Oncogene (1997) 15, 337 ± 345  1997 Stockton Press All rights reserved 0950 ± 9232/97 $12.00 -mediated accumulation of hypophosphorylated pRb after the G1 restriction point fails to halt progression

Steven P Linke1,2,*, Matthew P Harris3,*, Sarah E Neugebauer3, Kristie C Clarkin1, H Michael Shepard3, Daniel C Maneval3 and Geo€rey M Wahl1

1Gene Expression Laboratory, The Salk Institute for Biological Studies, 10010 North Torrey Pines Road, La Jolla, California 92037; 2Department of Biology, University of California - San Diego, La Jolla, California 92093; 3Canji, Inc, 3030 Science Park Road, Suite 302, San Diego, California 92121, USA

This study analyses whether the inability of p53 to and inactivate p53 (Momand et al., 1992). Mdm2 was induce G1 arrest after the restriction point relates to an later shown to interact with pRb and pRb-regulated inability to modulate pRb phosphorylation. Transient complexes as well (Martin et al., 1995; Xiao et al., p53 overexpression in normal human diploid ®broblasts 1995), although recent studies suggest that deletion of and p53-de®cient cells led to increased levels of Mdm2 has no additional e€ect on cell proliferation, the -dependent kinase inhibitor p21Cip1/Waf1/Sdi1 and cell cycle control or tumorigenesis when p53 is absent an accumulation of hypophosphorylated pRb in cells (Jones et al., 1996; McMasters et al., 1996). growing asynchronously and in cells synchronized in late p53 and pRb participate in several cell cycle

G1 or M. Similarly, g-irradiation of asynchronous, late- checkpoint responses that contribute to maintaining

G1, or ®broblasts led to an increase in genetic stability. DNA damage activates p53 (Kastan et hypophosphorylated pRb. Experiments with ®broblasts al., 1991), inducing a prolonged G0/G1 cell cycle arrest expressing the HPV16 E6 indicated that resembling senescence in human ®broblasts (Di accumulation of hypophosphorylated pRb required func- Leonardo et al., 1994; Linke et al., 1997). Ribonucleo- tional p53. Progression into and through S phase was not tide depletion activates a reversible p53-dependent altered by the presence of hypophosphorylated pRb in arrest in G0 or early G1, resembling quiescence (Linke late G1, consistent with the failure of p53 to mediate G1 et al., 1996). p53 has also recently been implicated in arrest in cells that are past the restriction point. These regulating replication (Fukasawa et al., data indicate that accumulation of hypophosphorylated 1996) and in preventing DNA rereplication when pRb has signi®cantly di€erent e€ects on cell cycle spindle assembly is blocked (Cross et al., 1995; Di progression in early G1 versus late G1 or S phase. Leonardo et al., 1997). Inactivation of p53 allows for the emergence of aneuploid cells and mutants with Keywords: p53; pRb; restriction point; DNA damage; structural chromosome changes such as gene amplifica- Glarrest tion (Livingstone et al., 1992; Yin et al., 1992). Similarly, inactivation of pRb compromises the DNA damage-dependent arrest response (Almasan et al., 1995; Demers et al., 1994; Slebos et al., 1994) and the Introduction checkpoint activated by spindle inhibitors (Di Leonar- do et al., 1997) and enables cells to enter the cycle Retinoblastoma (RB) and p53 gene de®ciencies occur when certain deoxyribonucleotides become limiting in diverse human neoplasms. p53 alterations have been after treatment with antimetabolites such as metho- reported in approximately half of human malignancies trexate (Almasan et al., 1995). However, inactivation of encompassing a wide spectrum of tissues (Greenblatt et pRb in cells expressing wild-type p53 can lead to p53- al., 1994) and although RB alterations are more tissue- dependent in vitro (Almasan et al., 1995) and speci®c (Horowitz et al., 1990), some member of the in vivo (Howes et al., 1994; Morgenbesser et al., 1994; pRb pathway (e.g., p16INK4a, , or pRb) is Pan and Griep, 1994). Thus, a collateral de®ciency in frequently disrupted in many tumors (Sherr, 1996). p53 pRb and p53 allows cells to enter S phase under was initially discovered as a cellular protein bound by suboptimal conditions that contribute to genetic SV40 T-antigen, and it was subsequently found that instability and accelerate tumorigenesis (Di Leonardo pRb is also inactivated by this viral oncoprotein et al., 1993; Hartwell, 1992; Hartwell and Kastan, (DeCaprio et al., 1988; Lane and Crawford, 1979; 1994).

Linzer and Levine, 1979). Replication of many DNA pRb modulation of G1-S phase progression is viruses requires inactivation of both p53 and pRb, regulated, at least in part, by cell cycle-dependent which is typically achieved by expression of one or phosphorylation (Buchkovich et al., 1989; Chen et al., more such oncoproteins (Levine et al., 1991). In 1989; DeCaprio et al., 1989). pRb which is found addition, regulation of the activity of p53 and pRb is predominantly in a hypophosphorylated form in early mediated by endogenous binding . For , binds to a subset of E2F complexes, thereby example, the Mdm2 protein was ®rst reported to bind inhibiting their transcriptional activation potential (Hiebert et al., 1992). As cells reach the restriction point in mid to late G , pRb is phosphorylated by Correspondence: GM Wahl 1 *These authors contributed equally cyclin/cyclin-dependent kinase (Cdk) complexes, most Received 1 August 1996; revised 14 April 1997; accepted 16 April notably /Cdk4 and cyclin D/Cdk6. In addition,

1997 it is further phosphorylated near the G1-S transition by p53 modulation of pRb phosphorylation state SP Linke et al 338 /Cdk2 and, perhaps, by /Cdk2 the inability of p53 to mediate cell cycle arrest beyond (reviewed by Bartek et al., 1996). Phosphorylation of this point relates to di€erences in its ability to pRb prevents its inhibitory association with E2F modulate pRb phosphorylation. We show that p53 complexes, allowing transcriptional activation of gene can be activated after the restriction point, leading to products required for S phase entry and DNA an accumulation of hypophosphorylated pRb.

