Small Contribution of G1 Checkpoint Control Manipulation to Modulation of P53-Mediated Apoptosis

Small contribution of G1 checkpoint control manipulation to modulation of p53-mediated apoptosis

Christine E Canman and Michael B Kastan

Johns Hopkins Oncology Center, 345 Ross Building, 720 Rutland Avenue, Baltimore, Maryland 21205, USA

Deregulation of the S-phase promoting E2F-1 transcrip- phase promoting E2F-1 transcription factor, has been tion factor has been shown to cooperate with p53 to linked to increasing the susceptibility of cells to p53- induce apoptosis. BaF3 cells undergo rapid, p53- mediated apoptosis (Lowe et al., 1993; Debbas and dependent apoptosis when irradiated in the absence of White, 1993; Lowe and Ruley, 1993; Morgenbesser et IL-3. Rapid apoptosis induced by ionizing radiation (IR) al., 1994; Pan and Griep, 1994; White, 1994; Wu and coincides with attenuated p21(WAF1/Cip1) induction. Levine, 1994; Qin et al., 1994). Failure to adequately induce p21 could result in The G1 cell cycle checkpoint activated in response to inappropriate release of E2F from Rb which may then DNA damage is mediated to a large extent by p53- cooperate with p53 to induce apoptosis in cells deprived dependent transcriptional activation of p21WAF1/CIP1. p21 of growth factor. We engineered BaF3 cells to express was originally identi®ed as a component of G1 cyclin- exogenous p21 and tested whether overexpressing p21 in cdk-PCNA quaternary complexes (Xiong et al., 1993b), cells irradiated in the absence of IL-3 protects from IR- whose activities are required for cell cycle progression induced apoptosis. Enforced p21 expression resulted in a from G1 into S (Sherr, 1994). p21 was later discovered consistent, but partial, protection of cells from under- to be a universal inhibitor of G1 cyclin-cdk activity, in going IR-induced apoptosis. However, deregulating E2F addition to a transcriptional target of p53 (Xiong et al., activity through expression of HPV E7 failed to sensitize 1993a; Harper et al., 1993; El-Deiry et al., 1993). The cells to IR-induced apoptosis in the presence of IL-3. ability of p21 to inhibit cdk activity appears to depend Together, these data strongly suggest that the IL-3- upon the stoichiometric relationship of the number of responsive factors which modulate p53-mediated apopto- p21 molecules bound to G1 cyclin-cdk complexes sis in BaF3 cells are largely independent of G1 cell cycle (Zhang et al., 1994; LaBaer et al., 1997). Therefore, checkpoint control mediated by p21. in cells exposed to ionizing radiation (IR), p53- dependent increases in p21 synthesis results in the Keywords: apoptosis; p53; p21; E2F inactivation of G1 cyclin-associated cdk2 complexes (Dullic et al., 1994; El-Deiry et al., 1994). Proof that p21 is a major eector of p53-mediated G1 arrest came from the observations that homozygous deletion of the Introduction p21 gene partially abrogates IR-induced G1 arrest in murine p217/7 embryonic ®broblasts and completely The decision of a cell to undergo p53-dependent G1 disrupts the G1 arrest response in a human tumor cell arrest or apoptosis appears to depend upon the line (Deng et al., 1995; Brugarolas et al., 1995; cellular context in which p53 is activated. It is well Waldman et al., 1995). established that certain cell types are extremely The S-phase promoting E2F family of transcription sensitive to p53-mediated apoptosis (e.g. lymphoid factors appear to be critical downstream targets of cells), whereas others typically undergo G1 arrest and p21 inhibitory activity, thereby implicating Rb and are relatively resistant to apoptosis (e.g. ®broblast related pocket proteins, p107 and p130, in the p53- cells). Only a few cellular factors that modulate this dependent DNA damage response pathway. Tran- decision have been identi®ed thus far. The presence or scriptionally active `E2F' is a heterodimer composed absence of speci®c growth factors or the actual levels of an E2F family member (E2F-1 through 5) and of p53 expressed within a cell can dictate which DP (DP-1 or 2) (Weinberg, 1995). The activity of pathway is chosen (Yonish-Rouach et al., 1991; these dimers is tightly regulated through their Canman et al., 1995; Lin and Benchimol, 1995; associations with hypophosphorylated Rb and re- Abrahamson et al., 1995; Chen et al., 1996; Ronen lated molecules in a cell cycle-dependent manner. In et al., 1996). The ability of cells to undergo p53- a simple model of cell cycle progression, sequential mediated apoptosis can also be greatly reduced phosphorylation of Rb by G1 cyclin-cdk results in through enhanced expression of the bcl-2 oncoprotein the release of active E2F, which then proceeds to and related proteins (Wang et al., 1995; Guillouf et transcriptionally activate genes necessary for S-phase al., 1995; Chiou et al., 1994). On the other hand, loss progression (Weinberg, 1995). Inhibition or repression of G1 cell cycle checkpoint control, such as through of E2F activity by Rb and related proteins appears inactivation of the retinoblastoma (Rb) tumor to be critical for p53-mediated G1 arrest, since suppressor gene product or overexpression of the S- expression of either the adenovirus E1A or human papillomavirus (HPV) E7 viral oncogenes, both of which disrupt E2F binding to Rb and related molecules, or overexpression of E2F-1 itself, can Correspondence: CE Canman Received 1 August 1997; revised 30 September 1997; accepted 30 overcome IR-induced G1 arrest (Slebos et al., 1994; September 1997 Demers et al., 1994; Lowe et al., 1993b; Degregori et G1 checkpoint control and p53-mediated apoptosis CE Canman and MB Kastan 958 al., 1995a,b). Consistent with this model, overexpres- Results sion of p21 signi®cantly inhibits E2F transcriptional activity, presumably by preventing phosphorylation of Withdrawal of IL-3 results in attenuated induction of Rb and related molecules by G1 cyclin-cdk's (Dimri et p21 protein by IR al., 1996; Shiyanov et al., 1996). Deregulation of E2F activity is strongly implicated The levels of p21 expressed in a cell are regulated by as a mechanism by which abrogation of Rb function both p53-dependent and independent mechanisms sensitizes cells to p53-mediated apoptosis. E1A (Macleod et al., 1995; Michieli et al., 1994). Growth expression dramatically sensitizes murine embryonic factor stimulation of the p21 promoter can occur in the ®broblasts and baby rat kidney cells to p53-mediated absence of p53 (Canman et al., 1995; Michieli et al., apoptosis (Lowe and Ruley, 1993; Debbas and White, 1994) and has been shown to be largely dependent 1993; Lowe et al., 1993). Abrogation of Rb function upon activation of the Ras/Raf/MAPK pathway (Lui leads to deregulation of cell growth and extensive p53- et al., 1996). To con®rm our previous ®ndings that the dependent apoptosis in the developing lens of mice, induction of p21 by IR is attenuated in cells withdrawn linking loss of E2F regulation with apoptosis of IL-3, a detailed