E2F-1 Induces the Stabilization of P53 but Blocks P53-Mediated Transactivation

E2F-1 induces the stabilization of p53 but blocks p53-mediated transactivation

John Nip1,4, David K Strom1,5, Christine M Eischen2, John L Cleveland2,3, Gerard P Zambetti2,3 and Scott W Hiebert*,1

1Department of Biochemistry, Vanderbilt University School of Medicine, Nashville, Tennessee, TN 37232, USA; 2Department of Biochemistry, St. Jude Children's Research Hospital, Memphis, Tennessee, TN 38105, USA; 3Department of Biochemistry, University of Tennessee, Memphis, Tennessee, TN 38163, USA

E2F-1 induces p53 accumulation and E2F-1 and p53 et al., 1995; Blake and Azizkhan, 1989; DeGregori et form a physical complex, which aects the ability of al., 1995; Dou et al., 1992; Hiebert et al., 1991; Ohtani E2F-1 to activate transcription. We mapped the domains et al., 1995; Pearson et al., 1991; Sala et al., 1994; on E2F-1 that interact with p53 and found two p53- Schulze et al., 1995). In cells, E2F-1 forms hetero- binding domains. To understand the functional con- dimers with DP-1, which markedly potentiates E2F sequences of the E2F-1/p53 association on p53 activities DNA binding and transactivation functions (Bandara we identi®ed the domains of E2F-1 that were responsible et al., 1993; Helin et al., 1993; Krek et al., 1993; La for the accumulation of p53. Unexpectedly, we found Thangue, 1994). that the E2F-1 transactivation domain was dispensable `E2F' is a central eector of the retinoblastoma for p53 induction. By contrast, further deletion of the (pRb) tumor suppressor pathway. When bound to DP-1 interaction/`marked' box domain eliminated p53 hypophosphorylated forms of pRb (Chellappan et al., accumulation. Radiolabeling pulse/chase analysis demon- 1991), or to the pRb family members p107 or p130, strated that E2F-1 caused post-translational stabilization E2F is prevented from activating transcription and/or of p53. Although E2F-1 caused the stabilization of p53, these complexes actively repress transcription (Brehm E2F-1 expression impaired p53-dependent transactiva- et al., 1998; Hiebert et al., 1992; Magnaghi-Jaulin et tion. Thus, the E2F-1 : p53 interaction may provide a al., 1998; Weintraub et al., 1992). Following mitogen checkpoint function to inactivate overactive E2F-1, but stimulation, pRb becomes phosphorylated by cyclin/ the association may also inactivate p53 transactivation to cyclin-dependent kinase (cdk) complexes, resulting in allow cell cycle progression. Oncogene (2001) 20, 910 ± the release of free E2F-1 (Bremner et al., 1995; 920. Dynlacht et al., 1994; Kato et al., 1993), which stimulates its targets and cell cycle progression. There- Keywords: E2F-1; p53; stability; cell cycle progression; fore, the activity of E2F-1 can also be indirectly apoptosis modulated by cdk inhibitors, including p15ink4b,p16ink4a, p21cip1/waf1, and p27Kip1 (Hannon and Beach, 1994; Hunter and Pines, 1994; Serrano et al., 1993; Sherr and Roberts, 1995), which prevent cyclin/cdk phos- Introduction phorylation of pRb. Inactivation of these tumor suppressors by mutation, or by the interaction of The E2F-1 transcription factor is a key regulator of cell pRb with DNA tumor virus oncoproteins such as cycle progression (Adams and Kaelin, 1995; Dobro- adenovirus E1A, leads to the activation of E2F. In wolski et al., 1994; Johnson et al., 1993; Shan and Lee, fact, the pRb pathway appears to be one of the most 1994). E2F-1 is the most characterized member of a frequently mutated targets in human cancer (Weinberg, family of six related transcription factors, and 1991). coordinately activates genes necessary for the G1 to S In tumors having mutations in the pRb pathway, the phase transition including DNA polymerase a, dihy- tumor suppressor p53 is also usually inactivated drofolate reductase, thymidine kinase, thymidylate (Schener et al., 1991), and DNA tumor viruses synthase, cyclins (D1, A, E), and E2F-1 itself (Almasan inactivate both pRb and p53 to induce transformation. Thus, one of the ®rst lines of defense in response to the disruption of the pRb pathway appears to be p53. Expression of E1A or E2F-1 leads to the induction of *Correspondence: SW Hiebert p53 (Debbas and White, 1993; Hiebert et al., 1995; Current addresses: 4Unilever Research US, 45 River Road, Lowe et al., 1993; Nip et al., 1997), and in some cell 5 Edgewater, NJ 07020, USA; University of South Carolina, Aiken, types, E1A- and E2F-1-induced apoptosis is p53- SC 29803, USA Received 14 August 2000; revised 8 December 2000; accepted 12 dependent, or is enhanced by overexpression of p53 December 2000 (Lowe and Ruley, 1993; Qin et al., 1994; Wu and Inactive E2F-1 stabilizes but inactivates p53 J Nip et al 911 Levine, 1994). Thus, E2F-1 appears to link the p53 and pRb tumor suppressor pathways. In support of this concept, the p19ARF gene, which induces the stabilization of p53 (Kamijo et al., 1998; Pomerantz et al., 1998; Zhang et al., 1998) is activated by E2F-1 in primary ®broblasts (Zindy et al., 1998). The p53 tumor suppressor functions as an essential component of the DNA damage response that reduces the probability of cells acquiring oncogenic mutations (Lane, 1992). p53 mediates G1 arrest or apoptosis of cells following DNA damage (Finlay et al., 1989; Kuerbitz et al., 1992; Lowe, 1995). However, indirect evidence suggests that p53 also plays a role in the cell cycle. p53 is induced in response to growth signals (Reich and Levine, 1984), hyperproliferative signals emanating from the overexpression of c-Myc, E2F-1, or E1A (Debbas and White, 1993; Hiebert et al., 1995; Lowe and Ruley, 1993; Zindy et al., 1998), and its levels oscillate during the cell cycle (Bischo et al., 1990; Shaulsky et al., 1990, 1991). Moreover, p53 can physically interact with E2F-1, as well as with its dimerization partner, DP-1 (O'Connor et al., 1995; Sorensen et al., 1996). Immortal 32D.3 myeloid cells contain wild-type p53 but do not express p19ARF. Regulated expression of E2F-1 in 32D.3 cells induces p53, but not apoptosis, when these cells are cultured in the presence of the survival factors IL-3 and serum (Hiebert et al., 1995). Moreover, E2F-1-induced apoptosis in these cells is p53-independent (Hiebert et al., 1995; Nip et al., 1997), Figure 1 Transcriptional activity of E2F-1 mutants. (a) Sche- allowing us to dissect the p53 response to E2F-1. matic diagram of full-length E2F-1 and deletion mutants. The Mapping studies demonstrated that the E2F-1 amino functional domains of full-length E2F-1 with the corresponding terminus contains two p53 interaction domains. amino acid positions are indicated. (b) Activation of the human ARF Functional studies indicated that the E2F-1-mediated p14 -promoter-luciferase plasmid by E2F-1 and E2F-1 mutants. MCF-7 cells were transfected with 10 ng of pCMV5 induction of p53 was due to stabilization of p53 containing wild-type E2F-1 or E2F-1 mutants, 1 mg of pCMV5- protein, and that this stabilization was independent of p14ARF-promoter-luciferase plasmid, and 200 ng of a pCMV5- cell cycle phase, transit through the cycle, and cell SEAP (internal control for transfection eciency) and processed viability. Surprisingly, p53 was also induced in cells for luciferase activity 48 h post-transfection as described in Materials and methods. Adjusted luciferase activity values re¯ect overexpressing a transactivation defective E2F-1 pro- a correction for transfection eciency as determined by SEAP tein that is incapable of promoting either cell cycle activity. Wild-type E2F-1 or E2F-1 mutants tested are indicated progression or apoptosis. Thus, p53 responds to on the abscissa. The basal activity of the p14ARF promoter was changes in the steady state levels of E2F-1, rather measured in the absence of E2F-1. Shown are the means and than its ability to stimulate transcription or induce cell average deviation of duplicate samples cycle progression. Finally, expression of E2F-1 impaired p53-mediated transactivation functions. Collec- tively, the data suggest a model whereby p53 is respect to their ability to transactivate E2F target stabilized by E2F-1 through associations that serve to genes, transcriptional reporter assays were performed keep E2F-1 activity in check, yet these associations also using the human p14ARF promoter (coupled to a impair the ability of p53 to activate transcription. luciferase reporter). Human p14 and its murine homologue p19, are induced, both at the mRNA and protein levels, by enforced expression of E2F-1 (Bates et al., 1998; DeGregori et al., 1997; Robertson and Results Jones, 1998) and E2F-2 (DeGregori et al., 1997). E2F-1 and E2F-1D24 activated the p14 promoter/luciferase Two domains of E2F-1 interact with p53 reporter in excess of 80-fold above basal levels (Figure Overexpression assays demonstrated that E2F-1 binds 1b). By contrast, E2F-1 mutants which lacked the p53 (O'Connor et al., 1995). To determine if E2F-1 transactivation or DNA binding domains (E2F-1D374, could physically interact with endogenous p53 in 32D.3 E2F-1D195, E2F-1D119) failed to activate the p14 myeloid cell lines, we used a panel of E2F-1 deletion promoter (Figure 1b). mutants that lack various functional domains of E2F-1 To address the in vivo physical interaction of these (Figure 1a). To characterize these E2F mutants with E2F-1 mutants with p53, we used clonal 32D.3 cell