replication (DeGregori et al., 1995; Farnham et al., Although these late-G1 cells contained predominantly 1993; Slansky et al., 1993). pRb then remains hypophosphorylated pRb, they progressed into S hyperphosphorylated under normal conditions until phase. The data imply that the barrier to arrest in

the transition from to (Durfee et late G1 is not due to an inability of p53 to receive al., 1993; Ludlow et al., 1993). damage signals or to mediate the accumulation of

The activity of the Cdks that phosphorylate pRb can hypophosphorylated pRb. Rather, cells in late G1 be modi®ed by Cdk inhibitors, such as p21Cip1/Waf1/Sdi1 appear to have activated an irreversible program to (), members of the INK family (reviewed by Hunter enter S phase that is refractory to the functions of both and Pines, 1994), and other kinases (reviewed by p53 and pRb. Morgan, 1995; Sherr, 1994). Since p21 is induced by p53 (el-Deiry et al., 1993; Harper et al., 1995), pRb phosphorylation can be a€ected by the diversity of Results stresses that activate p53 in G1, including DNA damage and ribonucleotide depletion (Demers et al., Overexpression of p53 in asynchronous cells causes cell 1994; Di Leonardo et al., 1994; Dulic et al., 1994; cycle arrest, induction of p21 and accumulation of Linke et al., 1996; Slebos et al., 1994). Thus the net hypophosphorylated pRb result of p53 activation in G1 is to maintain pRb in its hypophosphorylated form, preventing activation of We previously showed that transient overexpression of E2F-responsive gene products required for S phase p53 by adenoviral transduction in p53-de®cient cell progression (DeGregori et al., 1995; Slansky et al., lines resulted in a dramatic decrease in S phase 1993). progression as measured by [3H]thymidine uptake Normal human diploid ®broblasts (NDF) that are g- (Harris et al., 1996). The use of adenovirus for p53

irradiated in early to mid G1 undergo a dose-dependent gene transfer provides an attractive tool to assess the

permanent G0/G1 arrest (Di Leonardo et al., 1994). In e€ects of p53 expression on cell cycle regulation, as

contrast, NDF irradiated in late G1 fail to arrest. This high gene transfer eciencies are obtained in many cell suggests that a fundamental change occurs in the lines, and the episomal nature of the vector obviates transduction of DNA damage signals at this juncture. the need for selection procedures. The H358 tumor cell The mechanism by which the damage response line was chosen for initial studies because it lacks p53

pathway becomes desensitized in late G1 has not been protein and contains apparently normal pRb (Otterson elucidated. This study looks at the e€ects of p53 et al., 1994; Takahashi et al., 1989). NDF strain activation on pRb phosphorylation and cell cycle MRC-9 was studied in parallel to address the function

progression in late G1 and S phase to determine if of overexpressed p53 in normal human cells.

ac

b d ppRb ppRb — — pRb — pRb

— p53 — p53

— p21 — p21

Figure 1 p53 overexpression in asynchronous cells leads to cell cycle arrest accompanied by p21 induction and accumulation of hypophosphorylated pRb. Cultures of p53-de®cient H358 or NDF strain MRC-9 were infected with adenovirus containing wild- type p53 (V53) or control vector with an empty expression cassette (V), or they were left untreated (UN). Infections were at 36106 and 16107 infectious units/ml for H358 and MRC-9, respectively. After 24 h samples were pulse-labeled with BrdU for 1 h and harvested for cell cycle or Western blot analysis. Cells harvested for cell cycle analysis were stained with propidium iodide and anti- BrdU-FITC and analysed by bivariate ¯ow cytometry (a and c). Numbers represent the percentage of cells in each cell cycle phase. The lower-left region represents G0 and G1 phase cells, the upper region represents S phase cells, and the lower-right region represents G2 and M phase cells. b and d are the corresponding Western blots probed for pRb, p53 and p21. PAb1801 was used to detect p53 levels after V53 treatment due to its lower sensitivity (it does not detect p53 in wild-type ®broblasts) p53 modulation of pRb phosphorylation state SP Linke et al 339 Asynchronous H358 or MRC-9 were infected with has been reported that mimosine can arrest cells that either a recombinant adenovirus containing a wild-type have entered S phase, likely due to inhibition of p53 gene (V53) or a control vector containing an empty replication initiation (Mosca et al., 1992). This expression cassette (V). Both constructs contain the subpopulation of BrdU7 S phase cells is apparently cytomegalovirus immediate early promoter. The e€ects incapable of engaging in DNA replication even after of p53 overexpression on cell cycle distribution were mimosine removal. H358 cultures contained about measured by ¯ow cytometry 24 h after infection, and 20% of these BrdU7 S phase cells, while MRC-9 changes in p53, p21 and pRb were analysed by cultures contained less than 5% after mimosine Western blot (Figure 1). pRb undergoes a mobility treatment of asynchronous cultures. shift in SDS ± PAGE gels such that the hypophos- Mimosine-treated cells arrested with a G0/G1 DNA phorylated form migrates faster than phosphorylated content (Table 1), and about half of the pRb was forms (Buchkovich et al., 1989). hyperphosphorylated (Figure 2a). This is consistent