time course was conducted (Morgenbesser et al., 1994; Pan and Griep, 1994). comparing the absolute levels of p21 protein in Convincing evidence now suggests that p53 and E2F-1 irradiated and unirradiated cells cultured in either the can cooperate to induce apoptosis (Wu and Levine, absence or presence of growth factor. Withdrawal of 1994; Qin et al., 1994). These results would predict IL-3 was associated with rapid downregulation of p21 that elevated levels of p21 may protect from p53- levels as detected by immunoblot analysis (Figure 1). mediated apoptosis by virtue of its ability to indirectly When BaF3 cells were irradiated in the absence of IL- regulate E2F activity. In fact, disruption of p21 3, p53-dependent induction of p21 was signi®cantly less expression has been reported to sensitize a human as compared to cells irradiated in the presence of IL-3 tumor cell line and mouse embryonal ®broblasts to and only accumulated approximately to the basal levels p53-mediated apoptosis (Polyak et al., 1996; Gorospe observed in proliferating cells cultured in IL-3. et al., 1997). Another member of the cyclin-cdk inhibitor family, To search for molecular factors which may aect the p27Kip1, has been shown to be required for control of decision of cells to undergo G1 arrest versus apoptosis, the restriction point in the G1 phase of the cell cycle in we have been utilizing a model system where these murine ®broblasts (Coats et al., 1996). Withdrawal of pathways can be modulated by a growth factor. IL-3 mitogenic stimuli resulting in cell cycle arrest is dependent murine hematopoietic BaF3 cells undergo associated with induction of p27 (Nourse et al., p53-dependent G1 arrest when irradiated in the 1994). Although p21 is rapidly downregulated by IL- presence of IL-3 and rapid p53-dependent apoptosis 3 withdrawal in Baf-3 cells, p27 may be induced at the when irradiated in the absence of growth factor same time. This could keep the threshold needed for (Collins et al., 1992; Canman et al., 1995). We cyclin-cdk inhibition by p21 relatively constant in cells previously reported that IL-3 withdrawal results in irradiated in the absence of IL-3. However, immuno- rapid downregulation of p21 levels and that the levels blot analysis of p27 levels in BaF3 cells deprived of IL- of p21 induced by p53 in irradiated cells were 3 showed no change (TM Gilmer, unpublished signi®cantly less in the growth factor-deprived state observations). (Canman et al., 1995). Reduced transactivation of p21 correlated with apoptosis and an apparent failure to Characterization of BaF-3 derivatives which induce undergo G1 arrest. One potential explanation for the exogenous p21 rapid, irradiation-induced apoptosis observed in the absence of IL-3 was that accumulated p21 levels were Attenuated induction of p21 in response to IR insucient to inhibit G1 cyclin/cdk complexes. This correlates with an apparent failure of BaF3 cells to could contribute to the apoptotic pathway by resulting arrest in G1 and rapid apoptosis (Canman et al., 1995). in deregulation of active E2F in the presence of To test whether inadequate p21 induction is respon- induced p53. To directly test this model, BaF3 cells sible for IR-induced apoptosis when cells are cultured were engineered to express exogenous p21 protein in the absence of growth factor, an inducible system independent of p53 in order to determine whether was employed to restore p21 levels and enforce a G1 restoring p21 protein levels and G1 cell cycle cell cycle arrest under these conditions. The inducible checkpoint control protected from p53-mediated system utilizes the GalERVP transcriptional activator, apoptosis. Even though we were able to completely a fusion protein containing the yeast Gal4 DNA arrest cells in G1, overexpressing p21 only partially protected BaF3 cells from undergoing p53-mediated apoptosis when cultured in the absence of growth factor. Conversely, abrogration of G1 checkpoint + IL-3 –IL-3 control through expression of HPV E7 and resulting 2 4 6 8 2 4 6 8 (Hours) deregulation of E2F activity failed to sensitize BaF3 – + – + – + – + – + – + – + – + (4 Gy) cells to p53-mediated apoptosis in the presence of IL-3. These results suggest that inappropriate release of free p 21 E2F in the presence of active p53 is not sucient for activating the p53-dependent apoptotic pathway in Figure 1 Withdrawal of IL-3 is associated with attenuated p21 BaF3 cells. Other factors, most likely regulated by IL-3 induction by IR. Immunoblot analysis of whole cell lysates obtained from BaF3 cells irradiated in the presence or absence of signaling, determine the fate of these cells following IL-3 and incubated for the indicated times following exposure to activation of p53. IR and/or IL-3 deprivation G1 checkpoint control and p53-mediated apoptosis CE Canman and MB Kastan 959 binding domain, the estrogen receptor ligand binding Consistent with inhibition of G1 cyclin-cdk complexes domain, and the strong transactivation domain of by exogenous p21, Rb accumulated in its unpho- VP16 (Braselmann et al., 1993). GalERVP is activated sphorylated form following b-estradiol treatment following binding to its ligand, b-estradiol, and induces (Figure 2a). In addition, the Waf1.12 clone appeared the expression of a gene placed behind a minimal to induce greater levels of p21 in response to b- promoter containing four Gal4 DNA binding sites. estradiol such that by 9 h, the majority of Rb had BaF3 cells were infected with retrovirus encoding the accumulated in its hypophosphorylated form. GalERVP gene and then stably transfected with the The ability of exogenous p21 to induce cell cycle GalERVP responsive plasmid into which the cDNA arrest was directly evaluated by BrdUrd pulse labeling encoding murine p21 had been subcloned. in Waf1.12 and Waf2.18 cells (Figure 2b). Cells Two independent BaF3 clones (derived from cultured with and without b-estradiol were pulse separate transfections) expressing GalERVP and labeled with BrdUrd for 45 min, stained for BrdUrd stably transfected with the GalERVP-responsive and DNA content, and then analysed by FACS plasmid encoding murine p21 cDNA were character- analysis. By 9 h of b-estradiol treatment, BrdUrd ized. In response to b-estradiol treatment, both incorporation into DNA was strongly inhibited in Waf1.12 and Waf2.18 cells upregulate exogenous p21 both Waf1.12 and 2.18 cells. The induction of p21 was levels (Figure 2a). GalERVP cells, which only express also associated with a large increase in the G1/S ratio the GalERVP fusion protein, serve as a control. indicative of cell cycle arrest in G1. We also note that Inhibition of G1 cyclin-cdk activity by exogenous p21 prolonged treatment of control GalERVP cells with b- was assessed indirectly by analysing the phosphoryla- estradiol produced a slight increase in the G1/S ratio. tion status of Rb, a known substrate for cdk activity. Activation of the GalERVP transactivator for long

