Activation of the Wt1 Wilms' Tumor Suppressor Gene by NF-Кb

Mohammed Dehbi1, John Hiscott2 and Jerry Pelletier1,3

1Department of Biochemistry and 3Department of Oncology, McGill University, 3655 Drummond St., MontreÂal, QueÂbec, Canada, H3G 1Y6, and 2Lady Davis Institute for Medical Research, 3755 CoÃte Ste-Catherine, Montreal, QueÂbec, Canada, H3T 1E2

The Wilm's tumor suppressor gene, wt1, is expressed in the glutamine/proline rich amino (N-) terminus of a very de®ned spatial-temporal fashion and plays a key WT1, whereas the second alternatively spliced exon role in development of the urogenital system. Trans- results in the insertion of three amino acids, KTS, acting factors governing wt1 expression are poorly between the third and the fourth zinc ®ngers (for a de®ned. The presence of putative kB binding sites within review, see Reddy and Licht, 1996). WT1 has several the wt1 gene prompted us to investigate whether key structural features of a transcription factor such as members of the NF-kB/Rel family of transcription nuclear localization, a proline/glutamine N-terminal factors are involved in regulating wt1 expression. In rich region, and a DNA-binding domain consisting of transient transfection assays, ectopic expression of p50 four zinc ®ngers. Three of the four zinc ®ngers (II, III and p65 subunits of NF-kB stimulated wt1 promoter and IV) share extensive homology with those of the activity 10 ± 30-fold. Deletion mutagenesis revealed that early growth response (EGR) family of transcription NF-kB responsiveness is mediated by a short DNA factors and consequently, WT1 can bind to, and fragment located within promoter proximal sequences of regulate, the expression of numerous genes through the major transcription start site. Two kB-binding sites the GC-rich egr-1 consensus binding site are present in this region and form speci®c complexes (5'GCGGGGGCG3'), as well as through a TC-rich with puri®ed NF-kB proteins, as revealed by electro- site (5'TCC3')n (Rauscher et al., 1990; Wang et al, phoretic mobility gel shift assays. Ectopic expression of 1993). Genes harboring these elements are often p50 and p65 resulted in increased transcription of the involved in the regulation of cellular proliferation and endogenous wt1 gene, as revealed by nuclear run-on dierentiation (Reddy and Licht, 1996). experiments. Taken together, these results indicate that The wt1 gene is expressed in a very de®ned spatial- members of the NF-kB/Rel family are important for temporal manner being expressed mainly in the kidney, activating expression of wt1 and reside upstream of the gonads, spleen, and mesothelium (Pritchard-Jones et regulatory cascade leading to wt1 activation. al., 1990; Pelletier et al., 1991c). In the developing kidney, the gene is expressed during the mesenchymal Keywords: WT1; NF-kB; gene regulation; Wilms' to epithelial transition, suggesting that wt1 plays a tumor critical role during this developmental step (Pritchard- Jones et al., 1990; Pelletier et al., 1991c). Consistent with this idea is the failure of this system to develop in wt1-null mice (Kreidberg et al., 1993), as well as in Introduction many individuals with wt1 germline mutations (Pelletier et al., 1991a,b). Insights into the regulation Wilms' tumor (WT) or nephroblastoma is a renal of wt1 expression are thus essential for understanding malignancy of complex pathology that aects children, the role that WT1 plays during normal and abnormal usually before the age of 5 years and with an incidence development. The upstream regulatory regions of the of *1/10 000 (Matsunaga, 1981). The tumor is human and mouse wt1 genes are GC-rich and do not thought to originate from metanephric blastemal cells contain a CAAT or a TATA box (Pelletier et al., which have failed to undergo normal dierentiation, 1991c, Homann et al., 1993). Several ubiquitous and and instead, undergo neoplastic transformation with tissue-speci®c transcription factors, such as Sp1, WT1, subsequent progression to tumor formation. Most GATA-1, Pax-2 and Pax-8, have been shown to cases of WTs are unilateral and sporadic, however, a modulate the expression of the wt1 promoter in small number are bilateral and hereditary (5 ± 10%) transient transfection assays (Rupprecht et al., 1994; (Matsunaga, 1981). Mutations in the wt1 Wilms' tumor Wu et al., 1995; Dehbi and Pelletier, 1996; Dehbi et al., suppressor gene, residing at chromosome 11p13, are 1996; Cohen et al., 1997). associated with initiation of this malignancy (Haber et The NF-kB/Rel family of transcription factors play al., 1991; Park et al., 1993). a key regulatory role in the expression of a variety of

