Cloning and Characterization of a Novel Zinc Finger Protein that ...

0888-8809/03/$15.00/0 Printed in U.S.A.

Molecular Endocrinology 17(11):2303–2319 Copyright © 2003 by The Endocrine Society doi: 10.1210/me.2003-0158

Cloning and Characterization of a Novel Zinc Finger Protein that Modulates the Transcriptional Activity of Nuclear Receptors BENTE BØRUD, GUNNAR MELLGREN, JOHAN LUND,

AND

MARIT BAKKE

Department of Anatomy and Cell Biology and Hormone Laboratory (B.B., G.M., J.L., M.B.), Department of Clinical Biochemistry (G.M.), University of Bergen, N-5009 Bergen, Norway The orphan nuclear receptor steroidogenic factor-1 (SF-1) plays pivotal roles in the development and function of steroidogenic organs. It transcriptionally regulates an array of factors required for biosynthesis of steroid hormones and is also necessary for the expression of genes in the pituitary and the male reproductive tract. Here we describe the identification of a novel zinc finger protein that modifies the transcriptional potential of SF-1. This factor, which we call Zip67 (zinc finger protein 67 kDa), was cloned through a two-hybrid screen of a human testis cDNA library using the C-terminal part of SF-1 as the bait. Transient transfection experiments demonstrated that Zip67 represses SF1-dependent transcription in the context of both multimerized SF-1-binding sites and natural SF-1-

inducible promoters. The interaction between Zip67 and SF-1 was dependent on an intact activation function-2 domain of SF-1, and we propose a mechanism whereby Zip67 represses transcription through competition with p160 coactivators for binding to SF-1. Zip67 was detected in SF-1 expressing tissues such as testis, adrenal, ovary and spleen in addition to other tissues. In line with the broader expression pattern, we found that Zip67 also affected transcription mediated by several other nuclear receptors. In conclusion, we have isolated a novel zinc-finger protein that influences gene activation through interaction with the functionally important activation function-2 domain of nuclear receptors. (Molecular Endocrinology 17: 2303–2319, 2003)

S

mones, such as the cytochrome P450 steroid hydroxylase genes, the ACTH receptor gene and the gene encoding steroidogenic acute regulatory protein (StAR) (7–11). SF-1 is also expressed in Sertoli cells, where it synergizes with Sox9 and WT-1 in the regulation of the Mullerian inhibiting substance (MIS) gene (12, 13). Outside steroidogenic tissues, SF-1 is expressed in the gonadotropes of the pituitary, the VMH, and the spleen (14–16). In the gonadotropes, SF-1 regulates the genes encoding the GnRH receptor and the ␣- and ␤-subunits of the gonadotropins (17). No definitive target genes for SF-1 have been identified in the VMH or spleen. Based on its structural and functional characteristics, SF-1 belongs to the nuclear receptor (NR) family of transcription factors (8, 18–20). Nuclear receptors contain a characteristic zinc finger DNAbinding domain, an intervening hinge region and a C-terminal ligand-binding domain (LBD) that contains an activation function 2 (AF-2) domain. The AF-2 domain is required for ligand-dependent activation, and binding of ligand is believed to cause conformational changes that make the protein more accessible for interactions with coactivators (21). Although SF-1 can be activated by oxysterols in certain cellular contexts (22), a bona fide ligand for SF-1 has still not been identified, and SF-1 is therefore classified under the large subgroup of NRs designated orphan NRs (23). Nevertheless, similarly to other orphan NRs, SF-1 contains a well con-

TEROIDOGENIC FACTOR-1 (SF-1/AD4BP/NR5A1) is one of the major regulators of endocrine function (1). Targeted deletion of the gene that encodes SF-1 in mice leads to a severe and complex phenotype with ablation of the structures corresponding to adrenals, gonads, and the ventromedial hypothalamic nucleus (VMH). The phenotype is also associated with dysfunctional gonadotropes and male to female sex reversal of internal and external genitalia (2, 3). During the last years, human patients with mutations in the SF-1 gene have been described, and interestingly, the phenotype shows that SF-1 is essential for normal adrenal and gonadal development in humans also (4– 6). SF-1 is expressed throughout the zones of the adrenal cortex and in the steroidogenic cells of the gonads. In these cells SF-1 regulates a number of genes involved in the biosynthesis of steroid horAbbreviations: AF-2, Activation function 2; AhR, arylhydrocarbon receptor; Arnt, AhR nuclear translocator; CMV, cytomegalovirus; CREB, cAMP response element-binding protein; CRS2, cAMP-responsive sequence 2; ER, estrogen receptor; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; GR, glucocorticoid receptor; GRIP1, GR interacting protein 1; LBD, ligand-binding domain; LXR, liver X receptor; MIS, Mullerian inhibiting substance; NGFI-B, nerve growth factor inducible clone B; Ni-NTA, nickel-chelate-nitrilotriacetic acid; NR, nuclear receptor; PKA, protein kinase A; PPAR␥, peroxisomal proliferator-activated receptor ␥; RXR, retinoid X receptor; SF-1, steroidogenic factor-1; SRC-1, steroid receptor coactivator-1; TK, thymidine kinase; VMH, ventromedial hypothalamic nucleus.

2303

2304 Mol Endocrinol, November 2003, 17(11):2303–2319

served AF-2 domain that is involved in transcriptional control and interaction with coregulators. For instance, the coactivators steroid receptor coactivator-1 (SRC-1) (24, 25), transcription intermediary factor-2/glucocorticoid receptor (GR)-interacting protein 1 (GRIP1) and p/CIP [p300/CREB-binding protein (CBP)/cointegrator-associated protein] (26) all interact with the AF-2 domain of SF-1. In addition, SF-1 interacts with proteins that repress its transcriptional potential. The AF-2 domain is required for interaction with the corepressor RIP140 (27), whereas other repressors, such as the atypical NR Dax-1 (dosage-sensitive sex reversal-adrenal hypoplasia cogenita critical region on the X chromosome, gene 1) (12, 28), silencing mediator of retinoic acid and thyroid hormone receptor (29), and the DEAD box protein DP103 (30) appear to require domains N terminal to AF-2 for interactions with SF-1. Although it is possible that the activity of orphan receptors is regulated through the binding of as yet unidentified ligands, other modes of regulation, such as posttranslational modifications and selective interactions with cofactors, are likely to be decisive. In the case of SF-1, both acetylation and phosphorylation events are known to alter its transcriptional potential. Phosphorylation cascades that are initiated by hormonal regulators of steroidogenesis activate many SF-1 target genes, and considerable attention has therefore been focused at defining how SF-1 is altered by these events. Hammer and co-workers (29) demonstrated that SF-1 is directly phosphorylated by MAPKs at serine 203 in the hinge region and that this enhanced the recruitment of cofactors. Furthermore, it was proposed that phosphorylation of serine 203 mimics the stabilizing effect of ligand on the LBD of SF-1 (31). Protein kinase A (PKA) has also been implicated in SF-1 function, but SF-1 does not appear to be directly phosphorylated by PKA in vivo (29, 32, 33). However, a recent study demonstrates that the stability of SF-1 increases in the presence of PKA (33), and furthermore, several studies have pointed to a role for the PKA signaling pathway in SF-1-mediated transactivation (7, 26, 34–36). It was also reported that SF-1 is acetylated by GCN5 in vitro and that mutations in the putative acetylated residues decrease the transcriptional capacity of SF-1 (37). Clearly, therefore, various factors and signaling cascades within the cell can fine-tune the activity of SF-1, apparently in the absence of a specific ligand. With the aim to unravel novel aspects of how the transcriptional capacity of SF-1 can be modified, we performed a yeast two-hybrid screen to identify proteins that interact with the C-terminal region of SF-1. To this end, we identified a novel protein that is highly conserved between mouse and human and that represses SF-1-dependent transcriptional activity as well as the transcriptional activity of other NRs.

Børud et al. • Zip67, a Novel Repressor of Nuclear Receptors

RESULTS Cloning of Zip67 To identify factors that interact with SF-1, we used the LBD and hinge region of bovine SF-1 (amino acids 107–461) as the bait in a yeast two-hybrid screen of a human testis cDNA library. After repeated selection and retransformation, approximately 50 positive clones were sequenced. Among these was a 763-bp fragment of a hitherto unknown cDNA. Search of the human genome database with this fragment gave perfect match to chromosome 19, and further analyses using the NIX database (38) strongly indicated that the isolated fragment was part of a cDNA encoded by a gene consisting of nine exons located between bp 53,541 and 77,364 on chromosome 19 (Fig. 1A). The gene computation was strengthened by the prediction of a promoter region containing a TATA box approximately 100 bp upstream of the ATG start codon and a putative poly(A) tail downstream of the last exon. The fragment that was isolated in the two-hybrid screen corresponded to nucleotides 1,134–1,882 of the predicted cDNA (Fig. 1A). The full-length cDNA was subsequently cloned from a ␭ human testis cDNA library by a combinatorial approach of conventional hybridization screening and PCR (Fig. 1B). We named this novel factor Zip67, for zinc finger protein 67 kDa. Sequencing of the full-length human Zip67-cDNA revealed a perfect match to the predicted exons of the corresponding region of chromosome 19, and the intervening sequence completely matched the exonintron boundaries by the GT-AG rule (39) (see Table 1 for details). Importantly, the length of the Zip67 cDNA, as predicted from the NIX database, was in agreement with the mRNA size observed on Northern blots (data not shown). Translation of the full-length Zip67 cDNA resulted in a protein product of approximately 67 kDa, consistent with an open reading frame of 615 amino acids (Fig. 1C). We used the Compute pI/Mw tool from ExPASy (40) to calculate the theoretical molecular mass of the amino acid sequence, and Zip67 was predicted to be 67.234 kDa. Furthermore, an antibody raised against aa 445–458 (AIIYEIPKEPEKRR) in human Zip67 recognized a protein of 67 kDa in extracts prepared from mouse adrenocortical tumor cells (Y1) and COS-1 cells (Fig. 1D). The Pfam 6.6 protein family database (41) predicted five classic C2H2 zinc fingers in the C terminal of the Zip67 protein (with E values ranging from 2.1 e-7 to 0.005; Fig. 3), and the PSORT database (42) predicted two nuclear localization signals located between amino acids 107–118 and 445– 451 (Fig. 3). Apart from these domains, no other protein motifs with significant E values were recognized. Zip67 Interacts Directly with SF-1 in an AF-2-Dependent Manner To assert that the interaction between SF-1 and Zip67 was direct and independent of other factors present in