The number of cells in S phase in V53-infected with the arrest occurring in late G1, beyond the cultures was substantially reduced compared to restriction point. The low levels of hyperphosphory- cultures infected with the vector control or uninfected lated pRb and increased levels of p21 in the cultures. V53-treated H358 cells accumulated in G2/M, asynchronous controls in MRC-9 in Figure 2a whereas MRC-9 cells appeared to arrest in the probably resulted from contact inhibition of a fraction subsequent G1 (Figure 1a,c). H358 cells underwent of the cells, since they grow to a higher density than apoptosis at later time points (data not shown). These the drug or virus treated samples. Overexpression of e€ects on cell cycle occurred concomitantly with increases in p53 and p21 and an accumulation of hypophosphorylated pRb (Figure 1b,d). A dose- response analysis showed that induction of p21 and Table 1 Cell cycle distributions of H358 and MRC-9 cells after accumulation of hypophosphorylated pRb correlated treatment with cell cycle synchronizing agents and adenovirus a with the amount of p53 transgene expressed in both Treatment Time (h) %G0/G1 %S %G2/M cell types (data not shown). Vector treatment alone H358 had no e€ect in H358 and produced only slight Asynchronous 48 43 38 19 decreases in the percentage of S phase cells and subtle Mimosine 24 79 1 20 elevations in p21 and hypophosphorylated pRb in Mimosine 48 78 1 21 Mimosine+V53 48 76 1 23 MRC-9. Since cells with an intact p53 pathway arrest Colcemid 24 7 18 75 when even small numbers of DNA breaks are present Colcemid 48 5 6 89 (Di Leonardo et al., 1994; Huang et al., 1996), it is Colcemid+V53 48 5 1 94 possible that the slight changes observed in MRC-9 MRC-9 Asynchronous 48 79 13 8 resulted from activation of the endogenous p53 Mimosine 24 89 2 9 pathway by a small fraction of linear adenovirus Mimosine 48 87 1 12 DNAs with unprotected ends. Mimosine+V53 48 89 2 9 a Infections with V53 were done after 24 h of treatment with the synchronizing agent at 36106 and 16107 infectious units/ml for Overexpression of p53 in cells synchronized in late G1 or H358 and MRC-9, respectively, and the samples were harvested 24 h M leads to induction of p21 and an accumulation of later hypophosphorylated pRb p53 loses its ability to mediate G1 arrest in cells that have progressed into late G1, coinciding with the G1 ab restriction point (Di Leonardo et al., 1994). Two likely H358 MRC-9 H358 explanations are that p53 cannot receive DNA damage signals, or that it cannot induce downstream gene expression once cells traverse the restriction point. Several strategies were employed to investigate these UN Colc Colc+V Colc+V53 UN Mimo Mimo+V Mimo+V53 possibilities. UN Mimo Mimo+V Mimo+V53 We ®rst determined whether hypophosphorylated — ppRb — ppRb — — pRb pRb accumulated when p53 was overexpressed in late pRb

G1 phase or M phase. Asynchronous H358 or MRC-9 cells were synchronized in late G1 with the plant amino — p53 — p53 acid analog mimosine (Lalande, 1990). Cultures synchronized for 24 h were subsequently infected with — p21 — p21 V53 for an additional 24 h in the presence of mimosine. Samples were pulsed with bromodeoxyur- Figure 2 p53 overexpression in late-G1 or M-phase synchronized idine (BrdU) for 30 min and analysed by ¯ow cells leads to p21 induction and accumulation of hypophos- cytometry to determine cell cycle distributions. H358 phorylated pRb. Asynchronous H358 or MRC-9 cells were left untreated (UN), or they were treated with 200 mM mimosine and MRC-9 cells treated with 200 mM mimosine (Mimo) for 24 h to synchronize them in late G1. 100 mM colcemid displayed a signi®cant reduction in the percentage of (Colc) was used to synchronize H358 in . Cells were then

S phase cells and an increase in G0/G1 cells (Table 1). infected with V53 or vector control (V) and harvested 24 h later Mimosine treatment of asynchronous cultures resulted (MRC-9, 16107 infectious units/ml; H358, 66106 infectious units/ml). Western blots were probed for pRb, p53, and p21. in small populations of cells that did not incorporate PAb1801 was used to detect p53 levels after V53 treatment. (a) BrdU (BrdU7) yet had DNA contents between G /G 0 1 Western blot of late G1-synchronized H358 and MRC-9 cells. (b) and G2/M (i.e., they were presumably in S phase). It Western blot of H358 cells synchronized in M p53 modulation of pRb phosphorylation state SP Linke et al 340 p53 by infection with V53 during mimosine arrest a caused p21 induction and an accumulation of hypophosphorylated pRb (Figure 2a). No apoptosis was apparent during the duration of the study as

determined by a lack of cells with sub-G0/G1 DNA content after treatment with the cell synchronizing agents or infection (data not shown). H358 cultures were next treated with the mitotic spindle inhibitor colcemid (100 mM), which arrests cells in (Rieder and Palazzo, 1992) prior to when pRb becomes dephosphorylated at the transition b from metaphase to anaphase (Ludlow et al., 1993). The ppRb colcemid-treated cultures contained approximately — — pRb 90% 4N (G2/M) phase cells within 48 h. Infection

with V53 caused a slight increase in the G2/M fraction (Table 1). A signi®cant fraction of the pRb in — p53 colcemid-arrested cells was hyperphosphorylated, as in the mimosine-arrested cells. Overexpression of p53 led to an induction of p21 and an accumulation of hypophosphorylated pRb in H358 cells, as observed in — p21 asynchronous and mimosine-arrested cells (Figure 2b). Similar data were obtained in nocodazole-treated Figure 3 MRC-9 cells have a functional p53 DNA damage MRC-9 cells after p53 overexpression (data not response pathway. Asynchronous MRC-9 cells were left untreated shown). These data show that p53 overexpression can (0 h), or they were treated with 4 Gy g radiation and harvested 4 modulate p21 levels and pRb phosphorylation beyond and 24 h post-irradiation. (a) Samples were pulse-labeled with BrdU for 1 h, harvested, stained with propidium iodide and anti- the G1 restriction point. BrdU-FITC and analysed by bivariate ¯ow cytometry. The numbers represent the percentage of cells in each cell cycle g-irradiation of asynchronous NDF results in phase. (b) Western blot analysis of pRb, p53, and p21. DO-1 was accumulation of hypophosphorylated pRb and cell cycle used to detect endogenous levels of p53 in MRC-9 cells. DO-1 provides higher sensitivity to p53 than PAb1801, which was used arrest in the overexpression studies. b-actin con®rmed equal loading Since the level of p53 was supraphysiological after (data not shown) adenovirus infection, g radiation was used to provide a more physiologically relevant condition to determine whether p53 could be activated to modulate pRb lated pRb was predominant in the untreated sample, phosphorylation beyond the restriction point. A but pRb was completely hypophosphorylated 24 h baseline was obtained for these analyses by determin- after treatment with g radiation (Figure 4). The ing the e€ect of 4 Gy g radiation on the cell cycle accumulation of hypophosphorylated pRb in irra- distribution and expression of p53, p21 and pRb in diated cells took at least 4 h to occur (data not asynchronous MRC-9 cells. The fraction of cells in S shown), and the level of detectable pRb decreased