Oncogene Inactive E2F-1 stabilizes but inactivates p53 J Nip et al 912 lines expressing the E2F-1 deletion mutants for coimmunoprecipitation studies. E2F-1 or E2F-1 mutant proteins were expressed from the pMAM-neo vector at low basal levels and the expression was further augmented by the addition of dexamethasone to the culture media. Cell extracts were immunoprecipitated with an anti-p53 antibody and associated E2F- 1 was measured by immunoblot analysis. E2F-1 coimmunoprecipitated with p53, but was not immunoprecipitated by normal sheep serum (Figure 2a control lane), or an irrelevant monoclonal antibody directed to murine actin (data not shown). Likewise, E2F-1D374 protein coimmunoprecipitated with p53 and additional E2F-1D374 was coimmunoprecipitated with p53 following induction with dexamethasone (Figure 2a). In fact, while the panel of mutations aected each of the known E2F-1 functional domains, each mutant retained the ability to associate with p53 (Figure 2b), suggesting that E2F-1 contains more than one p53 binding domain. To de®ne the regions of E2F-1 responsible for p53 Figure 3 Regions of E2F-1 that bind p53. (a) Schematic interaction, in vitro association assays using GST-E2F- diagram of E2F-1 GST fusion proteins used and a summary of their ability to interact with p53. The functional domains and the 1 fusion proteins were performed. E2F-1 deletion corresponding amino acid positions of E2F-1 are indicated. (b) mutants were expressed as GST fusion proteins (Figure The N-terminus of E2F-1 contains the major p53-binding site. 3a) and tested for their ability to interact with in vitro GST fusion proteins were used to purify 35S-methionine labeled in 35S-labelled p53. Although the full-length GST-E2F-1 vitro translated p53 as described in Materials and methods. Shown are representative results of multiple experiments. fusion protein was produced poorly in our hands, a Interaction with GST protein alone was used as a negative GST-fusion protein (1 ± 190) containing the ®rst 190 control for p53 binding and is shown in the `GST Control' lane of amino acids of E2F-1 interacted with p53 and GST- each panel. Equal amounts of GST fusion proteins were used for E2F-1(190 ± 437) more weakly associated with p53 each sample (Figure 3b, left panel). The E2F-1 fusion protein containing amino acids 120 ± 320, which encompasses the DP-1 interaction domain, failed to bind p53, implying that the weak association with GST-E2F- 1(190 ± 437) was not mediated by DP-1 (Figure 3b). Further deletions of E2F-1 revealed that the ®rst 119 amino acids were sucient to allow p53 interaction (Figure 3b, center and right panels) and that the removal of most of the E2F-1 cyclin A/cdk2 binding domain (1 ± 119D24) abolished p53 association with E2F-1 (Figure 3b, right panel). Thus, the predominant domain for p53 association is in or around the cyclin A-binding domain of E2F-1, although a weaker interaction domain maps to the C-terminal domain (Figure 3b and O'Connor et al., 1995; Sorensen et al., 1996). A similar approach was used to pinpoint the domain of p53 required for binding to E2F-1 (Figure 4a). A series of GST-p53 fusion proteins were used to purify in vitro synthesized 35S-labeled E2F-1. Full-length p53 (1 ± 393) and p53 (1 ± 160) associated with E2F-1 whereas other portions of p53 failed to bind to E2F- 1 (Figure 4b). However, the best interaction was Figure 2 E2F-1 mutants associate with p53 in vivo.(a) E2F-1 consistently observed with amino acids 45 ± 145 of and E2F-1D374 interact with p53. (b) Mapping domains of E2F-1 p53 (C-terminal to the domain required for p53 required for association with p53. Coimmunoprecipitation experiments were performed using total cellular protein extracts binding to Mdm2). Given that two domains of E2F-1 from cell lines, cultured with (+) or without (7)of25mM contacted p53, this region may contain two E2F-1 dexamethasone. Control lane in (a) was E2F-1/DP-1-expressing binding domains, or the second domain may be too cellular lysate immunoprecipitated with normal sheep serum. weak to detect under these conditions. Coprecipitating E2F-1 proteins were identi®ed by immunoblot Both E2F-1 and p53 associate with Mdm2 (Martin analysis. There was cross-reactivity of the anti-mouse IgG et al., 1995; Momand et al., 1992). Because the E2F-1 antibody used for the immunoblotting, with sheep IgH in some experiments (see D195 Co-IP panel) binding motif mapped just C-terminal to the Mdm2

