Mdm2 inhibition of p53 induces E2F1 transactivation via p21 - Nature

ã

Oncogene (2002) 21, 4414 ± 4421 2002 Nature Publishing Group All rights reserved 0950 ± 9232/02 $25.00 www.nature.com/onc

Mdm2 inhibition of p53 induces E2F1 transactivation via p21 Mark Wunderlich1 and Steven J Berberich*,1 1

Wright State University, Department of Biochemistry and Molecular Biology, 3640 Colonel Glenn Hywy, Dayton, Ohio, OH 45435, USA

The transcription factor E2F1 functions as a key regulator for both cell-cycle progression and apoptosis. Mdm2, a major cellular regulator of the p53 tumor suppressor protein, is also closely involved in cell cycle and apoptosis. In addition to regulation of p53, Mdm2 has been reported to stimulate E2F1 transactivation by a mechanism that remains unclear. Here we examined how overexpression of Mdm2 alters E2F1/DP1 transactivation. Using a set of cell lines with diering p53 and Rb status we determined that Mdm2 induction of E2F1 transactivation was p53-dependent, resulting from release of repression by p53. While Mdm2 association with p53 was required to increase E2F1 transactivation, Mdm2 mediated degradation of p53 was not. p53 repression of E2F1 transactivation required a functional DNA binding and transactivation domain. Consistent with Mdm2 activation of E2F1 via an inhibition of p53 transactivation we demonstrate a concomitant reduction in p21 protein levels with Mdm2 overexpression. Furthermore, E2F1 repression by an Rb-phosphorylation mutant could not be reversed by Mdm2 overexpression. Mdm2 was also unable to enhance E2F1 transactivation in Mouse embryo ®broblasts lacking p21. Taken together, these results suggest that Mdm2 activation of E2F1 occurs through the repression of p53-dependent transcription of p21, a p53-target gene and cyclin dependent kinase inhibitor. Oncogene (2002) 21, 4414 ± 4421. doi:10.1038/sj.onc. 1205541 Keywords: Mdm2; E2F1; p53; p21

Introduction E2F1 is a member of a transcription factor family that regulates genes involved with cell proliferation, dierentiation and apoptosis. During G0 and G1, Rb binds to E2F/DP heterodimers and represses transactivation from responsive promoters by recruiting repressive factors (Cress and Nevins, 1994). One level of regulation of E2F transactivation occurs through its

*Correspondence: SJ Berberich; E-mail: [email protected] Received 28 December 2001; revised 27 March 2002; accepted 2 April 2002

association with hypophosphorylated pRB. When bound to pRB, the E2F/DP complex is unable to activate transcription. Late in G1, as cells approach S phase, cyclin-dependent kinases trigger phosphorylation of pRB which in turn causes release of E2F/DP heterodimers, activating cell cycle progression genes such as c-myc, DHFR, and cyclin E (Dyson, 1998). While pRB represents a major regulator of E2F1 transactivation, overexpression of the Mdm2 protein has been reported to activate E2F1 transactivation (Martin et al., 1995; Xiao et al., 1995). p53, an inducer of mdm2 expression, has also been reported to inhibit E2F1 transactivation via direct association (O'Connor et al., 1995). The mdm2 gene was initially described as the second of three genes found ampli®ed on murine double minute chromosomes isolated from a spontaneously transformed BALBc/3T3 cell line (Cahilly-Snyder et al., 1987). Deregulated expression of mdm2 can immortalize primary cells and transform certain cell lines (Fakharzadeh et al., 1991). Subsequently, the Mdm2 protein was shown to associate with the p53 tumor suppressor protein (Momand et al., 1992), downregulating p53 transactivation and triggering p53ubiquitin mediated proteosome degradation (Haupt et al., 1997; Kubbutat et al., 1997). The human homolog of mdm2, hdm2 has been found ampli®ed in approximately 7% of all human tumors, many of which possess wild-type p53 (Momand et al., 2000). While Mdm2 overexpression likely induces cellular transformation through Mdm2 association with p53, the ability of Mdm2 to stimulate the E2F1/DP1 transcription complex suggested that Mdm2 may function in tumor formation at multiple levels. While E2F1 overexpression can transcriptionally induce cell cycle progression in many quiescent cells (Johnson et al., 1993), E2F1 can also trigger apoptosis in manners that are both dependent and independent of p53 (Agah et al., 1997; Holmberg et al., 1998; Hunt et al., 1997; Nip et al., 1997; Qin et al., 1994; Wu and Levine, 1994). One p53-independent mechanism of E2F1 induced apoptosis may result from its ability to activate the p53-related gene, p73 (Irwin et al., 2000; Stiewe and Putzer, 2000). Although E2F1 overexpression can directly activate p73 and a number of proapototic genes (Muller et al., 2001), it has also been shown under certain conditions that the E2F1 transactivation domain is not required to induce apoptosis (Hsieh et al., 1997; Phillips et al., 1997). In

