Inhibition of the DNA-binding and transcriptional repression activity of ...

Oncogene (1997) 15, 2001 ± 2012  1997 Stockton Press All rights reserved 0950 ± 9232/97 $12.00

Inhibition of the DNA-binding and transcriptional repression activity of the Wilms' tumor gene product, WT1, by cAMP-dependent protein kinasemediated phosphorylation of Ser-365 and Ser-393 in the zinc ®nger domain Yoshimasa Sakamoto1, Mitsuaki Yoshida1, Kentaro Semba1,2 and Tony Hunter2 1

Department of Cellular and Molecular Biology, The Institute of Medical Science, The University of Tokyo, 4-6-1 Shirokanedai, Minato-ku, Tokyo 108, Japan; 2Molecular Biology and Virology Laboratory, The Salk Institute for Biological Studies, 10010 North Torrey Pines Road, La Jolla, California 92037-1099, USA

The Wilms' tumor suppressor gene, WT1, encodes a transcription factor in the zinc ®nger family, which binds to GC-rich sequences and functions as a transcriptional activator or repressor. The WT1 protein plays a crucial role in urogenital development in mammals and its function is thought to be conserved during vertebrate evolution. Although accumulating evidence suggests that WT1 regulates a subset of genes including growth factor and growth factor receptor genes, little is known about regulators or signal cascades that could modulate the function of WT1. In this study, we show that the WT1 protein expressed exogenously in ®broblasts was phosphorylated in vivo, and that treatment with forskolin, which activates the cAMP-dependent protein kinase (PKA) in vivo, induced phosphorylation of additional sites in WT1. We identi®ed the forskolin-induced phosphorylation sites as Ser-365 and Ser-393, which lie in the zinc ®nger domain in zinc ®ngers 2 and 3, respectively. PKA phosphorylated WT1 at Ser-365 and Ser-393 in vitro, as well as at additional sites, and this phosphorylation abolished the DNA-binding activity of WT1 in vitro. Using WT1 mutants in which Ser-365 and Ser-393 were mutated to Ala individually and in combination, we showed that phosphorylation of these sites was critical for inhibition of DNA binding in vivo. Thus, coexpression of the PKA catalytic subunit with wild type WT1 reduced the level of WT1 DNA-binding activity detected in nuclear extracts, and decreased transcriptional repression activity in vivo. In contrast to wild type WT1, all of the phosphorylation site mutants retained signi®cant DNAbinding activity and repression activity in the presence of PKA. Analysis of the mutants showed that phosphorylation of Ser-365 and Ser-395 had additive inhibitory eects on WT1 DNA-binding in vivo and that phosphorylation at both sites was required for neutralization of repression activity. Therefore, we conclude that PKA modulates the activity of WT1 in vivo through phosphorylation of Ser-365 and Ser-393, which inhibits DNA binding. This in turn results in a decrease in WT1 transcriptional repression. Our ®ndings provide the ®rst evidence that the function of WT1 can be modulated by its phosphorylation in vivo. Keywords: tumor suppressor gene; phosphorylation; transcription factor; DNA binding

Correspondence: T Hunter Received 6 February 1997; revised 30 June 1997; accepted 30 June 1997

Introduction Wilms' tumor is a pediatric kidney tumor that occurs in 1 in 10,000 children. The ®nding that patients with Wilms' tumor, aniridia, urogenital abnormality, mental retardation (WAGR) syndrome carry a cytogenetically visible deletion of chromosome 11p13 helped in the cloning of Wilms' tumor susceptibility gene (Coppes et al., 1993; Rauscher, 1993). The WT1 gene encodes a transcription factor which has four C2H2 zinc ®ngers at the carboxyl terminus and a proline/glutamine-rich region at the amino terminus (Call et al., 1990; Gessler et al., 1990). The WT1 gene is expressed in developing kidney, gonads and other organs (Pelletier et al., 1991b; Pritchard-Jones et al., 1990). In kidney development, its expression is ®rst detected in metanephric blastemal cells condensing and aggregating around the ureteric bud, where Wilms' tumor is known to arise. Since these cells are in the process of ceasing proliferation and assuming an epithelial character, the WT1 protein is thought to be involved in the mesenchymal-epithelial dierentiation program. A variety of WT1 mutations have been found, but in only 5 ± 10% of the Wilms' tumors (Coppes et al., 1993; Hu and Saunders, 1993). In contrast, the WT1 gene is mutated in almost 100% patients with DenysDrash syndrome (Pelletier et al., 1991a). WT1 knockout mice exhibit a failure of kidney and gonad development (Kreidberg et al., 1993) con®rming that WT1 is required for urogenital development in mammals. Recent observations suggest that the function of WT1 in urogenital development has been conserved during vertebrate evolution (Kent et al., 1995; Semba et al., 1996). The four zinc ®ngers at the carboxyl terminus of the WT1 protein recognize a variety of GC-rich DNA sequences. Initial studies showed that the WT1 isoforms lacking a 3 amino acid insertion, KTS, between the third and fourth zinc ®ngers bind to a 9 bp sequence, 5'-GCGGGGGCG-3', whereas the isoforms with KTS do not (Rauscher et al., 1990). A subsequent study showed that the isoforms with KTS required three more nucleotides, GTG, for ecient binding to DNA, consistent with each zinc ®nger recognising 3 bp (Drummond et al., 1994). Recently, Nakagama et al. identi®ed a 10 bp motif 5'GCGTGGGAGT-3' that is 20- to 30- times more ecient as a WT1 binding sequence than the 5'GCGGGGGCG-3' (Nakagama et al., 1995). The WT1 protein functions as a transcriptional activator or repressor depending on the promoter used for the assay (Reddy et al., 1995a,b). Its transcriptional

Phosphorylation of WT1 Y Sakamoto et al

2002

regulatory domain at the amino terminus is divided into subdomains that activate and repress transcription, respectively (Wang et al., 1993, 1995). A number of potential target genes containing WT1-binding sites, which can be regulated by WT1 in model systems, have been identi®ed, including growth factors and growth factor receptors that are supposedly important for normal and malignant cell growth or dierentiation, such as c-myc, bcl-2 (Hewitt et al., 1995), EGR-1 (Madden et al., 1991), EGF receptor (Englert et al., 1995), IGF II (Drummond et al., 1992), IGF-I receptor (Werner et al., 1994), inhibin-a (Hsu et al., 1995), PDGF A-chain (Gashler et al., 1992; Wang et al., 1992), PAX-2 (Ryan et al., 1995), TGF-b (Dey et al., 1994) and WT1 itself (Rupprecht et al., 1994). Recently Larsson et al. demonstrated that WT1 is colocalized with splicing factors (Larsson et al., 1995). Furthermore, in addition to DNA, WT1 also binds to RNA in vitro (Caricasole et al., 1996; Kennedy et al., 1996). Thus, WT1 may play a role in posttranscriptional processing of RNA. However, little is known about upstream regulators of WT1. In this paper we report that phosphorylation by cAMP-dependent protein kinase (PKA) is one of the mechanisms through which WT1 function can be regulated. Results To test the possibility that the WT1 protein is phosphorylated, we expressed the WT1 protein by transfecting a WT1 expression vector, pCMX-WT1 into COS-7 cells. Two days later, transfected cells were labeled with 32Pi for 4 h. Before lysis, half the cells were treated for 30 min with 10 mM forskolin, which activates adenylyl cyclase in vivo. Immunocomplexes made with anti-WT1 antisera were resolved on a 10% SDS-polyacrylamide gel. A phosphorylated 52 kDa protein was observed speci®cally in WT1-transfected cells compared with vector transfected control cells (Figure 1). Tryptic peptide mapping of 32P-labeled WT1 from untreated cells revealed that WT1 contained a single major phosphopeptide (peptide d) (Figure 2). Forskolin treatment stimulated the phosphorylation of CREB 2 ± 3-fold (data not shown), but caused no consistent increase in the extent of phosphorylation of WT1 (Figure 1). However, tryptic peptide mapping analysis of 32P-labeled WT1 protein from forskolintreated cells showed that three novel phosphopeptides were present (peptides a, b and c). In contrast, phosphorylation of peptide d was unaected (Figure 2). The tryptic phosphopeptide maps of rat and human WT1 were identical. Since forskolin treatment is known to activate protein kinase A (PKA) in vivo, we searched for candidate PKA phosphorylation sites in the WT1 sequence. Only two Ser residues, Ser-365 and Ser-393 in the zinc ®nger domain, are in the strict PKA consensus sequence, RRXS/T (Pearson and Kemp, 1991). To test whether these sites can be phosphorylated by PKA, we expressed the zinc ®nger domain of WT1 (145 amino acid residues) in bacteria as a His6tagged fusion protein (His-WT1ZnF), and phosphorylated puri®ed protein with PKA in the presence of [g-32P]ATP in vitro. The peptide maps of in vitro and in vivo phosphorylated WT1 proteins were very similar to

