The Wilms' Tumor Suppressor Gene WT1 Is Negatively Autoregulated*

Vol. 269,No. 8,Issue of February 25, pp. 6198-6206, 1994 Printed in U.S.A.

T H E JOURNAL OF BIOICGICAL CHEMISTRY 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc.

The Wilms’ Tumor Suppressor GeneWT1 Is Negatively Autoregulated* (Received for publication, August 13, 1993, and in revised form, November 2, 1993)

Harald D. RupprechtSO, Iain A. DrummondS, Stephen L. Maddenn, Frank J. Rauscher IIIn, and Vikas P. SukhatmeSII From the Wepartment of Medicine, Harvard Medical School and Beth Israel Hospital, Boston, Massachusetts 02215 and IThe Wistar Institute of Anatomy and Biology, Philadelphia, Pennsylvania 19107

TheWilms’tumor suppressorgene WT1 encodesa gene WT1 is alternatively spliced, giving rise to four different zinc-fingerDNA-bindingproteinthatfunctions as a transcripts, reflecting the absence or presence of two alternatranscriptional repressor. WTl is expressed in a dra- tive splice insertions (Haber et al., 1991). Splice insertion I matic spatial and temporal pattern during kidney devel- consists of a separate exon, encoding 17 amino acids, that is opment and is thought to be critical during mesenchy- inserted between the glutamine- andproline-rich amino terminus and thezinc-finger domains. Splice insertion I1 arises from mal-epithelialconversion. The WT1 protein bound multiple sites in the WT1 promoter and functionedas a the use of an alternative splice donor in exon 9 and results in powerful transcriptional repressor of its gene in vivo the insertion of 3 amino acids (lysine, threonine, and serine) (>BO-fold).The WT1 protein carrying an NH,-terminal between zinc fingers 3 and 4. The alternativesplices have been 17-amino acid insertion and a 3-amino acid insertion conserved between mouse and human genes (Buckler et al., (KTS) between zinc fingers 3 and 4, arising from the 1991). As shown by RNase protection analysis, the transcript most abundantof four alternatively spliced transcripts,containing both splice insertions is themost abundant in both was the most powerful repressor. Of importance, a sub- human andmouse species, whereas the transcript missing both set of WT1-bindingsites differs from the Egr-1 consensus splice insertions is the least common (Haber et al., 1991). sequence, which has been shown to bind onesplice varUsing recombinant proteinspanning thezinc-finger region of iant of the WT1 protein (WTl(-KTS)).We characterized WT1, Rauscher et al. (1990) demonstrated that WT1 binds to two of these sites and show that they bind both -KTS thesame sequence (GCGGGGGCG) recognized by Egr-1,a and +KTS forms ofthe WT1 zinc-finger protein and can closely related immediate early growth response zinc-finger conferrepression on aheterologouspromoter construct. Our data demonstrate that WT1, in addition to its protein (Sukhatme et al., 1988; Christy and Nathans, 1989). known effects on insulin-like growth factor 11, platelet- Cotransfection assays showed that the WT1 protein functions derived growth factor A, and Pax-2transcription, is a as a transcriptional repressorwhen bound to the Egr-1consenpowerful repressor ofits own gene. These observations sus site, and the repressor function has been mapped to the emphasize its critical role as a transcriptional regula- glutamine- and proline-rich NH2 terminus of WT1. The functional significance of WT1 DNA binding and transcriptional tory protein during normal kidney development. regulation is underscored by the observation that small deletions and point mutations in thezinc-finger region that abolish DNA binding havebeen detected in a number of Wilms’tumors Wilms’ tumor is anembryonic renal neoplasm that develops (Pelletier et al., 1991a, 1991b; Huff et al., 1991; Cowell et al., in 1in 10,000 children, occurring in both sporadic and familial 1990). forms (Matsunaga, 1981). It is thought to arise through aberThe Wilms’tumor gene is expressed in thedeveloping kidney, rant mesenchymal stem cell differentiation resulting from loss genital ridge, fetal gonads, and mesothelium (Pritchard-Jones of function of one or more tumor suppressorgenes (Cavenee et et al.,1990). Its expression inthe kidney has a dramatic spatial al., 1989). Cytogenetic deletion analysis of patients with WAGR and temporal pattern. WT1 mRNA is detected in the mesensyndrome (Wilms’ tumor, aniridia, genitourinary abnormalichyme and glomeruli of the mesonephros and in the metanephties, mental retardation) revealed that a Wilms’tumor suppresros, where its expression is most striking in the condensed sor gene lies at chromosomal position llp13 (Compton et al., blastemal cells of the renal vesicle. With glomerular matura1988; Francke et a l . , 1979; Gessler et al., 1989; Glaser et al., tion, the hybridization signal is attenuated, persistingonly in 1986). This led to the isolation of the candidate Wilms’ tumor the developing podocyte cells of the glomerular epithelium. suppressor gene WT1 (Call et al., 1990; Gessler et al., 1990) WT1 mRNA is not detected in the ureter, collecting ducts, or encoding a zinc-finger transcriptional regulatory protein. The remaining tubular segments of the nephron (Pritchard-Jones et predicted WT1 protein contains a region rich in glutamine and al., 1990). The observations suggest a primary role for WT1 in proline residues and four contiguous zinc fingers of the Cysznormal kidney development and a wider role in mesenchymalHisz class, a common DNA-binding motif. The Wilms’ tumor epithelial transitions. WT1 has recently been shown to regulate the transcription of several genes thought to be important in * This work was supported in part by National Institutes of Health kidney development: the insulin-like growth factor I1 gene, the Grants DK 45617 and DK 44921 (toV. P. S.). The costsof publication of insulin-like growth factor I receptor gene, the platelet-derived this articlewere defrayed in part by the payment of page charges. This article must thereforebe hereby marked “aduertisernent”in accordance growth factorA chain gene, and thePax-:! gene. (Drummond et al., 1992; Werner et al., 1993; Gashler et al., 1992; Wang et al., with 18 U.S.C. Section 1734 solely to indicate this fact. 9 Supported in part by the Verein zurFoerderung des Nephrolo- 1992). gischen Instituts (Nuernberg, Germany). In view of these findings, which demonstrate a criticalrole of ( 1 To whom correspondence should be addressed: Beth Israel Hospital, kidney develRenal Division,Dana 517, 330 Brookline Ave., Boston, MA 02215.Tel.: WT1 in regulating transcriptional events during opment and its own temporally limited expression, our goal 617-735-2105; Fax.: 617-278-7843.

