Transcriptional Regulation by the Wilms Tumor Protein, Wt1, Suggests a Role of the Metalloproteinase Adamts16 in Murine Genitourinary Development*

Background: ADAMTS16 is a mammalian metalloproteinase with unknown function. Results: Transcription of the Adamts16 gene is regulated by Wilms tumor protein Wt1, and knockdown of Adamts16 reduces branching morphogenesis in cultured embryonic kidneys. Conclusion: Adamts16 is a Wt1 target gene during murine genitourinary development. Significance: The findings provide novel insights into gene regulatory networks controlling kidney and gonad development. ADAMTS16 (a disintegrin and metalloproteinase with thrombospondin motifs) is a secreted mammalian metalloproteinase with unknown function. We report here that murine Adamts16 is co-expressed with the Wilms tumor protein, Wt1, in the developing glomeruli of embryonic kidneys. Adamts16 mRNA levels were significantly reduced upon transfection of embryonic murine kidney explants with Wt1 antisense vivo-morpholinos. Antisense knockdown of Adamts16 inhibited branching morphogenesis in kidney organ cultures. Adamts16 was detected by in situ mRNA hybridization and/or immunohistochemistry also in embryonic gonads and in spermatids and granulosa cells of adult testes and ovaries, respectively. Silencing of Wt1 by transfection with antisense vivo-morpholinos significantly increased Adamts16 mRNA in cultured embryonic XY gonads (11.5 and 12.5 days postconception), and reduced Adamts16 transcripts in XX gonads (12.5 and 13.5 days postconception). Three predicted Wt1 consensus motifs could be identified in the promoter and the 5′-untranslated region of the murine Adamts16 gene. Binding of Wt1 protein to these elements was verified by EMSA and ChIP. A firefly luciferase reporter gene under control of the Adamts16 promoter was activated ∼8-fold by transient co-transfection of human granulosa cells with a Wt1 expression construct. Gradual shortening of the 5′-flanking sequence successively reduced and eventually abrogated Adamts16 promoter activation by Wt1. These findings demonstrate that Wt1 differentially regulates the Adamts16 gene in XX and XY embryonic gonads. It is suggested that Adamts16 acts immediately downstream of Wt1 during murine urogenital development. We propose that Adamts16 is involved in branching morphogenesis of the kidneys in mice.

ADAMTS16 (a disintegrin and metalloproteinase with thrombospondin motifs) designates a growing family of secreted metalloproteinases in mammals and invertebrates (1,2). Their structural and functional properties are related to, although distinct from, the ADAM 3 group of proteinases. ADAMTSs have several functional domains, including a prodomain, an N-terminal catalytic domain, a disintegrin domain, and a C-terminal ancillary domain. The last one contains several thrombospondin-like repeats and is crucial for specific interactions with substrates (2,3). ADAMTS proteins lack the transmembrane domain present in ADAMs (4). To date, 20 ADAMTS family members have been identified, whose multiple functions include proteolytic degradation of proteoglycans, collagen processing, inhibition of angiogenesis, and the control of blood coagulation (1,2,(5)(6)(7)(8). Additional roles of ADAMTS molecules may extend to embryonic development, inflammatory responses, and fertility (9 -13). Mutations in ADAMTS genes can give rise to human disorders. For example, the Ehlers-Danlos syndrome type VIIC (MIM 2254109) is a recessive connective tissue disease, which results from mutations in the gene for the procollagen N-proteinase, ADAMTS2 (14). Mutations in the gene encoding ADAMTS13, which cleaves the von Willebrand factor, can lead to inherited thrombotic thrombocytopenia (15).
ADAMTS16 is a recently cloned family member, whose physiological functions are currently unknown. A full-length recombinant protein was capable of cleaving the proteinase inhibitor ␣ 2 -macroglobulin in vitro (16). ADAMTS16 mRNA was initially detected in human fetal lung and kidney but also in adult brain and ovary (1). The cell types that express ADAMTS16 in these tissues have not yet been identified. Later studies demonstrated ADAMTS16 transcripts in human cartilage (17) and pancreas (16). Expression of ADAMTS16 in cartilage and synovium was increased in patients suffering from osteoarthritis compared with the tissues of healthy individuals (17). Variants of the ADAMTS16 gene were recently linked to inherited arterial hypertension (18), and high levels of ADAMTS16 could be detected in a significant portion of esophageal squamous cell carcinomas (19).
Although a role of ADAMTS16 in development and disease is beginning to emerge, very little is known about the molecular mechanisms regulating its tissue-specific expression. ADAMTS16 mRNA levels were increased in differentiated luteinizing human granulosa cells following treatment with follicle-stimulating hormone (FSH) and forskolin (16). These observations suggested that ADAMTS16 expression is activated via the cyclic AMP pathway (16). The promoter of the ADAMTS16 gene is GC-rich and contains several consensus motifs for Egr1 and Sp1 (20). Both transcription factors were found to activate ADAMTS16 promoter constructs in transient co-transfection experiments (20). However, it is not known whether Egr1 and Sp1 are physiological regulators of ADAMTS16 in vivo; nor have other trans-acting factors been ascertained thus far.