Mdm2 alters E2F1 transactivation via p53 M Wunderlich and SJ Berberich

4415

cells harboring wild-type p53, E2F1 overexpression can also induce apoptosis through its transactivation of the ARF gene locus, which in turn activates p53 transactivation through ARF sequestration of Mdm2 (Sherr, 1998). Additionally, E2F1 and p53 are capable of forming a physical complex capable of blocking the transactivation capacity of both transcription factors (O'Connor et al., 1995) while simultaneously stabilizing p53 protein levels (Nip et al., 2001). Like p53, E2F1 protein levels are also stabilized following DNA damage (Blattner et al., 1999). Taken together, the important interactions between E2F1 and p53 led us to examine the eect of Mdm2 overexpression on E2F1 transactivation. Results To compare the eects of Mdm2 overexpression on E2F1 transactivation in cells with varying p53 and Rb status, the indicated cell lines were transiently transfected with a synthetic E2F1 plasmid reporter (3XE2FLUC) and either an empty vector (CMV) or mdm2 expression vector. Mdm2 induced E2F1 transactivation in human U2OS osteosarcoma and human U87 glioblastoma cells, both of which contain wildtype p53 and Rb (Figure 1a). Induction of E2F1 transactivation has been previously reported using U2OS cells thus the data from these transfections support the conclusion that Mdm2 overexpression results in increased E2F1 activity (Martin et al., 1995; Xiao et al., 1995). However, Mdm2 overexpression had no signi®cant eect on E2F transactivation in either human H1299 (p53 null, Rb wild-type) non-small cell lung carcinoma, human C33A (p53 mutant, Rb mutant) cervical carcinoma, or p537/7, mdm27/7 mouse embryo ®broblasts (2K0) suggesting that Mdm2 activation of E2F1 transactivation may be cell type dependent (Figure 1a). One possibility consistent with both the data presented in Figure 1 and that of Xiao et al. (1995) is that Mdm2 stimulation of E2F1 transactivation requires wild-type p53 and/or Rb. To test whether Mdm2 induction of E2F1 was dependent on the presence of wild-type p53 protein, a series of Mdm2 mutants were expressed in U2OS cells and their activity on E2F1 transactivation determined. As predicted from the data in Figure 1, all Mdm2 proteins that were capable of binding to p53, were capable of inducing E2F1 transactivation (Figure 2). Only Mdm2(p76) was unable to aect E2F1 transactivation. Mdm2(p76) encodes amino acids 50 ± 489 and was previously shown to be unable to associate with p53 (Haines et al., 1994). Since Mdm2 proteins capable of associating with p53 activated E2F1 transactivation in cell lines harboring wild-type p53, we next examined the domains of p53 that elicited repression of E2F1 transactivation. O'Connor et al. (1995) previously reported that p53 association with E2F1 resulted in an inhibition of E2F1 transactivation that was independent of pRB. Various forms of p53 were transfected with E2F1/DP1