The wt1 encodes a zinc ®nger protein of the Cys2- genes involved in several biological processes such as

His2 family and consists of 10 exons which are immune or in¯ammatory responses, growth, differen- alternatively spliced to produce four distinct mRNA tiation, and development (Baldwin, 1996; Verma et al., species. The ®rst alternatively spliced exon (#5) results 1995). Members of this family are characterized by the in the inclusion or exclusion of 17 amino acids from presence of a highly conserved domain of 300 amino acids at the N-terminal region called the Rel homology domain (RHD) and consisting of a DNA-binding Correspondence: J Pelletier, McIntyre Medical Sciences Building, Rm motif, a dimerization domain, and a nuclear localiza- 902, 3655 Drummond St., McGill University, Montreal, Quebec, tion signal. The prototype of this family, referred to as Canada, H3G 1Y6 Received 12 September 1997; revised 24 November 1997; accepted 26 NF-kB, consists of two subunits called p50 (NF-kB1) November 1997 and p65 (RelA) which form homo- and heterodimers. Transcriptional activation of wt1 by NF-kB MDehbiet al 2034 Members of the NF-kB transcription factor family speci®c cis-acting elements. In order to de®ne these, we have been identi®ed in various organisms, ranging generated a series of deletion mutants within the wt1 from ¯ies to mammals (Baldwin, 1996; Verma et al., promoter and examined the ability of p50 and p65 to 1995). In mammals, other members of this family activate wt1/CAT expression in K562 cells. All include p52 (NF-kB2), rel B, and the proto-oncogene promoter constructs contained the major wt1 tran- c-Rel. In most cells, NF-kB proteins are sequestered in scription initiation site (Figure 1a). In general, the the cytoplasm in a latent form by virtue of their transcriptional response of the deletion mutants association with the inhibitory protein, IkB. However, HindIII-CAT, SpeI-CAT, EcoRI-CAT, +23-CAT, stimulation of these cells by a wide variety of inducers +34-CAT, +118-CAT, and +261-CAT were similar including cytokines such as TNF-a and IL-1, mitogens, (Figure 2b, 20 ± 60-fold activation in the presence of and DNA damaging agents, leads to the rapid NF-kB). The highest response obtained was with phosphorylation and degradation of IkB, thus allow- +261-CAT (460-fold). Only construct +261/+360- ing nuclear translocation of NF-kB and modulation of CAT showed a transactivation response lower (only the appropriate target genes (Beg et al., 1993; Brown et ®vefold) than the parental plasmid, SalI-CAT, indicat- al., 1993; Sun et al., 1993). ing that this region does not confer full responsiveness The presence of putative kB binding sites to NF-kB (Figure 2b). These results suggest that the (5'GGGRNNYYCC3') in the human and murine wt1 region mediating NF-kB responsiveness is located promoter suggested to us that members of the NF-kB within a 173 bp region between nucleotides +318 family of transcription factors may be involved in and +491 (relative to the ®rst transcription initiation regulating wt1 expression. In this report, we demon- site). Similar results were also obtained in 293 and strate that overexpression of p50 and p65 strongly NIH3T3 cells (data not shown). transactivate the murine wt1 promoter. Likewise, ectopic expression of p50 and p65 leads to stimulation Identi®cation of NF-kB responsive elements of transcription of the endogenous wt1 gene, as revealed by nuclear run-on experiments. To identify the NF-kB responsive elements residing between nucleotides +318 and +491 of the wt1 promoter, we transferred a wt1 fragment spanning Results nucleotides +261 to +491 into a TATA-less, enhancer-less, version of pJFCAT1 (Figure 3a). Site- The murine wt1 promoter contains three transcription directed mutagenesis was then performed to eliminate start sites (Pelletier et al., 1991c). Of these, a major start each kB site identi®ed within the murine wt1 promoter site is conserved between human and mouse promoters, by visual inspection (Figure 1a). The NF-kB respon- and in the mouse occurs approximately 341 nucleotides siveness of these reporters in K562 cells is shown in downstream of the ®rst start site. Visual inspection of Figure 3b. Construct +261/pJFCAT1 was still kB the murine wt1 5' regulatory sequences lead us to responsive, although the fold activation was not as identify two potential decanucleotide kB-binding sites strong as with the enhancer containing expression present in the proximal promoter region: kB.d vector, +261-CAT (compare the response in Figure 3b (+376GGGGCTCTCC+385) and kB.u (+326GG- to that obtained with +261-CAT in Figure 2b). Site- GGGCAGCC+335) (Figure 1a). In the human wt1 directed or deletion mutagenesis of the kB.d site gene, the sequence of the kB.d site is conserved, whereas (+261-mkB.d/pJFCAT1 and +261-DkB.d/pJFCAT1, there is a single nucleotide dierence in the kB.u site respectively) essentially abolished NF-kB responsive- (5'GAGGGCAGCC3') (Homann et al., 1993). ness, whereas mutagenesis of the kB.u site had no eect (Figure 3b). These experiments identify the kB.d site as the functional kB responsive element. Transactivation of the wt1 promoter by ectopic expression of NFkB Interaction of p50 and p65 with kB.d Given that kB elements mediate activation by the large family of NF-kB/Rel transcription factors (Baldwin, To assess whether kB.d could interact directly with 1996), we tested whether members of NF-kB could NF-kB, we performed a series of mobility shift assays modulate activity of the wt1 promoter. To this end, using oligonucleotides containing either kB.u or kB.d expression vectors for p50 (NF-kB1) and p65 (relA) as probes in the presence of puri®ed GST-P50 and were co-transfected, in conjunction with SalI/CAT, in GST-p65 proteins (Figure 4). An oligonucleotide K562, 293, and NIH3T3 cells. As shown in Figure 1b containing the immunoglobulin kB enhancer served and c, coexpression of p50 and p65 reproducibly as a speci®c competitor and an oligonucleotide stimulated SalI-CAT expression 15 ± 20-fold relative containing a mutated copy of the Ig kB enhancer to pCAT. The fold activation observed with SalI-CAT served as a negative control for non-speci®c competi- was as strong as, if not better, than that observed with tion. An oligonucleotide containing the kB.d element the positive control IgkB-TATA-CAT, containing formed a single complex with NF-kB (Figure 4, lane immunoglobulin kB enhancer sites upstream of the b- 1). This complex is speci®c since it can be competed globin TATA box (Figure 1b, c). Consistent with the with excess of authentic wild-type oligonucleotide speci®city of the observed response being dependent on (compare lane 2 to 1) or an oligonucleotide containing expression of NF-kB, IgmkB-TATA-CAT, which an Ig kB site (compare lane 4 to 1). Oligonucleotides contains a mutated kB binding site, did not respond with mutated kB sites were not capable of competing to NF-kB (Figure 1b, c). NF-kB/kB.d complexes (compare lanes 3 and 5 to 1). These results indicate that p50 and p65 are capable An oligonucleotide containing the kB.u element was of activating transcription of the wt1 promoter through also capable of forming a speci®c complex with p50 Transcriptional activation of wt1 by NF-kB MDehbiet al 2035 a