In view of this background, our study was aimed at providing novel insights into the gene regulatory mechanisms underlying the characteristic expression pattern of ADAMTS16. For this purpose, we analyzed the cellular distribution of Adamts16 protein in various organs of mice and rats. Furthermore, we sought to determine transcription factors that are involved in the tissue-specific control of Adamts16. We report here that Adamts16 is a transcriptional downstream target gene of the Wilms tumor protein, Wt1. Importantly, knockdown of Adamts16 by transfection with antisense vivo-morpholinos impaired branching morphogenesis in murine embryonic kidney cultures. These findings suggest a role for Adamts16 in the development of the genitourinary system in mice.

EXPERIMENTAL PROCEDURES
Plasmids-A 3250-base pair (bp) DNA sequence (from Ϫ3000 bp to ϩ250 bp relative to the transcription start site) of the murine Adamts16 gene (NCBI accession number NM_172053) was cloned by PCR, utilizing a bacterial artificial chromosome (imaGenes, Berlin, Germany, clone RPCIB731I01114Q) as template. The amplified product was ligated into the SacI and HindIII restriction sites of the pGL3basic reporter plasmid (Promega, Mannheim, Germany). Adamts16 promoter constructs with gradually shortened sequences were generated by PCR. The PCR primers that were used for DNA amplification are listed in Table 1. All constructs were analyzed by automated DNA sequencing (Eurofins MWG, Ebersberg, Germany). The expression construct for the murine Wt1(ϪKTS) protein was kindly provided by Dr. Daniel Haber (Massachusetts General Hospital, Boston, MA).
Animals-A heterozygous (Wt1 ϩ/Ϫ ) mouse breeding pair (C57/BL6 (B6) strain) was obtained from the Jackson Laboratory (Bar Harbor, ME) and mated in the local animal facility in compliance with the current laws. The morning of vaginal plugs was considered as 0.5 days postconception (d.p.c.).
Organ Culture Experiments-Kidneys, gonads, and hearts were excised from embryos of timed pregnant mice (C57BL6 strain). The matched organs of each embryo were cultured separately on two polyethylene terephthalate Transwell filters with 0.4-m pore size (Corning, Inc.). This procedure allowed for a pairwise comparison between cultured explants from a single donor that had been transfected with either Wt1 antisense or mismatch vivo-morpholinos (see below). The Transwell filters were kept in DMEM nutrient with stable L-glutamine (PAA Laboratories) supplemented with 10% FCS (Biochrom KG), 100 IU/ml penicillin (PAA Laboratories), and 100 g/ml streptomycin (PAA Laboratories) at 37°C in a humidified atmosphere with 5% CO 2 . Embryonic hearts were kept in 96-well plates for 5 days to allow outgrowth of epicardium-derived cells. The epicardial origin of outgrown cells was proven by their characteristic cobblestone-like appearance and robust expression of Wt1 (data not shown). Prior to morpholino treatment of the cells, the hearts were removed from the tissue culture plates. In vitro cultured organs and epicardial cell monolayers were incubated for 72 h with either Wt1 antisense (CAGGTCCCGCACGTCGGAACCCAT) or mismatch (CAGCTCCGGCACCTCGCAACCGATG) vivo-morpholinos at 10 M each (Gene Tools, Philomath, OR) (24). Whole cell lysates and total RNA were prepared from cultured organs and cells as described below. Sex determination of the embryos was performed by PCR amplification of the Y-chromosomal gene Kdm5d from DNA using the following primers: mKdm5d-F, CTGAAGCTTTTGGCTTTGAG; mKdm5d-R, CCACTGC-CAAATTCTTTGG (25).
Cell Transfections and Reporter Gene Assays-COV434 cells were expanded to ϳ40% confluence in 24-well tissue culture plates. Firefly luciferase constructs, 200 ng each, harboring gradually shortened sequences of the murine Adamts16 promoter and the 5Ј-UTR were transiently transfected along with a Wt1(ϪKTS) expression construct (200 ng) and a Renilla luciferase plasmid (100 ng), using the Fugene6 reagent (4 l/well) as described in the supplier's manual (Roche Applied Science). As negative controls, transfections were performed with the pGL3basic reporter plasmid (Promega) and the empty pCB6 ϩ expression vector, respectively. After 24 h, the transfected cells were lysed in Reporter Lysis Buffer (Promega), and luciferase activities were measured in a luminometer (Microlite TLX1, MGM Instruments, Hamden, CT) as described (26,27). Data are presented as firefly luciferase activities normalized to Renilla luciferase activities.
Reverse Transcription (RT) Real-time PCR-Total RNA was isolated from primary tissues and cells with the RNeasy Micro Kit (Qiagen, Hilden, Germany) and from permanent cell lines with the TRIzol LS reagent (Invitrogen). First-strand cDNA synthesis was carried out using oligo(dT) primers and Super-Script TM II reverse transcriptase (Invitrogen). Real-time PCR amplification was performed with the SYBR Green PCR Master Mix and the StepOnePlus TM system (Invitrogen) as described in detail elsewhere (28). The PCR primers for realtime RT-PCR are listed in Table 1. Relative transcript levels were obtained by subtracting the threshold cycle (Ct) value of the housekeeping gene (␤-actin or Gapdh) from the corresponding Ct value of the gene of interest. Differences in mRNA levels were calculated according to the expression 2 ⌬⌬Ct .