6198

Autoregulation of the Wilms’ %mor Suppressor Gene

6199

(20,000 cpm) in the absence or presence of purified proteinat 4 “C for 30 min in binding buffer (20 m~ HEPES, pH 7.5,70 mM KCl, 12% glycerol, 0.05% Nonidet P-40, 100 p~ ZnSO,). Protein-DNA complexes were resolved on nondenaturing 5% polyacrylamide, 0.5 x Tridborate/EDTA gels at 4 “C. Dried gels were exposed overnight t o Amersham Hyperfilm-MP. DNase ZFootprinting-DNAfragments used for DNase I footprinting were the same as those used for EMSAs (probes I-IV). Single endlabeled probes were generated by PCR amplification with one labeled primer. Binding reactions were carried out exactly as described for EMSAs. 10 pl of cold 1 mM MgCl,, 0.5 mM CaCl, and 1pl of DNase I (50 pg/ml)wereadded for 2 min, and the reactions were then phenol/ chloroform-extracted and ethanol-precipitated. A/G ladders were prepared by the Maxam-Gilbert sequencing reaction as described (Sambrook et al., 1989). Samples were separated on urea-8% a polyacrylamide gel. Methylation Interference Assay-The oligonucleotides PWT-265 and MATERIALSANDMETHODS PWT-225 were annealed with their complementary strands and cloned Construction of Expression Vectors-pCMVhWT1 contains the huinto the EcoRV site of pBluescript SKII+. To create single end-labeled man WT1 cDNA driven by the cytomegalovirus immediate early pro- probes, the vectors were cut with either SmaI or SalI, alkaline phosmoter. The aminoacid insertions KTS (lysine, threonine, and serine) at phatase-treated, and labeled with T4 polynucleotide kinase. After a position 390 and VAAGSSSSVKWTEGQSN at position 248 were intro- second restriction digest, the resulting fragments were purified from a duced separately (pCMVhWT1+17AA andpCMVhWTl+KTS)orto3.5% agarosegel.The specific activity of the probeswas 400,000gether (pCMVhWT1+17AA+KTS) by site-directed mutagenesis as de- 500,000 c p d n g DNA. Preparation of methylated probe, binding reacscribed (Madden etal., 1991; Drummond et al., 1992). pCMVhWT1In tion, complex isolation, and piperidine cleavageof the complexes were TTL, a synthetic oligonucleotide containing stop codons in all three performed as described (Ausubel etal., 1989). Samples were separated reading frames was inserteda t a unique BamHI site (aminoacid 179) on a urea-8% polyacrylamide sequencing gel. a s described (Gashler et al., 1992). Construction ofReporterPlasmids-pGPmWT1 contains mouseWT1 RESULTS genomic sequences from-513 to +254 bp’ relative to the predominant WTl Negatively Regulates Its Own Promoter Actiuity-To intranscription start site. WT1 sequences were amplified by PCR from mouse genomic DNA using primers designed after the published mousevestigate whether WT1 might be subject to autoregulation, transient transfection assays were performed. Parts of the WT1 WT1 5”flanking region (-169 to +598 bp of the published sequence) (Pelletier et al., 1991~)and cloned into the HindIII site ofpGCAT-C promoter region linked to theCAT reporter gene were cotrans(Frebourg and Brison, 1988). This plasmid served as the startingpoint fected with WT1 expression vectors into 293 cells, a well charfor 5”deletion mutations. WT1 deletions at -394 and -128 bp were acterized adenovirus-transformed human fetalkidney cell line created by BspMI digest or BssHII and SalI digest, respectively, followed by religation of the vector. The deletions at -249 and -196 bp that has been shown to express low levels of WT1 mRNA (Pritchard-Jones et al. (1990)). 