Synergistic cooperation between the beta-catenin signaling pathway and steroidogenic factor 1 in the activation of the Mullerian inhibiting substance type II receptor.

Mullerian inhibiting substance type II receptor (MISRII) is a member of the transforming growth factor-beta superfamily. Mutations in mullerian inhibiting substance (MIS) or MISRII cause male sexual abnormalities, persistent mullerian duct syndrome, and pseudohermaphroditism. The spatial and temporal regulation of MIS and MISRII is important for its biological action. Male Wnt7a mutant mice do not undergo regression of mullerian ducts. Here we showed that the canonical Wnt signaling pathway regulated MISRII. The promoter MISRII was activated by beta-catenin expression, and this activation was dependent on TCF4-binding sites. The nuclear receptor superfamily member steroidogenic factor 1 (SF1) synergistically activated the MISRII promoter with beta-catenin. APC, a negative regulator of Wnt signaling, decreased SF1-mediated activation of the MISRII promoter in the colon carcinoma cell line SW480. We also showed a direct physical interaction between beta-catenin and SF1 by co-immunoprecipitation. Thus, our findings suggest that MISRII is a developmental target of Wnt7a signaling for mullerian duct regression during sexual differentiation.

Mullerian inhibiting substance type II receptor (MISRII) is a member of the transforming growth factor-␤ superfamily. Mutations in mullerian inhibiting substance (MIS) or MISRII cause male sexual abnormalities, persistent mullerian duct syndrome, and pseudohermaphroditism. The spatial and temporal regulation of MIS and MISRII is important for its biological action. Male Wnt7a mutant mice do not undergo regression of mullerian ducts. Here we showed that the canonical Wnt signaling pathway regulated MISRII. The promoter MISRII was activated by ␤-catenin expression, and this activation was dependent on TCF4-binding sites. The nuclear receptor superfamily member steroidogenic factor 1 (SF1) synergistically activated the MISRII promoter with ␤-catenin. APC, a negative regulator of Wnt signaling, decreased SF1-mediated activation of the MISRII promoter in the colon carcinoma cell line SW480. We also showed a direct physical interaction between ␤-catenin and SF1 by co-immunoprecipitation. Thus, our findings suggest that MISRII is a developmental target of Wnt7a signaling for mullerian duct regression during sexual differentiation.
One of the essential phases in testis development is differentiation of the wolffian ducts by androgen and regression of the mullerian ducts, which normally give rise to oviducts, uterus, and fallopian tubes in females (1). Sertoli cells secrete a transforming growth factor-␤ superfamily member, mullerian inhibiting substance (MIS), 1 also known as anti-mullerian hormone. Its type II receptor, MIS type II receptor (MISRII), also known as anti-mullerian hormone receptor, is expressed in the mesenchymal cells surrounding the mullerian ducts during the regression and in Sertoli cells in testes and granulosa cells in ovaries (2)(3)(4). The MIS type I receptor has not been identified; however, bone morphogenic protein 1a is a strong candidate to be this receptor (5). Interaction between MIS and its receptor initiates a signal that ultimately causes regression of the mullerian ducts through apoptosis by a paracrine mecha-nism (6). Mutations in MIS or its type II receptor cause the human male sexual abnormalities, persistent mullerian duct syndrome and pseudohermaphroditism (7), and deletion of MIS or the MISRII also results in development of pseudohermaphroditism in mice (8,9). Spatial and temporal regulation of MIS and its type II receptor is important for its biological action (10). One study showed that the MISRII promoter is up-regulated by steroidogenic factor 1 (SF1), a member of the nuclear hormone receptor family (11).
The Wnt molecules are a large family of secreted glycoproteins that play an important role in the developmental program in many organisms (12). Constitutive activation of Wnt signaling has been linked to developmental defects and tumorigenesis (13). Targeted disruption of Wnt7a in mice revealed that it is required for normal dorsoventral and anteroposterior polarity in the forming limb (14 -16). Wnt7a-deficient mice are infertile because of retained mullerian ducts and, ultimately, blocked sperm passages in males and abnormal differentiation of oviducts and uteri in females (15). There is no expression of MISRII in Wnt7a mutant mice. It is not clear, however, that MISRII is a direct target of the Wnt7a signaling pathway (15).
