JOURNAL OF VIROLOGY, Sept. 2002, p. 9526–9532 Vol. 76, No. 180022-538X/02/$04.00�0 DOI: 10.1128/JVI.76.18.9526–9532.2002Copyright © 2002, American Society for Microbiology. All Rights Reserved.

Induction and Bypass of p53 during Productive Infectionby Polyomavirus

Dilip Dey,† Jean Dahl, Sayeon Cho,‡ and Thomas L. Benjamin*Department of Pathology, Harvard Medical School, Boston, Massachusetts 02115

Received 1 March 2002/Accepted 19 June 2002

Lytic infection by polyomavirus leads to elevated levels of p53 and induction of p53 target genes p21Cip1/WAF1 (p21) and BAX. This is seen both in polyomavirus-infected primary mouse cell cultures and in kidneytissue of infected mice. Stabilization of p53 and induction of a p53 response are accompanied by phosphory-lation of p53 on serine 18, mimicking a DNA damage response. Stabilization of p53 does not depend on p19Arfinteraction with mdm2. Cells infected by a mutant virus defective in binding pRb and in inducing G1-to-Sprogression show a greatly diminished p53 response. However, cells infected by wild-type virus and blockedfrom entering S phase by addition of mimosine still show a p53 response. These results suggest a role of E2Ftarget genes in inducing a p53 response. Polyomavirus large T antigen coprecipitates with p53 phosphorylatedon serine 18 and also with p21Cip1/WAF1. Implications of these and other findings on possible mechanismsof induction and override of p53 functions during productive infection by polyomavirus are discussed.

The tumor suppressor p53 plays a pivotal role in carcino-genesis and is the most frequently mutated gene in humancancers (31, 34). Most DNA tumor viruses have mechanisms toinactivate p53. The large T antigen (large T) of simian virus 40(SV40) binds p53 and inactivates at least some of its functions(44). The E6 proteins of the highly oncogenic human papillo-mavirus type 16 (HPV-16) and HPV-18 promote the rapiddegradation of p53 through a ubiquitin-dependent proteolyticpathway (47). The oncogenic human adenoviruses act to blockp53 through dual functions of the E1B proteins acting directlyon p53 and on its downstream targets (16).

Thus far, none of these mechanisms have been ascribed tothe mouse polyomavirus. This is surprising in view of the effi-ciency and rapidity with which this virus induces tumors. p53 isnot stably upregulated in polyomavirus tumors, and most tu-mor cell lines examined show a normal p53 response to DNAdamage (18). Regulation of p53 occurs through a variety ofmechanisms that may operate differently in different cell types.The possibility that polyomavirus may have some way of coun-teracting p53 in various target cells was tested by determiningthe effect of the absence of p53 on tumor induction by thevirus. Tumors arose significantly more rapidly in p53�/� thanin p53�/� or p53�/� mice, supporting the generally held viewthat this virus has no effective way of blocking p53 functionsduring the course of tumor development (18). This result con-trasts with those in the SV40 system, where the large T antigenclearly binds p53 and where the absence of p53 in the host canretard rather than accelerate tumor development (28).

A single cycle of polyomavirus growth in mouse cells re-quires roughly 48 h and is dependent on cell cycle progression

from G0/G1 into S. In the absence of a counteracting mecha-nism(s) by the virus, induction of a p53 response leading eitherto cell cycle arrest or apoptosis would be expected to blockvirus growth. Expression of polyomavirus large T in NIH 3T3cells can overcome p53-dependent arrest by a mechanism de-pendent on large T interaction with pRb (20, 25). In REF52established rat fibroblasts, which express normal p53, polyoma-virus large T and/or small T can block signaling betweenp19ARF and p53 (39, 43). Middle T by itself fails to overridep53 (19) and also fails to transform REF52 cells unless accom-panied by large and/or small T (43). These studies were carriedout with subviral constructs or in nonpermissive cells, i.e.,under conditions where normal virus replication does not oc-cur. The present study was undertaken to gain a better under-standing of the effects of polyomavirus on p53 during lyticinfection of mouse cells and of how the virus might overrideeffects of p53.

MATERIALS AND METHODS

Cells and viruses. Primary baby mouse kidney (BMK) cells were preparedfrom 15-day-old baby mouse kidneys (63). INK-4a�/� mouse embryo fibroblasts(MEF) (third passage) were a generous gift of Ron DePinho and NormanSharpless of the Dana-Farber Cancer Institute, Boston, Mass. p53�/� MEF werederived from p53-deficient mice from our colony (18) or from James DeCaprioof the Dana-Farber Cancer Institute. Polyomaviruses were the wild-type labo-ratory strain RA (24), the highly virulent strain LID (4), and the RB1 mutantencoding large T defective for binding of pRb (26). Monolayers were maintainedin Dulbecco’s minimal essential medium supplemented with 1 and 10% fetalbovine serum for BMK and MEF, respectively. Cells were infected at a multi-plicity of infection (MOI) of 1 to 30 PFU. The percentage of cells infected wasdetermined by nuclear staining for large T by indirect immunofluorescence,using rat polyclonal anti-T antigen (T Ag) (55) and fluorescein isothiocyanate-conjugated anti-rat immunoglobulin G (IgG) (Jackson ImmunoResearch).