phase was reduced and the fractions in G0/G1 and G2/ reproducibly by 24 h after irradiation. M were increased within 24 h of irradiation (Figure 3), We then analysed the consequences of g-irradiation as observed in other NDF strains (Di Leonardo et al., on S phase cultures. Mimosine-arrested MRC-9 cells 1994; Dulic et al., 1994). This correlated with increases were released for 4 h, treated with 4 Gy g radiation or in the levels of p53 and p21 and an accumulation of left untreated, and then harvested 24 h later (total of hypophosphorylated pRb (Figure 3). The increase in 28 h after release). At the time of the treatment (4 h

G2/M phase cells is consistent with the presence of a after release from mimosine), 36% of cells showed

p53-independent G2 DNA damage checkpoint delay, BrdU incorporation (Figure 4). BrdU (10 mM) and which may lead to p53-dependent arrest in the nocodazole (0.05 mg/ml) were added immediately after

subsequent G0/G1 if unrepaired damage is present treatment to label all cells that progressed into S phase (Linke et al., 1997). and to prevent them from dividing. Hypophosphory- lated pRb accumulated in the irradiated cells but not the untreated cells (Figure 4). These results do not g-irradiation of NDF in late-G or early-S results in 1 relate to the nocodazole treatment, since cells accumulation of hypophosphorylated pRb but does not irradiated 4 h after release from mimosine and inhibit cell cycle progression allowed to progress in the absence of nocodazole also

We next determined whether g-irradiation in late G1 accumulated hypophosphorylated pRb (data not would result in p53 and p21 induction and an shown). Despite the presence of predominantly accumulation of hypophosphorylated pRb. MRC-9 hypophosphorylated pRb in the irradiated cells, nearly

cells were synchronized in G0 by contact inhibition, as many progressed to the subsequent G2/M (17%) as then released into 200 mM mimosine for 36 h to in the unirradiated control (20%) (Figure 4). The

synchronize them in late G1. The synchronized di€erence observed probably derives from a fraction of

population consisted of 89% cells with a G0/G1 DNA the mimosine-synchronized population being arrested

content and 11% with a G2/M DNA content. Two suciently early in G1 to be subject to g radiation-

samples were maintained in mimosine, one of which induced G1 arrest. This is consistent with the slight was treated with 4 Gy g radiation. Hyperphosphory- reduction in the overall percentage of BrdU+ cells p53 modulation of pRb phosphorylation state SP Linke et al 341 a Async Mimo Mimo->Noc γ a – + — radiation Mimo Mimo->Noc+BrdU – + – + γ ->Mimo –+–––+++ — radiation ppRb 0 0 — — time (hr) G G 24 24 4 8 24 4 8 24 — pRb — ppRb — pRb

MRC-9neo — p53 — p53

— p21 b — ppRb — — β-actin pRb

b

MRC-9E6 — p53

Figure 5 E6 expression in MRC-9 cells prevents modulation of pRb phosphorylation state after g-irradiation. Asynchronous cultures (Async), cultures synchronized in late G1 with mimosine Figure 4 g-irradiation of late-G1 or early S phase MRC-9 cells for 24 h and held in mimosine (Mimo), and cultures syn- leads to accumulation of hypophosphorylated pRb but not cell chronized with mimosine and released into nocodazole for 4 h cycle arrest. MRC-9 cells were synchronized in late G1 by (Mimo4Noc) were left untreated (7) or treated with 4 Gy g releasing G0-synchronized cells (G0) for 36 h in mimosine radiation (+), as in Figure 4. All samples were harvested 24 h (G0-4Mimo). Two samples were maintained in mimosine after treatment. pRb phosphorylation state and p53 levels were (Mimo), one left untreated (7) and the other treated with 4 Gy assessed by Western blot. b-actin con®rmed equal loading (not g radiation (+); both were harvested 24 h after the irradiation. shown). (a) MRC-9neo and (b) MRC-9E6 Six other samples were synchronized in early S phase by releasing them from mimosine for 4 h. The samples were then treated, and nocodazole and BrdU were added (Mimo-4Noc+BrdU). Untreated (7) and g-irradiated (+) samples were harvested 4, 8 contained nearly undetectable levels of p53 and lacked and 24 h after the irradiation (total of 28 h after release from a g radiation-induced G /G arrest, whereas MRC-9neo mimosine). (a) Western blot analysis of pRb, p53, and p21. All of 0 1 the lanes shown are from the same Western, but were rearranged cells contained detectable p53 and underwent a G0/G1 for clarity. b-actin con®rmed equal loading. (b) Cell cycle arrest after irradiation (Figure 5 and data not shown).

analysis. The numbers at the top-left of each cell cycle plot Treatment of asynchronous, late-G1, or early S-phase represent the total percentage of BrdU+ cells after release (cells MRC-9neo cells with 4 Gy g radiation resulted in an that entered the ®rst S, G2, and M phase, which lie above the accumulation of hypophosphorylated pRb within 24 h, horizontal line). The numbers at the top-right with pointers to the oval gates in the 24 h samples represent the percentage of cells as observed in the parental MRC-9 cells (Figure 5a). that progressed to G2 or M By comparison, pRb phosphorylation was not changed appreciably in irradiated MRC-9E6 cells (Figure 5b). These data support the hypothesis that the accumula- observed after the total 28 h release in the irradiated tion of hypophosphorylated pRb after g-irridation sample (39%) as compared with the untreated sample requires functional p53. (44%) (Figure 4). Irradiation of cells released into early S after an aphidicolin arrest also exhibited accumula- NDF released from late G1 with predominantly tion of hypophosphorylated pRb, indicating that the hypophosphorylated pRb progress through S phase observed e€ects are not due to the synchronizing agent (data not shown). Hyperphosphorylated pRb normally accumulates in