b 0.8 +/– 0.6 1 1 1 Fold Activation: 15.8 +/– 0.8 10.4 +/– 1.0

CMV/∆: + – ++–– CMV/p50 + CMV/p65: –––+++

Reporter plasmids: l-CAT

Sal B-TATA-CAT B-TATA-CAT κ µκ lg lg

Figure 1 Transactivation of the murine wt1 promoter by NF-kB. (a) Schematic representation of SalI-CAT and Ig kB-TATA-CAT reporter constructs used in this study. The murine wt1 promoter is delineated by a thick line, the CAT open reading frame by an open box, and transcription initiation sites by right-angled arrows. The relative positions of the putative NF-kB sites are represented by a darkened box (kB.u) and a hatched box (kB.d). The two kB sites in Ig kB-TATA-CAT and IgmkB-TATA-CAT are represented by patched or dark stippled boxes, respectively. The TATA box is symbolized by a shaded box. Nucleotide numbering is approximate with respect to the position of the ®rst transcription initiation site. (b) Autoradiograph of a representative TLC for CAT assays in which 293 cells were transfected with SalI-CAT, or TATA-CAT containing kB sites, and CMV/p50 and CMV/p65. The fold CAT activation is indicated above the lanes and the plasmids used in the transfections shown below the each reaction. (c) Summary of NF-kB responsiveness of SalI-CAT in 293, K562, and NIH3T3 cells. Values are representative of at least three separate experiments. Reporter constructs are indicated below the bar graphs. Open boxes signify response of each reporter to CMV/D, whereas blackened boxes signify responses to p50 and p65. CAT activity was normalized to b-galactosidase and the fold activation was calculated by expressing the conversion value obtained for each reporter in the presence of NF-kB, to the CAT conversion value obtained for that same reporter in the absence of NF-kB, which was set at 1 Transcriptional activation of wt1 by NF-kB MDehbiet al 2036 a a