Immunoprecipitation and immunoblotting. Lysates of infected cells were pre-pared at the indicated times with NP-40 lysis buffer (20 mM Tris, pH 7.5, 135 mMNaCl, 1 mM MgCl2, 0.1 mM CaCl2, 10% glycerol, 1% NP-40, 0.1 mM Na3VO4,50 mM �-glycerophosphate, 10 mM NaF, and the protease inhibitor CompleteMini from Roche) and were used for immunoblotting. For immunoprecipitationscells were lysed in buffer containing 20 mM Tris, pH 7.5, 135 mM NaCl, 1%glycerol, 1% NP-40, 0.1 mM Na3VO4, 10 mM �-glycerophosphate, 10 mM NaF,and the protease inhibitor Complete Mini from Roche. For immunoblotting, 70�g of protein was analyzed by sodium dodecyl sulfate-polyacrylamide gel elec-

* Corresponding author. Mailing address: Department of Pathology,Harvard Medical School, 200 Longwood Ave. D2-230, Boston, MA02115. Phone: (617) 432-1960. Fax: (617) 277-5291. E-mail:thomas_benjamin@hms.harvard.edu.

† Present address: Division of Biology, University of California atSan Diego, La Jolla, CA 92093-0366.

‡ Present address: Proteome Research Laboratory, Korea ResearchInstitute of Bioscience and Biotechnology, Taejon 305-33, Korea.

9526

trophoresis. For immunoprecipitations, extracts (0.5 to 1 mg of protein) wereincubated with antibody at 4°C overnight and the immune complexes wererecovered with protein A-Sepharose CL-4B (Amersham Pharmacia). The pro-teins were resolved on 10 or 12.5% polyacrylamide gels and were transferred tonitrocellulose membrane. Antibodies for immunoblotting were rabbit polyclonalanti-p53 (CM5) from Novocastra; monoclonal anti-p53 (Ab-3, Pab 240) fromOncogene Research; rabbit polyclonal anti-p21 (C-19), rabbit polyclonal anti-BAX (N-20), goat polyclonal anti-p19Arf (M-20), rabbit polyclonal anti-mdm2(H-221), mouse monoclonal antiactin (C-2) (all from Santa Cruz), and rat poly-clonal anti-T Ag ascites (55). Antibodies for immunoprecipitation were rabbitpolyclonal anti-p21 (C-19) from Santa Cruz, rabbit polyclonal anti-p53 or rabbitpolyclonal anti-phosphoserine-15 p53 from Cell Signaling, and monoclonal anti-p53 (Ab-1, Pab 421) from Oncogene Research.

To block immunoprecipitation by anti-phosphoserine-15 p53, 2.5 �l of anti-body and 2.5 �l of phosphoserine-15 blocking peptide (Cell Signaling) in 95 �lof lysing buffer were incubated on ice for 2 h before addition of cell extract.

Immunocytochemistry. Infected cells were fixed with 3.7% neutral bufferedformalin at room temperature for 20 min. After permeabilization with 0.3%Triton X-100 in phosphate-buffered saline (PBS) for 20 min, the cells wereblocked with 5% normal donkey serum (Jackson ImmunoResearch) in PBS.Cells were stained for T Ags with polyclonal anti-T ascites (55), followed byfluorescein isothiocyanate-conjugated anti-rat IgG (Jackson ImmunoResearch),and were stained for phosphoserine-15 p53 with a rabbit polyclonal antibodyfrom Cell Signaling, followed by tetramethyl rhodamine isocyanate-conjugatedanti-rabbit IgG (Jackson ImmunoResearch). Cells were counterstained with4�,6�-diamidino-2-phenylindole (Sigma) and were analyzed with a Nikon TE300fluorescence microscope. To block immunoreactivity, 2 �l of a 1:10 dilution ofthe phosphoserine-15 p53 antibody and 1 �l of blocking peptide (Cell Signaling)in 97 �l of Tris-buffered saline containing 0.1% Tween and 5% bovine serumalbumin were incubated at room temperature for 1 h before being added toslides.

Mice and virus injections. Newborn C3H/BiDa mice less than 18 h of age wereinjected intraperitoneally with the LID strain of polyomavirus at approximately106 PFU/animal. After 8 days, mice were sacrificed, kidney homogenates wereprepared in NP-40 lysis buffer, and 180 to 200 �g of protein was used forimmunoblotting.