late G1, enabling entry into and progression through S phase. Thus, treatments leading to accumulation of Modulation of pRb phosphorylation state after g- hypophosphorylated pRb, such as those described irradiation is abrogated by expression of the HPV16 E6 above, might be expected to alter progression into S oncoprotein phase. Since accumulation of hypophosphorylated pRb Pools of MRC-9 cells stably expressing HPV16 E6 and took more than 4 h to occur in MRC-9 cells after g- a neomycin-resistance marker (MRC-9E6) and isogenic irradiation, it was possible that the progression seen in controls expressing the neo marker alone (MRC-9neo) cells with predominantly hypophosphorylated pRb were next studied to determine whether p53 function may have been due to its delayed accumulation. was required for modulation of pRb phosphorylation Therefore, mimosine-arrested MRC-9 cells were state in response to g-irradiation after the restriction irradiated and then held in mimosine for an additional point. Cells expressing E6 have a non-functional p53 24 h to allow for the majority of pRb to become arrest pathway, as E6 hastens the degradation of p53 hypophosphorylated. Cells were then released and by ubiquitination (Werness et al., 1990). Although allowed to progress for 4 h in the presence of BrdU additional functions have been proposed for E6, to monitor their ability to enter S phase in the presence inactivation of p53 is sucient to abrogate g of predominantly hypophosphorylated pRb. Although radiation-induced arrest, and none of the other irradiated cells contained predominantly hypophos- putative E6 functions have been reported to a€ect the phorylated pRb (Figure 6a), their ability to enter S radiation response or pRb phosphorylation (e.g., phase was comparable to that of the untreated control Etscheid et al., 1994; Klingelhutz et al., 1996; Lechner (Figure 6b). Therefore, the presence of hypophos- and Laimins, 1994). Asynchronous MRC-9E6 cells phorylated pRb after the G1 restriction point has little p53 modulation of pRb phosphorylation state SP Linke et al 342 a unknown. pRb has a half-life of 4 to 6 h (Chen et al.,

Mimo Mimo->BrdU 1989), and de novo synthesis of pRb in S, G2, and M + – + — γ radiation presumably requires phosphorylation to generate and maintain the hyperphosphorylated species in these phases. We observed that 4 to 24 h was required for — ppRb the accumulation of hypophosphorylated pRb subse- — pRb quent to p53 overexpression or irradiation. Thus, it is reasonable to postulate that the accumulation of hypophosphorylated pRb we observed after irradia- tion may re¯ect both pRb protein turnover and b concomitant inhibition of pRb phosphorylation. We do not favor a model involving synthesis of hypophos- phorylated pRb in the absence of inhibition of

phosphorylation, since cells in late G1 or M treated only with synchronizing agents for 48 h maintained hyperphosphorylated pRb at similar levels as asyn- chronous controls. Despite the accumulation of hypophosphorylated

pRb following g-irradiation of late G1 or S phase cells, p53 levels did not increase to the same extent as in irradiated asynchronous cells and p21 increases over the untreated controls were not evident. A trivial explanation for this phenomenon is that the e€ect of mimosine on p53 levels could have obscured the response (Khanna and Lavin, 1993). Alternatively, it is possible that p53 may be activated after the restriction point by a mechanism that does not require Figure 6 MRC-9 cells progress into S phase in the presence of a signi®cant increase at the protein level. For example, hypophosphorylated pRb. Cells were synchronized in late G1 with phosphorylation of the p53 C-terminus enhances its mimosine and left untreated (7) or treated with 4 Gy g radiation sequence-speci®c DNA binding without an increase in (+). They were left in mimosine for an additional 24 h to allow time for pRb dephosphorylation, then released for 4 h into level (Hupp et al., 1992). In addition, p53 may have

medium containing BrdU (Mimo-4BrdU) or left in mimosine important targets other than p21 in late-G1 and/or S (Mimo). (a) pRb phosphorylation state was assessed by Western phase, such as other Cdk inhibitors or phosphatase + blot. (b) The percentages of BrdU cells were determined by cell activators, that prevent accumulation of hyperphos- cycle analysis and are presented relative to the untreated, released control for a typical experiment. The absolute percentages of phorylated pRb. Consistent with this, MEFs de®cient + BrdU cells was 1% for the irradiated sample held in mimosine, in p21 exhibit only a partial loss of p53-dependent G0/

22% for the sample released without treatment, and 18% for the G1 arrest when irradiated in early G1 (Deng et al., sample released after irradiation 1995). Also, In vitro phosphorylation of p53 by kinases prevalent in S phase and M phase results in di€erential target gene activation created by preferential binding to e€ect on the ability of cells to enter S phase even when certain promoters (Hecker et al., 1996; Wang and maintained in the cell for a prolonged period. Prives, 1995). These data suggest that speci®c molecular modi®cations to p53, mediated by kinases

present at the G1-S transition, could elicit molecular Discussion consequences that are distinctly di€erent from those produced by modi®cations that a€ect p53 function in

The results presented here show that the failure of p53 G0 or early G1 cells. Although cyclin D and cyclin E to prevent cell cycle entry after the restriction point levels were not signi®cantly altered after g-irradiation does not derive from an inability to modulate pRb (data not shown), the net e€ect of the p53-dependent phosphorylation state. When p53 was transiently inhibition of their associated kinase activities was to overexpressed or induced by treatment with g prevent the accumulation of hyperphosphorylated pRb. radiation, hypophosphorylated pRb was observed not The ability of a physiological event such as DNA only in asynchronous NDF but also in cell populations damage to produce a p53-dependent change in pRb

arrested in late G1 or M phase. This indicates that the phosphorylation in all cell cycle phases analysed damage signal received by p53 can cause accumulation suggests that such alterations would have conse- of hypophosphorylated pRb at various phases of the quences on cell cycle progression. However, we found

cell cycle. The data also reveal that cells can proceed that hypophosphorylated pRb present in late-G1 did into S with hypophosphorylated pRb, as long as it is not signi®cantly a€ect progression into S phase. This is