Electrophoretic Mobility Shift Assay (EMSA)-Non-radioactive EMSAs were performed with purified GST-tagged recombinant Wt1 proteins (30) and double-stranded oligonucleotides ( Table 2), which were selected on the basis of predicted Wt1 binding sites in the promoter of the murine Adamts16 gene. All oligonucleotides were 3Ј-digoxigenin (DIG)-labeled using the DIG Oligonucleotide 3Ј-End Labeling Kit (Roche Applied Sci- . Likewise, double-stranded oligonucleotides with mutations of the Wt1 binding motifs ( Table 2) were DIG-labeled. For competition experiments, we used an oligonucleotide with a previously identified Wt1 binding site in the promoter of the Ntrk2 gene (31). The binding reactions were carried out at room temperature for 1 h using 0.75 g of recombinant Wt1 protein and 1 pmol of labeled oligonucleotide in 1ϫ reaction buffer (10 mM Tris-HCl, pH 7.5, 50 mM KCl, 50 mM NaCl, 1 mM MgCl 2 , 1 mM EDTA, 5 mM DTT, 5 mM phenylmethylsulfonyl fluoride, PMSF). DNA-bound proteins were separated on a non-denaturing 6% polyacrylamide gel and transferred with a semidry blotting system (Bio-Rad) on a positively charged nylon membrane (Roche Applied Science). After UV crosslinking (Stratalinker UV cross-linker, Stratagene) and treatment with blocking reagent (Roche Applied Science), the membrane was incubated with anti-DIG-alkaline phosphatase antibody (catalog no. 11093274910, diluted 1:10.000, Roche Applied Science). Proteins were detected by nitro blue tetrazolium/5-bromo-4-chloro-3-indolyl phosphate staining (Roche Applied Science). Chromatin Immunoprecipitation (ChIP) Assays-UB27 and UD28 cells were kept for 72 h in the presence or absence of tetracycline (1 g/ml) to either inhibit or stimulate expression of Wt1(ϪKTS) and Wt1(ϩKTS) proteins, respectively. ChIP assays were performed with ϳ4 ϫ 10 6 cells. After fixation with 4% formaldehyde, the cells were disrupted by ultrasonication (Labsonic U, B. Braun) and centrifuged. The supernatants were diluted in immunoprecipitation buffer (0.01% SDS, 1.1% Triton X-100, 1.2 mM EDTA, 16.7 mM Tris-HCl, pH 8.1) and precleared for 1 h at 4°C with DNA-blocked protein G-agarose (Millipore). The following antibodies and control sera were used (0.6 g each) for overnight incubation at 4°C: normal rabbit IgG (catalog no. sc-2027, Santa Cruz Biotechnology, Inc.), rabbit polyclonal anti-Wt1 antibody (catalog no. sc-192, Santa Cruz Biotechnology, Inc.). Immunoprecipitates were bound to DNA-blocked protein G-agarose (Millipore) during a 1-h incubation at 4°C. Following several washing steps, the DNA was eluted in 1% SDS, 0.1 M NaHCO 3 , extracted in 25:24:1 phenol/chloroform/isoamyl alcohol according to the supplier's instructions (Invitrogen), and subsequently precipitated in 100% ethanol. This protocol was also applied to M15 cells, which were used at ϳ1 ϫ 10 6 cells for ChIP assays. Immunoprecipitation with an anti-histone H3 antibody (catalog no. 07-690, Millipore) was performed as a control. Immunoprecipi-tated DNA was amplified by quantitative PCR using primer pairs that flanked the identified Wt1 binding sites in the murine Adamts16 promoter (Table 1).