Mol Endocrinol, November 2003, 17(11):2303–2319

2305

yeast, we performed an in vitro protein-protein interaction assay using purified baculovirus-expressed (His)6-SF-1 fusion protein and in vitro translated Zip67 protein. When incubated together with (His)6-SF-1, radiolabeled Zip67 could be pulled down by nickelchelate-nitrilotriacetic acid (Ni-NTA) agarose, demonstrating a direct interaction between these two proteins (Fig. 2A). No interaction was observed with the Ni-NTA agarose alone. The part of SF-1 used in the two-hybrid screen contains the AF-2 domain that is known to be important for interactions with certain cofactors. To investigate the significance of this domain for the interaction with Zip67, we included a mutated form of SF-1 that carries a deletion in the AF-2 domain [(His)6-SF-1⌬AF-2] in the experiment. Interestingly, the interaction between Zip67 and SF-1 was nearly abrogated in the absence of a functional AF-2 domain (Fig. 2A). A yeast-mating assay confirmed the requirement for the AF-2 domain of SF-1. In these experiments the same region of SF-1 that was used in the initial two-hybrid screen (SF-1107–461), either wild type or a mutant carrying a deletion of the AF-2 domain, was fused to the LexA DNA-binding domain to make up the plasmids LexA/SF-1107–461 and LexA/SF-1107–461⌬AF-2. The cDNA encoding Zip67 (bp 1134–1882; the fragment isolated in the two-hybrid screen) was fused to the B42 transcriptional activation domain (B42/Zip67), and the bait and prey plasmids were transformed into MATa and MAT␣ strains, respectively. A liquid semiquantitative ␤galactosidase assay demonstrated a 16-fold increase in reporter gene activity when SF-1 and Zip67 were brought together in the diploid cell (Fig. 2B). Similarly to what we observed in the pulldown assay, the interaction between SF-1 and Zip67 was completely annulled when the AF-2 domain was deleted (Fig. 2B). Western blot analyses of protein extracts prepared from transformed yeast using an anti-LexA antibody, demonstrated that LexA/SF-1107–461 and LexA/SF1107–461⌬AF-2 were expressed at equal levels (data not shown) (26). LexA/SF-1 and B42/Zip67 showed very low or undetectable ␤-galactosidase activity when transformed alone or together with the empty vectors (Fig. 2B). Together these experiments indicate that the interaction between SF-1 and Zip67 is direct

Fig. 1. Chromosomal Localization and cDNA Sequence of Human Zip67 A, The gene encoding Zip67 is located on human chromosome 19 (between bp 52,860 and 75,217) as predicted by the NIX database. The gene, which contains 9 exons, encodes an mRNA of 2101 bp that gives rise to a protein product of 615 amino acids. The protein contains 5 zinc finger domains (u). The position of the 763-bp fragment that was originally iso-

lated in the yeast two-hybrid screen is indicated. B, The open reading frame and the corresponding amino acid sequence of Zip67 are shown. The asterisk indicates the stop codon. Also shown is the 3⬘-untranslated sequence containing the poly(A) sequence. C, Zip67 was translated in vitro from the plasmid pCMV-BK-Zip67 in the presence of [35S]methionine, and the resultant protein product was visualized by SDSPAGE, followed by autoradiography. The size of the protein (67 kDa) is in agreement with the length of coding region of the cDNA (1,845 bp) and the number of amino acids (615). D, Western blotting detected the Zip67 protein in both Y1 and COS-1 cells using an antibody raised against amino acids 445–458 (AIIYEIPKEPEKRR).



Table 1. Overview of the Exon Localization and the Exons/Intron Boundaries of Zip67 on Human Chromosome 19

a

Exon

From

1 2 3 4 5 6 7 8 9a Total

52,860 (ATG) 60,307 62,481 70,743 71,488 71,766 72,876 74,505 74,787

3⬘ Intron:5⬘ Exon

To

3⬘ Exon: 5⬘ Intron

bp

GCAG:GAAG ATAG:GGCG ACAG:TGGG GTAG:AGAA CCAG:GAGG CCAG:AACC CCAG:GTGT GCAG:GTGC

53,158 60,350 62,696 71,354 71,659 71,877 72,990 74,604 75,217

ACGG:GTGA AAAA:GTAA AAAG:GTAG AAAG:GTCT AGAA:GTGG CCAG:GTGA TCAG:GTCA TGCA:GTGA

299 44 216 612 172 112 115 100 431 2,101

The last exon (exon 9) contains a stop codon and a poly(A) site.

and independent of additional factors. Furthermore, although it cannot be excluded that other domains of SF-1 participate in the binding of Zip67, our results clearly point to an absolute requirement for the AF-2 domain. Cloning of the Mouse Homolog of Zip67 Search of the mouse expressed sequence tag database revealed the presence of fragments with high homology to the human Zip67 cDNA. Based on this information we constructed primers that corresponded to the most 5⬘ and 3⬘ sequences. Using these primers a fragment of the expected size was amplified from a mouse testis cDNA library, and sequencing revealed that we had isolated the mouse homolog of Zip67. The overall identity between human and murine Zip67 was 85% at the nucleotide level and 90% at the amino acid level (Fig. 3), demonstrating a high degree of evolutionary conservation and suggesting important physiological roles for this protein. Expression of Zip67 The expression profile of Zip67 was analyzed in mouse and human tissues. We performed quantitative RTPCR on mRNA isolated from various mouse tissues. The levels of Zip67 mRNA were measured using a standard curve generated from a linearized plasmid containing the full-length mouse Zip67 cDNA. The values were normalized against the content of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA in the same extracts. For both Zip67 and GAPDH the amplification crossed an intron/exon boundary, virtually eliminating the possibilities of chromosomal DNA artifacts. Agarose gel electrophoresis and melting point analyses of the PCR products revealed the presence of a single species in each sample (Fig. 4A). As shown in Fig. 4B, we identified the highest levels of Zip67 mRNA in the testis. Relatively high levels were also detected in the spleen. In addition, we found moderate levels of expression in lung, adrenal gland, uterus, and ovary, whereas the rest of the tissues examined (i.e. pancreas, heart, skeletal muscle, adi-

pose tissue, kidney, and liver) contained low to undetectable levels. To determine the expression profile of Zip67 in humans, we radiolabeled the human Zip67 cDNA and hybridized it to a Multiple Tissue Expression Array blot (Clontech, Palo Alto, CA). It became evident that in humans, Zip67 is expressed at high levels in the testis (Fig. 5). Furthermore, we found that Zip67 is abundantly expressed in certain areas of the brain (i.e. cerebellum, temporal lobe, and hippocampus). In addition, relatively high expression was found in the adrenal gland. After extensive washing of the filter, weak hybridization signals were still apparent in some of the negative controls. We therefore used a phosphorimaging system to obtain semiquantitative values of the hybridization signals relative to the negative controls. Such an analysis demonstrated that, similarly to those in the mouse, moderate levels of Zip67 were found in human spleen and uterus. In addition, Zip67 mRNA was apparent in thymus, pancreas, kidney, stomach, and rectum. Although the cellular localization needs to be examined further, it is clear that Zip67 is coexpressed with SF-1 in tissues in which this NR is known to play fundamental physiological roles (i.e. testis and adrenal gland) in both mice and humans. However, it also became evident that the expression patterns of these two factors do not overlap absolutely, as Zip67 does not appear to be present in human ovary, and Zip67 exhibits a broader expression pattern. Zip67 Represses Transcription Mediated by Nuclear Receptors To investigate the functional significance of the interaction between SF-1 and Zip67, we examined whether Zip67 could affect the transcription of reporter genes driven by SF-1-dependent promoters. We previously described an element, cAMP-responsive sequence 2 (CRS2), from the bovine cytochrome P450 17␣hydroxylase gene (CYP17) that is positively regulated by SF-1 (7). A reporter construct containing four copies of CRS2 placed upstream of a minimal thymidine kinase (TK) promoter and luciferase as the reporter gene (4CRS2-luc) was transfected into Y1 mouse adrenocortical cells. Y1 cells express relatively high lev-


Fig. 2. The Interaction between SF-1 and Zip67 Is Dependent on the AF-2 Domain of SF-1 A, [35S]Methionine-labeled, in vitro-translated human Zip67 (3 ␮l) was incubated with a (His)6-SF-1 fusion protein (5 ␮g) or a (His)6-SF-1⌬AF-2 fusion protein (5 ␮g) in which the AF-2 domain (LLIEML) of SF-1 is deleted. The protein mixtures or [35S]methionine-labeled Zip67 (3 ␮l) alone were then incubated with NiNTA agarose. After extensive washing, the proteins were analyzed by SDS-PAGE (6%). Thirty percent of the input radiolabeled Zip67 is shown in the first lane. B, A liquid semiquantitative ␤-galactosidase assay of the interaction between SF-1 and Zip67 in yeast. The prey plasmid carrying Zip67 fused to the B42 activation domain (B42/Zip67) was transformed into the yeast strain EGY191 (MAT␣) and mated to yeast (L40; MATa) carrying either wild-type SF-1 (LexA/SF-11107–461) or the AF-2 deletion mutant (LexA/SF-11107–461⌬AF-2) fused with the LexA DNA-binding domain. After mating, ␤-galactosidase activities were measured (RLU/OD600). As controls, yeast carrying either LexA/SF-1 or B42/Zip67 were mated with transformants carrying empty prey or bait vectors, respectively (marked ⫺). In addition, we included a negative control in which B42/Zip67 was combined with LexA/lamin. The figure shows the mean values of at least four independent colonies.