Oncogene Inactive E2F-1 stabilizes but inactivates p53 J Nip et al 913

Figure 5 Interactions between E2F-1, p53, and Mdm2. Coim- munoprecipitation experiments were performed using total cellular protein from lysates of 32D.3 (lane 1) or E2F-1.1/DP- Figure 4 Domains of p53 that bind E2F-1. (a) Diagram of GST- 1.2-expressing cell lines (lane 2), cultured with 25 mM dexametha- p53 fusion proteins and a summary of their ability to bind E2F-1. sone. Lysates were immunoprecipitated with either Ab-7 (p53), The major functional domains of p53 are as indicated. (b) Amino KH-20 (E2F-1), or 2A10 (Mdm2) and sequentially immuno- acids 45 ± 145 of p53 bind to E2F-1. GST fusion proteins bound blotted with antibodies against the other two proteins to identify to glutathione-agarose were used to purify 35S-methionine labeled coprecipitating proteins. Shown is a representative experiment of in vitro translated E2F-1. Native GST proteins were used as a three independent experiments. IgH represents the immunoglobu- negative control for E2F-1 binding (lane labeled GST Control). lin heavy chain The `Input E2F-1' lane represents 10% of the total labeled E2F-1 used per GST puri®cation assay. Equal amounts of GST fusion proteins were used for each sample. Shown are representative results of triplicate experiments 1 cells, whereas the radiolabeled p53 was undetectable in the control cells. Enforced inducible expression of high levels of DP-1 alone in these cells did not lead to binding domain on p53, we asked whether the p53 stabilization (data not shown). Thus, p53 is post- association between E2F-1 and p53 was stabilized by translationally stabilized approximately twofold in forming a trimeric complex with Mdm2. In 32D.3 cells, response to E2F-1/DP-1. a p53 and Mdm2 complex was detected, but only when One possible mechanism of p53 stabilization by E2F- E2F-1 induced higher levels of p53 (Figure 5, right, 1 could be through the E2F-1-mediated activation of middle and bottom panels, lane 2). Although p53 and the p19ARF gene (Bates et al., 1998; DeGregori et al., Mdm2 did interact, we did not observe a trimeric E2F- 1997; Robertson and Jones, 1998). p19ARF can bind to 1:Mdm2:p53 complex; E2F-1 was never detected in Mdm2 and prevent it from targeting p53 for ubiquitin- Mdm2 immune complexes or vice versa (Figure 5). mediated degradation, thus causing p53 stabilization Thus, it appears that Mdm2 and E2F-1 bind p53 (Pomerantz et al., 1998; Zhang et al., 1998). Speci®- independently. cally, p19ARF stabilizes p53 by sequestering Mdm2 in the nucleolus (Weber et al., 1999) and inhibiting the shuttling of p53 by Mdm2 from the nucleus to the E2F-1+DP-1 induces p19ARF-independent, cytoplasm (Tao and Levine, 1999), an event believed to post-translational stabilization of p53 be required for p53 degradation. However, parental The viral E1A oncoprotein induces post-translational 32D.3 cells or 32D.3 cells expressing E2F-1+DP-1 or stabilization of p53 (Lowe and Ruley, 1993) and E2F-1 the DNA-binding defective E2F-1, E2F-138, failed to also causes an accumulation of p53 (Hiebert et al., express p19ARF protein following induction of E2F-1 1995; Kowalik et al., 1995, 1998). To probe the and DP-1 with dexamethasone (Figure 6b, upper consequences of the E2F-1/p53 association for p53 panel). The lack of p19ARF was con®rmed by RNA function, we initially determined whether E2F-1 caused blot analysis and by reverse-transcriptase polymerase- a post-translational stabilization of p53. Pulse-chase chain reaction assays (data not shown). The levels of analysis indicated that the p53 protein that accumu- the pro-apoptotic protein Bax, were also not aected lated in response to the induction of E2F-1 plus DP-1 by the induction of E2F-1 (Figure 6b, middle panel). had a signi®cantly longer half-life (*90 min) than p53 As expected, p53 levels were elevated in E2F-1+DP-1 in control 32D.3 cells containing the empty pMAM- clones treated with dexamethasone (Figure 6b, lower Neo vector (*45 min) (Figure 6a). Four hours after panel), but were not elevated in E2F-138-expressing chase, p53 protein was still detectable in the E2F-1/DP- cells (Figure 6b, lower panel). As controls for

Oncogene Inactive E2F-1 stabilizes but inactivates p53 J Nip et al 914

Figure 7 E2F-1 induces p53 independent of cell cycle progression. E2F-1.1/DP-1.2 cells were arrested in the G1 phase of the cell cycle using 1 mM rapamycin. Representative ¯ow cytometric histograms of E2F-1.1/DP-1.2 cells are shown in (a). Untreated E2F-1.1/DP-1.2 cells are shown in the far left panel (Control). Following a 24 h incubation with rapamycin, dexamethasone (25 mM) was added to the cells where indicated for an additional 8 h incubation. (b) p53 protein was analysed by immunoblot using the Ab-7 antibody Figure 6 E2F-1 induces a p14ARF-independent, post-translational stabilization of p53. (a) Cell lines overexpressing E2F-1 together with DP-1 were treated with dexamethasone (25 mM) for 6 h and subsequently used for pulse-chase analysis as described in arrest (data not shown). As expected, addition of Materials and methods. p53 protein was immunoprecipitated dexamethasone to parental 32D.3 cells did not lead to from radiolabeled lysates (26107 c.p.m./sample) using p53- speci®c antibody PAb421. Aliquots of the samples were analysed any increases in the levels of p53 in the absence or at the indicated times (hours) following chase. (b) 32D.3 cells and presence of rapamycin (Figure 7b, lanes 5 ± 8). By their E2F-1 derivatives do not express p19ARF protein. E2F-1/DP- contrast, induction of E2F-1 levels with dexamethasone 1-expressing or parental 32D.3 cells were treated with dexametha- led to p53 accumulation whether or not these cells were sone for 16 h to induce E2F-1/DP-1 protein expression. Cells arrested with 1 mM rapamycin (Figure 7b, lanes 2 and were harvested and prepared for immunoblot analysis as previously described (Zindy et al., 1998). Sequential blotting for 4). Therefore, induction of p53 appears to be a direct Bax (antibody N-20) and p53 (antibody Ab-7) was also response to increases in E2F-1 levels, and is indepen- performed. Controls for p19ARF expression were lysates derived dent of transit through the G1 phase of the cell cycle. from p537/7, mutant p53, and p19ARF7/7 MEFs