In Situ mRNA Hybridization-In situ hybridization for Adamts16 mRNA was performed according to the protocol of Stricker et al. (32). Briefly, tissues were cryosectioned and fixed in freshly prepared paraformaldehyde (4% in PBS). Following acetylation with acetanhydride, the sections were prehybridized for 4 h at room temperature in hybridization buffer (5ϫ

Adamts16 and Wt1 Proteins Are Expressed in Gonads and
Kidneys-Although Adamts16 mRNA has been detected by RT-PCR in several organs, including kidneys and gonads (1), very little is known about the protein localization in these tissues. We used specific polyclonal rabbit antibodies (ab45048 (Abcam) and sc-50490 (Santa Cruz Biotechnology, Inc.)) to visualize Adamts16-expressing cells in the genitourinary system of mice and rats. Adamts16 and Wt1 proteins were readily visible in XY and XX embryonic gonads (Fig. 1, A (ac) and B (ac)) although more distinct in embryonic testes (Fig. 1A, FIGURE 1. Expression of Adamts16 and Wt1 in embryonic and adult gonads. A, specific antibodies against Wt1 (green) and Adamts16 (red) were used to detect both proteins in the XY gonads of embryonic (14 d.p.c.) mice (a-c) and adult rats (d-f). Note that two different antibodies were applied for the detection of Adamts16; the results shown in b, d, and e were obtained with antibody ab45048 (Abcam), and antibody sc-50490 (Santa Cruz Biotechnology, Inc.) was used for the stainings presented in c and f. Double immunolabeling identified Sertoli cells as the sites of Wt1 expression in adult testes (arrowheads in d and f). Antibody ab45048 detected Adamts16 protein mainly in the spermatids (asterisks in e), whereas antibody sc-50490 produced a signal in Sertoli cells (arrowheads in f) in addition to spermatids (asterisks in f). Cell nuclei were counterstained with DAPI (blue). B, Wt1 (green) and Adamts16 (red) proteins in XX gonads of embryonic (13 d.p.c.) mice (a and b) and adult rats (d-f). Data shown in b were obtained with antibody ab45048 (Abcam), whereas antibody sc-50490 (Santa Cruz Biotechnology, Inc.) was used for the stainings presented in d-f. Co-expression of Wt1 and Adamts16 was observed in granulosa cells of preovulatory tertiary (d) and secondary (asterisks in e) follicles and in mesothelial cells on the surface of the ovaries (arrowheads in e and f). Incubation of the tissue sections with normal sera produced no staining signal (c). O, oocyte; A, antrum folliculi. Cell nuclei were counterstained with DAPI (blue). C, expression of Adamts16 in developing (a) and adult (c and d) testis was confirmed by in situ hybridization using a specific antisense riboprobe. Note that Adamts16 transcripts were also present in granulosa cells of the ovaries (e and f). No hybridization signal was seen with the use of an Adamts16 sense probe (b). The data shown are representative for the five tissue samples from three different animals each. a-c). According to previous reports (33,34), Wt1 protein was present in developing and mature Sertoli cells (Fig. 1A, a and  d-f), whereas immunhistochemistry (Fig. 1A, b-f) and in situ hybridization (Fig. 1C, a, c, and d) detected Adamts16 mainly in germ cells of embryonic and adult testes. Expression of Adamts16 was predominant in elongated spermatids of adult testes (Fig. 1A, d-f), which was confirmed by in situ hybridization (Fig. 1C, c and d). Both proteins were also detected in differentiated follicular granulosa cells and in mesothelial cells covering the surface of mature ovaries (Fig. 1B, d-f). In situ hybridization confirmed Adamts16 transcripts in follicular granulosa cells of the ovaries (Fig. 1C, e and f). Adamts16 was demonstrated by immunostaining also in the developing glomeruli of embryonic kidneys (Fig. 2a), which are known sites of Wt1 expression (33, 34) (Fig. 2b). Co-expression of both pro-teins in glomerular progenitors was revealed by double immunolabeling with two different Adamts16 antibodies (Fig. 2, c and d, with corresponding single stainings in e-h). No fluorescence signal was obtained when normal sera instead of primary antibodies were used (Fig. 2, i and k). Adamts16 in murine embryonic kidneys (14 d.p.c.) was confirmed by in situ mRNA hybridization with a specific antisense riboprobe (Fig. 2l). No hybridization signal was obtained with the use of an Adamts16 sense probe (Fig. 2m).
Expression of Adamts16 in Developing Gonads and Kidneys Is Wt1-dependent-Co-expression of Adamts16 and Wt1 suggests a regulatory link between both proteins in different cell types of the genitourinary system. We therefore addressed the question of whether Wt1 is indeed necessary for normal expression of Adamts16 in embryonic gonads and kidneys. To  (c and d). The corresponding single stainings are shown in e and g and in f and h, respectively. Two different antibodies were applied for the detection of Adamts16: the results shown in a, c, and e were obtained with antibody ab45048 (Abcam), and antibody sc-50490 (Santa Cruz Biotechnology, Inc.) was used for the staining in d and f. Tissue sections were incubated with normal sera instead of primary antibodies as negative control (i and k). Cell nuclei were visualized with DAPI (blue). Note that the single immunostainings shown in a and b were performed on two different tissue sections. Adamts16 mRNA was detected in the kidneys of murine embryos (14.5 d.p.c.) by in situ hybridization using a specific antisense riboprobe (l). No hybridization signal was obtained with the use of an Adamts16 sense probe (m). a.g., adrenal gland. this end, Adamts16 transcripts were measured in isolated gonadal ridges of wild-type (Wt1 ϩ/ϩ ), heterozygous (Wt1 ϩ/Ϫ ), and Wt1-deficient (Wt1 Ϫ/Ϫ ) murine embryos at 13 d.p.c. Compared with the gonads of normal (Wt1 ϩ/ϩ ) embryos, Adamts16 mRNA levels were significantly lower (p Ͻ 0.05, ANOVA) in the developing gonadal ridges of Wt1 Ϫ/Ϫ mice (Fig. 3A). Later developmental stages could not be studied due to the embryonic lethality of Wt1-deficient mice (35). These observations suggest that Adamts16 mRNA levels in the developing gonads depend, at least to some extent, on Wt1. However, it is also conceivable that reduced Adamts16 transcripts were simply due to an arrest of gonad formation with subsequent apoptosis of the undifferentiated tissue in the absence of Wt1.