generated in late G1 (after the restriction point). These in agreement with studies showing that pRb, , data raise questions about the consequences of pRb antibodies to cyclin D1, or pRb phosphatases can

hypophosphorylation after the G1 restriction point and arrest cells when microinjected in early to mid G1, but

why cells become refractory to the e€ects of hypophos- not in late G1 or S (Alberts et al., 1993; Goodrich et phorylated pRb after this critical decision point in the al., 1991; Lukas et al., 1995). Our study extends these cell cycle (Pardee, 1989). ®ndings by providing evidence that pRb can be The mechanism by which p53 modulates pRb modi®ed to exist in a hypophosphorylated state by a phosphorylation state after cells commit to S phase is response to DNA damage path- p53 modulation of pRb phosphorylation state SP Linke et al 343 ways that occurs after the restriction point. Studies A replication-de®cient recombinant human adenovirus showing that DNA rereplication occurs in the presence encoding wild-type p53 under the control of the human of spindle inhibitors in cells with non-functional p53, cytomegalovirus immediate early promoter (V53) was used to but not in cells with functional p53, raises the exogenously express p53. An adenovirus consisting of the intriguing possibility that pRb might also participate human cytomegalovirus promoter with an empty expression cassette (V) was used as a control. Virus was produced and in this `spindle checkpoint' (Cross et al., 1995; Di titered using an infectious assay in 293 embryonic kidney Leonardo et al., 1997). Consistent with this possibility, cells, which provide E1 function in trans (Wills et al., 1994, we recently found that pRb-de®cient mouse embryo 1995). ®broblasts and HPV16 E7-expressing NDF rereplicate Previous studies have shown di€erential transduction when challenged with mitotic spindle inhibitors (Di eciencies between H358 and MRC-9 cells using a Leonardo et al., 1997). recombinant adenovirus encoding b-galactosidase (Harris et The inability of hypophosphorylated pRb to al., 1996). Therefore, the adenovirus concentrations used in infections were slightly higher for MRC-9 cells (16107 attenuate cell cycle progression in late G1 or S may be explained in several ways. For example, the infectious units/ml) than for H358 cells (3 ± 86106 infectious hypophosphorylated pRb produced in early G may units/ml). b-galactosidase expression studies at these concen- 1 trations resulted in comparable percentages of cells expressing not be identical to that generated later in the cycle with the transgene. Dose response experiments covered a regard to its phosphorylation, which may not be concentration range between 36105 and 16107 infectious revealed by SDS ± PAGE. It is also possible that the units/ml. high stability of the E2F-regulated biosynthetic enzymes required for S phase progression, such as DHFR and TS, makes the re-establishment of Generation of E6-expressing MRC-9 inhibition at the level of transcription insucient to Pools of MRC-9 expressing a neomycin resistance gene turn them o€ after they are produced in late G1. This is (MRC-9neo) or expressing neomycin resistance and the supported by the observation that phosphorylation of oncogenic HPV16 E6 protein (MRC-9E6) were prepared E2F-1 by cyclin A/Cdk2 in S phase inhibits its ability by retroviral gene transduction as described by Halbert et to bind DNA, shutting o€ transcription of E2F- al. (1992) using murine PA317 producer lines (provided by regulated genes well before the end of S phase Dr D Galloway, Fred Hutchinson Cancer Research Center, (Dynlacht et al., 1994; Krek et al., 1994). This process Seattle, WA). These MRC-9 pools were maintained as may be necessary to allow replication to occur through described above with the addition of 200 mg/ml G-418 these genes (Krek et al., 1995). Phosphorylation of E2F sulfate. complexes by cyclin A/CDk2 may also prevent reassociation with pRb, regardless of its phosphoryla- Flow cytometric cell cycle analysis tion state (Fagan et al., 1994; Xu et al., 1994). Alternatively, it is possible that events important for Cells were incubated in 10 mM BrdU to label S phase cells. initiation at DNA replication origins may occur prior After harvest by trypsinization, cells were ®xed in 70% to the mimosine block. This is supported by recent ethanol until the day of analysis. Cells were treated with 0.1 M HCl containing 0.5% Triton X-100 to extract evidence that E2F may directly regulate transcription histones, followed by boiling and rapid cooling to of the newly-identi®ed human origin recognition denature the DNA. Cells were then incubated with anti- complex gene HsOrc1 (Ohtani et al., 1996). In BrdU-FITC (PharMingen, San Diego, CA) and counter- addition, Mdm2 can bind to pRb (inhibiting its stained with propidium iodide containing RNase. Samples growth regulatory function) and to E2F (augmenting were run on a Becton-Dickinson FACScan using color its transactivation potential), so induction of Mdm2 in compensation when appropriate. Data analysis was done using SunDisplay3 (J Trotter, Salk Institute Flow late G1 could also prevent hypophosphorylated pRb from inhibiting E2F (Martin et al., 1995; Xiao et al., Cytometry Lab, La Jolla, CA). Cellular debris and ®xation artifacts were gated out. Events with a 2n DNA 1995). In any event, cells arrested in late G1 maintain content were gated as %G0/G1, events with a 4n DNA the ability to initiate DNA replication and complete + content were gated as %G2/M, and BrdU events with S phase, despite the presence of predominantly intermediate DNA content were gated as %S in pulse- hypophosphorylated pRb. The inability of hypophos- labeling experiments. In continuous-labeling experiments, phorylated pRb to induce arrest after cells have passed + BrdU G2/M events were also scored. the G1 restriction point emphasizes that traversing this checkpoint creates an irrevocable license to enter into and proceed through S phase, even under conditions Western blotting that can lead to genetic damage. Cell lysates were prepared using lysis bu€er (50 mM Tris- HClpH7.5,250mM NaCl, 0.1% NP-40, 50 mM NaF, Materials and methods 5mM EDTA, 10 mg/ml aprotinin, 10 mg/ml leupeptin and 1mM PMSF), and supernatants were collected after one freeze-thaw cycle. Samples were normalized for total Cell culture and treatments protein content by a Bradford assay (Bio-Rad), loaded NCI-H358 (non-small cell lung carcinoma) and MRC-9 onto pre-cast Tris-glycine gels (Novex), and separated by (human diploid embryonic lung ®broblast) cells were SDS ± PAGE. Typically, 5 to 10 mg of total protein was obtained from the ATCC and maintained in DMEM loaded per lane. After transfer onto 0.2 mm nitrocellulose supplemented with 10% fetal bovine serum and 1% membranes, proteins were detected using the following sodium pyruvate at 378Cand7%CO2. MRC-9 cells were antibodies: p53, PAb1801 (Novacastra) or DO-1 (Onco- used up to 25 ± 30 population doublings. g-irradiation was gene Science); pRb, mAb 3C8 (Wen et al., 1994); p21, Ab-1 performed at room temperature using a 60Co g-irradiator (Oncogene-Science); and b-actin, AC-15 (Sigma). Enhanced (Gammabeam 150-C) at a distance of 40 cm and a rate of chemiluminescence (Amersham) was used as a detection approximately 3.0 Gy/min (1 Gy=100 rads). system. p53 modulation of pRb phosphorylation state SP Linke et al 344 Acknowledgements manuscript. This work was supported, in part, by the The authors wish to thank Dr GW Demers and Dr R National Cancer Institute and the G Harold and Leila Y Bookstein for their advice in the preparation of this Mathers Charitable Foundation.