a GalERVP Waf1.12 Waf2.18 0 3 9 24 0 3 9 24 0 3 9 24 (Hours) P 21

Rb-p Rb

Figure 2 Induction of exogenous p21 results in the accumulation of hypophosphorylated Rb and cell cycle arrest in G1. (a) Following treatment with 1 mM b-estradiol for 0, 3, 9 or 24 h, whole cell lysates from GalERVP, Waf1.12, and Waf2.18 cells were resolved by SDS ± PAGE in either 12% (p21) or 7.5% polyacrylamide gels (Rb) and then subjected to immunoblot analysis. (b)A fraction of these cells were pulse-labeled with BrdUrd and analysed by FACS after staining for DNA content with propidium iodide and BrdUrd with FITC-conjugated antibodies directed towards BrdU. The ratio of cells residing in G1 versus S (G1/S) is noted in the lower right hand corner of each histogram G1 checkpoint control and p53-mediated apoptosis CE Canman and MB Kastan 960 a

c d

e GalERVP Waf1.12 U S A A B C C D D U S A A B C C D D E E E E E E — Rb-p — Rb

— p107-p — p107

Figure 3 Overexpressed p21 partially protects against IR-induced apoptosis in pre-synchronized BaF3 cells. Waf1.12 cells were synchronized in G1 by IL-3 deprivation for 15 h (a). Cells were then driven to enter the cell cycle by re-addition of IL-3 for 4 h. A fraction of the culture was treated with 1 mM b-estradiol for 1.5 h prior to manipulation of IL-3 or IR exposure. Cells (+/7 b- estradiol) were then exposed to either 0 or 4 Gy IR in the presence (b) or absence (c) of IL-3 and analysed by FACS 8 h post IR: cells were harvested and end-labeled in situ with biotinylated dUTP and FITC-conjugated avidin to detect apoptotic cells and then counterstained for DNA content with propidium iodide. (d) Summary of results analysing the protective eect of p21 against IR- induced apoptosis in pre-synchronized GalERVP (control), Waf1.12, and Waf2.18 cells cultured in the absence of IL-3. Data is expressed as the fraction of apoptosis speci®cally induced by IR in cells treated versus untreated with b-estradiol (mean+s.e., n=3 or 4). The amount of apoptosis induced by the synchronization procedure (no irradiation) alone was subtracted prior to data calculation. (e) Whole cell lysates were resolved by SDS ± PAGE in 7.5% polyacrylamide gels and subjected to immunoblot analysis. U, unsynchronized; S, synchronized (7IL-3, 15 h); synchronized cells 8 h post IR: (a) +IL-3/0 Gy; (b) +IL-3/4 Gy; (c) 7IL-3/ 0 Gy; (d) 7IL-3/4 Gy; (e) b-estradiol G1 checkpoint control and p53-mediated apoptosis CE Canman and MB Kastan 961 periods of time appears to have slight anti-proliferative the absence or presence of IL-3 (Figure 3e). In eects on BaF3 cells since parental BaF3 cells were summary, presynchronizing cells prior to IR exposure unaected by b-estradiol treatment (data not shown). resulted in greater protection from p53-dependent apoptosis by elevated p21 levels. It is recognized that synchronization of BaF3 cells Eects of induced p21 expression upon p53-dependent through IL-3 withdrawal complicates interpretation of apoptosis in unsynchronized versus synchronized cells the data discussed above, especially since this To test whether enforced G1 arrest in BaF3 cells procedure induces some degree of apoptosis regardless irradiated in the absence of IL-3 protects from p53- of exposure to IR. These results are further dependent apoptosis, Waf1.12 and 2.18 cells were complicated by the apparent eects of the GalERVP induced to express exogenous p21 2 h prior to IL-3 transcriptional activator alone on IR-induced apoptosis withdrawal and irradiation, a period of time sucient in BaF3 cells. We therefore sought to utilize a dierent to induce p21 protein levels in these cells. Cells were approach to elevate p21 in these cells prior to IR then harvested 8 h later and stained for DNA content exposure. Retroviral-mediated gene transfer is an with propidium iodide and for apoptosis through TdT- ecient, relatively non-toxic method used to introduce mediated end-labeling of apoptotic DNA fragments exogenous genes into cells. To circumvent the problem with biotinylated dUTP. Cells were then analysed by of low infection eciencies of hematopoietic cells and FACS. Although both clones underwent G1 arrest in to avoid drug selection of infected cells, the neomycin response to p21 induction, we failed to observe a resistance gene of the retroviral vector, LXSN, was consistent protective eect by exogenous p21 on IR- replaced with human CD4 cell surface marker cDNA induced apoptosis in either clone, especially the clone to permit selective detection of infected cells by FACS which induces less p21 (Waf2.18) (data not shown). analysis. This vector was then engineered such that it The failure of p21 to protect exponentially growing also encoded DNA sequences for either p21 (Waf1) or BaF3 cells from apoptosis could be due to the fact that HPV16 E6, the latter serving as a positive control for a signi®cant percentage of cells have already passed the complete protection of IR-induced apoptosis in BaF3 G1 cell cycle checkpoint at the time of exposure to IR cells (Canman et al., 1995). The morning after immediately following growth factor withdrawal. We infection, a period of time sucient to allow therefore presynchronized cells in G1 and re-examined expression of CD4 protein from integrated retrovirus, whether p21 protects from IR-induced apoptosis. Cells cells were washed free of IL-3 and irradiated. Eight were synchronized in G1 by withdrawing growth factor hours later, cells were harvested stained with anti-CD4 for 13 ± 15 h (Figure 3a), and then driven back into the phycoerythrin-conjugated antibody and FITC-conju- cell cycle through exposure to IL-3 for an additional 4 h prior to subsequent manipulation of IL-3 and/or irradiation. Some samples of the presynchronized cells were also induced to express p21 by adding b-estradiol 2 h prior to growth factor removal and irradiation. After 8 h, cells were harvested and stained for FACS analysis. At the time of harvesting, control (unirradiated) cells cultured in IL-3 were just beginning to re- enter into S (Figure 3b). The induction of exogenous p21 prevented cells from progressing into S as expected