To distinguish between these possibilities, we established in vitro cultures of embryonic gonads. For this purpose, the gonads were excised from normal murine embryos at 11.5, 12.5, and 13.5 d.p.c., respectively, and grown in a culture medium supplemented with either Wt1 antisense or mismatch vivomorpholinos. After 72 h, total RNA was isolated from the explants, and Adamts16 transcripts were measured by quantitative RT-PCR (Fig. 3, B-D). Down-regulation of Wt1 in response to Wt1 antisense morpholino treatment, which was assessed by immunoblot analysis, had differential effects on Adamts16 mRNA levels in female and male gonads (Fig. 3, C  and D). Thus, Adamts16 transcripts were significantly (p Ͻ 0.05, Student's paired t test) increased by knockdown of Wt1 in XY gonads obtained from mouse embryos at 11.5 and 12.5 d.p.c. but not at 13.5 d.p.c. (Fig. 3C). In contrast, Wt1 silencing significantly (p Ͻ 0.05, Student's paired t test) reduced Adamts16 mRNA levels in XX gonads at 12.5 and 13.5 d.p.c. but had no significant effect on embryonic day 11.5 (Fig. 3D). These findings suggest that Wt1 inhibits Adamts16 expression in XY gonads at early developmental stages (11.5 and 12.5 d.p.c.) and activates Adamts16 expression in XX gonads at later points in time (12.5 and 13.5 d.p.c.). Consistently, Adamts16 mRNA levels were increased in both XX and XY gonads treated with mismatch morpholinos, at 13.5 d.p.c. compared with earlier developmental stages (p Ͻ 0.05, ANOVA) (Fig. 3B). Based on differences in Ct values, Adamts16 mRNA levels were at least 10-fold higher in male and female gonads from 13.5 d.p.c. versus 11.5 d.p.c. embryos (Fig. 3B). At all developmental stages analyzed, Adamts16 transcripts were 2-3-fold higher in XX compared with XY gonads (p Ͻ 0.05, Student's paired t test) (Fig.  3B). Similar results were obtained when ␤-actin mRNA was used as a housekeeping gene for normalization (data not shown). Because Wt1 Ϫ/Ϫ mouse embryos lack kidneys (35), they are not useful for analyzing Wt1 downstream signaling pathways during nephrogenesis. To overcome this problem, we established in vitro organ cultures of normal embryonic kidneys, in which we knocked down Wt1 expression by antisense treatment (24). Densitometric analysis of the immunoblots using ImageJ software (National Institutes of Health) revealed a reduction of band intensities in response to Wt1 antisense treatment by 57 Ϯ 10% (p Ͻ 0.05, Student's paired t test, n ϭ 6). Embryonic kidney explants treated with Wt1 antisense vivomorpholinos had significantly (p Ͻ 0.05, Student's paired t test) lower Adamts16 mRNA levels than the corresponding mismatch controls (Fig. 4A). These findings indicate that normal expression of Adamts16 in embryonic kidneys and gonads requires Wt1.
Knockdown of Adamts16 Inhibits Branching Morphogenesis in Kidney Organ Culture-To explore the potential role of Adamts16 in kidney development, we transfected embryonic renal organ cultures (Fig. 4B) with either Adamts16 antisense or mismatch vivo-morpholinos. Knockdown efficiency was determined by immunoblot analysis of Adamts16 using the same antibody (ab45048, Abcam) as for the immunostainings. This antibody produced a signal of the predicted molecular mass (ϳ135 kDa) of the Adamts16 protein (Fig. 4C). Importantly, the protein band was clearly diminished in kidney organ cultures that had been treated with antisense vivo-morpholinos against Adamts16, indicating specificity of the antibody used (Fig. 4C). In vitro knockdown of Adamts16 resulted in an abnormal kidney morphology that was assessed by phase-contrast microscopy (Fig. 4B, a and b) and by immunostaining of the ureteric bud and the metanephric mesenchyme, respectively (Fig. 4B, c-f). Microscopic analysis of the immunolabeled whole-mount preparations revealed a significant (p Ͻ 0.05, Student's paired t test) reduction in the number of ureteric bud branching points upon Adamts16 knockdown (Fig. 4, B (g and  h) and C). These observations suggest a functional role of Adamts16 in branching morphogenesis of murine kidneys.