els of SF-1 endogenously, and luciferase activity can therefore also be detected in the absence of externally added SF-1 (Fig. 6A). However, the luciferase activity


2307

was further increased around 8-fold when SF-1 was coexpressed (Fig. 6A). Interestingly, when an expression plasmid encoding Zip67 was included in the experiments the luciferase activity decreased in both the presence and absence of overexpressed SF-1. These results clearly indicate, therefore, that Zip67 opposes the effects of SF-1 in vivo. To examine whether the effect of Zip67 could be reconstituted on an intact SF-1-responsive promoter, a fragment of the bovine CYP17 gene promoter was placed upstream of the reporter gene luciferase (CYP17–290-luc). This fragment, which spans from ⫺290 to ⫹17 relative to the start site of transcription, contains the CRS2 element described above. Sequence alignment with the human CYP17 gene suggests that another SF-1 binding element may be present around bp ⫺240 (43); however, the functional significance of this element has not been evaluated in the context of the bovine promoter. When CYP17– 290-luc was transfected into Y1 cells together with the expression plasmid encoding SF-1, we observed, as expected, an increase in luciferase activity (Fig. 6B). The induction was not as profound as for 4CRS2-luc, probably because of the fewer copies of SF-1 binding sites present in this native promoter. Again, coexpression of Zip67 led to decreased luciferase activity, and similarly as for 4CRS2-luc, the repressive effect of Zip67 was found in both the presence and absence of exogenously added SF-1. Similar results were observed when Y1 cells were transfected with the promoter from the cytochrome P450 side-chain cleavage gene (CYP11A). The CYP11A⫺186-luc plasmid contains the region between ⫺186 and ⫹12 of the bovine CYP11A gene linked to luciferase. This region of the CYP11A promoter is known to contain at least one SF-1-binding site located at ⫺39/⫺48 relative to the start site of transcription and will support SF-1 inducibility in COS-1 cells (data not shown) (44). In contrast to the CYP17 promoter construct, additional expression of SF-1 did not increase luciferase activity above background levels (Fig. 6C). Most likely the reason for this is that the endogenous levels of SF-1 in Y1 cells are sufficiently high for full activation of this construct. Analogously to what was observed for CYP17, Zip67 also repressed this natural SF-1-dependent promoter. To assure that the effect of Zip67 is dependent on SF-1, we used COS-1 cells that do not express endogenous SF-1 and in which 4CRS2-luc shows minimal activity without exogenous addition of SF-1 (7). The 4CRS2-luc construct was transfected into the COS-1 cells together with increasing amounts of Zip67 in the presence or absence of SF-1. As previously shown (7, 26), the transcriptional activity was induced by coexpression of SF-1 (Fig. 6D). Coexpression of Zip67 repressed transcriptional activity in presence of SF-1, whereas the effect was minimal in the absence of SF-1. These results thereby suggest that the repressive effect is dependent on SF-1 (Fig. 6D). Furthermore, SIP67 exerted no effect on transcription mediated by the cytomegalovirus (CMV)-␤-galactosi-



Fig. 3. Alignment of the Human and Mouse Zip67 Amino Acid Sequences The mouse and human Zip67 sequences are 85% identical at the nucleotide level and 90% identical at the amino acid level. The five predicted zinc finger regions are shaded, and the putative nuclear localization signals are given in bold. The underlined sequence corresponds to the fragment that was isolated in the initial two-hybrid screen.

dase reporter plasmid, arguing against a general repressive effect (data not shown). The rather broad expression pattern suggested a role for Zip67 that extends beyond that of being a cofactor exclusively for SF-1. To investigate whether Zip67 would repress transcription mediated by other NRs, we performed transient transfection experiments in COS-1 cells. Reporter constructs carrying binding sites for peroxisomal proliferator-activated receptor ␥ (PPAR␥), liver X receptor (LXR)␣, nerve growth factor inducible clone B (NGFI-B), GR and estrogen receptor (ER) linked to luciferase as the reporter gene were transfected together with expression plasmids encoding the corresponding NR. The cognate ligands were included as indicated in Fig. 7. Interestingly, coexpression of Zip67 led to a dose-dependent decrease in reporter gene activity for all of the NRs tested, suggesting that Zip67 functions as a general cofactor for this group of transcription factors.

The data presented in Fig. 2 demonstrated an absolute requirement for the AF-2 domain for the interaction between SF-1 and Zip67. To investigate whether Zip67 could inhibit other classes of transcription factors that lack this domain, we examined the effect of Zip67 on transcription mediated by the basic leucine zipper transcription factor cAMP response element-binding protein (CREB) and the helix-loop-helix PAS (Per-Arnt-Sim) domain factors arylhydrocarbon receptor (AhR)/AhR nuclear translocator (Arnt). COS-1 cells were transfected with an expression plasmid encoding CREB (Rous sarcoma virus-CREB) together with a reporter plasmid carrying CREB-binding elements (4CRE-luc). As transcription mediated by CREB is activated by PKA, an expression plasmid encoding the catalytic subunit of PKA (pCMV5-C␣) was included in the experiment (Fig. 8A). In Fig. 8B, HeLa cells were transfected with a reporter construct carrying xenobiotic response elements (2XRE-luc) and subsequently



2309

Fig. 4. Expression of Zip67 in Mouse Tissues mRNA was isolated from homogenized mouse tissues using the MagNA Pure LC. Real-time RT-PCR were performed using SYBR-Green I fluorescence to measure the product accumulation. A, Melting point analyses demonstrated single peaks and specific products for GAPDH (upper panel) and Zip67 (lower panel). Mouse testis and ovary cDNA and the GAPDH cDNA standards are presented in the figure to show that there are no melting point differences (upper panel). cDNA from mouse testis and ovary were also compared with the standard curves obtained from the pCR2.1-mZip67 plasmid (lower panel). B, mRNA quantification. The expression of Zip67 relative to the expression of GAPDH in various mouse tissues is shown. The figure shows the mean ⫾ SD of three or four separate experiments.

treated with 2,3,7,8-tetrachlorodibenzo-p-dioxin for 24 h to induce transcription mediated by AhR/Arnt. In these systems we did observe a decrease in luciferase activity when Zip67 was coexpressed. However, compared with NRs (Figs. 6 and 7), the effect was minor, and furthermore, a similar attenuation in reporter gene activity was found when transcription was driven by the TK promoter alone (i.e. TK-␤-galactosidase; data not shown). This indicates that the effect is independent of CREB and AhR/Arnt and instead is mediated through the TK promoter. Such a minor effect of Zip67 on basal transcription when overexpressed is not surprising, and it was also previously reported that NR coregulators can affect constitutive promoters (45). Zip67 Counteracts the Positive Effect of GRIP1 on SF-1-Dependent Transcriptional Activity To investigate whether Zip67 contains an intrinsic repressive function, we linked the Zip67 cDNA to the GAL4 DNA-binding domain. Surprisingly, as shown in Fig. 9, we found that this fusion protein slightly activated transcription when expressed together with a reporter plasmid that carries GAL4-binding elements. When a fusion protein between GRIP1 and GAL4 was included in the experiment, it became clear, however,

that Zip67 is a weak activator compared with this representative from the p160 family of coactivators. The fact that both p160 coactivators (24–26, 29) and Zip67 (Fig. 2) interact with SF-1 through the AF-2 domain led us to hypothesize that the mechanism by which Zip67 represses SF-1-dependent transcription might be through competition for the AF-2 domain. We therefore transfected COS-1 cells with the reporter plasmid 4CRS2-luc together with expression plasmids encoding SF-1, Zip67, and GRIP1. As shown in Fig. 10A, increasing amounts of Zip67 counteracted the positive effect of GRIP1 on SF-1-dependent transcription in a dose-dependent manner, and conversely, GRIP1 opposed the negative effect of Zip67 in this system. To examine whether Zip67 and GRIP1 could compete for SF-1 binding in a more direct mode, we performed an in vitro competition interaction analysis. In this experiment, [35S]methionine-labeled in vitro translated GRIP1 and Zip67 were incubated with a (His)6-SF-1 fusion protein, and the protein complexes were pulled down by Ni-NTA agarose. Although the bands in this experiment appeared relatively faint, it can be seen in Fig. 10B (lane 1) that GRIP1 interacts with SF-1 in this system, and furthermore, that the addition of increasing amounts of Zip67 causes a

2310



Fig. 5. Expression of Zip67 in Human Tissues A multiple tissue expression array containing poly(A)⫹ from various human tissues was hybridized with a cDNA probe. The diagram of the RNA sources is shown in the lower panel.

dose-dependent decrease in the binding of GRIP1 and a corresponding increase in the binding of Zip67 to SF-1 (Fig. 10B, lanes 2–4). Taken together, these results suggest that Zip67, although a weak activator in itself, will act as a general repressor of transcriptional activation by NRs by preventing the interaction with much more potent coactivators, such as GRIP1.