(+E2), as did exposing cells to 4 Gy IR (Figure 3b, right panel). Further analysis of unirradiated, growth factor deprived cells also indicates that these cells were still committed to enter into S phase by the brief 4 h treatment with growth factor (7E2) and induction of p21 prevented this cell cycle progression (+E2) (Figure 3c). Exposure to IR resulted in enhnaced apoptosis in cells primarily residing in G1 and early S and the induction of exogenous p21 blocked the appearance of the majority of this apoptotic population (circled in Figure 3c). The protective eect of p21 in presynchronized cells was more consistent in protecting against IR-induced apoptosis than had been observed in unsynchronized cells (Figure 3d). Again, Waf1.12 cells Figure 4 Infection of BaF3 cells with retrovirus encoding p21 displayed greater protection as compared to the partially protects from IR-induced apoptosis. BaF3 cells were Waf2.18 clone when exogenous p21 was induced just infected with retrovirus encoding CD4 alone or CD4 and either prior to and following exposure to IR. We also note p21 or HPV E6. The next day, cells were washed free of IL-3 and that activation of GalERVP appears to have some immediately irradiated. Eight hours later, cells were stained with PE-conjugated anti-CD4 and FITC-conjugated annexin to detect protective eect on IR-induced apoptosis in this assay. infected and apoptotic cells, respectively. The three dot plots To con®rm that induction of p21 was causing cell cycle show annexin V-FITC staining pro®les of CD4-positive cells arrest in this system, a sample of the cells shown in infected with either control, Waf1, or HPV E6 retrovirus Figure 3a, b and c were also immunoblotted for Rb following FACS analysis. The bar graph represents quantitative and p107. Induction of p21 resulted in the accumula- assessments (mean+s.e.) of four independent experiments where data are expressed as the percentage of CD4 positive/annexin V- tion of both Rb and p107 in their active, hypopho- positive cells observed in Waf1 or E6 infected cells as compared sphorylated forms in presynchronized cells cultured in to control G1 checkpoint control and p53-mediated apoptosis CE Canman and MB Kastan 962 gated annexin V to detect apoptotic cells, and then antisense oligonucleotides or antisense RNA expression immediately acquired by FACS. When CD4 positive were unsuccessful. We therefore chose to generate cels were analysed (Figure 4), cells which were infected BaF3 cells utilizing the GalERVP system, which induce with retrovirus encoding p21 were clearly protected the HPV16 E7 protein, known to disrupt E2F from IR-induced apoptosis, revealing a consistent, but complexes with Rb, p107, and p130. The b-estradiol- partial eect in unsynchronized cells. As expected, inducible expression of E7 in one such clone (E7 1.2) is BaF3 cells infected with HPV E6 retrovirus, which shown in Figure 5a. An E2F gel mobility shift assay targets p53 for proteolytic degradation thereby performed on nuclear extracts demonstrated that the abrogating p53 function, were almost completely majority of E2F DNA binding activity existed as either protected from IR-induced apoptosis. free E2F or as a multi-protein complex consisting of E2F, p107, cyclin A or E, and cdk2 (as determined by antibody supershift analysis, data not shown) typically Disruption of E2F/p107 complexes through expression of observed in proliferating cells (Shirododkar et al., HPV E7 fails to sensitize BaF3 cells to p53-mediated 1992; Lees et al., 1992; Schwarz et al., 1993) (Figure apoptosis 5b). Expression of HPV E7 clearly disrupted E2F/ If loss of G1 cell cycle checkpoint control caused by p107/cyclin/cdk2 complexes in E7 1.2 cells, even in the withdrawing IL-3 truly determined sensitivity to p53- absence of b-estradiol induction (Figure 5b). This mediated apoptosis in BaF3 cells, then downregulation indicates that even a relatively low level of E7 of p21 expression should promote IR-induced apopto- expression is sucient to eectively disrupt p107 sis even in the presence of growth factor. Repeated binding to E2F. The location of `free' E2F DNA attempts to downregulate p21 expression using either binding activity was identi®ed by treating extracts with