3 pg of plasmid pGPmWTl conwere createdby Bal-31 exonuclease digestion (Ausubel et al., 1989). The deletion end points were determinedby DNA sequencing. To create the taining mouse WT1 5’-untranslated regions from -513 to +254 3’-deletions, WT1 fragments were createdby PCR amplification of the bp relative to themajor transcription start site (Pelletieret al., promoter regions from -513 to +201 bp and -513 to +123 bp using 1 9 9 1 ~were ) cotransfected with15 pg of WT1 expression vectors pGPmWTl as a template andcloned into the HindIII site of pGCAT-C. carrying the four possible combinations of splice insertions unFor theWT1 deletion mutation spanningfrom -513 to +31 bp, 571-bp a der the control of the cytomegalovirus immediate early proSmaI fragment of pGPmWTl was cloned into the HindIII site of pGlA) or the emptyexpression vector pCB6+. CAT-C. Further digestof pGPmWT1-513+31 with BssHII and SalI and moter(Fig. religation of the vector yielded pGPmWT1-128+31. The heterologous pCMVhWTl+KTS (which contains a 3-amino acid insertion bepromoter constructs pBLPWT-265 and pBLPWT-225 were made by tween zinc fingers 3 and 4) repressed CAT activity to 45.2% of inserting the oligonucleotides PWT-265 (5”CAGCTAGGGTAG- basal levels, pCMVhWT1+17AA to 13.1%,and pCMVhWTl to CAGGGGGAGGCTTGCGGTACAGGC-3’)and PWT-225 (5”TGAGAGCACGTGGCGGGCCAGAGAGGAGGGTGTCT-3’)(Operon Technolo- 9.3%. The expression vector carrying both splice insertions (pCMVhWT1+17AA+KTS) repressed transcription to 8.2%of giesInc.,Almeda, CA), afterannealingwiththeircomplementary strands, into the BamHI site 5‘ of the herpessimplex virus thymidine basal levels (Fig. 1B). Interestingly, the WT1 transcript carrykinase promoter and CAT in pBLCAT2 (Luckow and Schiitz, 1987). ing both splice insertions is the most abundant of the four Dansfection and CATAssays-293 cells (ATCC CRL 1573) (Graham transcripts in mice and humans, accounting for -53% of the et al., 1977) were maintained in Dulbecco’s modified Eagle’s medium WT1 mRNA(Haber et al., 1991). Theobserved differences were with 10% fetalcalf serum (HyClone Laboratories). Cells were seeded at 2.5 x lo6 cells/100-mm dish the day before transfection. Transfection not due to differences in the expression levels of the various was doneby the calcium phosphate precipitation technique (Gorman et proteins since cell extracts of COS-1 cells transfected with the al., 1982) with the indicated amounts of reporter and expression plas- four vectors revealed comparable proteinlevels by immunopremids and 2 pg of pON260, an internalcontrol for transfection efficiency cipitation assays (Fig. 1C) and by Western blotting (data not with the P-galactosidase gene under control of the cytomegalovirus shown). In fact,densitometric scanning of aWestern blot immediate early promoter (Spaete and Mocarski, 1985). Results shown showed that the least effective repressor (pCMVhWTl+KTS) are averages of at least two separate experiments, with each sample was expressed at a slightlyhigher level (1.5-fold) than themost transfected in duplicate. For the assessment ofWTl protein expression, effective repressor (pCMVhWT1+17AA+KTS).An expression COS-1 cells were transfected with the WT1 expression vectors by the vector carrying anoligonucleotide insertion introducinga stop DEAE-dextranmethod,andimmunoprecipitationswere performed with a protein A-purified anti-WT-ZF antibody as previously described codon in all three reading frames (amino acid 179) served as a (Morris et al., 1991). negative control and did not significantly suppress CAT activElectrophoretic Mobility ShiftAssays (EMSAs)-WTl zinc-finger ity (Fig. 1B). When increasingamounts of the expression vector proteins were expressed in Escherichia coli as previouslydescribed pCMVhWT1+17AA+KTSwere cotransfected withthe WT1 pro(Rauscher et al., 1990). Probes for EMS& were generated by PCR amplification of cloned mouse WT1 sequences. Binding reactions were moter construct, CAT activity wasprogressively repressed (Fig. lD1. A maximum 58.8-fold repression was observed with 25 pg carried out ina total volume of 10 pl with 0.5 ng of end-labeled probe