DISCUSSION In this study we describe the protein Zip67, a novel transcriptional coregulator of SF-1 and other NRs. Zip67 cDNA was isolated through its ability to interact

32

P-labeled Zip67

with the LBD domain of SF-1 and encodes a 67-kDa protein containing five C2H2 zinc fingers in its Cterminal region. Several other C2H2 zinc finger proteins were previously described to directly interact with SF-1 and modulate its transcriptional activity (12, 46– 48). Similarly to Zip67, Sp1 and TReP-132 are known to affect the activity of steroid hydroxylase genes through interactions with SF-1 (i.e. CYP11A and CYP17) (44, 49). The Egr-1 protein interacts with SF-1 on the LH␤ promoter providing a synergistic effect on transcription and thereby affecting female reproductive capacity (48), and WT1 associates with SF-1 to promote MIS expression during testis development (12). The above-mentioned factors potentiate the tran-



2311

scriptional capacity of SF-1. In contrast, the major role of Zip67 appears to be that of a transcriptional repressor of NRs, at least in cellular contexts where p160 coactivators are not limiting. Electrophoretic mobility shift assays demonstrated that Zip67 is unable to interact directly with the SF-1-binding elements used in the transfection experiments, and moreover, that the binding of SF-1 was unaffected by the presence of Zip67 (data not shown). It appears, therefore, that Zip67 does not require contact with DNA to modulate SF-1-dependent gene activation. Analogous modes of protein-protein interaction between SF-1 and coregulators were previously described to take place on other promoter elements. For instance, the stimulation by WT1 via the SF-1-binding element from the MIS promotor occurs without direct interaction between WT1 and DNA (12). Similarly, the recruitment of TReP-132 to the human CYP11A1 gene is believed to happen solely through a protein-protein interaction with SF-1 (47). Furthermore, the homeobox transcription factor Ptx1 synergizes with SF-1 to promote expression of the proopiomelanocortin gene without interacting with DNA (50). In contrast, the promoters described to be regulated by SF-1 and Sp1 or Egr1 contain adjacent DNA-binding sites that are contacted by the involved factors (44, 46, 51). The presence of C2H2 zinc fingers in Zip67 suggests that it is able to act as a transcription factor through direct interaction with DNA. It is not unusual that transcription factors can be recruited to promoters by both protein-protein and protein-DNA interactions (12, 52–54), and it remains to be determined whether Zip67 recognizes DNA sequences through its zinc finger domains. The AF-2 domain is highly conserved between the NRs and comprises an amphipathic ␣-helix with the consensus motif ␾␾XE␾␾ (␾ indicates a hydrophobic amino acid, and X is any amino acid). Coactivators of NRs typically contain NR boxes that function as interaction domains. For instance, the p160 family interacts

Fig. 6. Zip67 Represses SF-1-Dependent Transcriptional Activation

A–C, Expression plasmids encoding SF-1 (pCMV-SF-1; 100 ng) or Zip67 (pCMV-BK-Zip67; 10–500 ng as indicated) were cotransfected with the reporter plasmids 4CRS2-luc (1.5 ␮g; A), CYP17–290-luc (1.5 ␮g; B), or CYP11A-186-luc (1.5 ␮g; C) into Y1 cells grown on 35-mm dishes by the calcium phosphate precipitation technique. After 24 h, the cells were harvested, and luciferase activities were determined. In all experiments a plasmid encoding ␤-galactosidase was cotransfected to control for transfection efficiencies. The figure shows the mean (⫾SD) luciferase activity (calculated relative to the ␤-galactosidase activity in each corresponding extract) from at least three independent experiments. D, Expression plasmids encoding SF-1 (pCMVSF-1; 100 ng) or Zip67 (pCMV-BK-Zip67; 100-1000 ng as indicated) were cotransfected with the reporter plasmid 4CRS2-luc (1.5 ␮g) into COS-1 cells. After 48 h, the cells were harvested, and luciferase activities were determined. The figure shows the mean (⫾SD) luciferase activity from at least three independent experiments.



with the AF-2 domain in a ligand-dependent manner through three leucine-rich motifs with the consensus sequence LXXLL in the NR box. Crystal structures revealed that the LXXLL motif forms a short ␣-helix that docks to a hydrophobic cleft on the surface of the LBD that consists of the AF-2 helix on one side and the helix 3 on the other side (55–57). The NR corepressors N-CoR and the silencing mediator of retinoic acid and thyroid hormone receptor (SMRT) contain sequences similar to the NR box, referred to as CoRNR boxes [i.e. LXX(I/H)IXXX(I/L) or (L/I)XX(I/V)I] (58, 59). Similarly to what is observed for the interaction between SF-1 and GRIP1, SRC-1, p/CIP, and TReP-132 (24–26, 47), both in vitro and in vivo experiments clearly demonstrate an absolute requirement for the AF-2 domain of SF-1 for the interaction with Zip67 (Fig. 2). Despite effort, we have not been able to map the domain(s) of Zip67 that mediates the contact with SF-1. The peptide sequence of Zip67 indicates the presence of three ␣helixes that resemble CoRNR boxes interspaced between the five C2H2 zinc fingers in the C-terminal part of the protein. However, mutational analysis suggests that these putative ␣-helixes are not of importance for interaction with SF-1, and it appears, therefore, that Zip67 interacts with the AF-2 domain in a different fashion than most coregulators. Interestingly, the coregulators PNRC and PNRC2 (proline-rich NR coregulator protein) interact with SF-1 and other NRs in an AF-2-dependent manner, but the interaction does not depend on LXXLL motifs, but, instead, on an Src ho-

Fig. 7. Zip67 Represses the Transcriptional Activation of NRs

A, Expression vectors encoding PPAR␥ (pcDNA3.1hPPAR␥; 150 ng) and Zip67 (pBK-CMV-Zip67; 50–1000 ng) were transfected into COS-1 cells as indicated in the figure together with the reporter construct 3PPRE-luc (0.9 ␮g). Twenty-four hours after transfection the cells were treated with 1 ␮M BRL49653. B, The expression vectors encoding LXR␣ (pCMX-LXR␣; 150 ng), RXR␣ (pCMV-RXR␣; 150 ng), and Zip67 (pBK-CMV-Zip67; 50–1000 ng) were transfected into COS-1 cells as indicated in the figure together with the reporter construct LXR␣-luc (1.0 ␮g). Twenty-four hours after transfection, the cells were treated with 20 ␮M 22(R)-hydroxycholesterol. C, Expression plasmids encoding NGFI-B (pCMX-NGFI-B; 100 ng) and Zip67 (pBK-CMV-Zip67; 1001000 ng) were cotransfected with the reporter plasmid NBRE-luc into COS-1 cells. D, Expression vectors encoding GR (pCMV5-GR; 100 ng) and Zip67 (pBK-CMV-Zip67; 50– 1000 ng) were transfected into COS-1 cells as indicated in the figure together with the reporter construct mouse mammary tumor virus-luciferase (1.0 ␮g). Twenty-four hours after transfection, the cells were treated with 0.1 ␮M dexamethasone. E, Expression vectors encoding ER (pSG5-ER␣; 100 ng) and Zip67 (pBK-CMV-Zip67; 10–1000 ng) were transfected into COS-1 cells as indicated in the figure together with the reporter construct estrogen response element-luciferase (1.0 ␮g). Twenty-four hours after transfection, the cells were treated with 0.1 ␮M 17␤-estradiol. Luciferase assays were performed 48 h after transfection. For all experiments the figure shows the mean ⫾ SD of triplicate transfections from representative experiments.



2313

Fig. 9. Zip67 Exhibits a Weak Intrinsic Activation Function Expression plasmids encoding GAL4 (pBIND or pM; 1.5 ␮g), GAL4-GRIP1 (pM. GRIP1; 1.5 ␮g), or GAL4-Zip67 (pBIND-Zip67; 1.5 ␮g) were cotransfected with the reporter plasmid pSGluc (1.5 ␮g) into Y1 cells. After 24 h, the cells were harvested, and the luciferase activities were determined. The figure shows the mean (⫾SD) luciferase activity from at least three independent experiments.

Fig. 8. Zip67 Does Not Affect Other Classes of Transcription Factors in the Same Fashion as Nuclear Receptors A, Expression vectors encoding CREB (Rous sarcoma virus-CREB; 100 ng), the catalytic subunit C␣ of PKA (pCMV5-C␣; 100 ng), and Zip67 (pBK-CMV-Zip67; 100–500 ng) were transfected into COS-1 cells as indicated in the figure together with the reporter construct 4CRE-luc (1.5 ␮g). Luciferase assays were performed 48 h after transfection. The figure shows the mean ⫾ SD of duplicate transfections from a representative experiment. B, The expression vector encoding Zip67 (pBK-CMV-Zip67; 100–500 ng) were transfected into HeLa cells as indicated in the figure together with the reporter construct 2XRE-luc (0.5 ␮g). Twenty-four hours after transfection the cells were treated with 10 nM 2,3,7,8tetrachlorodibenzo-p-dioxin. Luciferase assays were performed 48 h after transfection. The figure shows the mean ⫾ SD of triplicate transfections from a representative experiment.

mology domain 3 binding motif located within a proline-rich sequence (60, 61). However, Zip67 does not contain such a motif, and we are currently investigating the mechanism by which Zip67 interacts with the AF-2 domain. The results presented in Fig. 9 indicate that Zip67 contains an intrinsic activation function when linked to GAL4. However, these results further show that in comparison with the p160 coactivators (here exemplified by GRIP1), Zip67 appears to be a very weak

activator. Although it is possible that Zip67 acts as an activator under certain physiological conditions, our transient transfection experiments clearly point to a repressive function of Zip67 in a cellular context with biologically relevant amounts of p160 coactivators (Figs. 6 and 7). These results together with the findings that Zip67 counteracts the positive effect of GRIP1 in transient transfection experiments and that Zip67 and GRIP1 can displace each other from SF-1 (Fig. 10) obviously indicate a mechanism of action in which Zip67 represses SF-1-dependent transcription by hindering the binding of more potent coactivators. The zinc finger protein mZac1b appears to have similar properties, as it also exhibits an activation function when linked to GAL4, whereas it will either activate or repress NR function depending on the cellular context (52). Furthermore, RIP140, which has been demonstrated to interact with several NRs, including SF-1, seems to work in an analogous fashion (27, 62–67). RIP140 was also shown to act as a coactivator under certain experimental conditions (68), but the consensus is that this protein functions as a repressor by influencing the binding of endogenous coactivators as well as through recruitment of histone deacetylases (69). In transient transfection assays, RIP140 can antagonize SRC-1-mediated activation through competition with the coactivator for binding to the AF-2 domain (70). RIP140 also counteracts the stimulatory action of p160 coactivators on SF-1-dependet transcription, possibly by the same competition mechanism for binding to the AF-2 domain of SF-1 (27). Real-time PCR and dot-blot experiments (Figs. 4 and 5) demonstrated that Zip67 exhibits a broad ex-