The Wnt signaling pathway has been studied extensively, biochemically as well as genetically. The canonical Wnt/␤-catenin or noncanonical Wnt/Ca 2ϩ pathways transduce extracellular Wnt signals into the nucleus (17). The canonical Wnt pathway is mediated by stabilized ␤-catenin and nuclear TCF/LEF family members, a high mobility group box containing transcription factor, distantly related to SOX proteins (12). In the absence of Wnt signals, free ␤-catenin is phosphorylated by GSK3␤ in the cytosol. The tumor suppressor APC and axin are part of the large multiprotein complex that facilitates this phosphorylation process (18). Phosphorylated ␤-catenin is ubiquinated and ultimately degraded by the proteosome (19). In the presence of a Wnt signal, this phosphorylation of ␤-catenin is blocked, and the free cytosolic ␤-catenin is translocated to the nucleus and heterodimerized with one of the LEF/TCF family members and can activate Wnt-responsive genes (12). It is still unclear whether Wnt7a mediates its signal through the canonical or noncanonical pathway. In this study, we address the question of whether MISRII is a direct transcriptional target of the Wnt signaling pathway and, if so, whether the signal is mediated through the canonical or noncanonical Wnt signaling pathway. PCR products were subsequently subcloned into a modified pcDNA3 vector (20) harboring the 5Ј-untranslated region of the herpes simplex virus-thymidine kinase gene.
The 863-bp MISRII upstream sequence, which drove transcription of the luciferase gene, was amplified by PCR from human genomic DNA with MISRII promoter nucleotides Ϫ800 to ϩ63 and cloned in front of the luciferase gene in the vector pGL3basic (Promega). This construct was designated Ϫ800MISRPLuc. Deletion constructs were made by using this construct as a template. All site-directed mutagenesis of the promoter constructs was performed with the QuikChange site-directed mutagenesis kit (Stratagene). All constructs were confirmed by sequencing from both directions.
Cell Culture and Transfection-HeLa and SW480 cells were grown at 37°C in Dulbecco's modified Eagle's/F-12 medium supplemented with 10% fetal calf serum in an atmosphere containing 5% CO 2 . The cells were seeded at a density of 50,000 -70,000 cells/well in 12-well plates 16 -18 h before transfection. The cells were cotransfected with expression and reporter plasmids as indicated in the figure legends. The plasmid CMV-␤-galactosidase was cotransfected as an internal control to normalize for differences in transfection efficiency. The transfections were performed with LipofectAMINE Plus reagent (Invitrogen) according to the manufacturer's recommendations, and the cells were harvested after 40 -48 h. Luciferase activity was measured with a luciferase assay kit (Tropix) and a Lumat LB9507 luminometer (EG&G Berthold). ␤-Galactosidase was measured with the Galacto-Light plus kit (Tropix).
Gel Shift Assay-Gel shift reactions were performed in a total volume of 20 l on ice. Radiolabeled probes were prepared by end-labeling with [␥-32 P]ATP, and 100 pmol of each labeled probe and 2.5 l of in vitro translated protein were used for each reaction. For competition of wild-type and the WT1-binding site-specific oligonucleotides, a 100-fold excess of unlabeled oligonucleotides was added to the reaction mixture before addition of the labeled probe. Thirty minutes later, the reaction mixture was loaded onto a 5% polyacrylamide gel in Tris glycine buffer, and electrophoresis was performed at 150 V for 3 h.