Figure 1 Mdm2 overexpression results in increased E2F1 transactivation in cell lines with wild-type p53. Cells were transiently transfected with a luciferase reporter containing three E2F consensus DNA binding sites (3XE2FLUC; 2 mg) and 10 mg of the indicated expression vector. Each transfection also included a beta-galactosidase (pcDNA3.1-lacZ; 0.5 mg) expression plasmid for normalization of transfection eciency. (a) The eect of Mdm2 on E2F1 transactivation was monitored in U2OS, H1299, U87, MEF:2KO and C33A cells. Similar eects were seen with the DHFR-luc reporter construct (data not shown). An E2Fluciferase reporter containing mutant E2F consensus DNA binding sites did not show any eect with Mdm2 overexpression (data not shown). For each cell line, the Rel Luc. value of the CMV transfection was scaled to 100. (b) Table containing the p53 and Rb status of cell lines where the eect of Mdm2 overexpression on E2F1 transactivation was examined. An Mdm2 eect (yes) represents that the cell line showed activation of E2F1 transactivation with Mdm2 overexpression. All cell lines were tested for E2F1 transactivation with mdm2 in at least three independent experiments

expression vectors and the 3XE2FLUC reporter plasmid into H1299 cells devoid of p53. The coexpression of wild-type p53, a p53 lacking the last 13 amino acids (p53(1 ± 380)), or a p53 site-directed mutant p53(14/19), which was previously reported to induce p53 transactivation by prohibiting repression by Mdm2 (Lin et al., 2000), all resulted in a repression of E2F1 transactivation (Figure 3a). In contrast, both p53(mt 175) and p53(mt 22/23) mutants were not able to repress E2F1 transactivation. p53(mt 175) is a site directed mutant (His-175) form of p53, common to many human cancers that lacks DNA binding capacity but retains the ability to both form tetramers and bind Oncogene


4416

A

B

C

Figure 2 p53 binding domain of Mdm2 is required to induce E2F1 transactivation. U2OS cells were transfected with 3XE2Fluc (1 mg), pRc-HA-E2F1 (75 ng), pcDNA3.1-HA-DP1 (75 ng), pR-Luc (10 ng), 1 mg of pCMVEGFP and 10 mg of the indicated mdm2 expression vector. (a) Forms of Mdm2 capable of binding to p53 were able to activate E2F1 transactivation as reported by relative luciferase (Rel. Luc.(dual)). (b) Western blot analysis for Mdm2 (2A10 antibody) and GFP protein levels

Mdm2. p53(mt 22/23) is a double site-directed mutant that is transactivation-defective (Lin et al., 1994). Based on these results we believe that p53 repression of E2F1 transactivation requires both the DNA binding and transactivation domains. Since p53 overexpression has been also shown to cause global transcriptional repression (Mack et al., 1993) we examined whether p53 repression was speci®c to promoters containing E2F1 binding sites. Both a promoter containing mutant E2F1 DNA binding sites (3XE2F(mut)-luc) and a CMV driven promoter, lacking E2F1 responsive elements (pGL3-luc) showed no detectable p53 repression (Figure 3b,c) when tested under the identical conditions used to observe E2F1 transcriptional repression (Figure 3a). Taken together, these results are consistent with p53 overexpression resulting in a decrease in E2F1 transactivation. To complete the link between Mdm2 overexpression and p53 we examined whether p53 repression of E2F1 Oncogene

Figure 3 E2F1 repression by wild-type and mutant p53. p53 or various p53 mutant expression vectors were cotransfected into H1299 cells with 3XE2F-luc (1 mg), pcDNA3.1-lacZ (0.5 mg) and where indicated, E2F1 and DP1 expression vectors (0.1 mg each). (a) Upper panel shows E2F1 transactivation, as measured by relative luciferase. E2F1 transactivation was repressed by transfection of 0.4 mg of wild-type p53, p53 (1 ± 380) and p53 (mt 14/19). In contrast, E2F1 transactivation was not signi®cantly aected when transfected with 0.4 mg of either p53 (mt 175) or p53 (mt 22/23). Lower panel is a Western blot showing levels of p53 from the indicated transfections. Wild-type p53 was not able to repress transactivation from either a reporter containing mutant E2F1 DNA binding sites (b: 3XE2F(mut)-luc) or a CMV driven luciferase reporter (c: pGL3-luc) under transfection and assay conditions identical to those used in (a)

transactivation could be reversed by Mdm2 overexpression. p53 repression of E2F1 transactivation was reversible by Mdm2, Mdm2(-RF) and HPVE6 although was unaected by overexpression by MdmX (Figure 4a). While Mdm2 and HPVE6 are capable of degrading p53, the ability of Mdm2(-RF), which lacks