DMSO —

WT1

forskolin —

WT1

—

—

WT1

—

— Figure 1 Phosphorylation of WT1. Either pCMX or pCMXWT1 was transfected into COS-7 cells. Transfected cells were labelled with 32Pi for 4 h. Forskolin (10 mM) or its solvent DMSO was added directly to the labeling medium and then incubated for 30 min. Cell extracts were prepared and immunoprecipitated with an anti-WT1 antiserum. The immune complexes were analysed by electrophoresis on a 10% SDS-polyacrylamide gel. The dried gel was exposed to presensitized Kodak XAR ®lm at 7708C for 6 h. Bars starting at the top indicate molecular weight markers of 92, 66, 45 and 31 kDa, respectively

each other, although there was a cluster of peptides lying between the peptides b/c doublet and peptide d in the in vitro phosphorylated sample that were not observed in the in vivo labeled sample (Figures 2 and 3a). Therefore, we used in vitro phosphorylation of WT1 protein to map the PKA phosphorylation sites and to investigate the eects of phosphorylation. To test whether Ser-365 and Ser-393 were phosphorylated, we mutated Ser-365 and Ser-393 in His-WT1ZnF to Phe, which mimics the mutation observed at Ser-365 in a pre-B cell line (Call et al., 1990) and in a rat kidney tumor induced by N-nitroso-N'-methylurea (Sharma et al., 1994). Wild type and mutant His-WT1ZnF proteins phosphorylated by PKA were subjected to tryptic peptide mapping. Peptide b was lost when Ser-365 was mutated (Figure 3b), whereas peptide c was lost when Ser-393 was mutated (Figure 3c). In contrast, peptide e


2003

a

b

a

b

c

d

c

d

e

Figure 2 Tryptic peptide mapping of phosphorylated WT1. Human or rat WT1 proteins isolated from 32P-labeled cells were excised from a 10% SDS-polyacrylamide gel and subjected to tryptic peptide mapping analysis. Electrophoresis at pH 1.9 was in the horizontal (anode on the left) and chromatography in the vertical dimension. The origin is marked with an open arrowhead. The phosphopeptides induced by forskolin treatment are marked by arrowheads. Plates were exposed to presensitized ®lm for 14 days with an intensifying screen at 7708C. (a) Human WT1, mock treated; (b) human WT1, forskolin treated; (c) rat WT1, mock treated; (d) rat WT1, forskolin treated; (e) schematic illustration of in vivo phosphopeptide map. Cerenkov c.p.m. of the samples loaded were 161 (a), 191 (b), 177 (c), 176 (d)

was lost only when both residues were mutated (Figure 3d). We eluted peptides b and c from the cellulose plates, and redigested them with trypsin. Further incubation of peptide b or c with trypsin generated peptide e, indicating that both peptides b and c were partial digestion products (data not shown). Trypsin cleavage at the 72 position is known to be inhibited by phosphorylation of serine at position 0. Therefore, we conclude that peptides b, c and e correspond to RFpSR, KFpSR and FpSR, respectively (Figure 4). To con®rm this, we used a synthetic cyclic peptide, acetylYhCERRFSRSDQLKRPhC-NH2 (the two homocysteine residues, designated as hC, are joined by a disul®de bond) which is predicted to mimic the structure of zinc ®nger 2. Phosphorylation of this peptide by PKA followed by tryptic digestion showed that it generated spot b and a small amount of spot e as expected (data not shown).

e

Figure 3 Identi®cation of forskolin-induced phosphorylation sites in WT1. Bacterially-expressed mutant and wild type HisWT1ZnF proteins were phosphorylated in the presence of PKA catalytic subunit and [g-32P]ATP, and subjected to tryptic peptide mapping analysis as described in Figure 2. The position of the origin is marked with an open arrowhead. (a) Wild type HisWT1ZnF; (b) Phe-365 mutant His-WT1ZnF; (c) Phe-393 mutant His-WT1ZnF; (d) Phe-365/393 mutant His-WT1ZnF protein; (e) schematic illustration of in vitro phosphopeptide map. For all samples 3000 Cerenkov c.p.m. were loaded. Plates were exposed to presensitized ®lm overnight at 7708C with an intensifying screen

Figure 4 Alignment of phosphorylation sites in the WT1 zinc ®nger domain. An alignment of the four zinc ®ngers in human WT1 is shown together with the residue number for the beginning and end of each zinc ®nger. Ser-365 and Ser-393 are indicated by asterisks in zinc ®ngers 2 and 3, respectively. Closed arrows and open arrows indicate complete and partial cleavage sites for trypsin, respectively. The deduced identities of peptides b, c and e are shown. The conserved cysteine and histidine residues in each zinc ®nger motif are indicated by vertical lines