Cell cycle analysis. Cells were harvested by trypsinization and frozen at �80°Cin 250 mM sucrose–40 mM sodium citrate–5% dimethyl sulfoxide at 2 � 106

cells/ml (8). For analysis, thawed cells (200 �l) were washed in PBS containing2% calf serum and incubated with 300 �l of PBS–0.5% Triton X-100–0.5 mMEDTA containing 100 �g of propidium iodide/ml for 20 min. Eighty microlitersof 2 M HCl was added, incubation was continued for 30 min, and 300 �l of 1 MTris base was added. Ten thousand cells per sample were analyzed for total DNAon a FACSCalibur flow cytometer (Becton Dickinson) using the programs CellQuest and ModFit.

RESULTS

Stabilization of p53 and induction of p53 target genes dur-ing productive infection by polyomavirus. Extracts of wild-type-polyomavirus-infected BMK cells were prepared at vari-ous times after infection, separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and blotted with antibodiesto p53 and p53 target genes p21Cip1/WAF1 (p21) and BAX(Fig. 1). Levels of p53 in uninfected cultures were low andremained constant over the 40-h time course examined; p21was barely detectable. In infected cultures, levels of p53 andp21 began to rise around 12 h, coincident with the appearanceof large T. By 24 h postinfection, the amounts of p53 and p21rose 10- to 20-fold or more compared to those in uninfectedcultures. The slight declines in infected cultures at the latertime points coincide with the development of cytopathic effectsand loss of cells from the monolayers. Levels of BAX were lowand somewhat variable in uninfected cultures but were seen torise upon infection in most experiments but to a lesser degreethan those of p21. Induction of p53 and its targets followingpolyomavirus infection resembled that seen in uninfected cellstreated with 5 nM actinomycin D (Act D) for 24 h. When a

rabbit polyclonal antibody to mdm2 capable of recognizing the90-kDa form and the p60 cleavage product was used, little orno change in levels was seen in either infected cultures or inAct D-treated cells (not shown). The rise in levels of p53 ininfected BMK cells is thus accompanied by induction of somebut not all p53-responsive genes. Different patterns of p53response may occur among the many different cell types thatthe virus is known to lytically infect in vivo (15).

p21 is known to be induced at the transcriptional level byp53 (23, 38) but may also be regulated through E2F-1 in ap53-independent manner (30). To determine if the increase inp21 in virus-infected cells is p53 dependent, experiments werecarried out with p53�/� and p53�/� MEF. Results in Fig. 2show that induction of p21 following polyomavirus infection isdependent on p53. The induction of p21 by Act D is also p53dependent in these cells.

The p53 response in polyomavirus-infected cells is accom-panied by phosphorylation of p53 on serine 18 and does notdepend on the mdm2/p19Arf pathway. Regulation of p53 levelsoccurs primarily through posttranslational mechanisms (1).One of these depends on mdm2 binding to p53, leading to p53degradation via ubiquitin-dependent proteolysis (45, 52). Thispathway is opposed by the p19Arf product of INK-4a, whichbinds and sequesters mdm2 in nucleoli leading to accumula-tion of p53 (62). To determine if p53 accumulation followingpolyomavirus infection depends on p19Arf, normal and INK-

FIG. 1. Induction of p53 and its target genes. Confluent BMK cellswere either uninfected or infected with wild-type polyomavirus at anMOI of 20 to 30. Extracts were prepared at the times indicated andwere analyzed by immunoblotting. p53 was detected with anti-p53(CM5 from Novocastra). At 12 and 24 h postinfection, approximately20 and 85% of the cells were T Ag positive. Uninfected cells treatedwith 5 nM Act D for 24 h served as a control. LT, large T.

FIG. 2. Induction of p21 is p53 dependent. Early-passage p53�/�

and p53�/� MEF were either uninfected or infected with wild-typepolyomavirus at an MOI of 5 to 10 and were incubated for 24 h.Approximately 75% of the cells were T Ag positive. The conditions fortreatment with Act D are given in the legend for Fig. 1A. Extracts wereprepared and analyzed by immunoblotting. p53 was detected withanti-p53 (Ab-3 from Oncogene Research).

VOL. 76, 2002 POLYOMAVIRUS AND p53 9527

4a�/� MEF were infected and analyzed. Levels of p53 and p21were seen to increase in the same manner in both sets ofcultures (Fig. 3). These results, as well as the finding of little orno change in levels of mdm2 and p19Arf (not shown), indicatethat p53 accumulation in polyomavirus-infected BMK cells isnot primarily dependent on p19Arf.