b GalERVP E7 wt mt D – – – + + – + E2

a p107 GalER E7 1.2 – + – + E 2 Free E2F d E7

Figure 5 Expression of HPV16 E7 dissociates p107/cdk2/E2F complexes and abrogates IR-induced G1 arrest but fails to increase the incidence of IR-induced apoptosis in cells cultured in IL-3. (a) Immunoblot analysis of E7 expression in GalERVP (control) or E7 1.2 cells treated with 1 mM b-estradiol overnight. (b) Nuclear extracts were also isolated and subjected to EMSA utilizing an E2F- speci®c 32P-end-labeled probe. To con®rm binding speci®city, extracts were incubated with either a 50-fold excess of wild type (wt) or mutant (mt) un-labeled E2F probe. The location of free E2F, as determined by Deoxycholate treatment (d), and p107/cdk2/E2F (p107) complexes (as determined by antibody supershift analysis, not shown) are indicated on the right. (c) GalERVP or E7 1.2 cells were treated overnight with 1 mM b-estradiol and then exposed to 0 or 4 Gy IR. Ten hours later, cells were pulse-labeled with BrdUrd and analysed by FACS as in Figure 2. Data is expressed as the fraction of cells in G1 versus S (G1/S) from three independent experiments (indicated by the three dierent shades of grey). (d) GalERVP and E7 1.2 cells treated or untreated with b- estradiol overnight were exposed to 4 Gy IR Cells were harvested 32 h later, stained for DNA content and apoptosis as in Figure 3 and then subjected to FACS analysis G1 checkpoint control and p53-mediated apoptosis CE Canman and MB Kastan 963 deoxycholate, a mild detergent which dissociates higher order complexes of E2F found in association with Rb, a p107, or p130, but leaves E2F/DP DNA binding activity intact (Bagchi et al., 1990; Wu et al., 1995). As expected, a signi®cant increase in free E2F DNA binding activity was observed in E7 expressing cells. To con®rm that the p53-mediated G1 arrest pathway is abrogated in the E7 1.2 clone, cells were induced to express E7 by an overnight incubation with b- estradiol, irradiated, and then harvested 9 ± 10 h later after brief pulse-labeling with BrdUrd. Cells were stained and subjected to FACS analysis as in Figure b 2b. The fraction of cells found in G1 and S were calculated and expressed as the G1/S ratio. Expression of HPV E7 eectively abrogated the G1 arrest response to IR in BaF3 cells as indiciated by little change in the G1/S ratio in irradiated E7 1.2 cells (Figure 5c). This is in contrast to control GalERVP cells, which exhibited an increase in the G1/S ratio following IR exposure. However, E7 expression failed to increase the incidence of apoptosis induced by IR in E7 1.2 cells cultured in the presence of IL-3 (Figure 5d). Expression of E7 also failed to accelerate IR-induced apoptosis in BaF3 cells irradiated and cultured in the absence of IL-3, as well (data not shown). Therefore, this data strongly Figure 6 Infection of BaF3 cells with retrovirus encoding HPV suggests that apoptosis-protecting eect of IL-3 is E7 abrogates G1 arrest in response to IR but fails to promote dominant over the apoptosis-inducing eects of apoptosis in the presence of IL-3. Cells were infected with retrovirus encoding CD4 or CD4 and either HPV E6 or E7. (a) deregulated E2F activity in irradiated BaF3 cells. The next morning, a fraction of each population was irradiated in Since enough HPV E7 is expressed in these cells in the presence of IL-3, harvested and stained 8 h later for CD4 and the absence of b-estradiol to liberate E2F from binding apoptosis as in Figure 4. FACS analysis dot plots showing lack of to p107, the possibility exists that BaF3 clones could annexin V-FITC staining of CD4 positive cells infected with either control or E7 retrovirus are shown (compare staining have been selected for that are resistant to apoptosis in observed in Figure 4). (b) A portion of the infected cells were pre- response to deregulated E2F. We therefore infected synchronized in G1 through IL-3 withdrawal, re-exposed to BaF3 cells with retrovirus encoding both CD4 and growth factor for 2 h, and then irradiated. Seven hours later, cells HPV E7 so that it could be rapidly assessed whether were pulse-labeled with BrdUrd, stained with PE-conjugated anti- E7 promoted apoptosis in irradiated BaF3 cells CD4 and then FITC-conjugated anti-BrdU antibody. The percentage of unirradiated or irradiated CD4 positive cells without generating stable transfectants through clonal stained with anti-BrdU antibody (% cells in S phase) is shown selection. Again, E7 expression failed to promote IR- for each population of cells infected with the indicated retrovirus induced apoptosis in cells irradiated in the presence of IL-3 (Figure 6a). Cells infected with E7 retrovirus did lose the G1 cell cycle checkpoint demonstrating that E7 apoptosis induced by p53 and other stimuli (White, is functional in this assay; however, pre-synchroniza- 1996). However, we did ®nd that IL-3 withdrawal leads tion of cells was necessary to observe abrogation of to rapid downregulation of p21, as shown here, and p53-mediated G1 arrest using this double-staining this appears to attenuate the induction of p21 by p53 method. In this experiment, cells infected with in response to IR. Since the levels of p21 induced in control, HPV E6, or E7 retrovirus were synchronized these cells only approached those observed in in G1 through IL-3 withdrawal, re-exposed to growth proliferating cells cultured in IL-3, G1 cyclin-cdk factor for 2 h and irradiated. Seven hours later, cells activity would not be suciently inhibited, thereby were pulse-labeled with BrdUrd, harvested and then abolishing Rb regulatory control over E2F activity stained for CD4 and BrdUrd. Loss of the G1 cell cycle coincident with activation of p53 by IR. The checkpoint is demonstrated by a greater percentage of observations that the earliest cells undergoing apopto- CD4 positive cells infected with E7 retrovirus which sis in response to IR reside in S phase, supported this stained positive for BrdUrd following irradiation, as hypothesis (Canman et al., 1995). Given the substantial compared to control infected cells (Figure 6b). evidence suggesting that deregulation of E2F transcriptional activation cooperates with p53 to induce apoptosis, we asked whether attenuated p21 induction Discussion by p53 in growth factor-deprived BaF3 cells contributes to IR-induced apoptosis. We have previously shown that BaF3 cells irradiated in Two approaches were taken to ascertain whether an the presence of growth factor transiently arrest in G1 apparent loss of G1 cell cycle checkpoint control in and G2/M phases of the cell cycle, whereas BaF3 cells BaF3 cells irradiated in the absence of IL-3 contributes irradiated in the absence of IL-3 undergo rapid p53- to rapid p53-dependent apoptosis. First, BaF3 cells dependent apoptosis (Canman et al., 1995). IR-induced were genetically manipulated, by either an inducible apoptosis in the absence of IL-3 failed to temporally system or retroviral infection, so that p21 levels could correlate with any changes in the levels of bcl-2, bcl-xL, be controlled independently of p53. Inducing G1 arrest or bax, proteins known to either protect or promote through overexpression of p21 failed to consistently G1 checkpoint control and p53-mediated apoptosis CE Canman and MB Kastan 964 protect IL-3-deprived BaF3 cells from undergoing IR- dierent receptor systems, including those activated by induced apoptosis. This occurred despite Rb being cytokines (Kinoshita et al., 1995; Xia et al., 1995). These mantained in its hypophosphorylated form and results are supported by the ®ndings that expression of competent for inhibiting E2F transcriptional activa- constitutively activated Ras or Raf protect from tion. Consistent protection was observed if cells were apoptosis induced by growth factor deprivation and pre-synchronized in G1 prior to removal of IL-3 and p53 (Canman et al., 1995; Lin et al., 1995a; Kinoshita et exposure to IR; however, even in this setting, the al., 1995; Cleveland et al., 1994; Terada et al., 1995). protective eect of overexpressed p21 was only partial. Paradoxically, introduction of a dominant-negative Ras In contrast to the inducible system, increasing p21 protein failed to promote apoptosis in BaF3 cells levels through retroviral infection did reveal a cultured in IL-3, arguing against Ras-dependent path- consistent, but partial protective eect in unsynchro- ways playing a role in anti-apoptosis signaling in these nized cells. Based on both these results, one must cells (Terada et al., 1995). Recently, the Janus kinase conclude that deregulation of E2F activity, as a (Jak) family of kinases, which regulate the Stat (Signal consequence of attenuated p21 induction, makes a transducers and activators of transcription) family of relatively small contribution to increasing the suscept- transcription factors, have received attention as impor- ibility of IL-3-deprived BaF3 cells to p53-mediated tant mediators of cytokine signal transduction. Activa- apoptosis induced by IR. tion of Jak2 and Stat5 have been implicated in both In the second approach, E2F activity was deregu- survival and mitogenic signaling transduced by receptors lated through expression of the HPV E7 oncoprotein for erythropoietin and IL-3 and a chimeric receptor and we tested whether E7 promoted IR-induced consisting of the epidermal growth factor receptor fused apoptosis when cells were cultured in the presence of to the Jak2 tyrosine kinase domain (Zhang et al., 1995; IL-3. Although E7 dissociated the majority of E2F Mui et al., 1996; Nakamura et al., 1996). complexed with p107, resulting in substantial increases A major focus of future work will be to identify the in free E2F DNA binding activity, expression of E7 key survival signal transduction pathways activated by and loss of G1 checkpoint control failed to increase the IL-3 and their corresponding downstream targets. incidence of apoptosis in BaF3 cells irradiated and Survival signaling may be conveyed at the level of cultured in the presence of IL-3. This result contrasts transcriptional activation of essential genes or through with observations made in other systems where post-translational modi®cation of constitutively ex- deregulation of E2F-1 activity cooperates with p53 to pressed proteins, whose activities dictate the balance induce apoptosis (Lowe and Ruley, 1993; Debbas and between survival or death. The latter alternative has White, 1993; Lowe et al., 1993; Morgenbesser et al., recently been exempli®ed by the ®nding that IL-3 1994; Pan and Griep, 1994; Wu and Levine, 1994; Qin induces rapid phosphorylation of the apoptosis- et al., 1994; Hiebert et al., 1995). However, the results promoting BAD protein, thereby antagonizing its obtained in BaF3 cells versus other systems may be due ability to accelerate cell death through its interaction