Fig. 10. Zip67 and GRIP1 Counteracts Each Other’s Effects on SF-1-Dependent Transcription A, Expression plasmids encoding SF-1 (pCMV-SF-1; 100 ng), Zip67 (pCMV-BK-Zip67; 10–1000 ng), and GRIP1 (pSG5-HAGRIP1; 10–1000 ng) were cotransfected with the reporter plasmid 4CRS2-luc (1.5 ␮g) into COS-1 cells. After 48 h, the cells were harvested, and luciferase activities were determined. The figure shows the mean (⫾SD) luciferase activity from representative experiments. B, [35S]Methionine-labeled in vitro translated GRIP1 (3 ␮l) and increasing amounts of Zip67 (1, 3, or 10 ␮l) was incubated with a (His)6-SF-1 fusion protein (5 ␮g). The protein mixtures or [35S]methionine-labeled GRIP1 (3 ␮l) alone were then incubated with Ni-NTA agarose. After extensive washing, the proteins were analyzed by SDS-PAGE (6%). The experiment was repeated five times with equal results.

pression pattern, suggesting that it affects many biological systems. In line with this, we found that Zip67 repressed transcription mediated by an array of NRs. The expression data also demonstrated that Zip67 is present in the majority of SF-1-expressing tissues in both mice and humans (i.e. testis, adrenal, and spleen). Although this suggests a physiological connection between these factors, it will be important to determine the cellular localization of Zip67, because these tissues also consist of cells that do not express SF-1. In the adrenal, only the cortical part expresses SF-1 (71). We found the expression of Zip67 in Y1 cells, which are derived from the adrenal cortex, suggesting a functional correlation between these two proteins in this tissue. In testis and spleen, SF-1expressing cells are minor constituents (i.e. the Sertoli and Leydig cells of the testis and the red pulp of the spleen) (16, 72), and the relatively high levels of Zip67

observed in these tissues could possibly indicate that it is expressed in other cell types as well. Experiments are underway to determine the cellular localization of Zip67 in SF-1-expressing tissues as well as in other organs. Numerous studies have established the presence of steroidogenic enzymes in the brain and de novo steroidogenesis from cholesterol. Neurosteroids are believed to exert both genomic and nongenomic effects and to influence multiple brain functions (reviewed in Ref. 73). The expression data presented in Fig. 5 indicate relatively high levels of Zip67 in certain parts of the human brain (i.e. cerebellum, temporal lobe, and hippocampus), and intriguingly, these brain areas have been demonstrated to express several genes encoding steroidogenic enzymes (73). It is especially interesting to note that among the genes expressed are CYP11A1 and CYP17, which both contain SF-1 bind-


ing promoter elements that were affected by Zip67 in transient transfection experiments (Fig. 6). Importantly, Northern blot analyses of the human central nervous system have also revealed low levels of SF-1 expression at these sites, among other areas (72). These expression data might therefore indicate a role for Zip67 in the biosynthesis of neurosteroids. NRs are known to regulate a number of important functions in the brain, and interestingly, all of the NRs that we found to be modulated by Zip67 in transfection experiments are found at Zip67-expressing sites in the brain [i.e. ER (74), GR (75), NGFI-B (76), PPAR (77), and LXR (78, 79)]. Clearly, therefore, Zip67 might exert multiple functional roles in the brain through NRs other than SF-1. We did not examine the expression of Zip67 in mouse brain, and according to existing literature, SF-1 appears to be limited to the VMH in the mouse brain (71, 80). A recent study demonstrates a physiological role for SF-1 in energy homeostasis (81), but definitive targets genes for SF-1 in the VMH have yet to be identified. It will be of great interest to determine whether Zip67 is expressed and affects the function of SF-1 in the VMH. In conclusion, we have cloned and characterized a novel coregulator that represses transcription mediated by several NRs. Although we have found that Zip67 acts as a repressor on the investigated NR target promoters, further experiments are necessary to understand the full biological significance of Zip67. Interestingly, the interaction between Zip67 and the AF-2 domain might be of a novel type, as it does not appear to require NR or CoRNR boxes. We propose that the repressive properties of Zip67 are due to inhibition of coactivator recruitment. Possibly important, Zip67 possesses an intrinsic activation function when linked to GAL4, and this may imply that Zip67 can act as a transcriptional activator of still unknown promoters or in cellular contexts depleted of p160 coactivators.

MATERIALS AND METHODS Plasmid Constructs The expression plasmid pCMV5-SF-1 and the luciferase reporter plasmid pT81–4CRS2-luc, containing the SF-1-binding site from the proximal promoter region of the bovine CYP17 gene, were described previously (7). The bovine CYP17⫺290/⫹17 promoter fragment is inserted into the reporter pGL3-luc (pGL3-CYP17⫺290-luc) and was a gift from Dr. M. Waterman (Nashville, TN). The CYP11A reporter plasmid (pGL3-CYP11A⫺186-luc) contains the region between ⫺186 and ⫹12 of the bovine CYP11A gene (49). pHybLex/ Zeo-SF-1107–461 contains amino acids 107–461 of the bovine SF-1 cDNA linked to the LexA DNA-binding domain in the vector pHybLex/Zeo (26). pHybLex/Zeo-SF-1⌬AF-2 contains a 25-amino acid carboxyl-terminal deletion mutation (26). pHybLex/Zeo-Lamin (Invitrogen, Carlsbad, CA) was used as a negative control in the interaction studies. The pYESTrpZip67 plasmid was isolated from the human testis cDNA library in the yeast two-hybrid screen, and this fragment contains 763 bp (bp 1123–1871 of the final Zip67 cDNA)


2315

inserted into the BstXI site of the pYESTrp2 vector. The expression plasmid pBK-CMV-Zip67 contains the full-length human Zip67 cDNA inserted into the ␭ZAP Express EcoRI/ XhoI vector pBK-CMV (Stratagene, La Jolla, CA). pSG5-HAGRIP1 and pM.GRIP1 (GAL4-GRIP1 fusion) were provided by Dr. M. Stallcup (Los Angeles, CA). pSGluc (contains GAL4binding elements) and pBIND were obtained from Promega (Madison, WI). Zip67 was cloned in-frame with the GAL4 DNA-binding domain in the pBIND vector (pBIND-Zip67). The PPREx3-tk-luc (3PPRE-luc) reporter construct containing three copies of the peroxisome proliferate response element from the acyl-coenzyme A promoter and the expression plasmid pcDNA3.1-hPPAR␥1 were provided by Dr. K. Kristiansen (Odense, Denmark) (82). The pGL3basic-LXR␣ (⫺1515/ ⫹1822)-luc (LXR␣-luc), mouse mammary tumor virus-luc, the estrogen response element-TATA-luc reporter constructs, and the expression plasmids pCMV-retinoid X receptor ␣ (RXR␣), pCMX-LXR␣, pCMV5-GR, and pSG5-ER␣ were provided by Dr. H. I. Nebb (Oslo, Norway) (83). The NBRE-luc reporter construct contains three copies of the NGFI-B response element. The NGFI-B expression plasmid (pCMXNGFI-B) was provided by Dr. T. Perlmann (Stockholm, Sweden). The pCMV5-C␣ plasmid expressing the catalytic subunit of PKA was a gift from Dr. G. S. McKnight (Seattle, WA). The luciferase reporter plasmid pT81–4CRE-luc were constructed as previously described for 4CRS2-luc (7). The reporter construct carrying the xenobiotic response elements (pT81–2XRE-luc) was described previously (84). Cloning of Zip67 cDNA The yeast two-hybrid screen was performed according to the users manual for the LexA-based system Hybrid Hunter (Invitrogen) using a cDNA library from human testis cloned into the BstXI site of the pYESTrp2 vector (Invitrogen). The bait plasmid (pHybLex/Zeo-SF-1107–461) and the cDNA library were sequentially transformed into the yeast strain EGY191 (MAT␣ ura3 trp1 his3::2lexAop-LEU2; gift from Dr. E. Golemis, Philadelphia, PA) and plated on medium containing zeocin (200 mg/ml) and lacking the selection markers leucine and tryptophan. The yeast strains that were provided with the Hybrid Hunter kit [EGY48 (MAT␣ ura3 trp1 his3(6lexAopLEU2) GAL4) and L40 (MATa his3⌬200 trp1–901 leu2–3112 ade2 LYS2::(4lexAop-HIS3) URA3:: (8lexAop-lacZ) GAL4)] contain six or four copies of the LexA operon in front of the selection markers. Due to the intrinsic transcriptional activity of the C-terminal part of SF-1 we observed relatively high background activity in these strains. Instead, we therefore used the EGY191 strain, which contains only two copies of the LexA operon, and the background activity was reduced to a minimum. Surviving clones were replated several times, and clones that remained positive were grown in liquid medium to retrieve the cDNAs that interacted with the SF-1 bait plasmid. Yeast plasmid preparations were performed according to standard procedures. Full-length Zip67 cDNA was isolated through conventional hybridization to a human testis cDNA library using the ␭ZAP Express system (Stratagene). However, comparison of the isolated Zip67 cDNA with the predicted cDNA from the NIX database revealed that the most 5⬘ 27 nucleotides were not included in the fragment from the ␭ZAP library. Based on the predicted cDNA, an extended oligonucleotide and an internal primer were used to generate the remaining cDNA sequence using Advantage 2 polymerase (Clontech). PCR was performed in the presence of 1% dimethylsulfoxide to reduce secondary structures. The amplified fragment was cloned in-frame with the rest of the cDNA in the pBK-CMV vector. The primers used were: forward, 5⬘-GGG-GGG-ATG-GCG-GAG-CGG-GCG-CTA-GAGCCC-GAG-GCG-GAG-GCG-GAG-GCT-GAG-GCG-GGC-GCG-GGC-GGG-GAG-GCA-GCA-GCC-GAG-GAG-GGC-GCA-GCG-GGC-CGA-AAG-GCG-CGG-GGC-CGG-CC-⬘3; and reverse, 5⬘-CCG-CCG-CAG-GTG-GTT-GGA-TAA-ATA-G-⬘3. Three recent submissions to the sequence database have