Wt1 and Adamts16 Expression Levels Are Correlated in Permanent Cell Lines and Primary Epicardial Cells-Next we examined whether Wt1 and Adamts16 levels are associated in permanent cells derived from embryonic kidneys. M15 cells were originally established from the mesonephros of mice transgenically expressing the large T protein of polyoma virus under control of the early viral enhancer (21). The cells were transfected with a pool of four different siRNAs to inhibit Wt1 expression. Control experiments were performed by treatment with pooled non-targeting siRNAs. Silencing of Wt1 significantly (p Ͻ 0.001, Student's paired t test) reduced Adamts16 mRNA levels in M15 cells, indicating that Adamts16 expression in these cells depends on Wt1 (Fig. 5A). To analyze the expression of Adamts16 and Wt1 in primary cells, we established cultures from epicardial progenitor cells. The epicardium is a  microscopy (a and b) and immunofluorescent staining of the ureteric bud and the metanephric mesenchyme using antibodies against pancytokeratin (green) and Wt1 (red), respectively (c-f). Reconstructions of typical branching trees of the ureter are depicted in g and h. C, the number of branching points of the ureteric bud was significantly reduced in response to Adamts16 knockdown. Adamts16 knockdown in the kidney explants was assessed by immunoblot analysis, yielding a band of ϳ135 kDa (1,19). A representative Western blot is shown in the bottom panel in C. Data are presented as means Ϯ S.E., n ϭ 7; *, p Ͻ 0.02, Student's paired t test.

Wt1 Regulates Adamts16 in Developing Kidneys and Gonads
mesothelial tissue on the outer heart surface, which expresses Wt1 from the early developmental stages on throughout adulthood (34). We also detected Adamts16 protein by immunohistochemistry in the developing epicardium of murine embryos (data not shown). Transfection of primary epicardial cells with Wt1 antisense vivo-morpholinos strongly reduced Wt1 protein levels, and this effect was associated with a significant (p Ͻ 0.05, Student's paired t test) reduction of Adamts16 mRNA (Fig. 5B).
Importantly, the Wt1 mRNA is subject to alternative splicing. The usage of an alternative splice donor site at the end of exon 9 leads to the insertion of an additional 3 amino acids, lysine, threonine, and serine (KTS), between zinc fingers 3 and 4 of the Wt1 protein (36). Structural and functional data demonstrate that Wt1(ϪKTS) proteins, which lack the 3 amino acids, act as transcription factors (37,38). Compared with the Wt1(ϪKTS) isoforms, Wt1 proteins with the KTS insertion exhibit increased RNA binding affinity and play a presumed role in mRNA processing (39,40). The human osteosarcomaderived cell lines UB27 and UD28 express the Wt1(ϪKTS) and Wt1(ϩKTS) proteins, respectively, under control of a tetracycline-dependent promoter (23). A consistent rise of Wt1 proteins in UB27 and UD28 cells was achieved by removal of tetracycline from the tissue culture medium (Fig. 5, C and D). Induction of Wt1(ϪKTS) protein in UB27 cells resulted in a time-dependent increase of ADAMTS16 mRNA levels (Fig.  5C), whereas accumulation of Wt1(ϩKTS) protein in UD28 cells had no significant effect on ADAMTS16 transcripts (Fig.  5D). These results indicate that Adamts16 is co-regulated with Wt1 in primary epicardial cells and permanent cell lines. Restriction of the stimulatory effect on Adamts16 expression to the Wt1(ϪKTS) protein suggests that the underlying regulatory mechanism acts on the level of gene transcription.
Wt1 Protein Binds to the Promoter of the Adamts16 Gene in Vivo-ChIP was performed to examine whether Wt1 interacts with the promoter of the Adamts16 gene in M15 cells. The PCR primers for amplification of immunoprecipitated DNA bound to the proximal promoter and the 5Ј-flanking region of the Adamts16 gene, respectively (Fig. 6A). Compared with the use of normal rabbit serum, Adamts16 promoter DNA was enriched more than 2.5-fold with anti-Wt1 antibody (p Ͻ 0.01, Student's t test) (Fig. 6B). An ϳ2-fold enrichment of amplicons within the promoter of the Amhr2 gene (p Ͻ 0.05, Student's t test), a known downstream target of Wt1 (41), was obtained by immunoprecipitation with anti-Wt1 antibody (Fig. 6B). In contrast, amplified DNA fragments containing an actin promoter sequence were not enriched by immunoprecipitation of Wt1 protein (Fig. 6B). These results could be validated using UB27 cells with inducible expression of the Wt1(ϪKTS) protein.
Thus, Adamts16 promoter DNA accumulated ϳ3-fold in UB27 cells with Wt1(ϪKTS) expression compared with uninduced cells (p Ͻ 0.05, Student's t test). No differences in Wt1-bound actin promoter DNA were observed between unstimulated and
The Wt1(ϪKTS) Isoform Stimulates the Adamts16 Promoter-To study the molecular interaction of Wt1 protein with the Adamts16 promoter in detail, we co-transfected a firefly luciferase reporter construct, carrying ϳ3 kb of the upstream sequence of the murine Adamts16 gene, and in addition a Wt1(ϪKTS) expression plasmid into human granulosa cells (COV434). A nearly 8-fold stimulation of the Adamts16 promoter was measured in response to Wt1(ϪKTS) expression compared with co-transfection of the empty vector (p Ͻ 0.01, Student's t test) (Fig. 7A). Sequence analysis revealed the presence of three putative high affinity Wt1 binding sites in the promoter and 5Ј-flanking region of the murine Adamts16 gene. All three elements were necessary for maximal activation of the Adamts16 promoter by the Wt1(ϪKTS) protein (Fig. 7A). Truncated constructs with gradual deletion of the predicted binding motifs were successively less activated by co-transfection of the Wt1(ϪKTS) expression vector. Stimulation by the Wt1(ϪKTS) protein was completely lost in the absence of all three binding motifs (Fig. 7A). EMSAs verified that the Wt1(ϪKTS) protein indeed bound to each of these elements (Fig. 7, B-D). Interaction of recombinant Wt1(ϪKTS) protein could be competed with excess amounts of unlabeled oligonucleotide carrying the previously identified Wt1 binding site of the Ntrk2 gene promoter (Fig. 7, B-D) (31). Likewise, binding of Wt1 protein was abrogated by introducing single base pair mutations into the three oligonucleotides (Fig. 7, B-D). These results demonstrate that Wt1(ϪKTS) protein interacts with several cis-elements in the 5Ј-upstream sequence and stimulates the promoter of the Adamts16 gene. In contrast, recombinant Wt1(ϩKTS) protein did not interact with any of the identified binding sites (Fig. 7E).