to dierences in the amounts of free E2F-1 present. with Bcl-xL or Bcl-2 (Zha et al., 1996). Identi®cation of Perhaps greater levels of E2F activity, speci®cally E2F- these pathways are crucial not only to our under- 1, are required to override survival signaling mediated standing of tumor development, but also in the by the IL-3 receptor (Hiebert et al., 1995; Degregori et development of new therapeutic strategies for the al., 1997). Similar to the dominant eects of IL-3 treatment of cancer directed towards antagonizing observed here, the pro-apoptotic eects adenovirus inappropriate survival signaling. E1A and E2F-1 appear to be lessened by the presence of serum (Lowe and Ruley, 1993; Shan and Lee, 1994). Our results are also consistent with recently published Materials and methods data suggesting that the apoptosis-promoting activity of E2F-1 may actually be independent of its ability to Plasmid constructions induce progression into S phase (Degregori et al., 1997; cDNA encoding murine p21 or HPV16 E7 were subcloned Phillips et al., 1997). into GCpSP65, a mammalian expression vector containing Together, these data strongly suggest that the IL-3- four Gal4 DNA binding sites upstream of a minimal responsive factors which modulate p53-mediated apop- promoter (Braselmann et al., 1993). Murine p21 (WAF-1) tosis in BaF3 cells are largely independent of G1 cell cDNA was excised from pCMW35 (provided by Bert cycle checkpoint control. Multiple signal transduction Vogelstein, Johns Hopkins University, Baltimore, MD) with EcoRI and ®lled in with Klenow polymerase. HPV16 pathways are activated in response to IL-3 and three of E7 cDNA was excised from pCMVE7 (provided by these have been reported to transduce anti-apoptotic Kathleen Cho (Slebos et al., 1994)) with BamHI and also signals in other model systems. Phosphatidylinositol-3 ®lled in with Klenow polymerase. Both blunt-ended kinase (p110/PI-3 kinase) activity, and its target Protein fragments were ligated separately into the EcoRV site kinase B(Akt), play a critical role in the prevention of downstream of the minimal promoter within GCpSP65 apoptosis by Nerve Growth Factor, PDGF, and IGF-1 generating pGCmWaf-1 and pGC16E7. To construct (Yao and Cooper, 1995; Dudek et al., 1997; Kaufmann- LXSCD4, the neomycin phosphotransferase (neo)gene Zeh et al., 1997). However, PI-3 kinase activity appears was excised from LXSN (Miller et al., 1993) through to be dispensable for survival signal transduction digestion with HindIII and NaeI and replaced with an mediated by the IL-3 receptor since the PI-3 kinase EcoRI ± BamHI, 1.8 kb human CD4 cDNA fragment isolated from pGE1b/CD4 (gift of Chi Dang, Johns inhibitor, Wortmannin, failed to aect survival of IL-3- Hopkins University). Murine p21, was subcloned into the dependent FDCP myeloid cells (Minshall et al., 1996) EcoRI site, whereas HPV16 E7 and E6 (also provided by and BaF3 cells following irradiation (data not shown). Kathleen Cho) were subcloned into the BamHI site of Activation of the Ras/Raf/MAP kinase pathway has also LXSCD4 generating LXSCD4-Waf1, LXSCD4-E6 and been implicated in survival signaling mediated by several LXSCD4-E7 retroviral vectors. G1 checkpoint control and p53-mediated apoptosis CE Canman and MB Kastan 965

Cell culture, DNA transfection and retroviral infection 10 mM KCl, 2 mM MgCl2, and 25% glycerol) containing 0.1% Triton X-100 and the following protease inhibitors Murine, IL-3-dependent BaF3 cells (Palacios et al., 1985) (Sigma Chemical Company): 1 mM PMSF, 2 mg/ml were maintained in RPMI 1640 containing 10% fetal Pepstatin A, 20 mg/ml leupeptin and 10 mg/ml Aprotinin. bovine serum and 10% WEHI 3B-conditioned medium as a Nuclei were then pelleted and extracted for 30 min on ice source of murine IL-3. To remove IL-3, cells were washed with 150 ml of high salt buer (hypotonic buer adjusted two to three times with serum free RPMI, and then to 400 mM NaCl) containing protease inhibitors and 1 mM returned to medium supplemented with serum. In all DTT. Nuclear extracts were then cleared by centrifugation experiments, cells were irradiated immediately following at 13 000 g for 15 min at 48C. For E2F DNA binding removal of IL-3 and then returned to 378C. BaF3 cells assays, 10 mg of nuclear extracts (2 ± 4 ml) were incubated stably expressing the GalERVP transactivator were created with 0.4 ng of 32P end-labeled, E2F site-speci®c oligonu- by infecting cells with ecotropic retrovirus packaged in cleotide probe (Santa Cruz Biotechnology) in 30 ml BOSC 23 cells (Pear et al., 1993) following transient reactions containing 20 mM HEPES (pH 7.9) 40 mM transfection with the pMVGalERVP plasmid (Braselmann KCl, 6 mM MgCl ,1mM EDTA, 1 mM DTT, 5% et al., 1993). Following selection in 0.8 mg/ml G418, a 2 glycerol, 30 mgBSAand0.5mg sonicated salmon sperm neomycin-resistant polyclonal pool was subsequently DNA (Ikeda and Evans, 1993). Extracts were then transfected with either pGCmWaf-1 or pGC16E7 plus resolved by electrophoresis in 4% polyacrylamide gels pBabePuro (Dudek et al., 1997) by electroporation (250 V, containing 0.56TBE and subjected to phosphoimage 960 mFD) (BTX Electrocell Manipulator 600). Cells were analysis (Molecular Dynamics). Competitive binding then selected in 2 mg/ml puromycin and individual clones reactions included 20 ng of unlabeled wild-type or mutant were isolated by limiting dilution. BaF3 clones were E2F oligonucleotides (Santa Cruz Biotechnology). To screened for b-estradiol-inducible expression of murine liberate high order E2F complexes, nuclear extracts were p21orHPV16E7byimmunoblotanalysis.BaF3cellswere treated with 0.2% deoxycholate (DOC) for 3 min and infected with LXSCD4 retrovirus packaged in BOSC cells then NP40 was added to a ®nal concentration of 1% to and assessed for expression of the CD4 cell surface marker tirate out DOC and permit `free' E2F binding to probe the next morning. Typically 3 ± 7% of cells were positive (Bagchi et al., 1990; Wu et al., 1995). Antibodies used to for CD4 expression and subsequently used that day for identify proteins associated with E2F (data not shown) further experimentation. were monoclonal anti-Rb (Ab-1, Calbiochem) and rabbit polyclonal antibodies against p107 (C-18, Santa Cruz Flow cytometry Biotechnology) and human cdk2 (Upstate Biotechnology Incorporated). Cells were stained for both DNA content and apoptosis as previously described (Gorczyca et al., 1993; Canman et al., 1995). Brie¯y, DNA fragments were then end-labeled in situ Immunoblot analysis with biotinylated dUTP (Boehringer Mannheim) by terminal Cells were harvested and solubilized in Laemmli sample deoxynucleotidyl transferase (TdT) and then stained with buer as described previously (Kastan et al., 1992). FITC-conjugated avidin (Boehringer Mannheim). Cells were Equivalent amounts of protein were resolved by SDS ± then counter-stained for DNA content with propidium PAGE and transferred onto nitrocellulose membranes. iodide. For cell cycle analysis, cells were pulse-labeled with Monoclonal antibodies used for immunoblotting were 30 mM BrdUrd for 30 ± 45 min prior to harvesting. Cells were against murine p21 (Ab-6, Calbiochem), HPV16 E8 (8C9, then stained for both DNA content and BrdUrd incorpora- Zymed Laboratories Inc.) and Rb (G3-245, PharMingen). tion by the acid denaturation/protease method using anti- p107 was detected by immunoblotting with rabbit BrdU-FITC (Becton Dickinson) as described (Lowe et al., polyclonal anti-p107 (C-18, Santa Cruz Biotechnology). 1993b). To stain for both CD4 expression and apoptosis, After primary staining, blots were stained with horseradish cells were harvested and incubated with phycoerythrin(PE)- peroxidase-conjugated secondary antibodies and autora- conjugated anti-human CD4 monoclonal antibody (Phar- diographed using enhanced chemiluminescence (Amersham mingen), washed, and then resuspended in binding buer Life Sciences). containing Annexin V-FITC (Clontech Laboratories). To stain for both CD4 expression and BrdUrd incorporation, cells were pulse-labeled with BrdUrd, harvested, and stained with anti-CD4. Cells were then ®xed in PBS containing 1% formaldehyde for 15 min. An equal volume of PBS containing 2% Triton X-100 was added and cells were incubated Acknowledgements overnight at 48C. After a brief treatment with DNAse to We thank Bert Vogelstein for the murine p21 (WAF-1) expose BrdU epitope, cells were washed, and then stained cDNA, Alan Friedman for pMVGalERVP and GCpSP65 with anti-BrdU-FITC as described (Tough et al., 1994). For plasmids, Kathleen Cho for pCMV E6 and E7, and Chi all experiments, cells were analysed by ¯ow cytometry on a Dang for pGE1b/CD4. We also thank Tona M Gilmer for FACScan ¯ow cytometer (Becton Dickinson). sharing unpublished data and David Kirsch for helpful discussions. This work was supported by NIH Grants CA61949 (MBK) and CA60441 (CEC). CEC is also Electrophoretic mobility shift assay (EMSA) supported by an American Society of Hematology scholar Nuclear extracts were generated by lysing approximately fellowship award. MBK is the Steven Birnbaum Scholar of 107 cells in hypotonic buer (10 mM HEPES (pH 7.9) the Leukemia Society of America.