Chromatin Immunoprecipitation Assay-Formaldehyde was added to TCF4-expressing HeLa cells at a final concentration of 1%. Fixation was allowed to proceed at room temperature for 15 min and was stopped by addition of glycine to a final concentration of 0.125 M. The cells were then washed with phosphate-buffered saline and collected by centrifugation. The cells were incubated with buffer A (10 mM potassium acetate, 15 mM magnesium acetate, and 0.1 mM Tris (pH 7.4) with Roche protease inhibitor mixture) on ice for 20 min and homogenized with a Dounce homogenizer. The nuclei were collected by centrifugation, resuspended in sonication buffer, and incubated on ice for 15 min. The samples were sonicated on ice with an Ultrasonics sonicator at a setting of 10 for six 20-s pulses to an average length of ϳ1000 bp (confirmed by electrophoresis) and microcentrifuged. The chromatin solution was precleared with protein A-Sepharose (Pierce) for 15 min at 4°C. Immunoprecipitations (IPs) were performed overnight at 4°C with 1 g of anti-TCF4 antibody. After the final ethanol precipitation, each IP sample was resuspended in 30 l of TE. Total input chromatin samples were resuspended in 30 l of TE and further diluted 1:100. Each 50 l of PCR mixture contained 5 l of IP sample, 1.5 mM MgCl 2 , 50 ng of each primer, 300 M each dATP, dGTP, dCTP, and dTTP, 1ϫ PCR buffer (PerkinElmer Life Sciences), and 1.25 unit of TaqDNA polymerase (PerkinElmer Life Sciences). After 35 cycles of amplification, 5 l of each of the PCR products was subjected to electrophoresis on a 1.5% agarose gel, and the DNA was stained with ethidium bromide and visualized under UV light. The sequences of primers used for PCR were as follows: MISRP-F1, CAGGCCTCTGCAGTTATG; MISRP-R1, CATGGTGGTACAGCAAGG; MISRP-F2, CTGGGTTCTCAGCTGGGC-CTC; MISRP-R2, AGCCAGCACAGCTGCCCCTG; MISRII-int10F, TGCCCCATCTGCTCTCCTAATACA; MISRII-int10R, AGCCTCCCT-CCTCTCCCTCTTG.
Co-immunoprecipitation Assay-FLAG-tagged SF1 was transfected in SW480 cells, and whole cell extracts were diluted in modified RIPA buffer and co-immunoprecipitated with rabbit anti-SF1 (Upstate Biotechnology) or rabbit control IgG (Santa Cruz Biotechnology) antibody. Immunocomplexes were collected with Protein A/G-Sepharose (Pierce), separated by 4 -20% SDS-PAGE, and immunoblotted with rabbit anti-␤-catenin antibodies (Santa Cruz Biotechnology). Western blots were developed using SuperSignal West Femto Maximum Sensitivity Substrate (Pierce).

MISRII Promoter a Direct Target of the Wnt Signaling
Pathway-To determine whether ␤-catenin and TCF4, the intracellular mediators of the Wnt signaling pathway, can activate a luciferase reporter gene driven by the MISRII promoter, we transfected Ϫ800MISRPLuc into HeLa cells. Cotransfection of the luciferase reporter gene with constitutively active ␤-catenin resulted in ϳ5-fold greater activation of the reporter gene ( Fig. 1). The constitutively active ␤-catenin had mutated phosphorylation sites and therefore escaped APC-mediated degradation. TCF4 alone cannot activate this promoter, because ␤-catenin is not constitutively active in HeLa cells. Consistent with previous work on other Wnt-responsive genes, coexpression of TCF4 with ␤-catenin increased the reporter gene activity. However, coexpression of ␤-catenin and TCF4 did not dramatically increase luciferase activity. This was mostly because of the high level of expression of endogenous TCF4 in HeLa cells (21). In the negative control pGL3basic vector, ␤-catenin and TCF4 did not increase reporter activity (Fig. 1). This indicates that the MISRII promoter is responsive to the canonical ␤-catenin/TCF signaling pathway.
Identification of TCF-binding Sites in the MISRII Promoter-To identify the binding site in the MISRII promoter that is responsive to ␤-catenin, we generated a series of deletion mutants of the MISRII promoter based on known TCF/LEF-binding sites. Luciferase activity was up-regulated by severalfold by cotransfection of ␤-catenin and TCF4 with full-length MISRII promoter containing luciferase reporter genes (Fig. 2). Progressive deletion of sequences within the promoter gradually decreased the reporter gene activity. This implies that more than one TCF4-binding site is present in this promoter. In contrast, that the construct containing nucleotides Ϫ170 to ϩ63 was not responsive to ␤-catenin and TCF4, which indicates that sequences essential for activation of the MISRII promoter lie between bases Ϫ800 and Ϫ170 (Fig. 2).