DISCUSSION
ADAMTS16 is a mammalian metalloproteinase whose function is largely unknown. Our initial microarray gene expression analysis showed that Adamts16 transcripts were strongly reduced upon RNAi silencing of the Wilms tumor protein Wt1 in mesonephros-derived M15 cells. The following chain of evidence suggests that Adamts16 is indeed a molecular downstream target gene of the Wt1 transcription factor. First, Wt1 and Adamts16 mRNA levels were closely correlated in various tissues and cells . Second, a reporter construct containing the murine 5Ј-upstream sequence of the Adamts16 gene was stimulated ϳ8-fold by co-transfection of ovarian granulosa cells with a Wt1 expression vector (Fig. 7A). Moreover, Wt1(ϪKTS) protein bound to three presumed cis-elements in the promoter and the 5Ј-flanking region of the murine Adamts16 gene (Figs. 6 and 7, B-E). Importantly, knockdown of Wt1 in cultured embryonic murine kidneys and XX gonads (12.5 and 13.5 d.p.c.) by treatment with Wt1 antisense morpholinos significantly reduced Adamts16 transcripts, suggesting that Wt1 stimulates Adamts16 expression in these tissues (Figs. 3 and 4A).
Transcriptional regulation of Adamts16 by Wt1 may become relevant during various biological processes in the embryonic and adult organism. Thus, mammalian kidney formation depends on the reciprocal interaction between the metanephric mesenchyme and the invading ureteric bud (reviewed in Ref. 42). Signals emanating from the branching ureteric bud tips induce the surrounding mesenchymal cells to condensate and differentiate into comma-and s-shaped bodies, which eventually give rise to the glomeruli and build up parts of the tubular system of the kidneys (42). Epithelial transition of the metanephric mesenchyme in turn is a prerequisite for sustained branching of the ureteric bud tips during nephrogenesis (42). Metalloproteinases are considered as major candidates for the modulation of mesenchymal-epithelial interactions in embry-

Wt1 Regulates Adamts16 in Developing Kidneys and Gonads
onic kidneys and other tissues. Several matrix metalloproteinases and their inhibitors (i.e. MMP-2, MT1-MMP, and TIMP-2) are spatially and temporally regulated in the developing kidney (43). Moreover, targeted inactivation of the gene encoding ADAMTS1, which cleaves the major cartilage proteoglycan aggrecan, produced kidney abnormalities in mice (i.e. shrinkage of the renal parenchyma and enlargement of the caliceal space) (44). Our observation that vivo-morpholino knockdown of Adamts16 impaired branching morphogenesis in cultured murine embryonic kidneys suggests an involvement of this metalloproteinase in nephrogenesis (Fig. 4, B and C). Consistently, silencing of the nephrogenic Wt1 protein significantly reduced Adamts16 expression in kidney rudiments (Fig. 4A). Wt1 belongs to the Cys 2 -His 2 class of zinc finger proteins and promotes mesenchymal-to-epithelial transition during the early stages of kidney development (35,45). Kidney formation in Wt1-deficient mouse embryos, which are commonly lethal by 13.5 d.p.c., was disrupted due to apoptosis of the metanephric mesenchyme and a failure of ureteric bud branching (35). Subsequent in vitro organ culture experiments and transgenic animal studies proved a requirement of Wt1 not only for the early inductive events but also during the later phases of kidney formation (i.e. the differentiation and maturation of the glomeruli) (46) and for the integrity of mature kidneys (47)(48)(49). Interestingly, recent disruption of the Adamts16 gene in Dahl Salt-sensitive rats caused splitting and thickening of glomerular capillaries in addition to a significant reduction of arterial blood pressure (50). Renal damage was associated with increased proteinuria, suggesting that Adamts16 is indeed necessary for the structural and functional preservation of the kidneys in this model of genetic hypertension (50).