Phosphorylation of p53 occurs at multiple sites in responseto various stimuli, and these modifications are important inmediating p53’s stability and functions (1, 5, 21, 33, 37, 40,52–54). DNA damage from UV or ionizing radiation leads tophosphorylation on serines 18 and 23 of murine p53 involvingATM, ATR, CHK1/2, or other kinases (9, 29, 50, 51, 59).Phosphorylation of p53 at serines 15 and 20 (serines 18 and 23in mouse p53) is known to prevent the binding of mdm2 to p53and the ability of mdm2 to inhibit p53 transactivation (12, 52).Mutation of serines 15 and 20 to alanine prevents full stabili-zation of p53 in vivo (10, 60), and mutation of mouse serine 18to alanine in murine embryonic stem cells reduces p53 accu-mulation in response to DNA damage (9). A phosphopeptideantibody specific for p53 phosphorylated on serine 18 was usedin an immunoblot of extracts from uninfected, infected, andAct D-treated BMK cells. Polyomavirus infection clearly leadsto phosphorylation of p53 on serine 18, as does the DNAdamage induced by Act D (Fig. 4).

Induction of a p53 response in vivo. To determine whetherlytic infection in the intact host is also accompanied by a p53response, extracts of kidneys of 8-day-old neonatally infectedmice were prepared and analyzed. The virulent LID strain ofpolyomavirus was used because of its ability to cause rapid andextensive lytic damage in the kidney (4, 6). Extracts of kidneysfrom LID-infected mice showed significant elevations of p53compared to extracts from kidneys of uninfected mice (Fig. 5).Stabilization of p53 in the virus-infected kidney is accompaniedby induction of p21 and by phosphorylation of p53 on serine18. The p53 response induced by polyomavirus lytic infectionin vitro and in vivo resembles that induced by DNA damage inall respects examined thus far.

Induction of a p53 response does not require entry of cellsinto S phase and synthesis of viral DNA. In spite of inductionof p53 and p21, polyomavirus drives cells into S phase duringlytic infection in a manner dependent on large T binding topRb (20, 25). To determine whether p53 induction by the virusrequires override of the G1 block and synthesis of viral as wellas cellular DNA, levels of p53 and p21 were compared in cellsinfected by wild-type virus and the RB1 virus mutant. RB1encodes a large T defective in binding to pRb, is inefficient ininducing S-phase entry, and shows delayed and reduced levelsof viral DNA synthesis (25). BMK cultures were infected bywild-type or mutant virus at MOIs of around 1, leading toroughly two-thirds of the cells becoming T Ag positive in eachset of cultures. Cells infected by the wild-type virus progressedinto S phase, accompanied by a roughly 18-fold increase in thelevel of phosphorylated p53 and a 10-fold increase in p21. Incontrast, cells infected by RB1 at a matched MOI showedincreases of less than twofold in the same parameters (Fig. 6).At considerably higher MOIs, the RB1 mutant was able toinduce a larger p53 response. Binding of pRb by large T istherefore essential for efficient induction of a p53 response.These results are consistent with those reported earlier, indi-cating p21 induction by wild-type large T but not by a pRbbinding mutant (49).

To determine the efficiency of induction of a p53 response bywild-type virus at the cellular level and to confirm that the RB1mutant is impaired in inducing this response, infected BMKcells were examined by double indirect immunofluorescence

FIG. 3. Accumulation of p53 in polyomavirus-infected cells doesnot depend on the p19Arf/mdm2 pathway. Early-passage wild-typeMEF and INK-4a�/� MEF were infected as described for Fig. 1A.Extracts were prepared at the times indicated and were analyzed byimmunoblotting. p53 was detected with anti-p53 (CM5 from Novocas-tra).

FIG. 4. Phosphorylation of p53 on serine 18 (Pser18) is induced byinfection of BMK cells with polyomavirus. BMK cells were infected ortreated with Act D as described for Fig. 1A, and lysates were preparedat the times indicated. Immunoblot analysis was performed with apolyclonal antibody specific for p53 phosphorylated on serine 15(serine 18 in mouse).

FIG. 5. Induction of a p53 response in vivo. Newborn C3H/BiDapups were inoculated with a lethal dose of LID. After 8 days, pupswere sacrificed and kidney homogenates were prepared for immuno-blot analysis. p53 was detected with anti-p53 (CM5 from Novocastra)and p21 with C-19 (Santa Cruz). Immunoblot analysis for phospho-serine-18 (Pser18) p53 was performed in a separate experiment usinganti-phosphoserine-15 p53 from Cell Signaling.

FIG. 6. Induction of phosphoserine-18 (Pser18) p53 and p21 isdependent on binding of large T (LT) to pRb. Confluent BMK cellswere either uninfected or infected with wild-type polyomavirus ormutant RB1 at an MOI of 1. Cells were harvested at the times indi-cated and were analyzed by immunoblotting, T Ag staining, and fluo-rescence-activated cell sorting.