References

Abrahamson JLA, Lee JM and Bernstein A. (1995). Mol. Brugarolas J, Chandrasekaran C, Gordon JI, Beach D, Jacks Cell. Biol., 15, 6953 ± 6960. T and Hannon G. (1995). Nature, 377, 552 ± 557. Bagchi S, Raychaudhuri P and Nevins JR. (1990). Cell, 62, Canman CE, Gilmer TM, Coutts SB and Kastan MB. (1995). 659 ± 669. Genes Dev., 9, 600 ± 611. Braselmann S, Graninger P and Busslinger M. (1993). Proc. Chen X, Ko LJ, Jayaraman L and Prives C. (1996). Genes Natl. Acad. Sci. USA, 90, 1657 ± 1661. Dev., 10, 2438 ± 2451. G1 checkpoint control and p53-mediated apoptosis CE Canman and MB Kastan 966 Chiou S, Rao L and White E. (1994). Mol. Cell. Biol., 14, Macleod KF, Sherry N, Hannon G, Beach D, Tokino T, 2556 ± 2563. Kinzler K, Vogelstein B and Jacks T. (1995). Genes. Dev., Cleveland JL, Troppmair J, Pacleham G, Asleew DS, Lloyd 9, 935 ± 944. P, Gonzalez-Garcia M, Nunez G, Ihle JN and Rapp UR. Michieli P, Chedid M, Lin D, Pierce JH, Mercer WE and (1994). Oncogene, 9, 2217 ± 2226. Givol D. (1994). Cancer Res., 54, 3391 ± 3395. Coats S, Flanagan WM, Nourse J and Roberts JM. (1996). Miller AD, Miller DG, Garcia JV and Lynch CM. (1993). Science, 272, 877 ± 880. Methods in Enzymology, 217, 581 ± 599. Collins MKL, Marvel J, Malde P and Lopez-Rivas A. (1992). Minshall C, Arkins S, Freund GG and Kelly KW. (1996). J. J. Exp. Med., 176, 1043 ± 1051. Immunol., 156, 939 ± 947. Debbas M and White E. (1993). Genes Dev., 7, 546 ± 554. Morgenbesser SD, Williams BO, Jacks T and DePinho RA. Degregori J, Kowalik T and Nevins JR. (1995a). Mol. Cell. (1994). Nature, 371, 72 ± 74. Biol., 15, 4215 ± 4224. Mui AL-F, Wakao H, Kinoshita T, Kitamura T and Degregori J, Leone G, Ohtani K, Miron A and Nevins JR. Miyajima A. (1996). EMBO J., 15, 2425 ± 2433. (1995b). Genes Dev., 9, 2873 ± 2887. Nakamura N, Chin H, Miyasaka N and Miura O. (1996). J. Degregori J, Leone G, Miron A, Jakoi L and Nevins JR. Biol. Chem., 271, 19483 ± 19488. (1997). Proc. Natl. Acad. Sci. USA, 94, 7245 ± 7250. Nourse J, Firpo E, Flanagan WM, Coats S, Polyak K, Lee Demers GW, Foster SA, Halbert CL and Galloway DA. M-H, Massague J, Crabtree GR and Roberts JM. (1994). (1994). Proc. Natl. Acad. Sci. USA, 91, 4382 ± 4386. Nature, 372, 570 ± 573. Deng C, Zhang P, Harper JW, Elledge SJ and Leder P. Palacois R and Steinmetz M. (1985). Cell, 41, 727 ± 734. (1995). Cell, 82, 675 ± 684. Pan H and Griep AE. (1994). Genes Dev., 8, 1285 ± 1299. Dimri GP, Nakanishi M, Desprez P-Y, Smith JR and Pear WS, Nolan GP, Scott ML and Baltimore D. (1993). Campisi J. (1996). Mol. Cell. Biol., 16, 2987 ± 2997. Proc.Natl.Acad.Sci.USA,90, 8392 ± 8396. Dudek H, Datta R, Franke TF, Birnbaum MJ, Yao R, Phillips AC, Bates S, Ryan KM, Helin K and Vousden KH. Cooper GM, Sefal RA, Kaplan DR and Greenberg ME. (1997). Genes Dev., 11, 1853 ± 1863. (1997). Science, 275, 661 ± 665. Polyak K, Waldman T, He T-C, Kinzler KW and Vogelstein Dullic V, Kaufman WK, Wilson SJ, Tlsty TD, Lees E, B. (1996). Genes Dev., 10, 1945 ± 1952 Harper JW, Elledge SJ and Reed SI. (1994). Cell, 76, Qin X-Q, Livingston DM, Kaelin Jr, WG and Adams PD. 1013 ± 1023. (1994). Proc. Natl. Acad. Sci. USA, 91, 10918 ± 10922. El-Deiry WS, Tokino T, Velculescu VE, Levy DB, Parsons Ronen D, Schwartz D, Teitz Y, Gold®nger N and Rotter V. R, Trent JM, Lin K, Mercer WE, Kinzler KW and (1996). Cell Growth Di., 7, 21 ± 30. Vogelstein B. (1993). Cell, 75, 817 ± 825. Schwartz JK, Devoto SH, Smith EJ, Chellappan SP, Jakoi L El-Deiry WS, Harper JW, O'Connor