Interaction of the Wt1 transcription factor with the Adamts16 gene may also play a role in the development and function of the gonads. The mammalian gonads are formed by tissue that is derived from the intermediate mesoderm on either side of the embryonic midline axis (51). Establishment of the bipotential gonadal ridge occurs in a non-sex-specific manner in XX and XY individuals and is completed by ϳ12 d.p.c. in rats and mice (51). Wt1 acts on multiple steps during gonad development. Gonadal ridge formation in Wt1 null mice is initiated in both males and females but then degenerates due to apoptosis, leading to gonadal agenesis (35). The sex-determining factor Sry and the anti-Müllerian hormone receptor (Amhr) 2, which is critical for the regression of the female urogenital  1 and 2) was proven by an electrophoretic mobility shift assay. Mutation of the Wt1 consensus motifs abrogated Wt1(ϪKTS) binding (lanes 3 and 4). Interaction of Wt1(ϪKTS) protein with the wild-type oligonucleotides could be competed with unlabeled competitor DNA (lanes 5-7) corresponding to the previously identified Wt1 binding element in the Ntrk2 promoter (31). E, the Wt1(ϩKTS) isoform, which lacks 3 amino acids in its zinc finger domain (36), did not bind to the oligonucleotides (lanes 3, 6, and 9).

Wt1 Regulates Adamts16 in Developing Kidneys and Gonads
ducts, are among the genes that are regulated by Wt1 (41,52,53). Consistently, male-to-female sex reversal or genital tract abnormalities were observed in XY patients with heterozygous Wt1 germ line mutations (54), underscoring the role of Wt1 in testis determination. Our results indicate that Wt1 controls Adamts16 expression in the developing gonads in a time-dependent and sex-specific fashion. Thus, silencing of Wt1 significantly increased Adamts16 transcripts in in vitro cultured XY gonads of 11.5 and 12.5 d.p.c. embryos (Fig. 3C). In contrast, Adamts16 mRNA levels in female but not in male gonads (12.5 and 13.5 d.p.c.) were significantly reduced by Wt1 antisense inhibition (Fig. 3, C and D). These results suggest that Wt1 inhibits Adamts16 expression in XY gonads during the early developmental stages and stimulates Adamts16 expression in XX gonads at later points in time. Repression of Adamts16 by Wt1 in developing XY gonads is in line with our failure to clearly co-localize both proteins in the same cells in this tissue (Fig. 1, A and B). Further on, Wt1 may directly contribute to the developmental increase of Adamts16 mRNA levels in female gonads (Fig. 3B). Reporter gene assays and protein-DNA binding studies confirmed that Wt1(ϪKTS) protein indeed stimulates transcription of the Adamts16 gene in ovarian granulosa cells (Figs. 6 and 7). The molecular mechanisms underlying the variable stage-specific effects of Wt1 on Adamts16 expression in male and female gonads are currently unknown but may involve the physical interaction of Wt1 with other proteins in the control of gene transcription (55). Detection of Adamts16 in spermatids of mature testes (Fig. 1, A and C) points to a role in male germ cell maturation and/or function throughout adulthood. Generation of mouse mutants with conditional inactivation of the Adamts16 gene in the gonads will be helpful in clarifying this issue.
It has been shown previously that FSH induced the expression of ADAMTS16 in parietal granulosa cells of preovulatory human follicles (16). In contrast, FSH attenuated basal and nerve growth factor-stimulated Wt1 protein expression in isolated granulosa cells of 21-day-old mice (56). Recently, Wt1 has been reported to inhibit expression of the FSH receptor in immature rat granulosa cells through repression of the FSH receptor promoter (57). These combined observations suggest that Wt1 is an important element in the hormonally regulated transcriptional network in preovulatory follicles. Stimulation of Adamts16 expression by Wt1 in parietal granulosa cells may have a role in extracellular matrix (ECM) remodeling, which is a prerequisite for oocyte liberation during ovulation. Notably, various members of the ADAMTS protease family, in particular ADAMTS1, have been implicated in ECM transformation in the ovary (44, 58 -60). The involvement of other proteases and substrates in this process adds some functional redundancy, which may facilitate ovulation even in the absence of one or more constituents (61).
Another interesting aspect of our study pertains to the presumed origin of the follicular granulosa cells, which is still a matter of controversy. Indeed, three major sources for pregranulosa cells have been under discussion for many decades: 1) mesothelial cells on the surface of the developing ovary (62), 2) cells originating from a centrally located blastema (63), and 3) mesonephric cells derived from the rete ovarii system (64). The results of a more recent histological study in sheep favored the idea that Ͼ95% of the granulosa cells in newly formed primordial follicles arise from the ovarian surface epithelium (65). During this process, mesothelial cells on the outside of the ovary are expected to undergo epithelial-to-mesenchymal transition and migrate into the ovarian cortex, where they differentiate to follicular granulosa cells. Remarkably, Wt1 protein has been identified to act as a key molecule for epithelial-to-mesenchymal transition by down-regulating E-cadherin expression in the epicardium of mouse embryos and other mesothelial tissues (66). Thus, co-localization with Wt1 protein in the ovarian surface epithelium suggests that Adamts16 promotes a mesenchymal commitment of these cells, possibly through ECM remodeling.
In summary, our findings demonstrate that transcription of the gene encoding the Adamts16 metalloproteinase is regulated by the Wilms tumor protein Wt1 in a sex-and tissue-specific fashion. This regulatory link has a presumed role in the development of murine kidneys and gonads.