Figure 2 Transactivation of wt1/CAT promoter deletion mutants by NF-kB. (a) Schematic representation of wt1-CAT reporter constructs used in this study. The murine wt1 promoter is delineated by a thick line, the CAT open reading frame by an open box, and transcription initiation sites by right-angled arrows. The relative positions of the putative NF-kB sites are represented by a darkened box (kB.u) and a hatched box (kB.d). (b) Summary of NF-kB responsiveness of wt1-CAT reporters in K562 cells. Values are representative of at least three separate experiments. Reporter constructs are indicated below the bar graphs. Open Figure 3 Deletion mapping of NF-kB responsive elements in a boxes signify response of each reporter to CMV/D, whereas heterologous context. (a) Schematic representation of hetero- blackened boxes signify responses to p50 and p65. CAT activity logous wt1/pJFCAT1 fusions used in this study. The kB.u and was normalized to b-galactosidase and is expressed relative to the kB.d are represented by blackened and hatched boxes, respec- value obtained with CMV/D and SalI-CAT, which was set at 1 tively. Mutated kB.u and kB.d sites are represented by patched boxes. The CAT open reading frame is represented by an open box and the wt1 transcription initiation site is represented by a right-angled arrow. (b) Identi®cation of kB responsive element. Reporter plasmids used are indicated below the bar graphs. CAT and p65 (lanes 6 ± 10). However, probe titration values were normalized relative to b-galactosidase activity. The experiments indicated that the eciency of complex absence or presence of p50 and p65 expression vectors in the transfection mixture is symbolized by an open or darkened box, formation was at least 20-fold lower than with kB.d respectively. Each experiment has been reproduced a minimum of (MD, data not shown). These results are consistent three times and is expressed relative to that obtained with CMV/D with the idea that kB.d mediates the majority of kB- and +261/pJFCAT1, which was set at 1 responsiveness described above.

Activation of the endogenous wt1 promoter by NF-kB were generated in response to increase plasmid levels in To investigate whether ectopic expression of NF-kB the transfection mixture, indicating the linearity of the could modulate the activity of the endogenous wt1 assay. Taken together, these results suggest that promoter, we performed nuclear run-on transcription expression from the endogenous wt1 promoter can be analysis in NIH3T3. In this experiment, cells were positively modulated by the p50 and p65 subunits of transiently transfected with p65 alone or, p50 and p65, NF-kB. and nuclei were isolated and analysed for the expression of wt1. As shown in Figure 5, expression of p65 alone did not aect wt1 expression. Expression Discussion of NF-kB in NIH3T3 cells produced a 3 ± 4-fold increase in the transcription rate of wt1 (Figure 5). The analysis described in this report indicates that the Control hybridizations with RSV-LacZ and GAPDH wt1 promoter can be strongly transactivated by p50 indicated that this increase was not due to dierences and p65 subunits of NF-kB, and that this eect is in transfection eciency. Hybridization to p50 cDNA mediated by a binding site positioned immediately indicated that increased amounts of p50 transcripts downstream of the major transcription start site. This Transcriptional activation of wt1 by NF-kB MDehbiet al 2037 Probe: κB.d κB.u B B B B B.d B.u κ κ DNA competitor: µκ µκ B.d B.u – – lg lg lg lg µκ µκ (100x) κ κ (5 ug + 5 ug) (10 ug + 10 ug) CMV/p50+CMV/p65 CMV/p65 CMV/p50+CMV/p65 RSV-LacZ