A

B

Figure 4 Reversal of p53 repression of E2F1 transactivation by Mdm2 and E6 but not MdmX. (a) E2F1 transactivation was monitored in H1299 cells transfected with 3XE2F-luc (1 mg), pcDNA3.1-lacZ (0.5 mg) and the indicated expression vectors (E2F1, DP1 0.1 mg each; p53(wt) 0.4 mg; mdm2 expression vectors 6 mg). Each transfection also included 500 ng of a CMV-EGFP expression vector. Repression of E2F1 transactivation by p53 was reversed by Mdm2, Mdm2-RF and HPVE6 but not Mdm2 (p76) or MdmX. (b) Western blots for p53, Mdm2 (B3 polyclonal antibody), MdmX and EGFP. EGFP levels show that equivalent amounts of whole cell extract (based on transfection eciency) were loaded in each lane. p53 protein levels were reduced by Mdm2 and HPV as previously reported (Boyd et al., 2000; Geyer et al., 2000; Haupt et al., 1997; Kubbutat et al., 1997; Schener et al., 1990)

the last 13 amino acids and cannot degrade p53 (Jackson and Berberich, 2000), to reverse p53 repression of E2F1 argues that the inhibition of p53 transactivation, not degradation, is the critical mechanism to aect repression of E2F1 transactivation (Figure 4b). The inability of MdmX overexpression to reverse p53 repression of E2F1 transactivation, initially seemed inconsistent with the model that Mdm2 activates E2F1 transactivation through blocking p53 transactivation, however we have subsequently discovered that MdmX possesses a novel ability to repress E2F1 that is p53independent (MW and SJB, data not shown). Since Mdm2 inhibition of p53 transactivation was sucient for inducing E2F1 transactivation, we tested the possibility that p53 transactivation of the p21 gene was inhibiting E2F1 transactivation through inhibiting

phosphorylation of Rb. While a slight but reproducible increase in p21 protein levels was observed in transfections with E2F1/DP1 consistent with previous reports (Gartel et al., 1998; Hiyama et al., 1998), p53 signi®cantly induced endogenous p21 protein level (Figure 5). Consistent with the model that Mdm2 induction of E2F1 transactivation was p53 dependent, we observed a repression of p21 protein to near basal levels with Mdm2 when coexpressed with p53 and E2F1/DP1 (Figure 5). To con®rm that Rb was the indirect target that Mdm2 was aecting via p53 and p21, a non-phosphorylatable Rb (Rb.PSM) was cotransfected with E2F1/DP1 and the 3XE2FLUC reporter plasmid. Previous reports have demonstrated that Mdm2 preferentially associates with hypophosphorylated Rb (Hsieh et al., 1999). However, repression of E2F1 transactivation by the nonphosphorylatable Rb could not be reversed by Mdm2 overexpression in wild-type p53 containing U2OS cells while E1A was capable of reversing Rb.PSM repression of E2F1 transactivation (Figure 6a). Since these results are consistent with p53 eliciting its repression of E2F1 via p21 we tested whether p53 or Mdm2 could aect E2F1 transactivation in primary MEFs lacking p21 compared to wild-type MEFs. While cotransfection of E2F1 and DP1 triggered a signi®cant increase in luciferase activity by the 3XE2FLUC reporter plasmid, p53, Mdm2 had no eect on E2F1 transactivation in p217/7 MEFS (Figure 6b, left panel) but was capable of modulating E2F1 transactivation in wild-type MEFs (Figure 6b, right panel). However, Rb-PSM was able to repress E2F1 transactivation in both MEF cell types when coexpressed with E2F1 and DP1 (Figure 6b). Taken together, these results point to Mdm2 activation of E2F1 transactivation occurring via p53 and its ability to activate p21 and not through a direct Mdm2 : E2F1 or Mdm2 : Rb interaction as originally reported (Martin et al., 1995; Xiao et al., 1995).