confirmed the predicted cDNA sequence upon which we based our approach. Based on the human Zip67 cDNA we identified sequences in the mouse expressed sequence tag database that corresponded to the 5⬘and 3⬘ ends of the mouse Zip67 cDNA. The full-length mouse cDNA was subsequently amplified from a testis cDNA library (Clontech) using the following primers: forward, 5⬘-AAC-TTG-CCGCCT-ACC-TCA-TT-⬘3; and reverse, 5⬘-CCA-CTG-CCC-ACTTTG-TCA-CT-⬘3. In Vivo Protein Interaction Assay A two-hybrid mating assay was used to measure the interaction between the LexA fusion bait proteins SF-1107–461, SF-1107–461⌬AF-2 and lamin (negative control) and the B42 fusion prey protein Zip67. The LexA fusion proteins were introduced into the EGY191 strain and selected on plates containing zeocin (200 mg/ml), whereas the B42-fusion proteins were transformed into the L40 strain and selected on plates lacking tryptophan. Selection for the presence of both bait and prey plasmids (mating colonies) was carried out on plates with zeocin, lacking tryptophan. To examine the ␤galactosidase activities, strain EGY191 was also cotransformed with the 2lexAop-lacZ reporter plasmid pJK103. (The strains EGY191 and pJK103 were provided by Dr. E. Golemis, Philadelphia, PA.) The chemiluminescent reporter gene assay system from Galacto-Star (Tropix, Bedford MA) was used to detect ␤-galactosidase activity in yeast cell extracts. In Vitro Protein Interaction Assay The Bac-to-Bac Baculovirus Expression System (Invitrogen) was used to express the (His)6-containing wild-type SF-1 fusion protein and the (His)6-containing SF-1⌬AF-2 with an AF-2 core (LLIEML) deletion as previously described (26). The expressed (His)6-SF-1 and (His)6-SF-1⌬AF-2 proteins were purified using affinity chromatography with Ni-NTA agarose (Qiagen, Chatsworth, CA). [35S]Methionine-labeled Zip67 was prepared using the TNT-coupled reticulocyte lysate system (Promega) in the presence of [35S]methionine. For protein interaction assays, the [35S]methionine-labeled in vitro translated Zip67 (3 ␮l) was incubated with purified (His)6-SF-1 or (His)6-SF-1⌬AF-2 (5 ␮g) at 30 C for 1 h with occasional gentle mixing. The protein mixtures or the [35S]methionine-labeled coactivators (3 ␮l) alone were then incubated with Ni-NTA agarose (30 ␮l) in 20 mM Tris-HCl (pH 7.5), 50 mM KCl, 10% glycerol, 20 mM imidazole, 0.5% Nonidet P-40, 1 mM phenylmethylsulfonylfluoride, and 5 ␮g/ml aprotinin to a total volume of 300 ␮l on rotating wheel at room temperature for 1 h. After incubation, the Ni-NTA agarose beads were washed with 500 ␮l of the above buffer. The Ni-NTA agarose beads were subsequently boiled for 5 min in the presence of 15 ␮l sodium dodecyl sulfate loading buffer [100 mM Tris-HCl (pH 6.8), 100 mM dithiothreitol, 4% sodium dodecyl sulfate, 0.2% bromophenol blue, and 20% glycerol] and subjected to 6% SDS-PAGE, followed by autoradiography. The same method was used in the competition assay with [35S]methionine-labeled, in vitro-translated GRIP1 (3 ␮l) and increasing amounts of [35S]methionine-labeled, in vitro-translated Zip67 (1, 3, and 10 ␮l). Quantitative Real-Time RT-PCR Mouse tissues were homogenized, and mRNA from the different tissues was isolated using MagNA Pure LC mRNA Isolation Kit II and MagNA Pure LC according to the manufacturer’s protocol (Roche, Indianapolis, IN). Quantitative real-time RT-PCR was carried out using a LightCycler rapid thermal cycler system (Roche). Forward and reverse PCR primers designed from the mouse Zip67 sequence were as follows: forward, 5⬘-CCA-CTG-CCC-ACT-TTG-TCA-CT-3⬘;


and reverse, 5⬘-AAC-TTG-CCG-CCT-ACC-TCA-TT-3⬘. The predicted size of the Zip67 PCR product was 300 bp. Forward and reverse GAPDH primers were: forward, 5⬘-ACCACA-GTC-CAT-GCC-ATC-AC-3⬘; and reverse, 5⬘-TCCACC-ACC-CTG-TTG-CTG-TA-3⬘. The predicted size of the GAPDH PCR product was 480 bp. Reactions were performed using the LightCycler RNA Master SYBR-Green I Kit (Roche), and the PCR products were detected via intercalation of the fluorescent dye SYBR-Green. Zip67 and GAPDH standards were prepared by 10-fold serial dilutions of linearized pCR2.1-mZip67 plasmid and DNA fragments containing the full-length GAPDH cDNA sequence, respectively. Standards were used over the range of 100 pg/␮l to 0.01 pg/␮l. The negative controls were prepared by replacing the mRNA template with PCR-grade H2O. The protocols included a 20-min RT step at 61 C, a 5-sec denaturation step, and then 45 cycles consisting of denaturation at 95 C for 5 sec, annealing at 60 C (for both Zip67 and GAPDH) for 5 sec, and an extension phase at 72 C for 18 sec (Zip67) or 20 sec (GAPDH). Fluorescence was measured at the end of the 72 C extension phase. The quality of the RT-PCR products was controlled by melting point curve analysis. In addition, the amplification products were subjected to agarose gel electrophoresis (1.6%) and stained with ethidium bromide (0.5 ␮g/ml) to ensure specificity of amplification. Human Multiple Tissue Array A human multiple tissue expression array (Clontech) was probed with 32P-labeled human Zip67 cDNA according to the manufacturer’s protocol. The hybridization signals were quantified using a phosphorimager analyzer (Bas-5000, Fuji Photo Film Co., Ltd., Tokyo, Japan). Cell Culture and Transfections Y1 mouse adrenocortical tumor cells and HeLa cells were cultured in DMEM supplemented with 10% fetal calf serum, 100 U penicillin/ml, and 100 ␮g streptomycin/ml. For transfection experiments, Y1 cells were plated at a density of 2.5 ⫻ 105 cells/well onto six-well plates and transiently transfected the following day by using the calcium phosphate method according to standard procedures. Y1 cells were transfected with reporter plasmid (pT81–4CRS2-luc, pGL3-CYP17⫺290luc, or pGL-CYP11A⫺186-luc; 1.5 ␮g), pCMV-SF-1 (100 ng), and pBK-CMV-Zip67 (100–500 ng) as indicated in the figures. COS-1 and HeLa cells were seeded as described for Y1 cells, and transiently transfected the following day by the SuperFect transfection procedure according to the manufacturer’s protocol (Qiagen). In the control experiments, plasmids encoding other NRs, such as pcDNA3.1-hPPAR␥1 (0.15 ␮g), pCMV-RXR␣ (0.15 ␮g), pCMX-LXR␣ (0.15 ␮g), pCMXNGFI-B (0.1 ␮g), pCMV5-GR (0.15 ␮g), and pSG5-ER (0.15 ␮g), were transfected into COS-1 cells together with their respective reporters and in the presence of respective ligands as indicated in the figure. The total amount of plasmid was kept constant by compensating with empty expression vector (pCMV5). To control for transfection efficiency, an expression plasmid encoding ␤-galactosidase (pCMV-lacZ) was included in the transfection experiments. The cells were washed once with PBS 24 h (Y1) or 48 h (COS-1) after transfection, and assayed for luciferase activity. Forty microliters of the cell extracts were used for luciferase determination on a LUCY-1 luminometer (Anthos, Wals, Austria). The luciferase assay was performed in accordance with the protocol of the Luciferase Assay Kit (BIO Thema AB, Dalaro¨ , Sweden). All transfections were performed in triplicate and repeated three to five times.


Acknowledgments We thank Dr. E. Golemis for yeast strains, and Drs. K. Kristiansen, S. McKnight, H. I. Nebb, T. Perlmann, L. Poellinger, M. R. Stallcup, C. Totland, and M. R. Waterman for plasmids. Dr. Reidun Æsøy is acknowledged for assistance with the transfection experiments. We also thank T. Ellingsen, T. Hoang, A. M. Sellevold, and C. Cook for excellent technical assistance.

Received April 25, 2003. Accepted August 5, 2003. Address all correspondence and requests for reprints to: Dr. Bente Børud, Department of Anatomy and Cell Biology, Jonas Lies vei 91, N-5009 Bergen, Norway. E-mail: [email protected]. This work was supported by the Norwegian Cancer Society and the Norwegian Research Council.