9528 DEY ET AL. J. VIROL.

staining using anti-T and anti-phosphoserine-18 p53 antibod-ies. Images are shown in Fig. 7, and quantitation is given inTable 1. It is clear that a majority of wild-type-virus-infectedcells that become T Ag positive at 18 h postinfection also showinduction of phosphoserine-18 p53. The percentage of T Ag-positive cells that are also phosphoserine-18 p53 positive riseswith time after infection, from 62% at 18 h to 88% at 31 h.With the RB1 mutant at 18 h, only 3.4% of cells are doublypositive. The percentage rises with time but remains well belowthat achieved by the wild-type virus, consistent with the knowndefects and leakiness of this mutant (25).

Induction of p53 by the virus may depend on function(s) ofone or more E2F family members that are activated followinglarge T:pRb interaction but preceding actual S-phase entry. Totest this possibility, BMK cells were infected by wild-type virus

allowing release and activation of E2F and were incubated withmimosine to impose a block at the G1/S boundary (35). Underthese conditions, p53 and p21 were induced normally, alongwith phosphorylation of p53 on serine 18 (Fig. 8). Entry ofinfected cells into S phase and synthesis of viral DNA aretherefore not required for induction of a p53 response. Rather,the response depends on large T interaction with pRb and

FIG. 7. Immunofluorescence staining of BMK cells infected with wild-type or mutant RB1 virus for T Ag and phosphoserine-18 (Pser18) p53.Confluent BMK cells were infected with wild-type (WT) or RB1 virus at an MOI of 1 or 2 and incubated for 18 h. To block possible nonspecificstaining, antibody to phosphoserine-18 p53 was incubated with the antigenic peptide for 1 h before being applied to cells on coverslips. DAPI,4�,6�-diamidino-2-phenylindole.

FIG. 8. Induction of phosphoserine-18 (Pser18) p53 and p21 occursin cells blocked in S-phase entry by mimosine. Confluent BMK cellswere either uninfected or infected with wild-type polyomavirus at anMOI of 1. At 4 h postinfection mimosine was added to a concentrationof 0.2 or 0.5 mM. Cells were harvested at 24 h postinfection and wereanalyzed by immunoblotting, T Ag staining, and fluorescence-activatedcell sorting.

TABLE 1. Incidence of T Ag and phosphoserine-18 p53 (Pp53) inwild-type-virus-infected and mutant-RB1-infected BMK cells

No. of hpostinfection

Results for:

Wild-type Mutant RB1

% T Ag� % T Ag/Pp53� % T Ag� % T Ag/Pp53�

18 70 62 57 3.424 88 74 63 1231 91 88 70 18

VOL. 76, 2002 POLYOMAVIRUS AND p53 9529

most likely activation of E2F function(s). Cellular genes re-quired for viral DNA synthesis are known to be induced byE2F-1. Beside G1 cyclins, cyclin-dependent kinases, and vari-ous enzymes needed for DNA synthesis, these include DNApolymerase � (“primase”) and PCNA (17, 22). The latter as-semble along with RPA, topoisomerase I, and large T at theviral replication origin and are essential for the initiation re-action and replication (56).

Polyomavirus large T coprecipitates with phosphorylatedp53 and with p21. Earlier attempts at demonstrating directinteraction between p53 and polyomavirus large T were nega-tive using a monoclonal antibody that effectively brings downSV40 large T:p53 complexes (61). To determine if polyomavi-rus might interact with phosphorylated p53, immunoprecipita-tion of an infected cell extract was carried out using serine 18phosphopeptide antibody followed by blotting with anti-T. Thisantibody brought down significantly more large T than eithercontrol or other anti-p53 antibodies (Fig. 9A). The ability ofthe phosphopeptide antibody to coprecipitate large T was ef-fectively inhibited by preincubation with phosphopeptide (Fig.9B). Polyclonal rabbit anti-p53 gave various but generally lowamounts of large T by coimmunoprecipitation in comparison

with anti-phosphoserine-18 p53. Thus, while polyomaviruslarge T coprecipitates with p53 phosphorylated on serine 18,some form of interaction with unphosphorylated p53 or p53modified at other sites is also possible. Neither normal IgG northe mouse monoclonal antibody to p53 previously found to benegative (61) was found to coprecipitate polyomavirus large T.

The E7 protein encoded by the high-risk HPVs, in additionto binding pRb, also interacts with p21. This interaction con-tributes to the ability to override p53-mediated G1 arrest andto promote continued cell division in differentiating keratino-cytes (27, 32). The possibility that a similar interaction occursbetween polyomavirus large T and p21 was investigated usinganti-p21 as the precipitating antibody and extracts of infectedBMK cells. A substantial amount of large T was found tocoprecipitate with anti-p21 (Fig. 9C). Thus, large T may act ina manner similar to that of HPV E7 in binding p21 and block-ing its cell cycle inhibitory functions.