2004

Since phosphorylation of WT1 by PKA occurs in the zinc ®nger domain, we tested the eect of phosphorylation on DNA-binding activity. For this purpose, His-WT1ZnF was phosphorylated by PKA in vitro and then subjected to electrophoretic mobility shift assay (EMSA) using an oligonucleotide probe which contains a high anity WT1 binding site (Nakagama et al., 1995). Incubation of WT1 with PKA and ATP for 15 min completely impaired its DNA-binding activity as determined by EMSA (Figure 5). In contrast, PKA did not have any eect in the absence of ATP, strongly suggesting that phosphorylation caused the loss of DNA-binding activity. To exclude the possibility that the WT1 protein was degraded or denatured during the kinase reaction, we dephosphorylated WT1 by treatment with bacterial alkaline phosphatase (BAP) and then performed EMSA. A short incubation with BAP removed *35% of the labeled phosphate and restored some DNA-binding activity, whereas a longer incubation removed 460% of the phosphate and restored DNA-binding activity of WT1 more eciently (Figure 6). This eect of BAP was completely inhibited by inclusion of the phosphatase inhibitors, sodium phosphate and sodium orthovanadate. Thus, we conclude that phosphorylation by PKA impairs DNA-binding activity of WT1 in a reversible fashion. However, it was not possible to deduce from our in vitro analysis whether phosphorylation at Ser-365 and Ser-393 was solely responsible for the inhibition of DNA-binding, since PKA phosphorylated other unidenti®ed sites in the zinc ®nger domain in addition to Ser-365 and Ser-393, which were not observed in vivo (see Figure 3). Indeed, preliminary analysis of the double phenylalanine His-WT1ZnF mutant showed that its DNA-binding activity was still sensitive to inhibition by PKA-mediated phosphorylation, implying that phosphorylation of residues other than Ser-365 and Ser-393 could impair DNA binding. For this reason, we decided to examine whether activation of PKA had an inhibitory eect on WT1 activity in vivo, and whether mutation of Ser-365 and Ser-393 would relieve this eect. To test whether WT1 activity was aected by activated PKA, we transfected a metallothionein promoter-driven PKA catalytic subunit expression vector with or without a WT1 expression vector into CV-1 cells. Two days later, nuclear extracts were prepared from transfected cells and the DNAbinding activity of WT1 in each nuclear extract was tested by EMSA. We observed a speci®c WT1/DNA complex, which was abolished by anti-WT1 antibodies (C19) (Figure 7a). Coexpression of the catalytic subunit of PKA reduced DNA-binding activity of WT1 to 20%, although the relative amount of WT1 was slightly increased in PKA-transfected cells (Figure 7b ± d). Using this system we then tested the eects of mutating Ser-365 and Ser-393, singly or in combination. For this experiment, we created WT1 mutants in which Ser-365 and Ser-393 were replaced with Ala, which is a conservative, nonphosphorylatable substitution. Mutant AA, in which both phosphorylation sites are mutated to Ala, did not show any reduction of DNAbinding activity upon PKA coexpression. The single A365 and A393 mutants also showed signi®cant resistance to the inhibitory eect of PKA coexpression on DNA binding, but were not as resistant as the AA mutant (Figure 7b and c). The rate of in vitro

15 min

0 min – –

– +

+ –

+ +

– –

– +

+ –

+ : ATP + : PKA

WT1

F Figure 5 Inhibition of WT1 DNA-binding activity by PKA in vitro. His-WT1ZnF was incubated in the presence or absence of PKA with or without inclusion of ATP at 308C for 0 or 15 min as described in Materials and methods. After the kinase reaction, 5 ng of His-WT1ZnF was subjected to electrophoretic mobility shift assay (EMSA). Arrows indicate DNA/WT1 complex (WT1) or free probe DNA (F)

63 % 37 % – – –+ + + – – – ++ + – + +– + + – + + –+ + – – +– – + – – + – – +

: release of phosphate : PKA : BAP : BAP inhibitors

WT1

F Figure 6 Restoration of the DNA-binding activity by phosphatase treatment. His-WT1ZnF was incubated with or without PKA and [g-32P]ATP and then bound to Ni-NTA agarose beads as described in Materials and methods. His-WT1ZnF bound on the beads was then incubated in the presence of BAP, in the presence of BAP and its inhibitors, or in the absence of BAP at 308C for 2 (left hand six lanes) or 4 h (right hand six lanes). After incubation with BAP, the fusion protein was eluted from the beads with elution buer and then subjected to EMSA. Arrows indicate DNA/WT1 complex (WT1) or free probe DNA (F)

phosphorylation of Ser-365 or Ser-393 by PKA was not aected by mutation at the other phosphorylation site (data not shown). Because the Ser-365 to a Phe mutation has been found in a pre-B cell line (Call et al., 1990) and in a rat kidney tumor induced by N-nitroso-


2005

a —

WT1

wt

mock αWT

C

—

αWT

C

Ab

c WT1

F

b mock –

+

wt –

+

A365

A393

–

–

+

+

–

AA

WT1

+

PKA

d mock –

+

wt –

A365 +

–

+

A393 –

+

AA –

WT1 +

PKA

WT1 WT1

F Figure 7 Inhibition of WT1 DNA-binding activity by PKA in vivo. (a) WT1/DNA complex recognized by anti-WT1 antibody. Nuclear extracts were prepared from WT1 (wt) or vector transfected (mock) CV-1 cells. C19 anti-WT1 antibody (lanes aWT1), control antibody (lanes C) or no antibody (lanes 7) was added to nuclear extracts prior to addition of labeled WT1 probe and analysis by EMSA. Arrows indicate DNA/WT1 complex (WT1) or free probe DNA (F). (b) Eect of mutation of Ser-365 and Ser393 to Ala. Nuclear extracts were prepared from CV-1 cells transfected with wild type WT1 (wt), mutant WT1 (lanes A365, A393 and AA) or vector (mock) with (+) or without (7) PKA expression vector, and analysed for WT1-binding activity by EMSA. (c) Quantitative analysis of inhibitory eect of PKA on WT1 DNA-binding activity. The amount of WT1/DNA complex in each lane in (b) was measured using a Fujix BAS2000 phosphorimager, and then the relative inhibition of binding activity was calculated by dividing the amount of WT1/DNA complex in the presence of PKA by that in the absence of PKA (set at 1.0). (d) Expression of WT1 in nuclear extracts. The amount of WT1 protein in nuclear extracts of the indicated cells in (b) was determined by immunoblotting with C19 anti-WT1 antibodies as described in Materials and methods


2006

N'-methylurea (Sharma et al., 1994), we made and tested a Ser-365 to Phe WT1 mutant. The F365 mutant also showed signi®cant resistance to PKA-mediated inhibition of DNA-binding activity (Figure 8a). Therefore, we conclude that phosphorylation at both sites is required for complete inhibition of WT1 DNA binding activity by PKA in vivo. Next, we evaluated the eect of PKA on WT1mediated transcriptional repression. WT1 is known to repress transcription from the PDGF-A promoter

a mock –

+

wt –

F365 +

–

+

WT1 PKA

WT1

(Gashler et al., 1992). In our system, expression of wild type WT1 repressed transcription of a PDGF-A promoter driven luciferase reporter gene *5-fold (Figure 9). A similar degree of repression was observed when the A365, A393 and AA mutant forms of WT1 were expressed (Figure 9b). According to Gasler et al., regions 7578 to 7560, 7537 to 7496, 7440 to 7420 and 798 to 749 in the PDGF-A promoter were protected from DNase I by the WT1 zinc ®nger domain. Therefore, we tested WT1 activity on mutant promoters linked to a reporter gene in which the three putative promoter-distal WT1-binding sites (7578 to 7560, 7537 to 7496, 7440 to 7420) or the three promoter-proximal WT1-binding sites (798 to 749) were mutated, respectively. This mutational analysis showed that WT1-mediated transcriptional repression of the PDGF-A promoter was sequence-dependent and required the three promoterproximal WT1-binding sites (data not shown). To test whether PKA-mediated phosphorylation of WT1 could reverse WT1 repression, increasing amounts of the PKA expression vector were cotransfected into CV-1 cells with ®xed amounts of the wild type WT1 expression vector and the PDGF-A promoter reporter construct. Repression by wild type WT1 was signi®cantly reduced at the higher PKA expression vector levels (Figure 9a). In contrast, repression by the A365, A393 and the AA mutant forms of WT1 was not inhibited by coexpression of PKA (Figure 9b). Repression by the F365 WT1 mutant was also unaected by coexpression of PKA (Figure 9c). Thus, we conclude that expression of PKA abolished both DNA-binding and the transcriptional repression activity of WT1 through phosphorylation at both Ser-365 and Ser-393 in vivo. Discussion