REFERENCES 1. Parker KL, Schimmer BP 1997 Steroidogenic factor 1: a key determinant of endocrine development and function. Endocr Rev 18:361–377 2. Sadovsky Y, Crawford PA, Woodson KG, Polish JA, Clements MA, Tourtellotte LM, Simburger K, Milbrandt J 1995 Mice deficient in the orphan receptor steroidogenic factor 1 lack adrenal glands and gonads but express P450 side-chain-cleavage enzyme in the placenta and have normal embryonic serum levels of corticosteroids. Proc Natl Acad Sci USA 92:10939–10943 3. Luo X, Ikeda Y, Parker KL 1994 A cell-specific nuclear receptor is essential for adrenal and gonadal development and sexual differentiation. Cell 77:481–490 4. Achermann JC, Ito M, Hindmarsh PC, Jameson JL 1999 A mutation in the gene encoding steroidogenic factor-1 causes XY sex reversal and adrenal failure in humans. Nat Genet 22:125–126 5. Biason-Lauber A, Schoenle EJ 2000 Apparently normal ovarian differentiation in a prepubertal girl with transcriptionally inactive steroidogenic factor 1 (NR5A1/SF-1) and adrenocortical insufficiency. Am J Hum Genet 67: 1563–1568 6. Achermann JC, Ozisik G, Ito M, Orun UA, Harmanci K, Gurakan B, Jameson JL 2002 Gonadal determination and adrenal development are regulated by the orphan nuclear receptor steroidogenic factor-1, in a dosedependent manner. J Clin Endocrinol Metab 87: 1829–1833 7. Bakke M, Lund J 1995 Mutually exclusive interactions of two nuclear orphan receptors determine activity of a cyclic adenosine 3⬘,5⬘-monophosphate-responsive sequence in the bovine CYP17 gene. Mol Endocrinol 9:327–339 8. Lala DS, Rice DA, Parker KL 1992 Steroidogenic factor I, a key regulator of steroidogenic enzyme expression, is the mouse homolog of fushi tarazu-factor I. Mol Endocrinol 6:1249–1258 9. Honda S, Morohashi K, Nomura M, Takeya H, Kitajima M, Omura T 1993 Ad4BP regulating steroidogenic P-450 gene is a member of steroid hormone receptor superfamily. J Biol Chem 268:7494–7502 10. Naville D, Penhoat A, Marchal R, Durand P, Begeot M 1998 SF-1 and the transcriptional regulation of the human ACTH receptor gene. Endocr Res 24:391–395 11. Caron KM, Ikeda Y, Soo SC, Stocco DM, Parker KL, Clark BJ 1997 Characterization of the promoter region of the mouse gene encoding the steroidogenic acute regulatory protein. Mol Endocrinol 11:138–147


2317

12. Nachtigal MW, Hirokawa Y, Enyeart-VanHouten DL, Flanagan JN, Hammer GD, Ingraham HA 1998 Wilms’ tumor 1 and Dax-1 modulate the orphan nuclear receptor SF-1 in sex-specific gene expression. Cell 93:445–454 13. Arango NA, Lovell-Badge R, Behringer RR 1999 Targeted mutagenesis of the endogenous mouse Mis gene promoter: in vivo definition of genetic pathways of vertebrate sexual development. Cell 99:409–419 14. Ingraham HA, Lala DS, Ikeda Y, Luo X, Shen WH, Nachtigal MW, Abbud R, Nilson JH, Parker KL 1994 The nuclear receptor steroidogenic factor 1 acts at multiple levels of the reproductive axis. Genes Dev 8:2302–2312 15. Ikeda Y, Luo X, Abbud R, Nilson JH, Parker KL 1995 The nuclear receptor steroidogenic factor 1 is essential for the formation of the ventromedial hypothalamic nucleus. Mol Endocrinol 9:478–486 16. Morohashi K, Tsuboi-Asai H, Matsushita S, Suda M, Nakashima M, Sasano H, Hataba Y, Li CL, Fukata J, Irie J, Watanabe T, Nagura H, Li E 1999 Structural and functional abnormalities in the spleen of an mFtz-F1 genedisrupted mouse. Blood 93:1586–1594 17. Zhao L, Bakke M, Krimkevich Y, Cushman LJ, Parlow AF, Camper SA, Parker KL 2001 Steroidogenic factor 1 (SF1) is essential for pituitary gonadotrope function. Development 128:147–154 18. Heery DM, Kalkhoven E, Hoare S, Parker MG 1997 A signature motif in transcriptional co-activators mediates binding to nuclear receptors. Nature 387:733–736 19. Hong H, Kohli K, Garabedian MJ, Stallcup MR 1997 GRIP1, a transcriptional coactivator for the AF-2 transactivation domain of steroid, thyroid, retinoid, and vitamin D receptors. Mol Cell Biol 17:2735–2744 20. Voegel JJ, Heine MJ, Zechel C, Chambon P, Gronemeyer H 1996 TIF2, a 160 kDa transcriptional mediator for the ligand-dependent activation function AF-2 of nuclear receptors. EMBO J 15:3667–3675 21. Glass CK, Rosenfeld MG2000 The coregulator exchange in transcriptional functions of nuclear receptors. Genes Dev 14:121–141 22. Lala DS, Syka PM, Lazarchik SB, Mangelsdorf DJ, Parker KL, Heyman RA 1997 Activation of the orphan nuclear receptor steroidogenic factor 1 by oxysterols. Proc Natl Acad Sci USA 94:4895–4900 23. Mangelsdorf DJ, Evans RM 1995 The RXR heterodimers and orphan receptors. Cell 83:841–850 24. Ito M, Yu RN, Jameson JL 1998 Steroidogenic factor-1 contains a carboxy-terminal transcriptional activation domain that interacts with steroid receptor coactivator-1. Mol Endocrinol 12:290–301 25. Crawford PA, Polish JA, Ganpule G, Sadovsky Y 1997 The activation function-2 hexamer of steroidogenic factor-1 is required, but not sufficient for potentiation by SRC-1. Mol Endocrinol 11:1626–1635 26. Børud B, Hoang T, Bakke M, Jacob AL, Lund J, Mellgren G 2002 The nuclear receptor coactivators p300/CBP/ cointegrator-associated protein (p/CIP) and transcription intermediary factor 2 (TIF2) differentially regulate PKAstimulated transcriptional activity of steroidogenic factor 1. Mol Endocrinol 16:757–773 27. Mellgren G, Børud B, Hoang T, Yri OE, Fladeby C, Lien EA, Lund J 2003 Characterization of receptor-interacting protein RIP140 in the regulation of SF-1 responsive target genes. Mol Cell Endocrinol 203(1-2):91–103 28. Crawford PA, Dorn C, Sadovsky Y, Milbrandt J 1998 Nuclear receptor DAX-1 recruits nuclear receptor corepressor N-CoR to steroidogenic factor 1. Mol Cell Biol 18:2949–2956 29. Hammer GD, Krylova I, Zhang Y, Darimont BD, Simpson K, Weigel NL, Ingraham HA 1999 Phosphorylation of the nuclear receptor SF-1 modulates cofactor recruitment: integration of hormone signaling in reproduction and stress. Mol Cell 3:521–526


30. Ou Q, Mouillet JF, Yan X, Dorn C, Crawford PA, Sadovsky Y 2001 The DEAD box protein DP103 is a regulator of steroidogenic factor-1. Mol Endocrinol 15:69–79 31. Desclozeaux M, Krylova IN, Horn F, Fletterick RJ, Ingraham HA 2002 Phosphorylation and intramolecular stabilization of the ligand binding domain in the nuclear receptor steroidogenic factor 1. Mol Cell Biol 22: 7193–7203 32. Lund J, Bakke M, Mellgren G, Morohashi K, Doskeland SO 1997 Transcriptional regulation of the bovine CYP17 gene by cAMP. Steroids 62:43–45 33. Æsøy R, Mellgren G, Morohashi K, Lund J 2002 Activation of cAMP-dependent protein kinase increases the protein level of steroidogenic factor-1. Endocrinology 143:295–303 34. Jacob AL, Lund J 1998 Mutations in the activation function-2 core domain of steroidogenic factor-1 dominantly suppresses PKA-dependent transactivation of the bovine CYP17 gene. J Biol Chem 273:13391–13394 35. Reinhart AJ, Williams SC, Clark BJ, Stocco DM 1999 SF-1 (steroidogenic factor-1) and C/EBP␤ (CCAAT/enhancer binding protein-␤) cooperate to regulate the murine StAR (steroidogenic acute regulatory) promoter. Mol Endocrinol 13:729–741 36. Lopez D, Sandhoff TW, McLean MP 1999 Steroidogenic factor-1 mediates cyclic 3⬘,5⬘-adenosine monophosphate regulation of the high density lipoprotein receptor. Endocrinology 140:3034–3044 37. Jacob AL, Lund J, Martinez P, Hedin L 2001 Acetylation of steroidogenic factor 1 protein regulates its transcriptional activity and recruits the coactivator gcn5. J Biol Chem 276:37659–7664 38. Williams GW, Woollard PM, Hingamp P 1998 NIX: a nucleotide identification system at the HGMP-RC. http:// www.hgmp.mrc.ac.uk/NIX/ 39. Breathnach R, Benoist C, O’Hare K, Gannon F, Chambon P 1978 Ovalbumin gene: evidence for a leader sequence in mRNA and DNA sequences at the exon-intron boundaries. Proc Natl Acad Sci USA 75:4853–4857 40. Wilkins MR, Gasteiger E, Bairoch A, Sanchez JC, Williams KL, Appel RD, Hochstrasser DF 1999 Protein identification and analysis tools in the ExPASy server. Methods Mol Biol 112:531–552 41. Bateman A, Birney E, Durbin R, Eddy SR, Howe KL, Sonnhammer EL 2000 The Pfam protein families database. Nucleic Acids Res 28:263–266 42. Nakai K, Kanehisa M 1992 A knowledge base for predicting protein localization sites in eukaryotic cells. Genomics 14:897–911 43. Hanley NA, Rainey WE, Wilson DI, Ball SG, Parker KL 2001 Expression profiles of SF-1, DAX1, and CYP17 in the human fetal adrenal gland: potential interactions in gene regulation. Mol Endocrinol 15:57–68 44. Liu Z, Simpson ER 1997 Steroidogenic factor 1 (SF-1) and SP1 are required for regulation of bovine CYP11A gene expression in bovine luteal cells and adrenal Y1 cells. Mol Endocrinol 11:127–137 45. Koh SS, Chen D, Lee YH, Stallcup MR 2001 Synergistic enhancement of nuclear receptor function by p160 coactivators and two coactivators with protein methyltransferase activities. J Biol Chem 276:1089–1098 46. Sugawara T, Saito M, Fujimoto S 2000 Sp1 and SF-1 interact and cooperate in the regulation of human steroidogenic acute regulatory protein gene expression. Endocrinology 141:2895–2903 47. Gizard F, Lavallee B, DeWitte F, Teissier E, Staels B, Hum DW 2002 The transcriptional regulating protein of 132 kDa (TReP-132) enhances P450scc gene transcription through interaction with steroidogenic factor-1 in human adrenal cells. J Biol Chem 277:39144–39155 48. Lee SL, Sadovsky Y, Swirnoff AH, Polish JA, Goda P, Gavrilina G, Milbrandt J 1996 Luteinizing hormone defi-


49. 50.