wt1–/–

lacZ

GAPDH

p50

Figure 5 Nuclear run on experiments demonstrating the transcriptional response of wt1 to NF-kB. The nature of the immobilized plasmid on the ®lter paper is indicated to the right of the panel. Expression plasmids introduced by transfection into NIH3T3 cells are shown to the top of the panel. Following 12 3 4 5 67 8 910 hybridization and washing, the blot was exposed to Kodak X- Figure 4 Electrophoretic Mobility Shift Assays between NF-kB Omat ®lm for 48 h at 7708C with an intensifying screen and kB.d and kB.u. EMSA were performed as described in the Materials and methods. The nature of the DNA probe and competitor is indicated above each panel responsive to its own gene product, and this is thought to down-regulate the wt1 response in vivo (Rupprecht et al., 1994). A hematopoietic-speci®c site speci®cally interacts with puri®ed p50 and p65 as enhancer is located in the 3' end of the human wt1 assessed by EMSA and conforms to the kB consensus gene, greater than 50 kilobases (kb) downstream from sequence (5'GGGRNNYYCC3'). In contrast, the kB.u the promoter (Fraizer et al., 1994). The hematopoietic- element shares 80% similarity with the kB consensus speci®c transcription factor, GATA-1, has been shown sequence, cannot eciently form a complex with p50 to transactivate wt1 expression through this enhancer, and p65, and is not capable of mediating a kB response implying that GATA-1 plays a role in the regulation of when placed in a heterologous context. This site, unlike wt1 during hematopoiesis (Wu et al., 1995). the kB.d site, is not conserved between murine and It is unlikely that NF-kB complex expression is human wt1 5' regulatory sequences. We have also required for initiation of wt1 expression in the demonstrated the in vivo responsiveness of the developing urogenital system. Unlike wt1 null mice endogenous wt1 promoter to exogenously added p50 which show severe abnormalities of the heart, and p65, utilizing nuclear run-on assays (Figure 5). The diaphragm, and urogenital system (Kreidberg et al., intensity of the response obtained in vivo, is likely to be 1993), p50 null mice show no development abnormal- an underestimation, given that the nuclear run-ons ities, although they exhibit multiple defects in speci®c were performed on nuclei from transiently transfected and non-speci®c immune responses involving B cells and that the wt1 gene is usually not transcription- lymphocytes (Sha et al., 1995). Mice lacking p65 are ally active in NIH3T3 cells. Our results do not allow us embryonic lethal at 15 ± 16 days gestation and exhibit to infer conclusions regarding the role of additional kB massive degeneration of the liver by apoptosis (Beg et sites, not identi®ed in the current study, on in vivo wt1 al., 1995), but no reported renal or gonadal regulation. abnormalities. It is possible that other members of The wt1 gene shows a very de®ned spatial and the NF-kB/Rel family can compensate for p50 or p65 temporal expression in the developing urogenital function although in our hands, c-rel was unable to system, and is essential for mesenchymal-epithelial activate wt1 expression (MD, data not shown). transitions in this system (Pritchard-Jones et al., 1990; Alternatively, NF-kB regulation of wt1 expression Pelletier et al., 1991c). In addition, wt1 is also may be primarily restricted to the hematopoietic expressed in the hematopoietic system, although its system, in which case some the defects observed in role in this system is not de®ned (Reddy and Licht, p50 null mice could be due to depressed levels of WT1. 1996). Our studies, coupled with those of other groups, Studies on the regulation of wt1 gene expression has indicate that regulation of the wt1 promoter is been performed in several human hematopoeitic cell complex. Two developmentally regulated paired-box lines. Several investigators have noticed a marked family members, Pax-2 and Pax-8, have been shown to downregulation of wt1 in K562 and HL60 cells during activate wt1 expression through a 38 bp conserved dierentiation induced by TPA (Phelan et al., 1994; motif present upstream of the ®rst transcription start Sekiya et al., 1994). Interestingly, stimulation of HL60 site (Dehbi and Pelletier, 1996; Dehbi et al., 1996). In cells by TPA interferes with NF-kB activity (Hohmann