TCF-binding Sites Essential for ␤-Catenin-mediated Activation of the MISRII Promoter-Between nucleotides Ϫ800 and Ϫ170 within the MISRII promoter, there are four potential TCF-binding sites, designated TCF-RE 1-4, and one SF1-binding site (Fig. 3A). These binding sites are similar to the consensus TCF/Lef-binding sites. To determine the relative contribution of each of these sites to the ␤-catenin responsiveness of the MISRII promoter, we introduced point mutations in each of these binding sites. We measured luciferase reporter gene activity driven by each of these mutant-binding sites and by the wild-type promoter in the presence or absence of TCF4 and ␤-catenin. Mutation in any one of the binding sites did not reduce overall responsiveness to ␤-catenin. However, mutants TCF-M1 and TCF-M4 reduced MISRII promoter responsiveness to ␤-catenin significantly (Fig. 3B). Introduction of mutations in both TCF-binding sites 1 and 2 further reduced the reporter gene activity. However, this reporter construct still showed some responsiveness to ␤-catenin. Mutations in binding sites 2 and 3 did not induce any drastic changes in reporter gene activity, although these two binding sites did contribute some of the responsiveness to ␤-catenin. When mutations were made in all four TCF-binding sites, the reporter gene lost its responsiveness to ␤-catenin (Fig. 3B). We concluded from these analyses that the responsiveness of the MISRII promoter to ␤-catenin is dependent on all four TCF-binding sites and that binding sites 1 and 4 are most critical.
TCF4 Binding in Vitro to the TCF-binding Sites of the MISRII Promoter-TCF-binding sites 1 and 4 were further characterized by gel shift assays. The gel shift assay in Fig. 4 shows that TCF4 synthesized in vitro bound to a 32 P-labeled oligonucleotide probe for TCF-binding sites 1 and 4 (TCF-RE 1 and 4). Binding was competed by unlabeled wild-type probe but not by a WT1-binding site probe (Fig. 4). These data suggest that TCF-binding sites 1 and 4 in the MISRII promoter are essential for the binding of TCF4 to the promoter as well as for ␤-catenin-mediated transactivation of the promoter.
To fully address the possibility that regulation of MISRII gene expression is mediated by ␤-catenin, we used the chromatin immunoprecipitation assay to determine whether TCF4 can bind the endogenous MISRII promoter. PCR analysis of formaldehyde-cross-linked chromatin from HeLa cells immunoprecipitated with antibodies specific to TCF4 revealed that TCF4 can indeed bind to the MISRII promoter sequences, whereas rabbit IgG antibody controls did not (Fig. 5). Primers for MISRII intron 10 were used as a negative control to show the specificity of the chromatin immunoprecipitation assay. The expected 436-bp bands from MISRII intron 10 were not present in the immunoprecipitated DNA by anti-TCF4 and control antibody, which implies that TCF4 antibodies specifically immunoprecipitated only the MISRII promoter-bound TCF4-DNA complex, not a big fragment of genomic DNA (Fig. 5).
SF1 and ␤-catenin-TCF4 Synergistically Activated the MISRII Promoter-There is an SF1-binding site in the MISRII promoter. Addition of exogenous SF1 increases luciferase activity driven by the MISRII promoter in the teratocarcinoma cell line NT2D1 (11). However, SF1 does not activate luciferase in R2C2, a rat Leydig cell line, although endogenous SF1 protein occupies the SF1-binding sites in these cells (22). This suggests that SF1 alone cannot activate this promoter in the R2C2 cells and that a necessary cofactor is missing. To address this possibility, we transfected SF1 with various combinations of reporter and effector plasmids into HeLa cells. In the absence of ␤-catenin expression, SF1 had no effect on the MISRII promoter, even in the presence of TCF4 (Fig. 6). In the presence of ␤-catenin, however, SF1 synergistically activated this promoter. Furthermore, coexpression of ␤-catenin, TCF4, and SF1 resulted in superinduction of this promoter (Fig. 6A). However, mutation in the SF1-binding site abrogated this synergistic activation (Fig. 6B). This result suggests that ␤-catenin is the necessary factor for activation of the MISRII promoter.