4417

Figure 5 Induction of p21 protein levels correlates with repression of E2F1 transactivation. H1299 cells were transfected with the indicated plasmids as described in Figure 4 and whole cell extracts made 24 h post transfection. One hundred micrograms of each extract were separated by 10% SDS ± PAGE and Western blotted with antibodies to p21 (upper panel) and b-actin (lower panel). Endogenous p21 protein levels show a slight increase in cells transfected with E2F1/DP1 but are signi®cantly induced in transfections containing wild-type p53. Mdm2 overexpression blocks p53 induction of p21 protein levels but has no signi®cant eect on p21 levels when cotransfected with E2F1/DP1 Oncogene


4418

Figure 6 Inhibition of E2F1 transactivation by Mdm2 overexpression requires p21. (a) Inhibition of E2F1 transactivation by Rb.PSM, a phosphorylation mutant, cannot be reversed by Mdm2 overexpression. U2OS cells were transfected with 3XE2F-luc (2 mg), pcDNA3.1-lacZ (0.5 mg) and where indicated, E2F1/DP1 (0.25 mg), Rb.PSM (50 ng) and 10 mg of either a mdm2 or E1A expression vector. While E1A overexpression can reverse Rb.PSM inhibition of E2F1 transactivation, Mdm2 overexpression was unable to aect E2F1 repression by Rb.PSM. (b) Overexpression of p53 or Mdm2 can alter E2F1 transactivation in primary Mouse embryo ®broblasts (MEFs) but not primary MEFs lacking p21. (Left panel): p217/7 MEFS were transfected with 3XE2F-luc (1 mg), pcDNA3.1-lacZ (1.0 mg) and where indicated, E2F1/DP1 (0.1 mg each), p53 (0.4 mg), mdm2 (6 mg), HPV-E6 (6 mg) and Rb.PSM (0.4 mg). Only Rb.PSM could repress E2F1 transactivation in p217/7 MEFs. (Right panel): Wild-type primary MEFs were transfected with the same ratios of plasmids used in the p217/7 MEF experiments except that pR-Luc (5 ng) replaced pcDNA3.1lacZ (1.0 mg) and the cellular extracts were assayed using the Dual Luciferase method (Promega). These cells show that E2F1 transactivation can be repressed by exogenous p53 and stimulated by Mdm2

Discussion This study examined how Mdm2 overexpression aected E2F1 transactivation. While Mdm2 overexpression was able to increase E2F1 transactivation as originally reported (Martin et al., 1995; Xiao et al., 1995), this ability was limited to cell lines possessing wild-type p53 and Rb (Figure 1). Using various mutants of Mdm2 (Figure 2) and p53 (Figure 3), the conclusions drawn from the experiments suggest that Mdm2 activation of E2F1 transactivation is actually a reversal of p53-dependent repression of E2F1. Conclusions about the necessary protein : protein interactions can also be drawn from the data Oncogene

presented. Each of the nuclear factors has been reported to form complexes with each other. Since only Mdm2(p76) protein was unable to block p53dependent repression of E2F1 transactivation (Figure 2), association with p53 seems to be a critical determinant. The N-terminal region of Mdm2 has also been reported to facilitate Mdm2 association with E2F1. However, the data presented in this study is not consistent with that model since in the absence of p53, Mdm2 does not aect E2F1 transactivation. The degradation of p53 observed in transfections with Mdm2 and HPV-E6 demonstrate that elimination of p53 protein can derepress E2F1 transactivation (Figure 4). However, the ability of Mdm2(-RF)


to also derepress E2F1 transactivation without altering p53 protein levels (Figure 4) led us to conclude that degradation of p53 is not required to block p53 repression of E2F1 transactivation. Interestingly, even though full-length MdmX could not reverse the repression of E2F1 transactivation elicited by p53, we observed that an N-terminal MdmX could reverse p53 repression of E2F1 transactivation (data not shown). This dierence in eects between the two MdmX proteins results from a unique ability of MdmX to repress E2F1 transactivation independent of p53 (M Wunderlich and SJ Berberich, manuscript in preparation). To examine the role of p53 transactivation in the repression of E2F1 transactivation further a set of sitedirected p53 mutants were examined. Based on the inability of p53(mt175) to repress E2F1 transactivation, a functional DNA binding domain seems to be a critically important domain in triggering p53 repression of E2F1 transactivation. The need for p53 DNA binding activity suggested that p53 transactivation may be involved in the repression of E2F1 transactivation. Consistent with that hypothesis, a p53 mutant defective in p53 transactivation and the ability to bind Mdm2, p53(mt22/23), could not repress E2F1 transactivation. However, a p53 which is unable to bind Mdm2 but transcriptionally active, p53(mt 14/19), repressed E2F1 transactivation The somewhat decreased ability of p53(mt14/19) to repress E2F1 transactivation relative to wild-type p53 is consistent with a decreased ability of p53(mt14/19) to activate transcription of a synthetic p53 responsive luciferase reporter construct under identical conditions (data not shown). A