Induction of p53 by E2F-1 is independent of E2F-1 transactivation functions immunoblotting, MEFs de®cient in p53 or harboring E2F-1D374, a truncated form of E2F-1, lacks the mutant p53 expressed high levels of p19ARF, whereas transcriptional transactivation domain, including the MEFs from p19ARF7/7 mice did not (Figure 6b, pRb and CREB-binding protein (CBP) interaction upper panel). Thus, p53 stabilization by E2F-1 can also domains (Trouche et al., 1996; Trouche and Kouzar- occur in the absence of the p19ARF protein. ides, 1996). To assess whether E2F-1 transactivation functions were required for the induction of p53, the E2F-1D374-expressing cell lines were used. E2F-1D374 E2F-1-mediated induction of p53 is independent of cell protein expressed in these cells bound to consensus cycle progression E2F-1 DNA binding sites as measured by electro- To determine if E2F-1-induced accumulation of p53 is phoretic mobility shift assay (data not shown). In due to a direct response to E2F-1 levels, or is mediated contrast to E2F-1- (Hiebert et al., 1995) or E2F-1/DP- by events associated with inappropriate cell cycle 1-expressing cells (Figure 8a, right panel), cells progression, 32D.3 cells expressing E2F-1/DP-1 were expressing E2F-1D374 arrested in G1 when cultured treated with the immunosuppressant rapamycin to in media lacking IL-3, con®rming that the E2F-1 arrest cells in the G1 phase. Cell cycle arrest was transactivation functions are required for E2F-1 to con®rmed by propidium iodide staining and ¯ow promote continued cell cycle progression (Figure 8a, cytometric analysis for DNA content (Figure 7a). center panel). Immunoblot analysis of cell lysates Following cell cycle arrest, the cells were treated with showed an induction of E2F-1D374 protein upon dexamethasone to induce E2F-1 and DP-1 levels and dexamethasone addition (Figure 8b, upper panel). aliquots of these cells were lysed and p53 levels were Unexpectedly, p53 levels simultaneously accumulated analysed by immunoblotting. Induction of E2F-1 plus (Figure 8b, lower panel). The ability of this transacti- DP-1 failed to override rapamycin-induced cell cycle vation-de®cient E2F-1 mutant to also induce p53

Oncogene Inactive E2F-1 stabilizes but inactivates p53 J Nip et al 915 The hypoxia-inducible factor 1a protein and p53 physically interact and this interaction may contribute to stabilization of p53 protein (An et al., 1998). To determine whether the interaction of E2F-1 with p53 contributes to p53 stabilization, we induced the expression of E2F-1 or the E2F-1 mutant proteins in 32D.3 cells by the addition of dexamethasone and assessed their ability to induce p53. Deletion of the major interaction domain on E2F-1, the cyclin A binding domain, amino acids 79 ± 103, impaired the ability of E2F-1 to stabilize p53 (Figure 8c, upper left panel, bottom row), suggesting that physical interaction contributes to p53 stabilization. However, two DNA binding-de®cient mutants of E2F-1, E2F-138 and E2F-1D119 that interact with p53, failed to stabilize p53, indicating that an intact E2F-1 DNA binding domain was required (Figure 8c, bottom rows, upper and lower right panels). The lack of function of a full- length E2F-1 with a linker insertion in the DNA binding domain, suggests a mechanism other than (or in addition to) blocking MDM2 binding to p53. Similarly, the E2F-1 mutant (E2F-1D195) that lacks the C-terminal DP-1 heterodimerization and transactivation domains, also failed to induce p53. Therefore, p53 stabilization by E2F-1 requires the DNA binding domain, the D24 domain, and sequences between E2F- 1 residues 195 and 374 (Figure 8c, lower left panel, bottom row) for p53 stabilization.

E2F-1 impairs p53-dependent transactivation The association of p53 with E2F-1 blocks the ability of E2F-1 to bind DNA and to transactivate gene expression (Sorensen et al., 1996). However, the physiological eects of E2F-1/p53 complex formation on p53 transcriptional activity have not been tested. In some cell types, the apoptosis-stimulating gene Bax is Figure 8 Induction of p53 by E2F-1 is independent of E2F-1 transactivation functions. (a) E2F-1D374 overexpressing cells regulated by p53 (Miyashita and Reed, 1995); however, arrest in the G1 phase of the cell cycle in the absence of IL-3. we did not detect any alterations in Bax protein levels Cells were divided to a density of 0.56106 cells/mL and cultured in E2F-1 expressing cells (Figure 6b). Therefore, we in the absence of IL-3 for 16 h prior to ¯ow cytometric analysis. also assessed the levels of the p21Cip1/Waf1 CDK inhibitor Apoptotic cells in the E2F-1.1/DP-1.2 sample were gated out. Shown are the results of duplicate experiments using a in response to E2F-1-mediated p53 accumulation. Like representative clone. (b) E2F-1 transactivation domain is not Bax, the levels of p21 were unchanged following the required for induction of p53. Representative clones of E2F- induction of E2F-1 (Figure 9a). However, a tempera- 1D374 were cultured in the absence (7) or presence (+) of 25 mM ture-sensitive mutant of p53 was able to induce p21 dexamethasone and whole cell lysates of these cells were expression in 32D cells at the permissive temperature. immunoblotted for E2F-1 (antibody KH20, upper panel), and p53 (Ab-7, lower panel). (c) E2F-1 DNA binding is required for Because DNA damage causing only a moderate p53 stabilization. Representative clones of E2F-1D24, E2F-138, induction of p53 did lead to elevation of p21 levels E2F-1D195, and E2F-1D119 were cultured in the absence (7)or (Hiebert et al., 1995), we conclude that E2F-1-mediated presence (+) of 25 mM dexamethasone and total cellular protein induction of p53 failed to induce p53 target gene extracts from these cells were processed for SDS ± PAGE and immunoblotting. Detection of E2F-1 proteins was carried out expression. using KH-20 antibodies for all E2F-1 proteins except E2F-1D24, The lack of p21 or Bax induction in response to which was detected by a rabbit polyclonal antibody directed E2F-1 suggested that E2F-1 might impair the ability of against the pRb-binding domain of E2F-1. Blots were reprobed p53 to activate transcription. Therefore, we tested the with Ab-7 for the detection of p53 eects of E2F-1 and E2F-1 mutants on p53-mediated transactivation in transient expression assays with an idealized p53-reporter plasmid. Using p53-null 10-1 underscores the concept that cell cycle progression is cells, when expressed separately, p53 activated this not required for p53 induction by E2F-1, and suggests promoter and E2F-1 and its deletion mutants had little that p53 responds directly to changes in steady state or no eect (Figure 9b). However, when co-expressed, levels of E2F-1. E2F-1 displayed a dose-dependent impairment of p53-

Oncogene Inactive E2F-1 stabilizes but inactivates p53 J Nip et al 916 induce p53, did not signi®cantly aect p53 function, suggesting that the inhibition of p53 function was not due to E2F-1 binding to p300. These results are summarized in Figure 9c. Thus, E2F-1 may bypass the p53 checkpoint, at least in part, through canceling p53- mediated transactivation functions.