DISCUSSION

Mice developing polyomavirus tumors show no evidence ofeffects attributable to the virus leading to a p53 block (18).Polyomavirus tumors fail to show stabilization of p53 in themanner of SV40-induced tumors, and the majority of primarypolyomavirus tumor-derived cell lines retain p53 and show anormal p53 response to DNA damage (18). Thus, in the con-text of tumor cell growth, polyomavirus appears neither toelicit nor to block p53.

In contrast, results of the present study clearly show a p53response elicited by polyomavirus during productive viral in-fection, both in cell culture and in the mouse. Lytic infectionand virus amplification are essential steps leading to the induc-tion of tumors. These findings thus raise questions at twolevels. First, what is the mechanism of p53 induction operatingin productively infected cells but not in nonproductively in-fected (i.e., tumor) cells, and second, how is the virus able tocircumvent the cell cycle arrest and proapoptotic functions ofp53 during its own replication?

In polyomavirus-infected primary BMK cells, p53 begins toaccumulate along with expression of the T Ags and prior to theonset of viral DNA replication. By 20 to 24 h postinfection,levels of p53 are 10- to 20-fold higher than those found inmock-infected cells. Coincident with its accumulation, p53 be-comes phosphorylated on serine 18, leading to its stabilization.p53 target genes are selectively induced. p21 is highly induced,BAX is induced to a lesser degree, and mdm2 remains essen-tially unchanged. This pattern of induction is similar to thatseen in the same cells following Act D treatment. This suggeststhat lytic infection by polyomavirus in some manner triggers aDNA damage-like response. The latter typically involves acti-vation of ATM, ATR, DNA-PK, and the CHK1 and CHK2kinases, which are known to phosphorylate p53 on N-terminalserine residues. Though the specific pathway leading to p53phosphorylation has not been determined, some aspect of thispathway upstream of p53, or possibly the p53 response itself,may be important to the virus.

Results suggest that the viral trigger for induction of a p53response in a lytic infection may be viral DNA replicationinitiation complexes. These complexes, best characterized for

FIG. 9. Polyomavirus large T (LT) coprecipitates with p53 phos-phorylated on serine 18 and with p21. Confluent BMK cells wereinfected at an MOI of 5 to 10 and incubated for 24 h. (A) Immunecomplexes were prepared with rabbit polyclonal anti-p53 (Cell Signal-ing), rabbit polyclonal anti-phosphoserine-18 (Pser18) p53, or mousemonoclonal anti-p53 (Pab 421). (B) Immune complexes were preparedwith rabbit polyclonal anti-phosphoserine-18 p53 preincubated with orwithout blocking peptide. (C) Immune complexes were prepared withrabbit polyclonal anti-p21 antibody. The corresponding normal IgGswere used as controls. Immunoblots were performed with rat poly-clonal anti-T Ag ascites (55).

9530 DEY ET AL. J. VIROL.

SV40 (56), consist of large T bound to the viral origin alongwith replication proteins encoded by the host. The viral DNAis partially unwound at the origin, mimicking cellular DNAorigins that have fired (“theta” structures). Results from twoexperiments support this interpretation and at the same timeindicate an important role for E2F-1. The large T mutant RB1,which is defective in binding pRb and in inducing cellular aswell as viral DNA synthesis, is also defective in inducing a p53response. Wild-type virus, on the other hand, is able to inducep53, even when blocked from inducing synthesis of viral andcellular DNA by addition of mimosine to infected cultures.Under these conditions, wild-type large T is expected to bindpRb leading to activation of E2F-1-responsive genes. The lat-ter include DNA polymerase � and PCNA (17, 22), whichfunction along with T antigen in viral DNA initiation andreplication complexes (56). Thus, cells harboring free viralDNA with fired origins but unable to enter S phase may re-spond with a DNA damage response.

Two other mechanisms of induction are possible but lesslikely. One is that the introduction of linear fragments of cel-lular DNA via pseudovirions (42) would mimic the generationof double-strand breaks by ionizing radiation and thereby in-duce a p53 response. This seems unlikely to be the only mech-anism, since the RB1 mutant is impaired in its ability to inducethe response. A second possibility is that induction of p19Arfby E2F-1 (3) would lead to p53 stabilization by interferencewith mdm2-dependent p53 degradation (62). In an establishedline of rat fibroblasts, middle T can induce a p53-dependentcell cycle block via p19Arf, a block that large and/or small Tcan overcome (39). In the mouse system, however, stabilizationof p53 occurs following polyomavirus infection of INK-4a�/�

MEF and is therefore not dependent solely on the p19Arf/mdm2 pathway.