These results are consistent with a model (Figure 7) where p53 activation of the p21WAF1 gene results in the induction of a protein capable of inhibiting phosphorylation of pRB (Harper et al., 1993) leading to repression of E2F1 transactivation (Figure 5). Based on this ®nding, we propose although wild-type p53 may associate with E2F1 (O'Connor et al., 1995), it is the ability of p53 to transactivate genes that is required to repress E2F1 transactivation. The ability of p21 to inhibit cyclin:cdk2/4 complexes which phosphorylate Rb makes it the most likely target for p53 to aect E2F1 transactivation in vivo. Consistent with that concept, when we transfected U2OS cells with a Rb protein with point mutations that inactivate seven cdk phosphorylation sites at the C-terminus that regulate association with E2F1, E2F1 transactivation was repressed (Figure 6). Mdm2 overexpression in U2OS cells expressing the mutant Rb was unable to reverse repression of E2F1 while E1A produced a partial reversal of Rb.PSM repression (Figure 6a). E1A binds to Rb within the E2F1 binding domain and prevents active, hypophosphorylated Rb from binding to and inactivating E2F1. To demonstrate that p21 was critical regulator in this pathway, we examined whether p53 could repress E2F1 transactivation in primary mouse embryo ®broblasts lacking p21 (Figure 6b, left panel). Consistent with the model, E2F1 transactivation was not repressed by p53, Mdm2 or HPV-E6 however it was repressed by a non-phosphorylatable form of Rb, which acts downstream of p21 (Figures 6b and 7). In contrast, E2F1 transactivation was repressed by p53 and

4419

B

Figure 7 Model of Mdm2 regulation of E2F1 Transactivation via p21. (a) Expression of wild-type p53 leads to activation of the p21 gene that encodes a cyclin-dependent kinase inhibitor capable of inhibiting cyclin:cdk2/4 complexes responsible for maintaining hyperphosphorylated Rb. Under these conditions, E2F1 transactivation will be repressed. (b) Upon overexpression of Mdm2, p53 transactivation is blocked leading to a reduction in p21 protein levels and a concomitant increase in hyperphosphorylated Rb as cells transverse from G1 into S phase. Under these conditions, E2F1 target genes will be activated Oncogene


4420

stimulated by Mdm2 in primary MEFs where the p53/p21/Rb pathway is fully intact (Figure 6b, right panel). These results provide a pathway by which Mdm2 indirectly targets Rb phosphorylation state and association with E2F1 by directly downregulating p53 transactivation and thereby p21 protein levels.

Materials and methods Cell lines and antibodies U2OS, H1299, C33A and U87 cells were purchased from the American Type Culture Collection. Mouse embryo ®broblast (MEF) 2KO cells, lacking both the p53 and mdm2 genes, were provided by Guillermina Lozano (MD Anderson). Primary MEFs lacking p21 were a kind gift from Stephen Jones (University Massachusetts) and primary wild-type MEFs were obtained from Bio Whittaker. All cells were maintained in full growth media consisting of Dulbecco Modi®ed Eagle Medium with 10% fetal bovine serum and 10 mg/ml gentamicin. Antibodies against p53 (FL-393, Santa Cruz), Mdm2 (2A10, polyclonal B3), p21 (C-19, Santa Cruz), b-actin (mouse anti b-actin, Sigma), and GFP (mouse anti GFP, Zymed) were used as recommended by the manufacturer. The polyclonal anti-MdmX antibody has been previously described (Jackson and Berberich, 1999). Horseradish peroxidase-conjugated anti-mouse or anti-rabbit secondary antibodies (Promega) were used with SuperSignal substrate (Pierce) in chemiluminescence-Western blot experiments. Plasmids The following plasmids have been described previously: pcDNA3.1-mdm2, pcDNA3.1-mdm2(p76), pcDNA3.1mdm2(RF), pcDNA3.1-mdmx, pRChp53, pcDNA3.1-lacz (Jackson and Berberich, 1999, 2000; Jackson et al., 2001). pGL3LUC (Promega), pR-Luc (Promega) and pEGFPN1 (Clontech) were purchased from the indicated vendors. pRcHA-E2F1 and pcDNA3.1-HA-DP1 were kind gifts of Dr David Livingston. p53(mt14/19) and p53(mt 22/23) were obtained from Dr Lin (University of Michigan). Dr. Gerald Zambetti (St Jude) kindly provided p53(380). p53(mt 175) was obtained from Dr Bert Vogelstein (John Hopkins) and p1A-E1A from Dr Michael Cole (Princeton University). Lastly, 3XE2FLUC, 3XmtE2FLUC, and Rb.PSM were generous gifts of Dr Erik Knudsen (University of Cincinnati).