F

b cytoplasmic fraction mock – +

wt – +

F365 – +

nuclear fraction mock wt – + – +

F365 – +

WT1 PKA

WT1 Figure 8 Eect of the Phe-365 mutation on PKA-regulated WT1 binding activity. (a) Eect of mutation of Ser-365 to Phe. Nuclear extracts were prepared from CV-1 cells transfected with wild type WT1 (wt), Phe-365 mutant WT1 (F365) or vector (mock) with (+) or without (7) PKA expression vector and analysed for WT1-binding activity by EMSA. (b) Expression of wild type and Phe-365 mutant WT1 in nuclear and cytoplasmic extracts. The amount of wild type and Phe-365 mutant WT1 proteins in nuclear and cytoplasmic extracts was determined by immunoblotting using C19 anti-WT1 antibodies as described in Materials and methods. Nuclear cytoplasmic extracts loaded on the gel are from the same number of cells

In this paper, we show (1) WT1 was phosphorylated in vivo when expressed exogenously in ®broblasts, and that forskolin treatment, which activates PKA, induced phosphorylation at two serines, (2) the forskolininduced phosphorylation sites, which can be phosphorylated by PKA in vitro, are localized in the zinc ®nger domain mapping to Ser-365 and Ser-393 in zinc ®ngers 2 and 3, respectively, (3) phosphorylation by PKA abolished the DNA-binding activity of WT1 in vitro, (4) expression of the catalytic subunit of PKA impaired the DNA-binding activity and transcriptional repression activity of WT1 in vivo. During the preparation of our manuscript, Ye et al. (1996) reported that PKA- and PKC-mediated phosphorylation of WT1 inhibits DNA-binding activity in vitro and that transcriptional repression by WT1 is inhibited by PKA activation in vivo. Our results extend theirs by identi®cation of two speci®c PKA phosphorylation sites in the zinc ®nger domain, and by showing that the inhibitory eect of PKA on WT1 function in vivo required that both PKA phosphorylation sites be intact. There have been no prior reports of WT1 phosphorylation in vivo. Previously, Morris et al. failed to detect phosphorylation of WT1 when it was expressed exogenously in COS-1 cells in a system similar to our own (Morris et al., 1991). We do not


a

b

c

Figure 9 Requirement for both Ser-365 and Ser-393 for PKA repression of WT1 (a) Dose-dependent suppression of WT1 activity by coexpression of PKA. The activity of the PDGF-A promoter luciferase reporter plasmid in cells expressing wild type WT1 (closed bars) and the AA mutant (open bars) in the presence of the indicated amounts of PKA expression vector was compared to that in the absence of WT1 is shown. The fold activation shown in the ordinate was calculated by setting the luciferase

know the reason for this discrepancy, but the level of WT1 phosphorylation is relatively low and is hard to detect unless conditions are optimized. The in vivo phosphopeptide map of WT1 after forskolin treatment was essentially identical to the in vitro map of WT1 phosphorylated by PKA, suggesting that PKA is one of the protein kinases that is responsible for phosphorylation of WT1 in vivo. However, peptide d, which contains a site that can be phosphorylated in vitro when a high level of PKA is used, was constitutively phosphorylated and its phosphorylation did not seem to be further induced by forskolin treatment. This site may not be a good phosphorylation site even in vitro, because when we used reduced amount of PKA, we did not detect peptide d (data not shown). These results suggest that other protein kinase(s) can phosphorylate the site in peptide d. For example, Ca2+/calmodulin-dependent protein kinases also phosphorylate Ser or Thr residues that are preceded by basic amino acid residues (Pearson and Kemp, 1991). We have not mapped this site but it must lie within residues 302 ± 449, since His-WT1ZnF containing these residues generates peptide d upon PKA phosphorylation. Phosphorylation of WT1 by PKA abolished its DNA-binding activity in vitro. This reaction was reversible, because BAP treatment restored the DNA-binding activity of phosphorylated WT1. Thus, WT1 function could in principle be regulated reversibly in vivo in response to changes in cAMP levels. We tested the eect of phosphorylation on WT1 mutants in which either or both of the PKA phosphorylation site Ser residues are mutated. Using EMSA and a PDGF-A promoter luciferase reporter gene assay, we demonstrated that expression of the PKA catalytic subunit inhibited the activity of WT1 in vivo by two criteria. First, coexpression of PKA greatly decreased the DNA-binding activity of WT1 detected by EMSA without changing the amount of WT1 in the nuclear fraction. Since the PKA catalytic subunit is translocated into the nucleus, we presume that WT1 is phosphorylated in the nucleus. However, we reproducibly detected a slight increase in the level of WT1 in the cytoplasmic fraction when PKA was coexpressed (Figure 8b), in agreement with the recent report by Ye et al. (1996). This may result from partial inhibition of nuclear translocation of WT1 due to its phosphorylation. However, it should be noted that the amount of WT1 retained in the cytoplasmic fraction was very small compared to that in the

activity of the reporter plasmid in the absence of WT1 to 1.0, and then dividing the activity in the presence of WT1 by that in its absence. All the data are the means with standard deviations of three independent assays. (b) Requirement for both Ser-365 and Ser-393 for PKA repression. The fold activation of the reporter plasmid in cells expressing wild type WT1, or the A365, A393 or AA mutants in the absence (closed bars) or the presence (open bars) of PKA expression vector (5 mg) is shown. The fold activation was calculated as in (a). All the data are the means with standard deviations of three independent assays. (c) Eect of the Phe-365 mutation on PKA repression. The fold activation of the reporter plasmid in cells expressing wild type WT1 (closed bars) or the F365 mutant (open bars) in the presence of the indicated amounts of PKA expression vector is shown. Fold activation was calculated as in (a). All the data are the means with standard deviations of three independent assays

2007


2008

nuclear fraction, which did not change signi®cantly upon coexpression of PKA. Second, coexpression of PKA also inhibited the transcriptional repression activity of WT1 measured using a PDGF-A promoter luciferase reporter construct. In contrast to wild type WT1, all of the phosphorylation site mutants retained signi®cant DNA-binding activity and repression activity in the presence of PKA. Therefore, we conclude that PKA modulates the activity of WT1 in vivo through phosphorylation of Ser-365 and Ser393, which inhibits DNA binding. This in turn results in a decrease in WT1 transcriptional repression. What is the mechanism by which phosphorylation of Ser-365 and Ser-393 inhibits DNA binding? The crystal structure of the EGR-1 zinc ®nger motifs, which are similar to those of WT1, suggests that Ser-365 as well as Ser-393 could make contact with DNA through hydrogen bonding to the phosphate backbone (Pavletich and Pabo, 1991). Phosphorylation at one or both serines might be anticipated to disturb such interactions, resulting in inhibition of DNA-binding activity. We found that stoichiometric phosphorylation of WT1 by PKA in vitro at both Ser-365 and Ser-393 abolished DNA binding. The eects of mutating either Ser-365 or Ser-393 singly in preventing PKA-mediated inhibition of DNA binding in vivo were not as pronounced as that of mutation of both serines. This suggests that the presence of a single phosphate at either Ser-365 or Ser-393 weakens but does not abolish DNA binding, whereas the presence of phosphates at both sites totally inhibits DNA binding. The fact that the eects of the two phosphates on DNA binding are additive is what one might expect, given that the zinc ®ngers of this type of protein each bind DNA independently (Pavletich and Pabo, 1991). However, mutation of either Ser alone was as eective as the double mutation in counteracting the ability of PKA to reverse WT1 repression of the PDGF-A promoter. This suggests that a single phosphate at either Ser-365 or Ser-393 does not weaken DNA binding in vivo suciently to compromise the repression function of WT1. PKA-resistant WT1 mutants would be expected to be functionally dominant over wild type WT1, and thus mutation of either Ser-365 or Ser-393 in the WT1 gene would be sucient to abrogate the ability of PKA to inhibit WT1 repression in vivo. The importance of PKA-mediated phosphorylation of WT1 is underscored by the ®nding that Ser-365 is mutated to Phe in two dierent types of transformed cell, namely in N-nitroso-N'-methylurea-induced rat embryonic kidney tumors, which resemble Wilms' tumor (Sharma et al., 1994), and in a human pre-B cell line (Call et al., 1990). Analysis of Phe-365 mutant WT1 shows that it has a dierent DNA-binding preference to that of wild type WT1 (Bickmore et al., 1992). Our results provide insights into the phenotype of this mutant. PKA-dependent phosphorylation of Ser-365 would in principle be abolished by mutation of Ser-365 to any other residue. The fact that a Ser to Phe mutation is a highly nonconservative mutation that has occurred twice suggests that a Phe may have been selected for at position 365. Phe-365 may serve not only to prevent PKA-mediated phosphorylation, but may also be important in altering DNA-binding speci®city. In this connection, it is interesting to note that we have found that mutation of Ser-393 to Phe,