addition, multiple sites for Sp1 are present throughout et al., 1992). Our studies have not addressed the the wt1 promoter, and this factor has been shown to possible regulation of the human wt1 gene by NF-kB, transactivate wt1 expression (Homan et al., 1993; although sequence comparison of the wt1 5' regulatory Cohen et al., 1997). The wt1 promoter is also sequences between human and murine promoters Transcriptional activation of wt1 by NF-kB MDehbiet al 2038 revealed that kB.d element is conserved. Moreover, we 5' ATAAGCTTGTCGACACACCGAGGGGACTCATTA- note the presence of 2 kB-like sequences in the human C3', +34; 5'ATAAGCTTGT CGACCTTACCTAGACGC- promoter, arranged as a tandem repeat between CTTGCC3', +118; 5'ATAAGCTTGTCGACGCTGAGA- nucleotides 7189 and 7166, upstream of the ®rst GCACGTGGCCG3', +261; 5'ATAAGCTTGTCGACG- transcription initiation site (5'GGGTGTCTCCGAGA CCGAGCATCCTGGCTCC3', +318; 5'ATAAGCTTT- GGGAAGCTGGGGGC3' respectively, in conjunction GGGACGCTCC3', kB sites are underlined) (Hofmann with the antisense oligonucleotide: 5'GGAAGATCTTCC- et al., 1993). This organization is similar to the 2 kB GGTTCCTGGGCTGCCGTCC3'. The same strategy was elements found in the HIV-1 LTR where both sites used to generate +261/+360-CAT, by using the +261 have been shown to be important for NF-kB activation sense oligonucleotide (indicated above) and the +360 (Nabel and Baltimore, 1987). Whether the kB sites antisense oligonucleotide; 5'GGAAGATCTTGAACTCC- present in the human wt1 promoter act in similar TTACCCCAGTGCT3'. The resulting products were fashion remains to be established, as does the potential digested with HindIII and BglIIandinsertedintothe relationship between antagonism of NF-kB in differ- HindIII-BamHI sites of pCAT-Enhancer (Dehbi and entiating HL60 cells and wt1 downregulation. Pelletier, 1996). The sequence integrity of each clone was The data presented herein raise several questions directly con®rmed by sequencing (Sanger et al, 1977). Plasmid constructs carrying only the proximal region of about the signi®cance of wt1 activation by NF-kBand the wt1 promoter (+261/+491) driving expression of the whether inducers of NF-kB can modulate the CAT reporter gene, in an enhancer-less context, were expression of wt1 gene. We have tested whether the generated by cloning the +261/+491 PCR product into the cytokine TNF-a, a potent inducer of NF-kB, can SalI-BglII sites of pJFCAT1 (Fridovich-Keil et al., 1991). activate wt1 gene expression. Although we have failed Plasmids +261-mkB.d/pJFCAT1 and +261-mkB.u/pJFCAT1 to detect an eect of TNF-a on transcription of the were made by overlap PCR mediated mutagenesis and endogenous wt1 promoter in 293, K562, and NIH3T3 changed the upstream (5'GGGGCA3') or downstream cells, we did observe a 4 ± 5-fold stimulation of +261- (5'GGGGCT3') kB core sequences to an XbaI site CAT expression in these same cells following treatment (5'TCTAGA3'). The kB.d sequence was deleted by PCR with TNF-a (MD, data not shown). We note that the using primers to loop out the kB.d site (along with ¯anking sequence [5'AGGCAGCGCCCACACCCGGGGCTCTCCG- kB response obtained with +261-CAT however, was CAACCC3']) to generate plasmid +261-D kB.d/pJFCAT1. dependent on the presence of an SV40 enhancer Plasmids CMV/p50 (NF-kB1) and CMV/p65 (relA) have element within the parental plasmid, suggesting been previously described (Hiscott et al., 1993). possible synergy between NF-kB and trans-acting factor(s) on the enhancer (MD, data not shown). Whether a similar situation is also occurring in vivo at Cell culture, transfections, and CAT assay the wt1 gene remains to be determined. The reasons for Two hundred and ninety-three kidney cells, NIH3T3 the discrepancy between the in vivo and in vitro ®broblast, and erythroleukemic K562 cells, were grown at experiments with TNF-a are not clear, but may relate 378Cina-MEM, DMEM, and RPMI 1640, respectively. to the chromatin context of the endogenous wt1 gene Media was supplemented with 10% heat-inactivated fetal in our tissue culture assays, which prevents activation calf serum, penicillin, and streptomycin. For transactiva- by endogenous levels of NF-kB. tion assays, K562, 293, and NIH3T3 cells were transfected It is also possible that NF-kB complexes are with 10 mg of reporter plasmid, and 1.15 mg of each CMV/ p50 (NF-kB1) and CMV/p65 (relA) expression vector, or involved in sustaining the wt1 activation response, 2.3 mg of the empty expression vector, CMV/D (Dehbi and possibly involving synergy with other transcription Pelletier, 1996). Transfection eciency was controlled by factors. Such cooperativity has recently been described cotransfecting with 3 mg of pRSV-lacZ. Thirty-six to forty- by Perkins et al. (1993), who demonstrated that four hours post-transfection, cells were harvested and activation of the HIV-1 LTR depends on an assayed for b-galactosidase and CAT activity (Gorman, interaction between p65 and Sp-1. Transactivation of 1985). Regions of the TLC plate containing 14C-acetylated the wt1 promoter is also positively regulated by Sp-1 and unacetylated chloramphenicol were isolated by (Cohen et al., 1997; Homan et al., 1993). Whether scraping and were quantitated by counting in a liquid transactivation of wt1 by NF-kB complexes is Sp-1 scintillation counter. dependent remains to be determined. In summary, p50 and p65 subunits of NF-kB are potent inducers of wt1 Electrophoretic mobility shift assays (EMSA) activity. The role that such a response plays in The p50 and p65 subunits were expressed as GST-fusion regulating wt1 activity, either in the urogenital and/or proteins and puri®ed from bacterial lysates by anity hematopoeitic system, remains to elucidated. chromatography using glutathione-Sepharose 4B beads as recommended by the manufacturer (Pharmacia). Twenty- ®ve nanograms of each puri®ed protein was used in EMSA Materials and methods with oligonucleotides derived from the murine wt1 promoter and consisted of either downstream [kB.d (+376/+385)] or upstream [kB.u (+326/+335)] se- Plasmid constructs quences. Probes were labeled by ®lling in the ends with The wt1-chloramphenicol acetyl transferase (CAT) reporter the Klenow fragment of DNA polymerase I in the presence plasmids, SalI-CAT, SpeI-CAT, HindIII-CAT and EcoRI- of a-[32P]dCTP (43000 Ci/mmol; Dupont-New England CAT have been previously described and contain varying Nuclear). EMSA were performed essentially as described portions of the wt1 promoter fused to the CAT gene previously (Dehbi and Pelletier, 1996). In competition (Figure 1a) (Dehbi and Pelletier, 1996). The reporter experiments, proteins were preincubated with an excess of plasmids +23-CAT, +34-CAT, +118-CAT, +261-CAT, unlabeled wild-type (kB.d, kB.u or IgkB) or mutated +318-CAT, and +261/+360-CAT were generated by (mkB.d, mkB.u or IgmkB) oligonucleotides for 15 min, PCR-mediated mutagenesis using SpeI-CAT plasmid as prior to addition of labeled probe. The following template with the following sense oligonucleotides: +23; oligonucleotides were used in competition assays: kB.d Transcriptional activation of wt1 by NF-kB MDehbiet al 2039 (5'TCGACCCGGGGCTCTCCGCA3'), mkB.d (5'TCGAC- both alternatively spliced exons], RSV-LacZ, GAPDH, and CCGGCGCTCTCCGCA3'), kB.u (5'TCGAGCTGGGGG- CMV/p50 cDNAs immobilized onto Nytran membranes CAGCCAGG3'), mkB.u (5'TCGAGCTGGCCGCAGCC- (Schleicher & Schuell), and hybridized with 1.06106 AGG3'), IgkB(5'TCGAGTGGGGACTTCCTCT3'), and c.p.m./ml of radiolabeled transcripts for 48 h at 658C. IgmkB(5'TCGAGTGGCGACTTCCTCT3') [mutated nu- Blots were washed to a stringency of 0.16SSC/0.1% SDS cleotides are underlined]. at 558C for 30 min and exposed against X-OMAT (Kodak) ®lm at 7708C for 48 h. Nuclear run-on transcription assay Nuclear run-on transcription assays were performed from NIH3T3 cells transiently transfected with 10 mg CMV/p65, 10 mgCMV/D,or5mg each of CMV/p50 and CMV/p65. Acknowledgements Dierences in transfection eciency were monitored by co- We would like to thank Nabeel Bardeesy for his transfecting RSC-LacZ. Forty-eight hours after transfec- stimulating discussions and critical reading of the manu- tion, nuclei (46107/assay) were isolated and resuspended in script. MD holds a Terry Fox fellowship from the National 100 ml of run-on transcription buer containing a-[32P]UTP Cancer Institute of Canada. This work is supported by (6000 Ci/mmol, Dupont-NEN) as described previously grants from the National Cancer Institute and Kidney (Dehbi et al., 1996). Slot blots were prepared with 2 mg Foundation of Canada to JP. JP is a Medical Research of linearized WT1 (7/7) [the murine wt1 cDNA lacking Scientist.