Transfections Cells were transiently transfected according to a standard calcium phosphate procedure (Chen and Okayama, 1987). Cells were plated at a density of 2 ± 56105 per 60 mm plate. Twenty-four hours later, cells were refed with full growth media. Three hours following refeeding, the cells were exposed to the DNA-calcium phosphate mixture for 16 ± 24 h. Cells were then washed with PBS and refed with full growth media. After an additional 24 h, cells were harvested, frozen, and lysed in PBSA (phosphate-buered saline containing 5 mM EDTA, 5% Triton X-100, and 1% of a protease inhibitor cocktail). For Figure 5, cells were transfected with Lipofectamine 2000 (Gibco ± BRL) according to the manufacturer's protocol. Luciferase, dual luciferase and b-gal assays For luciferase assays, a portion of the whole cell extracts were mixed with 50 ml luciferase assay substrate (Promega). Relative luciferase units were measured by a luminometer (Zylux) and normalized to b-gal activity. Relative luciferase (Rel. Luc.) values are averages of two to three independent samples assayed in parallel. For dual luciferase assays, the cell pellets were extracted with passive lysis buer and analysed for ®re¯y and Renilla luciferase following manufacturer's protocols (Promega). Relative dual luciferase (Rel. Luc. (dual)) values are averages of three independent samples assayed in parallel. Error bars represent average deviation of the independent measurements. Western blot analysis Whole cell extracts, normalized based on equivalent bgalactosidase activity or protein amounts as determined by Bradford assay, were separated by 10% SDS ± PAGE and subsequently transferred to a polyvinylidene di¯uoride membrane (Millipore) using the Transblot system (BioRad). Proteins were detected by incubating the membrane with the appropriate primary and secondary antibodies followed by chemiluminescence detection documented with exposure to X-ray ®lm.

Acknowledgments We are indebted to the many colleagues who generously provided reagents, cell lines and plasmids. This work is supported by a NIH grant (CA64430) to SJ Berberich.

References Agah R, Kirshenbaum LA, Abdellatif M, Truong LD, Chakraborty S, Michael LH and Schneider MD. (1997). J. Clin. Invest., 100, 2722 ± 2728. Blattner C, Sparks A and Lane D. (1999). Mol. Cell. Biol., 19, 3704 ± 3713. Boyd SD, Tsai KY and Jacks T. (2000). Nat. Cell. Biol., 2, 563 ± 568. Cahilly-Snyder L, Yang-Feng T, Francke U and George DL. (1987). Som. Cell Mol. Gen., 13, 235 ± 244. Chen C and Okayama H. (1987). Mol. Cell. Biol., 7, 2745 ± 2752. Cress WD and Nevins JR. (1994). J. Virol., 68, 4213 ± 4219. Dyson N. (1998). Genes Dev., 12, 2245 ± 2262. Oncogene

Fakharzadeh SS, Trusko SP and George DL. (1991). EMBO J., 10, 1565 ± 1569. Gartel AL, Serfas MS, Gartel M, Goufman E, Wu GS, ElDeiry WS and Tyner AL. (1996). Exp. Cell Res., 227, 171 ± 181. Geyer RK, Yu ZK and Maki CG. (2000). Nat. Cell. Biol., 2, 569 ± 573. Haines DS, Landers JE, Engle LJ and George DL. (1994). Mol. Cell. Biol., 14, 1171 ± 1178. Harper JW, Adami GR, Wei N, Keyomarsi K and Elledge SJ. (1993). Cell, 75, 805 ± 816. Haupt Y, Maya R and Oren M. (1997). Nature, 387, 296.