was to investigate whether the WT1 gene is subject to autoregulation. Several linesof evidence suggested such a negative feedback loop. The presence of multiple degenerate Egr-l-binding sites (GNGNGGGNG) (Christy and Nathans, 1989; Pelletier et al., 1991c), which are potential targets for the WT1 protein (Rauscher et al., 1990), and thefinding that a subset of Wilms’ tumors shows elevated WT1 mRNA levels compared to human fetal kidney (Haber et al., 1990; Pritchard-Jones et al., 1990) suggested that a loss of function of the WT1 protein might lead toa failure todown-regulate its own gene. Our data indicate that the temporal and spatial down-regulation of WT1 in thedeveloping nephron may result in partfrom action of its own gene product on its transcriptional regulation.

of expression vector. ’ The abbreviations used are: bp, base paifis); PCR, polymerase chainMultiple BindingSitesfor WTl in Its Own Promoterreaction; CAT, chloramphenicol acetyltransferase; EMSAs, electropho- Sequence analysis of the G/C-rich WT1 promoter revealed the retic mobility shift assays; HSV, herpes simplex virus. presence of several partially overlapping, degenerate Egr-l-

6200

Autoregulation of Wilms' the

lhmor Suppressor Gene

A pGPrnWl

E

C

REPORTER-PLASMID

E'%

r [p -513

b b

"

EXPRESSION-PLASMIDS pCMVhWTl

L

On-IE

I

.