We also examined the effects of SF1 on TOPflash, a Wntresponsive reporter that contains three multimerized TCF-binding sites. In contrast to its synergistic activation of the MISRII promoter, SF1 had no effect on ␤-catenin/TCF-mediated induction of luciferase activity from the TOPflash reporter (Fig. 7). Thus, SF1 is not a general cofactor for ␤-catenin/TCF, and SF1 only increases ␤-catenin signaling from the MISRII promoter.
Dominant Negative Action of ⌬N TCF4 -To further characterize the ␤-cateninand SF1-mediated transactivation of the MISRII promoter, we used a dominant negative form of TCF4, ⌬N TCF4, which lacks the N-terminal ␤-catenin-interacting domain but has the intact DNA-binding domain. Activation of the MISRII promoter was abolished by the dominant negative TCF4 construct (Fig. 8). ␤-Cateninand SF1-mediated syner- FIG. 5. TCF4 binding to the MISRII promoter in vivo in HeLa cells. Cross-linked chromatin from HeLa cells was incubated with antibodies to the N-terminal region of TCF4. The immunoprecipitated DNA was analyzed by PCR with primers spanning TCF-binding sites 1 and 4 (␣TCF4). PCR was performed with DNA obtained from IP with rabbit IgG as control (PIS). As a negative control, PCR was performed with primers for MISRII intron 10 and the same DNA obtained from IP with TCF4-specific antibody and with rabbit IgG (PIS) (fifth and sixth lanes). As a positive control input DNA was amplified by PCR with primers for MISRII intron 10 along with primers spanning TCF-binding sites 1 and 4 (middle panel). The first lane contains a 100-bp DNA ladder. gistic activation of the MISRII promoter was lost by addition of dominant negative TCF4 (Fig. 8). These observations further show that TCF4-binding sites are essential for ␤-cateninas well as SF1-mediated activation of the MISRII promoter.
APC Down-regulation of MISRII Promoter Activation in a Colon Cancer Cell Line-In the colon cancer cell line SW480, the nuclear level of ␤-catenin is increased because of an inactivating mutation in the tumor suppressor gene APC, which leads to constitutive activation of the Wnt-responsive genes by endogenous ␤-catenin (23). We transfected the MISRII promoter-driven luciferase reporter gene with or without APC and SF1 into SW480 cells. Although SF1 could not activate this promoter without coexpression of ␤-catenin in HeLa cells (Fig.   6A), SF1 was sufficient to activate this promoter in SW480 cells (Fig. 9), probably because of the presence of endogenous ␤-catenin. This activation was reduced by transient expression (by cotransfection) of the wild-type APC. Because of the nature of transient transfection, adding APC did not dramatically decrease reporter gene activity.
Interaction between ␤-Catenin and SF1-We used in vivo co-immunoprecipitation experiments to determine whether synergistic activation of the MISRII promoter by ␤-catenin and SF1 involves a direct physical interaction between them. We transfected FLAG-tagged SF1 into SW480 cells, in which a significant amount of free ␤-catenin is present both in the cytoplasm and in the nucleus (21). IP of SF1 using anti-SF1 antibody followed by immunoblotting with anti-␤-catenin antibody revealed a ϳ100-kDa band corresponding to ␤-catenin (Fig. 10). We found no evidence for a direct interaction between SF1 and TCF4 by cotransfection or by mixing in vitro translated proteins followed by IP and immunoblotting (data not shown).