References

Alberts AS, Thorburn AM, Shenolikar S, Mumby MC and Hartwell LH and Kastan MB. (1994). Science, 266, 1821 ± Feramisco JR. (1993). Proc. Natl. Acad. Sci. USA, 90, 1828. 388 ± 392. Hecker D, Page G, Lohrum M, Weiland S and Scheidtmann Almasan A, Yin Y, Kelly RE, Lee EY, Bradley A, Li W, KH. (1996). Oncogene, 12, 953 ± 961. Bertino JR and Wahl GM. (1995). Proc. Natl. Acad. Sci. Hiebert SW, Chellappan SP, Horowitz JM and Nevins JR. USA, 92, 5436 ± 5440. (1992). Genes Dev., 6, 177 ± 185. Bartek J, Bartkova J and Lukas J. (1996). Curr. Opin. Cell. Horowitz JM, Park SH, Bogenmann E, Cheng JC, Yandell Biol., 8, 805 ± 814. DW, Kaye FJ, Minna JD, Dryja TP and Weinberg RA. Buchkovich K, Du€y LA and Harlow E. (1989). Cell, 58, (1990). Proc. Natl. Acad. Sci. USA, 87, 2775 ± 2779. 1097 ± 1105. Howes KA, Ransom N, Papermaster DS, Lasudry JG, Chen PL, Scully P, Shew JY, Wang JY and Lee WH. (1989). Albert DM and Windle JJ. (1994). Genes Dev., 8, 1300 ± Cell, 58, 1193 ± 1198. 1310. Cross SM, Sanchez CA, Morgan CA, Schimke MK, Ramel Huang L, Clarkin KC and Wahl GM. (1996). Proc. Natl. S, Idzerda RL, Raskind WH and Reid BJ. (1995). Science, Acad. Sci. USA, 93, 4827 ± 4832. 267, 1353 ± 1356. Hunter T and Pines J. (1994). Cell, 79, 573 ± 582. DeCaprio JA, Ludlow JW, Figge J, Shew JY, Huang CM, Hupp TR, Meek DW, Midgley CA and Lane DP. (1992). Lee WH, Marsilio E, Paucha E and Livingston DM. Cell, 71, 875 ± 886. (1988). Cell, 54, 275 ± 283. Jones SN, Sands AT, Hancock AR, Vogel H, Donehower DeCaprio JA, Ludlow JW, Lynch D, Furukawa Y, Grin J, LA, Linke SP, Wahl GM and Bradley A. (1996). Proc. Piwnica WH, Huang CM and Livingston DM. (1989). Natl. Acad. Sci. USA, 93, 14106 ± 14111. Cell, 58, 1085 ± 1095. Kastan MB, Onyekwere O, Sidransky D, Vogelstein B and DeGregori J, Kowalik T and Nevins JR. (1995). Mol. Cell. Craig RW. (1991). Cancer Res., 51, 6304 ± 6311. Biol., 15, 4215 ± 4224. Khanna KK and Lavin MF. (1993). Oncogene, 8, 3307 ± Demers GW, Foster SA, Halbert CL and Galloway DA. 3312. (1994). Proc. Natl. Acad. Sci. USA, 91, 4382 ± 4386. Klingelhutz AJ, Foster SA and McDougall JK. (1996). Deng C, Zhang P, Harper JW, Elledge SJ and Leder P. Nature, 380, 79 ± 82. (1995). Cell, 82, 675 ± 684. Krek W, Ewen ME, Shirodkar S, Arany Z, Kaelin WJ and Di Leonardo A, Khan SS, Linke SP, Greco V, Seidita G and Livingston DM. (1994). Cell, 78, 161 ± 172. Wahl GM. (1997). Cancer Research, 57, 1013 ± 1019. Krek W, Xu G and Livingston DM. (1995). Cell, 83, 1149 ± Di Leonardo A, Linke SP, Clarkin K and Wahl GM. (1994). 1158. Genes. Dev., 8, 2540 ± 2551. Lalande M. (1990). Exp. Cell. Res., 186, 332 ± 339. Di Leonardo A, Linke SP, Yin Y and Wahl GM. (1993). Cold Lane DP and Crawford LV. (1979). Nature, 278, 261 ± 263. Spring Harb. Symp. Quant. Biol., 58, 655 ± 667. Lechner MS and Laimins LA. (1994). J. Virol., 68, 4262 ± Dulic V, Kaufmann WK, Wilson SJ, Tlsty TD, Lees E, 4273. Harper JW, Elledge SJ and Reed SI. (1994). Cell, 76, Levine AJ, Momand J and Finlay CA. (1991). Nature, 351, 1013 ± 1023. 453 ± 456. Durfee T, Becherer K, Chen PL, Yeh SH, Yang Y, Kilburn LinkeSP,ClarkinKC,DiLeonardoA,TsouAandWahl AE, Lee WH and Elledge SJ. (1993). Genes Dev., 7, 555 ± GM. (1996). Genes Dev., 10, 934 ± 947. 569. Linke SP, Clarkin KC and Wahl GM. (1997). Cancer Res., Dynlacht BD, Flores O, Lees JA and Harlow E. (1994). 