Discussion

Mapping studies indicated that E2F-1 contains two p53-binding domains (Figures 2 and 3; O'Connor et al., 1995; Sorensen et al., 1996). The functional outcome of inducing the E2F-1/p53 association in immortalized myeloid cells was p53 post-transcriptional stabilization that was independent of p19ARF. E2F-1-dependent induction of p53 was also independent of cell cycle progression. p53 accumulation was even observed with forms of E2F-1 that lack transcriptional activity and are defective in promoting cell cycle progression. These data imply that p53 responds directly to changes in steady state levels of E2F-1, rather than through E2F-1-mediated events caused by the induction of inappropriate cell cycle progression. Furthermore, E2F-1-mediated stabilization of p53 requires domains of E2F-1 that physically contact p53 in addition to the E2F-1 DNA binding domain. The outcome of the physical association of E2F-1 and p53 is that p53-dependent transactivation is impaired, suggesting that this allows E2F-1 to bypass the p53 checkpoint and promote S phase entry. E2F-1-transactivation defective mutants, such as D374, can physically interact with p53 and cause accumulation of p53 (Figure 8), yet unlike E2F-1, these mutants fail to induce cell cycle progression or impair p53-mediated transactivation functions (Figure Figure 9 E2F-1 impairs p53-dependent transactivation func- 9). We have also tested the eect of wild-type and tions. (a) E2F-1 does not aect the levels of endogenous p21. Immunoblot analysis of p21 in the absence or presence of mutant E2F-1 on p53-dependent transrepression of the dexamethasone to induce E2F-1 or E2F-1+DP-1 expression. (b) MAP4 and topoisomerase IIa promoters (Murphy et E2F-1 blocks p53-dependent transactivation, but E2F-1 mutants al., 1996, 1999; Wang et al., 1997). However, in both can cooperate with p53. 10(1) cells lacking p53 were transiently cases E2F-1 alone activated these promoters. Thus, transfected with 1 mg of the p53-dependent reporter plasmid 50-2- while p53 still moderately repressed in the presence of luciferase, 200 ng of pBIG-p53 and 200 ng of the given pCMV- E2F-1 expression plasmids. One hundred ng of Rous-sarcoma E2F-1 in these assays, we could not assess whether virus LTR-renilla luciferase plasmid was included as an internal E2F-1 aected p53-mediated repression or simply control for transfection eciency. Values shown are representa- activated these promoters (data not shown). However, tive results of an experiment performed in duplicate and corrected we did note that in 32D.3 cells expressing wild-type for expression of renilla luciferase. (c) Summary of p53 interaction and stabilization by E2F-1 mutants E2F-1, but not the D374 mutant, that levels of the endogenous topoisomerase IIa were signi®cantly re- duced (data not shown). Thus, it appears that E2F-1 may suppress p53-dependent transactivation (Figure 9), dependent transactivation (up to ®vefold, Figure 9b). but not p53-dependent transrepression. Given that p53- The ability of E2F-1 to inhibit p53-dependent dependent repression has been associated with p53- transactivation required the E2F-1 transactivation induced cell death (Sabbatini et al., 1995), this result domain as the D195 and D374 mutants had no eect, could explain how E2F-1 and p53 can cooperate to or actually cooperated with p53 to activate transcrip- induce apoptosis in some cell types (Wu and Levine, tion (Figure 9b, note that at lower levels of E2FD374 it 1994). Alternatively, E2F-1 may directly or indirectly also cooperated with p53, data not shown). By modulate topoisomerase IIa levels, independent of p53. contrast, E2F-1D24, which lacks one of the p53 Transactivation-defective forms of E2F-1 have binding sites and was impaired in p53 stabilization, diering activities when overexpressed in divergent cell had only small eects on p53 function. Further, the types including the induction of apoptosis or the DNA binding de®cient E2F-1D119, that failed to stimulation of cell cycle progression (Hsieh et al.,

Oncogene Inactive E2F-1 stabilizes but inactivates p53 J Nip et al 917 1997; Phillips et al., 1997; Zhang et al., 1999). In apoptosis (Lin et al., 1994), which can also be immortalized 32D.3 myeloid cells, only wild-type E2F- suppressed by 19K (Han et al., 1996; Sabbatini et al., 1 induced apoptosis when cultured in the absence of 1995), but also to prevent p53 from binding E2F-1 to IL-3 (Nip et al., in preparation) and E2F-1 mutants inhibit cell cycle progression. that lacked transactivation functions failed to induce cell cycle progression. In addition, while wild-type E2F-1 blocked p53-mediated transactivation, mutants that bound and stabilized p53, but lacked the Materials and methods activation domain, cooperated with p53 to activate transcription (Figure 9). The ability of E2F-1 to bind Cell lines and cell cycle analysis p53 and modulate its activity may explain some of the The culture of parental 32D.3 myeloid progenitor cells and diverse phenotypes ascribed to E2F-1 and its deriva- establishment of E2F-1 mutant-expressing cell lines was tives. performed as previously described (Hiebert et al., 1995; Nip In MEFs, the cell cycle checkpoint function of p53 is et al., 1997). The 32D.3 cell lines overexpressing E2F-1, DP- associated with the co-stimulation of the p19ARF tumor 1, and E2F-138 have been described (Hiebert et al., 1995; Nip suppressor, yet as demonstrated here, E2F-1 induction et al., 1997). Mouse embryo ®broblasts (MEF) (p53 null, p19ARF null, or mutant p53) were isolated and characterized ARF of p53 can be p19 -independent, as 32D.3 cells lack as previously described (Zindy et al., 1998). 32D.3 cells detectable levels of the p19ARF protein or mRNA overexpressing the E2F-1D374 truncation mutant (encom- (Figure 6 and data not shown). The overall cellular passing aa 1 ± 374 of E2F-1) (Helin et al., 1993) and the E2F- response to E2F-1 overexpression, therefore, appears 1D24 mutant (deleting aa 79 ± 103) (Krek et al., 1994, 1995), to be an attempt to inactivate E2F-1 by p53. This were generated as described earlier (Hiebert et al., 1995; Nip model is supported by the observations that p53 can et al., 1997). To generate the E2F-1D195 construct, an XbaI/ also interact with DP-1, and studies demonstrating that MslI fragment encoding the ®rst 195 amino acids of E2F-1, p53 blocks E2F-1-dependent transactivation (O'Con- containing an added in-frame stop codon along with ¯anking nor et al., 1995; Sorensen et al., 1996). In addition, the XbaI-linkers, was subcloned into an engineered XbaI site in major in vitro p53-interaction domain of E2F-1 pMAM-Neo. A PCR strategy was used to construct E2F- 1D119 ± 181, an E2F-1 DNA binding domain deletion encompasses the cyclin A/cdk2 binding region and mutant, through the ligation of two separately synthesized removal of this motif disrupts the proper regulation of PCR fragments corresponding to aa 1 ± 118 and aa 182 ± 437 E2F-1 transcription functions (Krek et al., 1995). Thus, of E2F-1 and subcloned into pMAM-Neo. The primer pairs it is possible that p53 also competes for cyclin A used to amplify the fragments were as follows: E2F 1 ± 118 binding to disrupt E2F-1-dependent transcriptional (5'-ATTCTAGAGCTAGCATGGCCTTGGCCGGG-3';5'- control. ATGGATCCTTTCACACCTTTTCCTGGATG-3'); E2F A physical interaction between p53 and hypoxia- 182 ± 437 (5'-ATGGATCCTCCAAGAACCACATCCAG-3'; inducible factor 1a protein has been hypothesized to 5'-ATTCTAGATCAGAAATCCAGGGGGGT-3'). Sequen- contribute to p53 stabilization (An et al., 1998). Two cing (both strands) con®rmed the accuracy of the PCR domains of E2F-1 were identi®ed that contact p53, and product. Cell lines transfected with the pMAM-Neo vector constructs were treated with 25 mM dexamethasone to induce DP-1 also interacts with p53 (O'Connor et al., 1995; E2F-1 protein expression. Single cell cloning of transfectant Sorensen et al., 1996). Given that the domains that pools was performed by limiting dilution. were required for p53 stabilization include a p53 Cell cycle analysis was performed by ¯ow cytometry as interaction domain, the E2F-1 DNA binding domain described previously (Krishnan, 1975; Nip et al., 1997). Ten- and the DP-1 interaction motif, it is possible that for thousand total events were analysed per sample. The ModFit E2F-1 to stabilize p53 it must contact p53 and DP-1 computer program (Verity Software House, Topsham, ME, (and perhaps DNA) at the same time. Nevertheless, USA) was used to generate the histograms and to determine our studies indicate that p53 responds to the levels of the percentage of cells within the G1 phase of the cell cycle. E2F-1, not to its transcriptional activation functions. DNA tumor viruses target both p53 and pRb for Antibodies and drugs inactivation, leading to the induction of cell cycle Immunoblots were performed as described previously (Nip et progression. The targeting of p53 was thought to be al., 1997), using the following antibodies: a rabbit polyclonal due to the ability of p53 to induce apoptosis. However, anity-puri®ed antibody directed against the pRb-binding in our cell system E2F-1-induced apoptosis is indepen- domain of E2F-1 (Hiebert et al., 1995); a mouse monoclonal dent of p53 (Hiebert et al., 1995; Nip et al., 1997), antibody directed to the N-terminal region (aa 1 ± 89) of E2F- allowing us to assess its additional roles in regulating 1 (KH-20) (Helin et al., 1993); a sheep polyclonal antiserum E2F-1-dependent cell cycle control. The adenovirus directed to murine p53 (Ab-7, Oncogene Science, Cambridge, E1B gene encodes two proteins that have been MA, USA); a mouse monoclonal antibody directed to human postulated to aect apoptosis, 55K and 19K. The p53 and reacting with murine p53 (PAb 421); a mouse E1B 19K protein is homologous to Bcl-2 and inhibits monoclonal antibody (2A10) made against human Mdm2 protein (aa 294 ± 339) and cross-reacting with murine Mdm2 cell death (Han et al., 1996), whereas the 55K protein (Olson et al., 1993); a rabbit antibody directed to the carboxy binds to the N-terminal p53 transactivation domain terminus of murine p19ARF (Quelle et al., 1995), a rabbit and inhibits the ability of p53 to activate transcription. polyclonal antibody (N-20) made to the N-terminus of Thus, it is possible that adenovirus has evolved the human Bax (Santa Cruz Biotechnology Inc, Santa Cruz, E1B 55K protein not only to block p53-mediated CA, USA), anti-p21 Ab-5 (CalBiochem) and normal sheep