References

Baldwin Jr AS. (1996). Ann. Rev. Immunol., 14, 649 ± 681. Pelletier J, Bruening W, Li FP, Haber DA, Glaser T and Beg AA, Finco TS, Nantermet PV and Baldwin AS. (1993). Housman DE. (1991b). Nature, 353, 431 ± 434. Mol. Cell. Biol., 13, 3301 ± 3310. Pelletier J, Schalling M, Buckler A, Rogers A, Haber DA and Beg AA, Sha WC, Bronson RT, Ghosh S and Baltimore D. Housman D. (1991c). Genes & Dev., 5, 1345 ± 1356. (1995). Nature, 376, 167 ± 170. Perkins ND, Edwards NL, Duckett CS, Agrano AB, Brown K, Park S, Kanno T, Franzoso G and Siebenlist U. Schmid RM and Nabel GJ. (1993). EMBO J, 12, 3551 ± (1993). Proc. Natl. Acad. Sci., 90, 2532 ± 2536. 3558. Cohen HT, Bossone SA, Zhu G, McDonald GA and Phelan SA, Lindberg C and Call KM. (1994). Cell, Growth, & Sukhatme VP. (1997). J. Biol. Chem., 272, 2901 ± 2913. Di., 5, 677 ± 686. DehbiM,GhahremaniM,LechnerM,DresslerGand Pritchard-Jones, Fleming S, Davidson D, Bickmore W, Pelletier J, (1996). Oncogene, 13, 447 ± 453. Porteous D, Gosden C, Bard J, Buckler A, Pelletier J, Dehbi M and Pelletier J. (1996). EMBO J., 15, 4297 ± 4306. Housman D, van Heyningen V and Hastie N. (1990). Fraizer GC, Wu Y-J, Hewitt SM, Maity T, Ton CCT, Hu V Nature, 346, 194 ± 197. and Sauders GF. (1994). J. Biol. Chem., 269, 8892 ± 8900. Rauscher III FJ, Morris JF, Tournay OE, Cook DM and Fridovich-Keil JL, Guda JM, Bouvard Bryan L and Pardee Curran T. (1990). Science, 250, 1259 ± 1262. AB. (1991). Biotech, 11, 572 ± 579. Reddy JC and Licht JD. (1996). Biochem. Biophys. Acta Gorman C. (1985). in DNA cloning,eds.GloverDM(IRL 1287, 1±28. Press Ltd., Oxford) Vol II, pp. 143 ± 190. Rupprecht HD, Drummond IA, Madden SL, Rausher III FJ Haber DA, Buckler AJ, Glaser T, Call KM, Pelletier J, Sohn and Sukhatme VP. (1994). J. Biol. Chem., 269, 6198 ± RL, Douglass EC and Housman DE. (1991). Cell, 61, 6206. 1257 ± 1269. Sanger F, Nicklen S and Coulson AR. (1977). Proc. Natl. HiscottJ,MaroisJ,GaroufalisJ,D'AddarioM,RoulstonA, Acad. Sci. USA, 74, 5463 ± 5467. Kwan I, Pepin N, Lacoste J, Nguyen H, Bensi G and Sekiya M, Adachi M, Hinoda Y, Imai K and Yachi A, Fenton M. (1993). Mol. Cell. Biol., 13, 6231 ± 6240. (1994). Blood, 83, 1876 ± 1882. Homann W, Royer HD, Drechsler M, Schneider S and Sha WC, Liou H-C., Toumanen EI and Baltimore D. (1995). Royer-Pokora B. (1993). Oncogene, 8, 3123 ± 3132. Cell, 80, 321 ± 330. Hohmann H-P, Remy R, Aigner L, Brockhaus M and van Sun S-C, Ganchi PA, Ballard DW and Greene WC. (1993). Loon APGM. (1992). J. Biol. Chem., 267, 2065 ± 2072. Science, 259, 1912 ± 1915. Kreidberg JA, Sariola H, Loring JM, Maeda M, Pelletier J, Verma IM, Stevenson JK, Schwarz EM, Van Antwerp D and Housman D and Jaenisch R. (1993). Cell, 74, 679 ± 691 Miyamoto S. (1995). Genes & Dev., 9, 2723 ± 2735. Matsunaga E. (1981). Hum. Genet., 57, 231 ± 246. Wang Z-Y, Qui Q-Q, Enger KT and Deuel TF. (1993). Proc. Nabel G and Baltimore D. (1987). Nature, 326, 711 ± 713. Nat. Acad. Sci. USA, 90, 8896 ± 8900. Park S, Bernard A, Bove K, Sens DA, Hazen-Martin DJ, Wu Y-J, Frasier GC and Saunders G. (1995). J. Biol. Chem., Julian Garvin A and Haber DA. (1993). Nature Genet. 5, 270, 5944 ± 5949. 363 ± 367. Pelletier J, Bruening W, Kashtan CE, Mauer SM, Manivel JC, Striegel JE, Houghton DC, Junien C, Habib R, Fouser L, Fine RN, Silverman BL, Haber DA and Housman D. (1991a). Cell, 67, 437 ± 447.