which is the equivalent Ser in zinc ®nger 3, reduces the DNA-binding activity of recombinant WT1 in vitro, whereas mutation of Ser-365 to Phe in zinc ®nger 2 does not (data not shown). Our experimental analysis of the Phe-365 mutant WT1 shows that the Phe-365 mutation does not aect WT1-mediated repression of the PDGF-A promoter, but is sucient to render WT1 insensitive to negative regulation by PKA. Is mutation of Ser-365 to Phe and the loss of PKA regulation important in the transformed phenotype? WT1 repression is thought to be critical for shutting down expression of growth-promoting genes; the mutational loss of WT1 activity that is observed in Wilms' tumors could prevent the shutdown of such genes, contributing to transformation. In contrast, the loss of PKA regulation through mutation of Ser-365 to Phe would, if anything, make WT1 a better repressor, and on the face of it this would not be consistent with the Phe-365 mutation playing a role in the transformed phenotype. However, the growth suppression function of WT1 is complex and seems to be cell-type dependent. All four of the WT1 splice variants suppressed the colony forming ability of RM1, a Wilms' tumor cell line, and CV-1 cells (Haber et al., 1993; Kudoh et al., 1995). In contrast, in the case of adenovirus-transformed baby rat kidney cells (Ad-BRK), expression of the splice variant lacking both alternative splices increased tumor growth rate, whereas other splice variants strongly suppressed the tumorigenic phenotype (Menke et al., 1996). Furthermore, WT1 is necessary for proliferation of some leukemic cells (Algar et al., 1996; Yamagami et al., 1996). Thus, such growth stimulatory eects of WT1 may be enhanced by the Phe-365 mutation. In addition, WT1 also has a transcriptional activator function (Wang et al., 1993), which in principle would also be downregulated by PKAdependent phosphorylation. Thus, the Phe-365 mutation could result in continued expression of genes that are activated by WT1, which would normally be downregulated by activation of PKA. A variety of hormonal signals can activate PKA in vivo in principle leading to phosphorylation of WT1, but without knowing more about which signals regulate WT1 and the nature of its target genes it is hard to determine the exact eect of the Phe-365 mutation on WT1 function. WT1 cDNAs have been cloned from various vertebrates including human, mouse, rat, chicken, alligator, Xenopus laevis and zebra®sh (Call et al., 1990; Gessler et al., 1990; Kent et al., 1995; Semba et al., 1996; Sharma et al., 1992). All these vertebrate WT1 proteins have Ser residues corresponding to both Ser-365 and Ser-393. Thus, functional regulation of WT1 by PKA-dependent phosphorylation may be conserved during vertebrate evolution. In fact, the phosphopeptide map of rat WT1 was identical to that of human (Figure 2). A search for the corresponding serine residues in the Genbank database showed that Ser-365 and the preceding two basic residues are conserved in the ®rst zinc ®nger of the early growth response gene (EGR) family, EGR-1 and EGR-2 (Joseph et al., 1988; Suggs et al., 1990). The members of the Spl family, Spl, Sp2, Sp3 and Sp4 also have a similar sequence, KRFT in their second zinc ®nger instead of RRFS in WT1 (Hagen et al., 1992;


Kadonaga et al., 1987; Kingsley and Winoto, 1992). Whether these corresponding Ser or Thr residues can be phosphorylated and aect DNA-binding activity remains to be tested. It will obviously be important to determine when and where PKA-mediated phosphorylation of WT1 occurs. For this purpose, antibodies that speci®cally recognize only the phosphorylated form of WT1 will be invaluable. It will also be important to identify extracellular signals that modulate the phosphorylation state of the WT1 protein. The WT1 gene is highly expressed in granulosa and Sertoli cells (Hsu et al., 1995; Pelletier et al., 1991b). These cells respond to follicle-stimulating hormone (FSH) with an increase in cAMP. FSH is known to play a key role in the menstrual cycle. In this cycle, granulosa cells in a group of primordial follicles start to grow under the in¯uence of FSH. FSH acting together with luteinizing hormone (LH), causes follicle cells to produce increasing amounts of estrogen, which regulates further progression through the menstrual cycle. FSH and LH, which activate the cAMP/PKA pathway, also increase ovarian inhibin expression and secretion, which in turn suppresses pituitary FSH secretion (see e.g. Aloi et al., 1995). Since inhibin-a is one of the putative WT1 target genes (Hsu et al., 1995), FSH may regulate the expression of inhibin-a in part by suppressing the activity of WT1 through phosphorylation.

Materials and methods Materials To construct the WT1 expression vector, pCMX-WT1, the 2 kbp ClaI-DraI fragment of pCMVhWT1(7/7), which contains the entire coding sequence of the human WT1 cDNA without either of the two alternative splices (Madden et al., 1991) was cloned between the ClaI and SmaI sites of pBluescript SK(+) to make pSK-WT1. pCMX-WT1 was then generated by cloning the KpnI ± XbaI (followed by treatment with the Klenow fragment of E. coli DNA polymerase I) fragment of pSK-WT1 between the KpnI and EcoRV sites of pCMX (Umesono et al., 1991). Since Ser-365 was mutated to Phe in the original WT1 cDNA in pCMVhWT1, we corrected this residue to Ser by site-directed mutagenesis using Kunkel's method (Kunkel et al., 1987). The PDGF-A promoter luciferase reporter plasmid (pPDGF-A-luc) was constructed by cloning a SacI ± HindIII fragment of pACCAT~SacI (Gashler et al., 1992) between the SacI and HindIII sites of pGL2-Basic vector (Promega). The expression vector, pRSV-x was constructed by cloning a 0.6 kbp NdeI (followed by treatment with Klenow fragment)-HindIII fragment of pRSVneo (Gorman, 1985) between SpeI (followed by treatment with Klenow fragment)-HindIII sites of pCMX. pRSV-hWT1 was then generated by cloning 2 kbp HindIII fragment of pRcCMVWT1 (from Dr T Akiyama, Osaka Univ.) into HindIII site of pRSV-x. For in vitro phosphorylation experiments, we generated a (histidine)6-tagged WT1 zinc ®nger protein (His-WT1ZnF) expression vector, which expresses residues 302 ± 449 of human WT1 and has six histidine residues at its amino terminus, by inserting the 1 kbp A¯III (followed by treatment with Klenow fragment)-BamHI fragment of pCMX-WT1 between the BamHI (followed by treatment with Klenow fragment) and BglII sites of pRSETC (Invitrogen). The fusion protein was puri®ed using a Ni-NTA column (Qiagen) according to Rauscher et al., 1990). Mutation of Ser-365 and