Oncogene Inactive E2F-1 stabilizes but inactivates p53 J Nip et al 918 serum (Oncogene Science, Cambridge, MA, USA). Immune Luminometer (Analytical Luminescence Laboratory, San complexes were detected using the SuperSignal1 Chemilumi- Diego, CA, USA). Transfection eciency was determined nescent detection system (Pierce, Rockford, IL, USA). by quantitation of secreted alkaline phosphatase (SEAP) The dexamethasone (Sigma, St. Louis, MO, USA), and activity following co-transfection of a pCMV5-SEAP reporter rapamycin (Calbiochem, San Diego, CA, USA) were added construct as previously described (Berger et al., 1988). The to the cell culture medium as indicated in the results and adjusted luciferase activity values were normalized with ®gure legends. respect to SEAP activity.

Cell cycle arrest and p53 induction experiments Glutathione S-transferase (GST) fusion protein expression, purification, and in vitro glutathione-sepharose pull-down E2F-1/DP-1 overexpressing cells were treated with the experiments indicated drugs for 16 ± 30 h, at 378C and an aliquot analysed by ¯ow cytometry to con®rm cell cycle arrest. The doses of GST-E2F fusion proteins were generated by subcloning PCR- the drugs used in this study were chosen to give maximal cell ampli®ed, E2F-1 deletion and truncation mutants, into the cycle arrest and minimal cytotoxicity. Following cell cycle prokaryotic GST gene fusion expression vector, pGEX-4T-1 arrest, E2F-1 was induced by addition of dexamethasone (Pharmacia Biotech, Piscataway, NJ, USA). The E2F-1 mutant (25 mM) for 8 h. Cells (16107) were collected, washed in PBS, PCR-generated DNA fragments contained engineered EcoRI/ and processed for Western blot analysis of p53 protein as XhoIorEcoRI/SalI restriction site pairs. The fragments were described below. subcloned into pGEX-4T-1 digested with EcoRI/XhoI. The For half-life determinations, cells (0.36106/mL) in replete following oligonucleotide primer pairs were used to generate medium were treated with dexamethasone (25 mM) for 6 h. the E2F-1 mutants: 1 ± 190 (5'-TCCCCGGAATTCATGGCC- They were then cultured in methionine- and cysteine-free TTGGCCGGG-3',5'-CGGCCGCTCGAGCCACTGGATG- complete RPMI medium for 30 min and metabolically TGGTTCTT-3'; 120 ± 320 (5'-TCCCCGGAATTCAAATCC- labeled with 1 mCi 35S-Translabel (ICN; Irvine, CA, USA) CCGGGGGAGAAG-3',5'-CGGCCGCTCGAGCTCCTCA- for 2 h. Cells were washed extensively and cultured in GAAGTGACCTC-3'); 190 ± 43 (5'-TCCCCGGAATTCTGG- complete media containing excess unlabeled methionine and CTGGGCAGCCACACC-3',5'-CGGCCGGTCGACATCA- cysteine (chase). At the indicated intervals, cells were GAAATCCAGGGGGGT-3'); 1 ± 64 (5'-TCCCCGGAATT- harvested for immunoprecipitation analysis. Radiolabeled CATGGCCTTGGCCGGG-3',5'-CGGCCGCTCGAGCAG- cellular proteins were quantitated by trichloroacetic acid CAGGTCAGGGTCGCA-3'); 1 ± 119 and 1 ± 119D24 (5'- (TCA) precipitation (Olson and Levine, 1994) and normal- TCCCCGGAATTCATGGCCTTGGCCGGG-3',5'-CGGC- ized prior to analysis (26107 c.p.m./sample). Following CGCTCGAGCACACCTTTTCCTGGATG-3'). The PCR electrophoresis, the gels were treated with sodium salicylate DNA template was the pMAM-Neo vector containing full- and glycerol to amplify the 35S signals. The gels were dried length E2F-1 for all the mutants except 1 ± 119D24, where the under vacuum at 808C and exposed to X-ray ®lm (Kodak template was pMAM-Neo-E2F-1D24. GST-p53 fusion pro- XAR5). The samples were quantitated using a BioImage teins were constructed as previously described (Chen et al., Visage 110 Kodak camera and BioImage whole band 1993). analyser software (version 3.3). The GST fusion proteins were expressed and puri®ed as previously described (Hiebert, 1993). The puri®ed fusion proteins were isolated by incubation with glutathione Coimmunoprecipitation experiments and electrophoretic sepharose 4B beads for 30 min at 48C. The beads were mobility shift assays (EMSA) washed three times with PBS and resuspended in TENN Cells were lysed in TENN buer (50 mM Tris-Cl pH 7.4, buer. Bound GST fusion protein was analysed by 5mM EDTA, 150 mM NaCl and 0.5% NP-40) containing electrophoresis through a 10% SDS-polyacrylamide gel proteinase inhibitors, sonicated, and clari®ed by centrifuga- followed by Coomassie Blue R-250 staining and methanol/ tion at 10 000 g for 5 min. Two mg of total cellular protein acetic acid destaining. was incubated with the appropriate primary antibody over- E2F-1 or p53 was synthesized in vitro using the TNT night at 48C. Subsequently, immune complexes were isolated reticulocyte lysate-coupled transcription/translation system by the addition of Protein G Sepharose 4B (Sigma, St. Louis, with and without 35S-methionine (Promega, Madison, WI, MO, USA). The beads were washed three times in TENN USA). For the glutathione-sepharose pull-down experiments, buer, resuspended in SDS sample buer and analysed by 35S-labeled p53 or E2F-1 (10 mLofa50mL reaction) and immunoblot analysis (Nip et al., 1997). equal amounts of glutathione sepharose-bound GST fusion proteins were mixed together in 200 mL TENN buer containing protease inhibitors and incubated for 1.5 h at Luciferase reporter assay 48C. The sepharose beads were washed three times in TENN The cDNA for all the E2F-1 mutants described above were lysis buer, resuspended in SDS sample buer, heated to subcloned into the pCMV5 plasmid. Ten ng of pCMV5 1008C, vortexed, and analysed on a 10% SDS-polyacrylamide constructs containing E2F-1 or E2F-1 mutants were gel. Following electrophoresis, the gel was destained and transfected along with 1 mg of a vector containing the human subjected to ¯uorography. The gel was dried and exposed to p14ARF promoter (Robertson and Jones, 1998) coupled to a X-ray ®lm (Kodak X-OMAT AR) at 7808C. luciferase reporter gene into MCF-7 breast carcinoma cells, using the SuperFect transfection reagent (Qiagen, Inc., Santa Clarita, CA, USA). The transfection was performed according to the manufacturer's protocol. Forty-eight hours post- Acknowledgments transfection, cells were lysed and processed for luciferase We thank Yue Hou, Dana King, Hui Yang, Chunying activity using the Luciferase assay reagent (Promega, Yang, and Elsie White for expert technical assistance. We Madison, WI, USA) according to manufacturer's instruc- also thank Dr Kristian Helin for the pCMV-E2F-1D374 tions. Light intensity was measured using a Monolight 2010 plasmid and for the KH20 antibody, Dr William Kaelin for