PM, Velculescu VE, and Nevins JR. (1993). EMBO J., 12, 1013 ± 1020. Canman CE, Jackman J, Pietenpol JA, Burrell M, Hill Shan B and Lee W-H. (1994). Mol. Cell Biol., 14, 8166 ± 8173. DE, Wang Y, Wiman KG, Mercer WE, Kastan MB, Kohn Sherr CJ. (1994). Cell, 79, 551 ± 555. KW, Elledge SJ, Kinzler KW and Vogelstein B. (1994). Shirododkar S, Ewen M, DeCaprio JA, Morgan J, Cancer Res., 54, 1169 ± 1174. Livingston DM and Chittenden T. (1992). Cell, 68, 157 ± Gorczyca W, Gong J and Darzynkiewicz Z. (1993). Cancer 166. Res., 54, 2591 ± 2597. Shiyanov P, Bagchi S, Adami G, Kokontis J, Hay N, Arroyo Gorospe M, Cirielli C, Wang X, Seth P, Capogrossi MC and M, Morozov A and Raychaudhuri P. (1996). Mol. Cell. Holbrook NJ. (1997). Oncogene, 14, 929 ± 935. Biol., 16, 737 ± 744. Guillouf C, Grana X, Selvakumaran M, De Luca A, Slebos RJ, Lee MH, Plunkett BS, Kessis TD, Williams BO, Giordano A, Homan B and Liebermann DA. (1995). Jacks T, Hedrick L, Kastan MB and Cho KR. (1994). Blood, 85, 2691 ± 2698. Proc.Natl.Acad.Sci.USA,91, 5320 ± 5324. Harper JW, Adami GR, Wei N, Keyomarsi K and Elledge Terada K, Kaziro Y and Satoh T. (1995). J. Biol. Chem., 270, SJ. (1993). Cell, 75, 805 ± 816. 27880 ± 27886. Hiebert SW, Packman G, Strom DK, Haner R, Oren M, ToughDFandSprentJ.(1994). J.Exp.Med.,179,1127 ± 1135. Zambetti G and Cleveland JL. (1995). Mol. Cell. Biol., 15, Waldman T, Kinzler KW and Vogelstein B. (1995). Cancer 6864 ± 6874. Res., 55, 5187 ± 5190. Ikeda M-A and Nevins JR. (1993). Mol. Cell. Biol., 13, Wang Y, Okan I, Szekely L, Klein G and Wiman KG. (1995). 7029 ± 7035. Cell Growth & Dierentiation, 6, 1071 ± 1075. Kastan MB, Zhan Q, El-Deiry W, Carrier F, Jacks T, Walsh Weinberg RA. (1995), Cell, 81, 323 ± 330. WV, Plunkett BS, Vogelstein B and Fornace Jr, AJ. White E. (1994). Nature, 371, 21 ± 22. (1992). Cell, 71, 587 ± 597. White E. (1996). Genes Dev., 10, 1±15. Kaumann-Zeh A, Rodriguez-Viciana P, Ulrich E, Gilbert Wu C-L, Zukerberg LR, Ngwu C, Harlow E and Lees J. C, Coer P, Downward J and Evan G. (1997). Nature, 385, (1995). Mol. Cell. Biol., 15, 2536 ± 2546. 544 ± 548. Wu X and Levine AJ. (1994). Proc. Natl. Acad. Sci. USA, 91, Kinoshita T, Yokota T, Arai K and Miyajima A. (1995). 3602 ± 3606. EMBO J., 14, 266 ± 275. Xia Z, Dickens M, Raingeaud J, Davis RJ and Greenberg LaBaer J, Garrett MD, Stevenson LF, Slingerland JM, ME. (1995). Science, 270, 1326 ± 1331. Sandhu C, Chou HS, Fattaey A and Harlow E. (1997). XiongY,HannonGJ,ZhangH,CassoD,KobayashiRand Genes Dev., 11, 847 ± 862. Beach D. (1993a). Nature, 366, 701 ± 704. Lees E, Faha B, Dulic V, Reed SI and Harlow E. (1992). Xiong Y, Zhang H and Beach D. (1993b). Genes Dev., 7, Genes Dev., 6, 1874 ± 1855. 1572 ± 1583. Lin H-JL, Eviner V, Prendergast GC and White E. (1995). Yao R and Cooper GM. (1995). Science, 267, 2003 ± 2006. Mol. Cell. Biol., 15, 4536 ± 4544. Yonish-Rouach E, Resnitzky D, Lotem J, Sachs L, Kimchi A Lin Y and Benchimol S. (1995). Mol. Cell. Biol., 15, 6045 ± and Oren M. (1991). Nature, 352, 345 ± 347. 6054. Zha J, Harada J, Yang E, Jockel J and Korsmeyer SJ. (1996). Lowe SW and Ruley HE. (1993). Genes Dev., 7, 535 ± 545. Cell, 87, 619 ± 628. Lowe SW, Ruley HE, Jacks T and Housman DE. (1993). Zhang J, Hannon GJ and Beach D. (1994). Genes Dev., 8, Cell, 74, 957 ± 967. 1750 ± 1758. Liu Y, Martindale JL, Gorospe M and Holbrook NJ. (1996). Zhaung H, Niu Z, He T-C, Patel SV and Wojchowski DM. Cancer Res., 56, 31 ± 35. (1995). J. Biol. Chem., 270, 14500 ± 14504.