Ser-393 to Ala or Phe was accomplished by oligonucleotidemediated site-directed mutagenesis (Kunkel et al., 1987). To prepare antisera against WT1, a glutathione-Stransferase (GST)-WT1 fusion protein expression vector was constructed by inserting the 220 bp NcoI ± SacI fragment of rat WT1 into the vector pGEX-KG (Guan and Dixon, 1991). This vector expresses residues 182 ± 256 of rat WT1 (Sharma et al., 1992). This amino acid sequence is identical to both mouse and human WT1 proteins except that residue 250 in human WT1 is Met in rat WT1. The GST-WT1 fusion protein was expressed and puri®ed using glutathione-agarose (Sigma), and then used to immunize rabbits. Before immunoprecipitation of WT1 from 32P labeled cells, we con®rmed that this antiserum recognized in vitro translated WT1 protein and that its activity was blocked by an excess of GST-WT1 but not by GST alone. The C19 anti-WT1 antibody, made against the C-terminal 19 residues, was purchased from Santa Cruz Biotechnology, Inc. DNA transfection, metabolic labeling and immunoprecipitation COS-7 cells (from Jackie Dyck, Salk Institute) and CV-1 cells were grown in DMEM with 10% calf serum (CS) or 10% fetal calf serum (FCS), respectively. One6105 cells per 60 mm dish were seeded one day prior to DNA transfection. Ten mg DNA per dish was transfected by calcium phosphate-mediated transfection (Gorman et al., 1982). Twenty four h after transfection, transfected cells were shifted to DMEM with 0.5% CS. The next day, cells were metabolically labeled by incubating with 1 mCi/ml 32Pi (ICN) in phosphate-free DMEM including 15 mM HEPES (N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid) pH 7.4 plus 0.5% dialyzed CS for 4 h. Forskolin (10 mM) (Sigma) was added directly to the labeling medium and then incubated for 30 min. Following forskolin treatment, cells were lysed in 200 ml of sodium dodecyl sulfate (SDS)-boiling lysis buer (10 mM sodium phosphate pH 7.2, 0.5% SDS, 1% Trasylol (Miles, Inc., West Haven, CN), 1 m M EDTA, and 1 mM dithiothreitol (DDT)) and then boiled for 5 min. After dilution of the lysates with 800 ml of RIPA (-SDS) buer (1% NP40, 1% sodium deoxycholate, 0.15 M NaCl, 2 mM EDTA, 10 mM sodium phosphate, pH 7.2, 1% Trasylol, 50 mM sodium ¯uoride, 100 mM sodium orthovanadate) lacking SDS, immunoprecipitation of WT1 with anti GST-WT1 antiserum was performed according to Meisenhelder and Hunter (1992). The immune complexes were solubilized in Laemmli sample buer and analysed by eletrophoresis on a 10% SDS-polyacrylamide gel. Tryptic peptide mapping analysis To prepare the in vitro phosphorylated WT1 proteins, the four His-WT1ZnF proteins (1 mg) were phosphorylated by 0.5 mg of recombinant PKA catalytic subunit (kindly provided by Susan Taylor) in 10 ml kinase buer (20 mM Tris-HCl pH 7.5, 100 mM NaCl, 12 mM MgCl2, 1 mM DTT) in the presence of 100 mM ATP and 10 mCi of [g-32P]ATP at 308C for 60 min. The speci®c activity of the PKA was 25 mmol/min/mg. The reactions were terminated by the addition of 10 ml of 26Laemmli sample buer and then resolved by 10% SDS-polyacrylamide gel. 32P-labeled WT1 proteins were excised from the SDS-polyacrylamide gel and then subjected to peptide mapping analysis according to Boyle et al. (1991). Tryptic peptide samples were electrophoresed for 25 min at 1.0 kV in pH 1.9 buer using the HTLE7000 apparatus (CBS Scienti®c, Del Mar, CA); the plates were air dried and then placed in tanks for ascending chromatography using phosphochromo buer. After ascending chromatography, the plates were air dried and then exposed to presensitized Kodak XAR ®lm with an intensifying screen at 7708C for 14 days (in vivo) or for 2 days (in vitro).

2009


2010

Electrophoretic mobility shift assay Fifteen ng of each oligonucleotide, 5'-GATCCGCCCACTCCCACGCGCCGGG-3' and 5'-GATCCCCGGCGCGTGGGAGTGGGCG-3' were annealed to generate a high anity WT1-binding site based on that reported by Nakagama et al. (1995). This oligonucleotide contains one WT1 binding site which replaces the EGR-1/Zif268 binding site in the EGR-1/Zif268 5' ¯anking region (Christy and Nathans, 1989). The annealed fragment was then endlabeled by incubating in 10 ml of the labeling buer (50 mM Tris-HCl pH 7.6, 10 mM MgCl2, 5 mM DTT) containing two units of T4 polynucleotide kinase (NEB) and 50 mCi of [g-32P]ATP (Amersham) at 378C for 30 min. After dilution to 100 ml with TE (10 mM Tris-HCl pH 7.6, 1 mM EDTA), the labeled fragment was puri®ed through Sephadex G-50 (Pharmacia Biotech). The His-WT1ZnF protein (120 ng) was incubated in 25 ml of kinase buer (20 mM Tris-HCl pH 7.5, 100 mM NaCl, 12 mM MgCl2, 1 mM DTT) in the presence or absence of 100 mM ATP and 0.5 mg of recombinant PKA at 308C for 0 or 15 min. After the reaction, 95 ml of the binding buer (25 mM HEPES-KOH pH 7.5, 100 mM KCl, 10 mM ZnSO4, 0.1% NP40, 1 mM DTT, 5% glycerol and 1mg of poly (dI-dC). poly (dI-dC)) was added to terminate the reaction. Five ml of this mixture (5 ng of His-WT1ZnF) and end-labeled probe 10 000 c.p.m.) were incubated in 10 ml binding buer at room temperature for 20 min. Protein-DNA complexes were resolved on a nondenaturing 5% polyacrylamide gel (acrylamide/bisacrylamide ratio 30:1) in 0.56Tris/borate/ EDTA buer at 48C. To prepare nuclear extracts from WT1-transfected cells, CV-1 cells seeded at 66105 per 90 mm dishes were transfected with 20 mg of pRSV-hWT1 and 20 mg of pMT-PKA, a metallothionein promoter-driven PKA catalytic subunit expression vector (from Dr Marc Montminy, Salk Institute). Four hours after transfection, the medium was changed to DMEM containing 10% FCS. After additional 20 h, the medium was changed to DMEM containing 0.5% FCS. The next day, nuclear extracts were prepared basically according to Andrews and Faller (1991). Transfected cells were washed twice with PBS(7), and then lysed in 200 ml of modi®ed buer A (10 mM HEPES-KOH pH 7.9, 1.5 mM MgCl2, 10 mM KC1, 0.5 mM DTT, 0.05% NP40, 0.2 mM PMSF, 2 mg/ml aprotinin, 1 mg/ml leupeptin, 1 mg/ml pepstatin, 25 nM staurosporine) on ice for 10 min. Samples were centrifuged to precipitate crude nuclei. Supernatants (crude cytoplasmic extracts) were saved for immunoblotting. The nuclear fraction was suspended in 100 ml of modi®ed buer C (20 mM HEPESKOH pH 7.9, 25% glycerol, 700 mM NaCl, 1.5 mM MgCl2, 0.5 mM DTT, 0.2 mM PMSF, 2 mg/ml aprotinin, 1 mg/ml leupeptin, 1 mg/ml pepstatin, 25 nM staurosporine) and then incubated on ice for 30 min. Supernatants (crude nuclear extracts) were stored in aliquots at 7708C. For EMSA, 2 ml of nuclear extract and end-labeled probe (10 000 c.p.m.) were incubated in 15 ml of the binding buer II (50 mM HEPESKOH pH 7.5, 50 mM KCl, 5 mM MgCl2, 10 mM ZnSO4, 1 mM DTT, 20% glycerol and 1 mg of poly (dI-dC)⋅poly (dI-dC)) on ice for 30 min. For inhibition of WT1/DNA complex formation by antibody, 200 ng of C19, anti-WT1 antibody or the same amount of an irrelevant antibody was added prior to additon of probe. Protein-DNA complexes were resolved on a nondenaturing 5% polyacrylamide gel (acrylamide/ bisacrylamide ratio, 30:1) in 0.56Tris/borate/EDTA buer at room temperature (Nakagama et al., 1995). After electrophoresis, gels were dried and then exposed to Kodak XAR ®lm for 16 h with an intensifying screen at 7708C.