51.

52. 53. 54. 55.

56.

57.

58.

59. 60. 61.

62.

63. 64.

65.

66.

ciency and female infertility in mice lacking the transcription factor NGFI-A (Egr-1). Science 273:1219–1221 Ahlgren R, Suske G, Waterman MR, Lund J 1999 Role of Sp1 in cAMP-dependent transcriptional regulation of the bovine CYP11A gene. J Biol Chem 274:19422–19428 Tremblay JJ, Marcil A, Gauthier Y, Drouin J 1999 Ptx1 regulates SF-1 activity by an interaction that mimics the role of the ligand-binding domain. EMBO J 18: 3431–3441 Halvorson LM, Ito M, Jameson JL, Chin WW 1998 Steroidogenic factor-1 and early growth response protein 1 act through two composite DNA binding sites to regulate luteinizing hormone ␤-subunit gene expression. J Biol Chem 273:14712–14720 Huang SM, Stallcup MR 2000 Mouse Zac1, a transcriptional coactivator and repressor for nuclear receptors. Mol Cell Biol 20:1855–1867 Galvin KM, Shi Y 1997 Multiple mechanisms of transcriptional repression by YY1. Mol Cell Biol 17:3723–3732 Wilhelm D, Englert C 2002 The Wilms tumor suppressor WT1 regulates early gonad development by activation of Sf1. Genes Dev 16:1839–1851 Darimont BD, Wagner RL, Apriletti JW, Stallcup MR, Kushner PJ, Baxter JD, Fletterick RJ, Yamamoto KR 1998 Structure and specificity of nuclear receptor-coactivator interactions. Genes Dev 12:3343–3356 Nolte RT, Wisely GB, Westin S, Cobb JE, Lambert MH, Kurokawa R, Rosenfeld MG, Willson TM, Glass CK, Milburn MV 1998 Ligand binding and co-activator assembly of the peroxisome proliferator-activated receptor-␥. Nature 395:137–143 Shiau AK, Barstad D, Loria PM, Cheng L, Kushner PJ, Agard DA, Greene GL 1998 The structural basis of estrogen receptor/coactivator recognition and the antagonism of this interaction by tamoxifen. Cell 95:927–937 Perissi V, Staszewski LM, McInerney EM, Kurokawa R, Krones A, Rose DW, Lambert MH, Milburn MV, Glass CK, Rosenfeld MG 1999 Molecular determinants of nuclear receptor-corepressor interaction. Genes Dev 13: 3198–3208 Hu X, Lazar MA 1999 The CoRNR motif controls the recruitment of corepressors by nuclear hormone receptors. Nature 402:93–96 Zhou D, Chen S 2001 PNRC2 is a 16 kDa coactivator that interacts with nuclear receptors through an SH3-binding motif. Nucleic Acids Res 29:3939–3948 Zhou D, Quach KM, Yang C, Lee SY, Pohajdak B, Chen S 2000 PNRC: a proline-rich nuclear receptor coregulatory protein that modulates transcriptional activation of multiple nuclear receptors including orphan receptors SF1 (steroidogenic factor 1) and ERR␣1 (estrogen related receptor ␣-1). Mol Endocrinol 14:986–998 Ikonen T, Palvimo JJ, Janne OA 1997 Interaction between the amino- and carboxyl-terminal regions of the rat androgen receptor modulates transcriptional activity and is influenced by nuclear receptor coactivators. J Biol Chem 272:29821–29828 L’Horset F, Dauvois S, Heery DM, Cavailles V, Parker MG 1996 RIP-140 interacts with multiple nuclear receptors by means of two distinct sites. Mol Cell Biol 16:6029–6036 Cavailles V, Dauvois S, L’Horset F, Lopez G, Hoare S, Kushner PJ, Parker MG 1995 Nuclear factor RIP140 modulates transcriptional activation by the estrogen receptor. EMBO J 14:3741–3751 Masuyama H, Brownfield CM, St-Arnaud R, MacDonald PN 1997 Evidence for ligand-dependent intramolecular folding of the AF-2 domain in vitamin D receptor-activated transcription and coactivator interaction. Mol Endocrinol 11:1507–1517 Subramaniam N, Treuter E, Okret S 1999 Receptor interacting protein RIP140 inhibits both positive and negative gene regulation by glucocorticoids. J Biol Chem 274:18121–18127


67. Sugawara T, Abe S, Sakuragi N, Fujimoto Y, Nomura E, Fujieda K, Saito M, Fujimoto S 2001 RIP 140 modulates transcription of the steroidogenic acute regulatory protein gene through interactions with both SF-1 and DAX-1. Endocrinology 142:3570–3577 68. Joyeux A, Cavailles V, Balaguer P, Nicolas JC 1997 RIP 140 enhances nuclear receptor-dependent transcription in vivo in yeast. Mol Endocrinol 11:193–202 69. Wei LN, Hu X, Chandra D, Seto E, Farooqui M 2000 Receptor-interacting protein 140 directly recruits histone deacetylases for gene silencing. J Biol Chem 275: 40782–40787 70. Treuter E, Albrektsen T, Johansson L, Leers J, Gustafsson JA 1998 A regulatory role for RIP140 in nuclear receptor activation. Mol Endocrinol 12:864–881 71. Ikeda Y, Shen WH, Ingraham HA, Parker KL 1994 Developmental expression of mouse steroidogenic factor-1, an essential regulator of the steroid hydroxylases. Mol Endocrinol 8:654–662 72. Ramayya MS, Zhou J, Kino T, Segars JH, Bondy CA, Chrousos GP 1997 Steroidogenic factor 1 messenger ribonucleic acid expression in steroidogenic and nonsteroidogenic human tissues: Northern blot and in situ hybridization studies. J Clin Endocrinol Metab 82: 1799–1806 73. Stoffel-Wagner B 2001 Neurosteroid metabolism in the human brain. Eur J Endocrinol 145:669–679 74. Wehrenberg U, Prange-Kiel J, Rune GM 2001 Steroidogenic factor-1 expression in marmoset and rat hippocampus: co-localization with StAR and aromatase. J Neurochem 76:1879–1886 75. Meijer OC 2002 Coregulator proteins and corticosteroid action in the brain. J Neuroendocrinol 14:499–505 76. Zetterstrom RH, Solomin L, Mitsiadis T, Olson L, Perlmann T 1996 Retinoid X receptor heterodimerization and developmental expression distinguish the orphan nuclear receptors NGFI-B, Nurr1, and Nor1. Mol Endocrinol 10:1656–1666


2319

77. Kainu T, Wikstrom AC, Gustafsson JA, Pelto-Huikko M 1994 Localization of the peroxisome proliferator-activated receptor in the brain. Neuroreport 5:2481–2485 78. Kainu T, Kononen J, Enmark E, Gustafsson JA, PeltoHuikko M 1996 Localization and ontogeny of the orphan receptor OR-1 in the rat brain. J Mol Neurosci 7:29–39 79. Whitney KD, Watson MA, Collins JL, Benson WG, Stone TM, Numerick MJ, Tippin TK, Wilson JG, Winegar DA, Kliewer SA 2002 Regulation of cholesterol homeostasis by the liver X receptors in the central nervous system. Mol Endocrinol 16:1378–1385 80. Shinoda K, Lei H, Yoshii H, Nomura M, Nagano M, Shiba H, Sasaki H, Osawa Y, Ninomiya Y, Niwa O, et al. 1995 Developmental defects of the ventromedial hypothalamic nucleus and pituitary gonadotroph in the Ftz-F1 disrupted mice. Dev Dyn 204:22–29 81. Majdic G, Young M, Gomez-Sanchez E, Anderson P, Szczepaniak LS, Dobbins RL, McGarry JD, Parker KL 2002 Knockout mice lacking steroidogenic factor 1 are a novel genetic model of hypothalamic obesity. Endocrinology 143:607–614 82. Helledie T, Antonius M, Sorensen RV, Hertzel AV, Bernlohr DA, Kolvraa S, Kristiansen K, Mandrup S 2000 Lipidbinding proteins modulate ligand-dependent trans-activation by peroxisome proliferator-activated receptors and localize to the nucleus as well as the cytoplasm. J Lipid Res 41:1740–1751 83. Tobin KA, Steineger HH, Alberti S, Spydevold O, Auwerx J, Gustafsson JA, Nebb HI 2000 Cross-talk between fatty acid and cholesterol metabolism mediated by liver X receptor-␣. Mol Endocrinol 14:741–752 84. Berghard A, Gradin K, Pongratz I, Whitelaw M, Poellinger L 1993 Cross-coupling of signal transduction pathways: the dioxin receptor mediates induction of cytochrome P-450IA1 expression via a protein kinase C-dependent mechanism. Mol Cell Biol 13:677–689