Induction of p53 during lytic infection potentially could leadto G1 arrest or to apoptosis. Either cellular response wouldeffectively block virus replication. However, several mecha-nisms are available to the virus to bypass or block p53 re-sponses. First, the binding of pRb by large T would be expectedto act downstream of p21, bypassing the G1 arrest broughtabout by inhibition of the G1 cyclin–cyclin-dependent kinasesacting on pRb. Induction of p21, which can be brought aboutby the virus either via p53 or possibly E2F-1 (30), may con-ceivably serve a positive function for the virus by titrating thelevels of free p21 so as to aid in the assembly and promoterather than inhibit the activity of these kinases (30, 36, 64).

Coprecipitation of large T with phosphorylated p53 mayindicate direct or indirect association. Only a small fraction oflarge T appears to be involved. Interaction could be mediatedby the CBP/p300 coactivators of p53 which also bind polyoma-virus large T (11, 43a). Alternatively, large T and p53 may bebrought together on the viral DNA, which contains p53 andlarge T binding sites adjacent to the core origin (42a). Theanti-p53 monoclonal antibody Pab 421 fails to bring downcomplexes with polyomavirus large T (reference 61 andpresent results). This antibody binds to p53 at the C terminus(57) and effectively brings down complexes with SV40 large T,which binds to the central DNA-binding domain of p53 (2, 58).This suggests that polyomavirus large T may interact differ-ently from SV40 large T by binding at the C terminus of p53such that these complexes would not be recognized by the Pab

421 monoclonal antibody. Large T does not block the induc-tion of p21 by p53. However, the binding of large T to p21provides a plausible mechanism for counterbalancing the cellgrowth arrest functions of p53 in a manner similar to thatshown for HPV E7 (27, 32).

The proapoptotic functions of p53 mediated by BAX couldwell be overridden by middle T activation of the phosphatidyl-inositol 3-kinase/Akt pathway, which is known to protectagainst apoptosis in virus-infected cells (13, 41, 46). Akt canblock apoptosis at several levels, including phosphorylation ofBAD, caspases, and p21 (7, 14, 65). Both the hamster andmouse polyomaviruses encode middle T proteins that activatethis antiapoptotic pathway (13, 48), distinguishing them fromother DNA tumor viruses that lack a middle T and that targetp53 more directly.

ACKNOWLEDGMENTS

D. Dey and J. Dahl contributed equally to this work.We thank James DeCaprio for p53 �/� fibroblasts, Ron DePinho

and Norman Sharpless for providing the INK-4a�/� MEF, and WalterEckhart and Carol Prives for helpful discussions. The technical assis-tance of John Carroll and Rebecca Dowgiert is also gratefully acknowl-edged.

This work has been supported by grants R35-44343, PO1 CA-50661,and RO1 CA-90992 from the National Cancer Institute.

REFERENCES

1. Appella, E., and C. W. Anderson. 2001. Post-translational modifications andactivation of p53 by genotoxic stresses. Eur. J. Biochem. 268:2764–2772.

2. Bargonetti, J., I. Reynisdottir, P. N. Friedman, and C. Prives. 1992. Site-specific binding of wild-type p53 to cellular DNA is inhibited by SV40 Tantigen and mutant p53. Genes Dev. 6:1886–1898.

3. Bates, S., A. C. Phillips, P. A. Clark, F. Stott, G. Peters, R. L. Ludwig, andK. H. Vousden. 1998. p14ARF links the tumour suppressors RB and p53.Nature 395:124–125.

4. Bauer, P. H., R. T. Bronson, S. C. Fung, R. Freund, T. Stehle, S. C. Harrison,and T. L. Benjamin. 1995. Genetic and structural analysis of a virulencedeterminant in polyomavirus VP1. J. Virol. 69:7925–7931.

5. Bean, L. J. H., and G. R. Stark. 2001. Phosphorylation of serines 15 and 37is necessary for efficient accumulation of p53 following irradiation with UV.Oncogene 20:1076–1084.

6. Bolen, J. B., S. E. Fisher, K. Chowdhury, T. C. Shan, J. E. Williams, C. J.Dawe, and M. A. Israel. 1985. A determinant of polyomavirus virulenceenhances virus growth in cells of renal origin. J. Virol. 53:335–339.

7. Cardone, M. H., N. Roy, H. R. Stennicke, G. S. Salvesen, T. F. Franke, E.Stanbridge, S. Frisch, and J. C. Reed. 1998. Regulation of cell death pro-tease caspase-9 by phosphorylation. Science 282:1318–1321.

8. Chang, T. H., F. A. Ray, D. A. Thompson, and R. Schlegel. 1997. Disregu-lation of mitotic checkpoints and regulatory proteins following acute expres-sion of SV40 large T antigen in diploid human cells. Oncogene 14:2383–2393.

9. Chao, C., S. Saito, C. W. Anderson, E. Appella, and Y. Xu. 2000. Phosphor-ylation of murine p53 at ser-18 regulates the p53 responses to DNA damage.Proc. Natl. Acad. Sci. USA 97:11936–11941.