Hiyama H, Iavarone A and Reeves SA. (1998). Oncogene, 16, 1513 ± 1523. Holmberg C, Helin K, Sehested M and Karlstrom O. (1998). Oncogene, 17, 143 ± 155. Hsieh JK, Chan FS, O'Connor DJ, Mittnacht S, Zhong S and Lu X. (1999). Mol. Cell, 3, 181 ± 193. Hsieh JK, Fredersdorf S, Kouzarides T, Martin K and Lu X. (1997). Genes Dev., 11, 1840 ± 1852. Hunt KK, Deng J, Liu TJ, Wilson-Heiner M, Swisher SG, Clayman G and Hung MC. (1997). Cancer Res., 57, 4722 ± 4726. Irwin M, Marin MC, Phillips AC, Seelan RS, Smith DI, Liu W, Flores ER, Tsai KY, Jacks T, Vousden KH and Kaelin WG, Jr. (2000). Nature, 407, 645 ± 648. Jackson M and Berberich SJ. (2000). Mol. Cell. Biol., 20, 1001 ± 1007. Jackson MW and Berberich SJ. (1999). DNA Cell. Biol., 18, 693 ± 700. Jackson MW, Lindstrom MS and Berberich SJ. (2001). J. Biol. Chem., 10, 10. Johnson DG, Schwarz JK, Cress WD and Nevins JR. (1993). Nature, 365, 349 ± 352. Kubbutat MHG, Jones SN and Vousden KH. (1997). Nature, 387, 299. Lin J, Chen J, Elenbaas B and Levine AJ. (1994). Genes Dev., 8, 1235 ± 1246. Lin J, Jin X, Page C, Sondak VK, Jiang G and Reynolds RK. (2000). Cancer Res., 60, 5895 ± 5901. Mack DH, Vartikar J, Pipas JM and Lamininns LA. (1993). Nature, 363, 281 ± 283.

Martin K, Trouche D, Hagemeier C, Sorensen TS, La Thangue NB and Kouzarides T. (1995). Nature, 375, 691 ± 694. Momand J, Wu HH and Dasgupta G. (2000). Gene, 242, 15 ± 29. Momand J, Zambetti GP, Olson DC, George D and Levine AJ. (1992). Cell, 69, 1237 ± 1245. Muller H, Bracken AP, Vernell R, Moroni MC, Christians F, Grassilli E, Prosperini E, Vigo E, Oliner JD and Helin K. (2001). Genes Dev., 15, 267 ± 285. Nip J, Strom DK, Eischen CM, Cleveland JL, Zambetti GP and Hiebert SW. (2001). Oncogene, 20, 910 ± 920. Nip J, Strom DK, Fee BE, Zambetti G, Cleveland JL and Hiebert SW. (1997). Mol. Cell. Biol., 17, 1049 ± 1056. O'Connor DJ, Lam EW, Grin S, Zhong S, Leighton LC, Burbidge SA and Lu X. (1995). EMBO J., 14, 6184 ± 6192. Phillips AC, Bates S, Ryan KM, Helin K and Vousden KH. (1997). Genes Dev., 11, 1853 ± 1863. Qin XQ, Livingston DM, Kaelin WG, Jr. and Adams PD. (1994). Proc. Natl. Acad. Sci. USA, 91, 10918 ± 10922. Schener M, Werness BA, Huibregtse JM, Levine AJ and Howley PM. (1990). Cell, 63, 1129 ± 1136. Sherr CJ. (1998). Genes Dev., 12, 2984 ± 2991. Stiewe T and Putzer BM. (2000). Nat. Genet., 26, 464 ± 469. Wu X and Levine AJ. (1994). Proc. Natl. Acad. Sci. USA, 91, 3602 ± 3606. Xiao ZX, Chen J, Levine AJ, Modjtahedi N, Xing J, Sellers WR and Livingston DM. (1995). Nature, 375, 694 ± 698.

4421

Oncogene