Oncogene Inactive E2F-1 stabilizes but inactivates p53 J Nip et al 919 the pRC-CMV-E2F-1D24 plasmid (Krek et al., 1994, 591 (SW Hiebert), NIH grants DK44158 and CA76379 (JL 1995), Drs Jiandong Chen and Tom Shenk for the GST- Cleveland), CA63230 (G Zambetti), CA09346 (CM p53 vectors, Dr Jennifer Pietenpol for pBIG-p53, and Dr Eischen), NIH Core grant 5 P30 CA21765, by the Peter Jones for the p14ARF reporter plasmid. We also thank American Lebanese Syrian Associated Charities of St Jude the Vanderbilt-Ingram Cancer Center sequencing facility Children's Research Hospital and by the Vanderbilt- for support. This work was supported by National Ingram Cancer Center. J Nip was a Leukemia and Institutes of Health (NIH) grants AG13726, CA64140 Lymphoma Society of America Special Fellow (Grant and CA77274, and American Cancer Society grant JFRA- No. 3827-99).

References

Adams PD and Kaelin Jr WG. (1995). Sem. Cancer Biol., 6, Hunter T and Pines J. (1994). Cell, 79, 573 ± 582. 99 ± 108. Johnson DG, Schwarz JK, Cress WD and Nevins JR. (1993). Almasan A, Yin Y, Kelly RE, Lee EY, Bradley A, Li W, Nature, 365, 349 ± 352. Bertino JR and Wahl GM. (1995). Proc. Natl. Acad. Sci. Kamijo T, Weber JD, Zambetti G, Zindy F, Roussel MF and USA, 92, 5436 ± 5440. Sherr CJ. (1998). Proc. Natl. Acad. Sci. USA, 95, 8292 ± An WG, Kanekal M, Simon MC, Maltepe E, Blagosklonny 8297. MV and Neckers LM. (1998). Nature, 392, 405 ± 408. Kato J, Matsushime H, Hiebert SW, Ewen ME and Sherr CJ. Bandara LR, Buck VM, Zamanian M, Johnston LH and La (1993). Genes Dev., 7, 331 ± 342. Thangue NB. (1993). EMBO J., 12, 4317 ± 4324. Kowalik TF, DeGregori J, Leone G, Jakoi L and Nevins JR. Bates S, Phillips AC, Clark PA, Stott F, Peters G, Ludwig (1998). Cell Growth Dier., 9, 113 ± 118. RL and Vousden KH. (1998). Nature, 395, 124 ± 125. Kowalik TF, DeGregori J, Schwarz JK and Nevins JR. BergerJ,HauberJ,HauberR,GeigerRandCullenBR. (1995). J. Virol., 69, 2491 ± 2500. (1988). Gene, 66, 1±10. Krek W, Ewen ME, Shirodkar S, Arany Z, Kaelin Jr WG Bischo JR, Friedman PN, Marshak DR, Prives C and and Livingston DM. (1994). Cell, 78, 161 ± 172. Beach D. (1990). Proc. Natl. Acad. Sci. USA, 87, 4766 ± Krek W, Livingston DM and Shirodkar S. (1993). Science, 4770. 262, 1557 ± 1560. Blake MC and Azizkhan JC. (1989). Mol. Cell. Biol., 9, Krek W, Xu G and Livingston DM. (1995). Cell, 83, 1149 ± 4994 ± 5002. 1158. Brehm A, Miska EA, McCance DJ, Reid JL, Bannister AJ Krishan A. (1975). J. Cell. Biol., 66, 188 ± 193. and Kouzarides T. (1998). Nature, 391, 597 ± 601. Kuerbitz SJ, Plunkett BS, Walsh WV and Kastan MB. Bremner R, Cohen BL, Sopta M, Hamel PA, Ingles CJ, (1992). Proc. Natl. Acad. Sci. USA, 89, 7491 ± 7495. Gallie BL and Phillips RA. (1995). Mol. Cell. Biol., 15, La Thangue NB. (1994). Curr. Opin. Cell. Biol., 6, 443 ± 450. 3256 ± 3265. Lane DP. (1992). Nature, 358, 15 ± 16. Chellappan SP, Hiebert S, Mudryj M, Horowitz JM and Lin J, Chen J, Elenbaas B and Levine AJ. (1994). Genes Dev., Nevins JR. (1991). Cell, 65, 1053 ± 1061. 8, 1235 ± 1246. Chen J, Marechal V and Levine AJ. (1993). Mol. Cell. Biol., Lowe SW. (1995). Curr. Opin. Oncol., 7, 547 ± 553. 13, 4107 ± 4114. Lowe SW and Ruley HE. (1993). Genes Dev., 7, 535 ± 545. Debbas M and White E. (1993). Genes Dev., 7, 546 ± 554. Lowe SW, Ruley HE, Jacks T and Housman DE. (1993). DeGregori J, Kowalik T and Nevins JR. (1995). Mol. Cell. Cell, 74, 957 ± 967. Biol., 15, 4215 ± 4224. Magnaghi-Jaulin L, Groisman R, Naguibneva I, Robin P, DeGregori J, Leone G, Miron A, Jakoi L and Nevins JR. Lorain S, Le Villain JP, Troalen F, Trouche D and Harel- (1997). Proc. Natl. Acad. Sci. USA, 94, 7245 ± 7250. Bellan A. (1998). Nature, 391, 601 ± 605. Dobrowolski SF, Stacey DW, Harter ML, Stine JT and Martin K, Trouche D, Hagemeier C, Sorensen TS, La Hiebert SW. (1994). Oncogene, 9, 2605 ± 2612. Thangue NB and Kouzarides T. (1995). Nature, 375, 691 ± Dou QP, Markell PJ and Pardee AB. (1992). Proc. Natl. 694. Acad. Sci. USA, 89, 3256 ± 3260. Miyashita T and Reed JC. (1995). Cell, 80, 293 ± 299. Dynlacht BD, Flores O, Lees JA and Harlow E. (1994). Momand J, Zambetti GP, Olson DC, George D and Levine Genes Dev., 8, 1772 ± 1786. AJ. (1992). Cell, 69, 1237 ± 1245. Finlay CA, Hinds PW and Levine AJ. (1989). Cell, 57, 1083 ± Murphy M, Ahn J, Walker KK, Homan WH, Evans RM, 1093. Levine AJ and George DL. (1999). Genes Dev., 13, 2490 ± Han J, Sabbatini P, Perez D, Rao L, Modha D and White E. 2501. (1996). Genes Dev., 10, 461 ± 477. Murphy M, Hinman A and Levine AJ. (1996). Genes Dev., Hannon GJ and Beach D. (1994). Nature, 371, 257 ± 261. 10, 2971 ± 2980. Helin K, Wu CL, Fattaey AR, Lees JA, Dynlacht BD, Ngwu NipJ,StromDK,FeeBE,ZambettiG,ClevelandJLand C and Harlow E. (1993). Genes Dev., 7, 1850 ± 1861. Hiebert SW. (1997). Mol. Cell. Biol., 17, 1049 ± 1056. Hiebert SW. (1993). Mol. Cell. Biol., 13, 3384 ± 3391. O'Connor DJ, Lam EW, Grin S, Zhong S, Leighton LC, Hiebert SW, Blake M, Azizkhan J and Nevins JR. (1991). J. Burbidge SA and Lu X. (1995). EMBO J., 14, 6184 ± 6192. Virol., 65, 3547 ± 3552. Ohtani K, DeGregori J and Nevins JR. (1995). Proc. Natl. Hiebert SW, Chellappan SP, Horowitz JM and Nevins JR. Acad.Sci.USA,92, 12146 ± 12150. (1992). Genes Dev., 6, 177 ± 185. Olson DC and Levine AJ. (1994). Cell Growth Di., 5, 61 ± Hiebert SW, Packham G, Strom DK, Haner R, Oren M, 71. Zambetti G and Cleveland JL. (1995). Mol. Cell. Biol., 15, Olson DC, Marechal V, Momand J, Chen J, Romocki C and 6864 ± 6874. Levine AJ. (1993). Oncogene, 8, 2353 ± 2360. Hsieh JK, Fredersdorf S, Kouzarides T, Martin K and Lu X. Pearson BE, Nasheuer HP and Wang TS. (1991). Mol. Cell. (1997). Genes Dev., 11, 1840 ± 1852. Biol., 11, 2081 ± 2095.