57, 1171 ± 1179. Genes Dev., 8, 1772 ± 1786. Linzer DI and Levine AJ. (1979). Cell, 17, 43 ± 52. el-Deiry WS, Tokino T, Velculescu VE, Levy DB, Parsons R, Livingstone LR, White A, Sprouse J, Livanos E, Jacks T and Trent JM, Lin D, Mercer WE, Kinzler KW and Vogelstein Tlsty TD. (1992). Cell, 70, 923 ± 935. B. (1993). Cell, 75, 817 ± 825. Ludlow JW, Glendening CL, Livingston DM and DeCarprio Etscheid BG, Foster SA and Galloway DA. (1994). Virology, JA. (1993). Mol. Cell. Biol., 13, 367 ± 372. 205, 583 ± 585. Lukas J, Parry D, Aagaard L, Mann DJ, Bartkova J, Strauss Fagan R, Flint KJ and Jones N. (1994). Cell, 78, 799 ± 811. M, Peters G and Bartek J. (1995). Nature, 375, 503 ± 506. Farnham PJ, Slansky JE and Kollmar R. (1993). Biochim. Martin K, Trouche D, Hagemeier C, Sorensen TS, La TN Biophys. Acta, 1155, 125 ± 131. and Kouzarides T. (1995). Nature, 375, 691 ± 694. Fukasawa K, Choi T, Kuriyama R, Rulong S and Vande McMasters KM, Montes de Oca Luna R, Pena JR and WG. (1996). Science, 271, 1744 ± 1747. Lozano G. (1996). Oncogene, 13, 1731 ± 1736. Goodrich DW, Wang NP, Qian YW, Lee EY and Lee WH. Momand J, Zambetti GP, Olson DC, George D and Levine (1991). Cell, 67, 293 ± 302. AJ. (1992). Cell, 69, 1237 ± 1245. Greenblatt MS, Bennett WP, Hollstein M and Harris CC. Morgan DO. (1995). Nature, 374, 131 ± 134. (1994). Cancer Res., 54, 4855 ± 4878. Morgenbesser SD, Williams BO, Jacks T and DePinho RA. Halbert CL, Demers GW and Galloway DA. (1992). J. (1994). Nature, 371, 72 ± 74. Virol., 66, 2125 ± 2134. Mosca PJ, Dijkwel PA and Hamlin JL. (1992). Mol. Cell. Harper JW, Elledge SJ, Keyomarsi K, Dynlacht B, Tsai L, Biol., 12, 4375 ± 4383. Zhang P, Dobrowolski S, Bai C, Connell-Crowley L, Ohtani K, DeGregori J, Leone G, Herendeen DR, Kelly TJ Swindell E, Fox MP and Wei N. (1995). Mol. Biol. Cell, 6, and Nevins JR. (1996). Mol. Cell. Biol., 16, 6977 ± 6984. 387 ± 400. Otterson GA, Kratzke RA, Coxon A, Kim YW and Kaye FJ. Harris MP, Sutjipto S, Wills KN, Hancock W, Cornell D, (1994). Oncogene, 9, 3375 ± 3378. Johnson DE, Gregory RJ, Shepard HM and Maneval DC. Pan H and Griep AE. (1994). Genes Dev., 8, 1285 ± 1299. (1996). Cancer Gene Therapy, 3, 121 ± 130. Pardee AB. (1989). Science, 246, 603 ± 608. Hartwell L. (1992). Cell, 71, 543 ± 546. p53 modulation of pRb phosphorylation state SP Linke et al 345 Rieder CL and Palazzo RE. (1992). J. Cell. Sci., 102, 387 ± Werness BA, Levine AJ and Howley PM. (1990). Science, 392. 248, 76 ± 79. Sherr CJ. (1994). Cell, 79, 551 ± 555. Wills KN, Huang WM, Harris MP, Machemer T, Maneval Sherr CJ. (1996). Science, 274, 1672 ± 1677. DC and Gregory RJ. (1995). Cancer Gene Ther., 2, 191 ± Slansky JE, Li Y, Kaelin WG and Farnham PJ. (1993). Mol. 197. Cell. Biol., 13, 1610 ± 1618. Wills KN, Maneval DC, Menzel P, Harris MP, Sutjipto S, Slebos RJ, Lee MH, Plunkett BS, Kessis TD, Williams BO, Vaillancourt MT, Huang WM, Johnson DE, Anderson Jacks T, Hedrick L, Kastan MB and Cho KR. (1994). SC, Wen SF, Bookstein R, Shepard HM and Gregory R. Proc. Natl. Acad. Sci. USA, 91, 5320 ± 5324. (1994). Hum. Gene Ther., 5, 1079 ± 1088. TakahashiT,NauMM,ChibaI,BirrerMJ,RosenbergRK, Xiao ZX, Chen J, Levine AJ, Modjtahedi N, Xing J, Sellers Vinocour M, Levitt M, Pass H, Gazdar AF and Minna JD. WR and Livingston DM. (1995). Nature, 375, 694 ± 698. (1989). Science, 246, 491 ± 494. Xu M, Sheppard KA, Peng CY, Yee AS and Piwnica WH. Wang Y and Prives C. (1995). Nature, 376, 88 ± 91. (1994). Mol. Cell. Biol., 14, 8420 ± 8431. Wen SF, Nodelman M, Nared HK, Duncan J, Geradts J and Yin Y, Tainsky MA, Bischo€ FZ, Strong LC and Wahl GM. Shepard HM. (1994). J. Immunol. Methods, 169, 231 ± 240. (1992). Cell, 70, 937 ± 948.