17 promoter, EMSA was performed using normal theca cell NEs and a radiolabeled probe corresponding to sequences within 113/ 75 bp of the CYP17 promoter. Competition analysis was performed by addition of 50-fold excess of unlabeled 113- to 75-bp oligonucleotide, NF-1 consensus oligonucleotide, or 113/ 75-bp oligonucleotide containing mutations in the NF-1 site (Fig. 9B). Incubation with unlabeled WT 113/ 75-bp, or NF-1 consensus oligonucleotide resulted in a reduction in the intensity of all protein-dna complexes. In contrast, addition of 50-fold excess of the mutant 113/ 75 oligonucleotide partially competed for the formation of all but two protein-dna complexes, indicated by arrows. Supershift analysis utilizing the NF-1C specific antisera (8199) resulted in the formation of supershifted complexes, as well as the dimunition of two specific complexes. Addition of NF-1 antisera competed with the same two specific complexes. Together, these data verify that NF-1C in theca cell NEs also binds to sequences within 113 to 75 bp of the CYP17 promoter. The Proximal NF-1 Site Is Not Required for Increased Promoter Function in PCOS To evaluate the extent to which these NF-1 sites contribute to CYP17 promoter function, the promoter activity of constructs containing mutations in: 1) the distal site (site A) between 174/ 162 bp (Mut A); 2) the more proximal site (site B) from 102/ 90 (Mut B); or 3) both NF-1 sites (Mut A/B), was examined within the context of the WT 188 CYP17 promoter (WT) (Fig. 10A). To independently examine NF-1C-dependent regulation of the two NF-1 sites, we compared the

12 Wickenheisser et al. Regulation of CYP17 Transcription in Human Theca Mol Endocrinol, March 2004, 18(3): Fig. 9. Identification of a More Proximal NF-1 Site within 113 and 75 bp of the CYP17 Promoter A, Normal theca cells were transfected with the 188 Mut 4/6 CYP17 construct, containing mutations to both half-sites of the NF-1 site located between 174 and 162 bp of the CYP17 promoter, in the presence of empty pcdna, NF-1C2, or NF-1C-DBD expression vector (1.0 g/dish). The promoter function of 188 Mut 4/6 CYP17 was decreased in the presence of NF-1C2 (*, P 0.05). Cotransfection of NF-1C-DBD did not significantly affect either 188 Mut 4/6 CYP17 promoter function. B, Schematic of the 113- to 75-bp region of the CYP17 promoter. The NF-1 site within 102 and 90 bp of the CYP17 promoter is boxed. The nucleotides mutated in the Mut 113/ 75 oligonucleotide are underlined. The sequence of the consensus NF-1 oligonucleotide used in these studies is also presented. EMSA utilizing a 32 P-labeled probe corresponding to 113/ 75 bp of the CYP17 promoter and 3 g of normal theca cell NEs. For competition analysis, 50-fold excess of unlabeled WT ( 113/ 75), mutant 113/ 75, or a consensus NF-1 site oligonucleotide were included in the reaction. The positions of protein-dna complexes containing NF-1C proteins are indicated by arrows. In the presence of antisera specific for NF-1C (8199), the NF-1 complexes are supershifted (*), whereas, addition of the antibody to all NF-1 family members competes for NF-1 complex formation. Fig. 10. Increased CYP17 Promoter Function Requires only the Distal NF-1 site Located within 174 to 162 bp A, A schematic of the CYP17 promoter constructs used in these studies containing mutations in either the distal NF-1 site (Mut A), the proximal NF-1 site (Mut B), or both NF-1 sites (Mut A/B) within the context of 188/ 44 bp of the CYP17 promoter ( 188). B, A comparison of WT, Mut A, Mut B and Mut A/B promoter activities in the presence of pcdna, NF- 1C2, or NF-1C-DBD expression vectors (1.0 g/dish) after transient transfection into normal theca cells. In the presence of NF-1-C2, the promoter activities of WT, Mut A, and Mut B were decreased ( c P 0.05). In the presence of NF-1C-DBD, the promoter activities of WT and Mut B were increased ( d P 0.05). C, A comparison of the promoter activities of the NF-1 mutant CYP17 promoter constructs after transient transfection into normal and PCOS theca cells under basal conditions. Data represent the mean SEM of independent experiments performed with theca cells isolated from five independent normal and five independent PCOS patients. The promoter activities of the WT and Mut B construct were increased after transfection into PCOS theca cells as compared with normal theca cells ( a P 0.05). promoter activities of the WT, Mut A, Mut B, and the double mutant (Mut A/B) after cotransfection with pcdna vectors encoding NF-1C2 or DBD/NF1C in normal theca cells. As shown in Fig. 10B, NF-1C2