Oncogene Inactive E2F-1 stabilizes but inactivates p53 J Nip et al 920 Phillips AC, Bates S, Ryan KM, Helin K and Vousden KH. Shaulsky G, Gold®nger N, Peled A and Rotter V. (1991). (1997). Genes Dev., 11, 1853 ± 1863. Cell Growth Dier., 2, 661 ± 667. Pomerantz J, Schreiber-Agus N, Liegeois NJ, Silverman A, Sherr CJ and Roberts JM. (1995). Genes Dev., 9, 1149 ± 1163. Alland L, Chin L, Potes J, Chen K, Orlow I, Lee HW, SorensenTS,GirlingR,LeeCW,GannonJ,BandaraLR Cordon-Cardo C and DePinho RA. (1998). Cell, 92, 713 ± and La Thangue NB. (1996). Mol. Cell. Biol., 16, 5888 ± 723. 5895. Qin XQ, Livingston DM, Kaelin Jr WG and Adams PD. Tao W and Levine AJ. (1999). Proc. Natl. Acad. Sci. USA, (1994). Proc. Natl. Acad. Sci. USA, 91, 10918 ± 10922. 96, 6937 ± 6941. Quelle DE, Zindy F, Ashmun RA and Sherr CJ. (1995). Cell, Trouche D, Cook A and Kouzarides T. (1996). Nucleic Acids 83, 993 ± 1000. Res., 24, 4139 ± 4145. Reich NC and Levine AJ. (1984). Nature, 308, 199 ± 201. Trouche D and Kouzarides T. (1996). Proc. Natl. Acad. Sci. Robertson KD and Jones PA. (1998). Mol. Cell. Biol., 18, USA, 93, 1439 ± 1442. 6457 ± 6473. Wang Q, Zambetti GP and Suttle DP. (1997). Mol. Cell. Sabbatini P, Chiou SK, Rao L and White E. (1995). Mol. Biol., 17, 389 ± 397. Cell. Biol., 15, 1060 ± 1070. Weber JD, Taylor LJ, Roussel MF, Sherr CJ and Bar-Sagi Sala A, Nicolaides NC, Engelhard A, Bellon T, Lawe DC, D. (1999). Nature Cell Biol., 1, 20 ± 26. Arnold A, Grana X, Giordano A and Calabretta B. Weinberg RA. (1991). Science, 254, 1138 ± 1146. (1994). Cancer Res., 54, 1402 ± 1406. Weintraub SJ, Prater CA and Dean DC. (1992). Nature, 358, Schener M, Munger K, Byrne JC and Howley PM. (1991). 259 ± 261. Proc. Natl. Acad. Sci. USA, 88, 5523 ± 5527. Wu X and Levine AJ. (1994). Proc. Natl. Acad. Sci. USA, 91, Schulze A, Zerfass K, Spitkovsky D, Middendorp S, Berges 3602 ± 3606. J, Helin K, Jansen-Durr P and Henglein B. (1995). Proc. Zhang HS, Postigo AA and Dean DC. (1999). Cell, 97, 53 ± Natl. Acad. Sci. USA, 92, 11264 ± 11268. 61. Serrano M, Hannon GJ and Beach D. (1993). Nature, 366, Zhang Y, Xiong Y and Yarbrough WG. (1998). Cell, 92, 704 ± 707. 725 ± 734. Shan B and Lee WH. (1994). Mol. Cell. Biol., 14, 8166 ± 8173. Zindy F, Eischen CM, Randle DH, Kamijo T, Cleveland JL, Shaulsky G, Ben-Ze'ev A and Rotter V. (1990). Oncogene, 5, Sherr CJ and Roussel MF. (1998). Genes Dev., 12, 2424 ± 1707 ± 1711. 2433.

Oncogene