DISCUSSION
Wnt genes play an important role in organogenesis, and aberrant Wnt signaling is believed to be involved in many types of cancer in humans (13). Although there are many Wnt genes in mammals, only a few physiological target genes are known to date (12). In this study, we have shown that MISRII, an important component in male sexual differentiation, is a direct target of the canonical Wnt signaling pathway. Overexpression of mutant ␤-catenin that cannot be phosphorylated or degraded activated the MISRII promoter in HeLa cells. ␤-Catenin itself is not a DNA-binding factor, but it binds to the TCF/LEF family of DNA-binding factors and acts as a co-activator (12). In the absence of nuclear ␤-catenin, TCF keeps the target genes in a repressed condition by virtue of its association with groucho and related co-repressors (24). Although TCF4 is abundantly expressed in HeLa cells (21), expression of TCF4 alone was not sufficient to activate the MISRII promoter, the addition of ␤-catenin was necessary to activate the MISRII promoter in these cells. However, addition of TCF4 with ␤-catenin did not robustly activate the MISRII promoter because the levels of endogenous TCF4 in this cell line were already saturating. TCF family members bind to SOX-related binding sites, progressive deletions in the MISRII promoter revealed four TCF-binding sites, two of which were essential for TCF binding as well as ␤-catenin-mediated activation of this promoter.
Barbara et al. (11) reported that SF1 can activate the MISRII promoter in the teratocarcinoma cell line NT2D1. SF1 can also activate the MISRII promoter in some other cell lines. In our study, however, SF1 alone did not activate the MISRII promoter in HeLa cells. This finding indicated that some factor is missing in HeLa cells. Our results indicate that ␤-catenin and TCF4 are the missing factors. It is not surprising that ␤-catenin and SF1 synergistically activate the MISRII promoter. ␤-Catenin interacts with a number of factors to exert its biological effects, the nuclear hormone receptor RXR and androgen receptors among them (25,26).
MISRII fulfills the requirements of a direct physiological target gene of Wnt signaling. In Wnt7a mutant mice, which do not express MISRII, the mullerian ducts do not regress (15). The MISRII promoter is regulated in a canonical ␤-catenin/ TCF4-dependent manner. Moreover, ␤-catenin is in the nucleus with the lymphoid enhancer factor during the critical period of mullerian duct regression, this nuclear accumulation of ␤-catenin is independent of Wnt7a action (27). To our knowledge, however, there are no other signaling pathways that can activate ␤-catenin/TCF signaling.
Transforming growth factor signaling has been shown to cooperate with Wnt signaling in Xenopus and Drosophila systems. Because MIS is a member of the transforming growth factor superfamily, the signal MIS may cooperate with Wnt in the activation of the MIS receptor. Transforming growth factor transduces its signal through SMAD-related transcription factors. However, we did not find SMAD-binding sites within the 800 nucleotides of the MISRII promoter. On the other hand, we did find that SF1 could cooperate with ␤-catenin to activate this promoter. SF1 is an orphan nuclear receptor for which there is no known ligand. It is known that phosphorylation modulates the function of SF1 and is mediated by the mitogenactivated protein kinase signaling pathway (28). It is likely that MIS-mediated signal may involve interaction with mitogen-activated protein kinase signaling (29). This postulated MIS-mediated mitogen-activated protein kinase activation may modulate the function of SF1.
Wnt molecules play an important role in sexual determination and differentiation in species ranging from Drosophila to mice (30,31). Our results also showed that Wnt signaling is important for proper sexual differentiation in human males.
Regression of the mullerian ducts is tightly regulated by the temporal and spatial expression of genes involved in this process during embryogenesis. Although both male and female embryos of Wnt7a mutant mice have defects in proper sex organ development, mullerian duct regression is important only for the male embryo. Wnt7a is expressed in both male and female gonads, but SF1 expression is sexually dimorphic. Initially, SF1 is expressed in both male and female embryonic gonads; later, expression of SF1 is up-regulated in male gonads and down-regulated in female gonads in mice. This higher level of expression of SF1 cooperates with the intracellular Wnt mediators ␤-catenin and TCF and up-regulates expression of MISRII during the critical period of male sexual differentiation.