mixtures were saved in the tube containing 190 ml of the binding buer. Sixty ml of 50% suspension of Ni-NTA agarose beads (Qiagen) and 240 ml of the binding buer without DTT were then added to the reaction mixture (total volume was 500 ml), which was incubated on a rotary shaker for 1 h at 48C. His-WT1ZnF bound to Ni-NTA agarose beads was recovered by the brief centrifugation, washed three times with binding buer without DTT and ®nally suspended in 150 ml of the binding buer without DTT. Fifty ml of these suspensions, which contained 6 mg of His-WT1ZnF, were incubated in 65 ml of the binding buer with or without two units of bacterial alkaline phosphatase (BAP) (Pharmacia Biotech) and phosphatase inhibitors, 100 mM sodium phosphate and 1 mM sodium orthovanadate at 308C for 2 or 4 h. After the BAP reaction, the beads were washed three times and then the fusion proteins were eluted with binding buer containing 250 mM imidazole. Two percent of the eluted fraction (equivalent to 120 ng His-WT1ZnF, if 100% of the protein was recovered) was subjected to EMSA. To monitor the dephosphorylation, an identical kinase reaction but containing 25 mCi of [g-32P]ATP was performed in parallel. Luciferase assays CV-1 cells were seeded at 36105 cells per 60 mm dish the day before transfection, and then transfected by calcium phosphate coprecipitation method with 1 mg of pPDGF-Aluc, 10 mg of pRSV-hWT1 or pRSV-x and the indicated amount of pMT-PKA. pMT-PKA mutant vector which expresses kinase-negative PKA was added to adjust the total amount of DNA to 21 mg. Four hours after transfection, the medium was changed to DMEM containing 10% FCS. After an additional 20 h, the medium was changed to DMEM containing 0.5% FCS. The next day, cells were washed with PBS(7), and lysed in Lysis Buer (Luciferase Assay System, Promega). Luciferase activity of each extract was measured with Luciferase Assay System (Promega) by using a Berthold Lumat luminometer, LB9501. The values obtained were normalized for protein concentration measured with Bio-Rad Protein Assay reagent (Bio-Rad). In our assay system, coexpression of 5 mg of PKA, for example, resulted in 4 ± 6-fold increase of reporter activity, even though no canonical cAMP response element (CRE) is present in the PDGF-A promoter fragment used in the reporter plasmid. For this reason, the fold activation was calculated by setting the luciferase activity of the reporter plasmid in the absence of WT1 to 1.0, and then dividing the activity in the presence of WT1 by that in its absence. Immunoblotting Crude nuclear extracts (5 ml) or cytoplasmic extracts (10 ml), which correspond to the same number of cells, were resolved on a 10% SDS-polycrylamide gel and then blotted onto Immobilon (Millipore). After blocking with skim milk, the membrane was incubated with 5% skim milk and C19 anti-WT1 antibody diluted 1:400 (Sant Cruz Biotechnology, Inc.) in PBS(7) for 1 h at room temperature. After three washes with PBS(7) containing 0.05% Tween 20, the membrane was incubated with 5% skim milk and horse radish peroxidase-conjugated protein A diluted 1:4000 (Amersham) in PBS(7) for 1 h at room temperature. After three washes with PBS(7) containing 0.05% Tween 20, WT1 protein was detected by the ECL detection system (Amersham).

Phosphatase treatment Twenty mg of His-WT1ZnF was incubated in 200 ml kinase buer containing 100 mM ATP with or without 10 mg of PKA at 308C for 3 h. Ten ml aliquots of the reaction

Acknowledgements We thank Nobumoto Watanabe, Jill Meisenhelder and other members of the Hunter and Eckhart labs at the Salk


Institute and in the Department of Cellular and Molecular Biology, IMSUT for helpful discussions, Steve Koerber and Carl Hoeger (Salk Institute) for the design and synthesis of the cyclic WT1 zinc ®nger peptide, Tetsuhiro Kudoh and Tetsu Akiyama (Osaka University) for helpful discussions and for providing WT1 expression vectors, Saraswati Sukumar (Johns Hopkins School of Medicine) for providing the rat WT1 cDNA, Susan Taylor (UC San Diego) for recombinant PKA catalytic subunit and Marc

Montminy (Salk Institute) for the PKA expression vector. KS was a recipient of the National Cancer InstituteJapanese Foundation for Cancer Research Training Program, and was supported in part by the J Aron Charitable Foundation, Inc. KS is greatly indebted to his parents for their continuous support and encouragement. TH is an American Cancer Society Research Professor. This work was supported by USPHS grants CA14195 and CA39760.

References Algar EM, Khromykh T, Smith SI, Blackburn DM, Bryson GJ and Smith PJ. (1996). Oncogene, 12, 1005 ± 1014. Aloi JA, Dalkin AC, Schwartz NB, Yasin M, Mann B, Haisenleder DJ and Marshall JC. (1995). Endocrinology, 136, 1227 ± 1232. Andrews NC and Faller DV. (1991). Nucleic Acids Res., 19, 2499. Bickmore WA, Oghene K, Little MH, Seawright A, van Heyningen V and Hastie ND. (1992). Science, 257, 235 ± 237. Boyle WJ, van der Geer P and Hunter T. (1991). Meth. Enzymol., 201, 110 ± 149. Call KM, Glaser T, Ito CY, Buckler AJ, Pelletier J, Haber DA, Rose EA, Kral A, Yeger H, Lewis WH, Jones C and Housman DE. (1990). Cell, 60, 509 ± 520. Caricasole A, Duarte A, Larsson SH, Hastie ND, Little M, Holmes G, Todorov I and Ward A. (1996). Proc. Natl. Acad. Sci. USA, 93, 7562 ± 7566. Christy B and Nathans D. (1989). Proc. Natl. Acad. Sci. USA, 86, 8737 ± 8741. Coppes MJ, Campbell CE and Williams BR. (1993). FASEB J., 7, 886 ± 895. Dey BR, Sukhatme VP, Roberts AB, Sporn MB, Rauscher FJ and Kim SJ. (1994). Mol. Endocrinol., 8, 595 ± 602. Drummond IA, Madden SL, Rohwer-Nutter P, Bell GI, Sukhatme VP and Rauscher FJ. (1992). Science, 257, 674 ± 678. Drummond IA, Rupprecht HD, Rohwer-Nutter P, LopezGuisa JM, Madden SL, Rauscher FJ and Sukhatme VP. (1994). Mol. Cell. Biol., 14, 3800 ± 3809. Englert C, Hou X, Maheswaran S, Bennett P, Ngwu C, Re GG, Garvin AJ, Rosner MR and Haber DA. (1995). EMBO J., 14, 4662 ± 4675. Gashler AL, Bonthron DT, Madden SL, Rauscher FJ, Collins T and Sukhatme VP. (1992). Proc. Natl. Acad. Sci. USA, 89, 10984 ± 10988. Gessler M, Poustka A, Cavenee W, Neve RL, Orkin SH and Bruns GA. (1990). Nature, 343, 774 ± 778. Gorman C. (1985). High eciency gene transfer into mammalian cells. DNA cloning: a practical approach. Oxford, IRL Press Limited. Gorman CM, Moat LF and Howard BH. (1982). Mol. Cell. Biol., 2, 1044 ± 1051. Guan KL and Dixon JE. (1991). Anal. Biochem., 192, 262 ± 267. Haber DA, Park S, Maheswaran S, Englert C, Re GG, Hazen-Martin DJ, Sens DA and Garvin AJ. (1993). Science, 262, 2057 ± 2059. Hagen G, Muller S, Beato M and Suske G. (1992). Nucleic Acids Res., 20, 5519 ± 5525. Hewitt SM, Hamada S, McDonnell TJ, Rauscher FJ and Saunders GF. (1995). Cancer Res., 55, 5386 ± 5389. Hsu SY, Kubo M, Chun SY, Haluska FG, Housman DE and Hsueh AJ. (1995). Mol. Endocrinol., 9, 1356 ± 1366. Hu V and Saunders GF. (1993). Biochim. Biophys. Acta, 1155, 295 ± 306. Joseph LJ, Le Beau MM, Jamieson GA, Acharya S, Shows TB, Rowley JD and Sukhatme VP. (1988). Proc. Natl. Acad. Sci. USA, 85, 7164 ± 7168.

Kadonaga JT, Carner KR, Masiarz FR and Tjian R. (1987). Cell, 51, 1079 ± 1090. Kennedy D, Ramsdale T, Mattick J and Little M. (1996). Nature Genet., 12, 329 ± 331. Kent J, Coriat A-M, Sharpe PT, Hastie ND and van Heyningen V. (1995). Oncogene, 11, 1781 ± 1791. Kingsley C and Winoto A. (1992). Mol. Cell. Biol., 12, 4251 ± 4261. Kreidberg JA, Sariola H, Loring JM, Maeda M, Pelletier J, Housman D and Jaenisch R. (1993). Cell, 74, 679 ± 691. Kudoh T, Ishidate T, Moriyama M, Toyoshima K and Akiyama T. (1995). Proc. Natl. Acad. Sci. USA, 92, 4517 ± 4521. Kunkel TA, Roberts JD and Zakour RA. (1987). Meth. Enzymol., 154, 367 ± 382. Larsson SH, Charlieu JP, Miyagawa K, Engelkamp D, Rassoulzadegan M, Ross A, Cuzin F, van Heyningen V and Hastie ND. (1995). Cell, 81, 391 ± 401. Madden SL, Cook DM, Morris JF, Gashler A, Sukhatme VP and Rauscher FJ. (1991). Science, 253, 1550 ± 1553. Meisenhelder J and Hunter T. (1992). Mol. Cell. Biol., 12, 5843 ± 5856. Menke AL, Riteco N, van Ham RC, de Bruyne C, Rauscher FJ, van der Eb AJ and Jochemsen AG. (1996). Oncogene, 12, 537 ± 546. Morris JF, Madden SL, Tourney OE, Cook DM, Sukhatme VP and Rauscher FJ. (1991). Oncogene, 6, 2339 ± 2348. Nakagama H, Heinrich G, Pelletier J and Houssman DE. (1995). Mol. Cell. Biol., 15, 1489 ± 1498. Pavletich NP and Pabo CO. (1991). Science, 252, 809 ± 817. Pearson RB and Kemp BE. (1991). Meth. Enzymol., 200, 62 ± 81. 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 DE. (1991a). Cell, 67, 437 ± 447. Pelletier J, Schalling M, Buckler AJ, Rogers A, Haber DA and Housman D. (1991b). Genes Dev., 5, 1345 ± 1356. Pritchard-Jones K, Fleming S, Davidson D, Bickmore W, Porteous D, Gosden C, Bard J, Buckler A, Pelletier J, Housman D, van Heyningen V and Hastie N. (1990). Nature, 346, 194 ± 197. Rauscher FJ. (1993). FASEB J., 7, 896 ± 903. Rauscher FJ, Morris JF, Tournay OE, Cook DM and Curran T. (1990). Science, 250, 1259 ± 1262. Reddy JC, Hosono S and Licht JD. (1995a). J. Biol. Chem., 270, 29976 ± 29982. Reddy JC, Morris JC, Wang J, English MA, Haber DA, Shi Y and Licht JD. (1995b). J. Biol. Chem., 270, 10878 ± 10884. Rupprecht HD, Drummond IA, Madden SL, Rauscher FJ and Sukhatme VP. (1994). J. Biol. Chem., 269, 6198 ± 6206. Ryan G, Steele-Perkins V, Morris JF, Rauscher FJ and Dressler GR. (1995). Development, 121, 867 ± 875. Semba K, Saito-Ueno R, Takayama G and Kondo M. (1996). Gene, 175, 167 ± 172. Sharma PM, Bowman M, Yu BF and Sukumar S. (1994). Proc. Natl. Acad. Sci. USA, 91, 9931 ± 9935.

2011


2012

Sharma PM, Yang X, Bowman M, Roberts V and Sukumar S. (1992). Cancer Res., 52, 6407 ± 6412. Suggs SV, Katzowitz JL, Tsai-Morris C and Sukhatme VP. (1990). Nucleic Acids Res., 18, 4283. Umesono K, Murakami KK, Thompson CC and Evans RM. (1991). Cell, 65, 1255 ± 1266, Wang ZY, Madden SL, Deuel TF and Rauscher FJ. (1992). J. Biol. Chem., 267, 21999 ± 22002. Wang ZY, Qiu QQ, and Deuel TF. (1993). J. Biol. Chem., 268, 9172 ± 9175.

Wang ZY, Qiu QQ, Gurrieri M, Huang J and Deuel TF. (1995). Oncogene, 10, 1243 ± 1247. Werner H, Rauscher FJ, Sukhatme VP, Drummond IA, Roberts CT and LeRoith D. (1994). J. Biol. Chem., 269, 12577 ± 12582. Yamagami T, Sugiyama H, Inoue K, Ogawa H, Tatekawa T, Hirata M, Kudoh T, Akiyama T, Murakami A and Maekawa T. (1996). Blood, 87, 2878 ± 2884. Ye Y, Raychaudhuri B, Gurney A, Campbell CE and Williams BRG. (1996). EMBO J., 15, 5606 ± 5615.