10. Chehab, N. H., A. Malikzay, E. S. Stavridi, and T. D. Halazonetis. 1999.Phosphorylation of Ser-20 mediates stabilization of human p53 in responseto DNA damage. Proc. Natl. Acad. Sci. USA 96:13777–13782.

11. Cho, S., Y. Tian, and T. L. Benjamin. 2001. Binding of p300/CBP co-activatorby polyoma large T antigen. J. Biol. Chem. 276:33533–33539.

12. Craig, A. L., L. Burch, B. Vojtesek, J. Mikutowska, A. Thompson, and T. R.Hupp. 1999. Novel phosphorylation sites of human tumour suppressor pro-tein p53 at Ser20 and Thr18 that disrupt the binding of mdm2 (mouse doubleminute 2) protein are modified in human cancers. Biochem. J. 342:133–141.

13. Dahl, J., A. Jurczak, L. A. Cheng, D. C. Baker, and T. L. Benjamin. 1998.Evidence of a role for phosphatidylinositol 3-kinase activation in the block-ing of apoptosis by polyomavirus middle T antigen. J. Virol. 72:3221–3226.

14. Datta, S. R., A. Brunet, and M. E. Greenberg. 1999. Cellular survival: a playin three Akts. Genes Dev. 13:2905–2927.

15. Dawe, C. J., R. Freund, G. Mandel, K. Ballmer-Hofer, D. A. Talmage, andT. L. Benjamin. 1987. Variations in polyoma virus genotype in relation totumor induction in mice. Characterization of wild type strains with widelydiffering tumor profiles. Am. J. Pathol. 127:243–261.

16. Debbas, M., and E. White. 1993. Wild-type p53 mediates apoptosis by E1A,which is inhibited by E1B. Genes Dev. 7:546–554.

VOL. 76, 2002 POLYOMAVIRUS AND p53 9531

17. DeGregori, J., T. Kowalik, and J. Nevins. 1995. Cellular targets for activationby the E2F1 transcription factor include DNA synthesis- and G1/S-regulatorygenes. Mol. Cell. Biol. 15:4215–4224.

18. Dey, D. C., R. P. Bronson, J. Dahl, J. P. Carroll, and T. L. Benjamin. 2000.Accelerated development of polyoma tumors and embryonic lethality: dif-ferent effects of p53 loss on related mouse backgrounds. Cell Growth Differ.11:231–237.

19. Doherty, J., and R. Freund. 1999. Middle T antigen activation of signaltransduction pathways does not overcome p53-mediated growth arrest. J. Vi-rol. 73:7882–7885.

20. Doherty, J., and R. Freund. 1997. Polyomavirus large T antigen overcomesp53 dependent growth arrest. Oncogene 14:1923–1931.

21. Dumaz, N., and D. W. Meek. 1999. Serine15 phosphorylation stimulates p53transactivation but does not directly influence interaction with HDM2.EMBO J. 18:7002–7010.

22. Dyson, N. 1998. The regulation of E2F by pRB-family proteins. Genes Dev.12:2245–2262.

23. el-Deiry, W. S., T. Tokino, V. E. Velculescu, D. B. Levy, R. Parsons, J. M.Trent, D. Lin, W. E. Mercer, K. W. Kinzler, and B. Vogelstein. 1993. WAF1,a potential mediator of p53 tumor suppression. Cell 75:817–825.

24. Feunteun, J., L. Sompayrac, M. Fluck, and T. L. Benjamin. 1976. Localiza-tion of gene functions in polyoma virus DNA. Proc. Natl. Acad. Sci. USA73:4169–4173.

25. Freund, R., P. H. Bauer, H. A. Crissman, E. M. Bradbury, and T. L. Ben-jamin. 1994. Host range and cell cycle activation properties of polyomaviruslarge T-antigen mutants defective in pRB binding. J. Virol. 68:7227–7234.

26. Freund, R., R. T. Bronson, and T. L. Benjamin. 1992. Separation of immor-talization from tumor induction with polyoma large T mutants that fail tobind the retinoblastoma gene product. Oncogene 7:1979–1987.

27. Funk, J. O., S. Waga, J. B. Harry, E. Espling, B. Stillman, and D. A.Galloway. 1997. Inhibition of CDK activity and PCNA-dependent DNAreplication by p21 is blocked by interaction with the HPV-16 E7 oncoprotein.Genes Dev. 11:2090–2100.

28. Herzig, M., M. Novatchkova, and G. Christofori. 1999. An unexpected rolefor p53 in augmenting SV40 large T antigen-mediated tumorigenesis. Biol.Chem. 380:203–211.

29. Hirao, A., K. Y. Y., S. Matsuoka, A. Wakeham, J. Ruland, H. Yoshida, D.Liu, S. J. Elledge, and T. W. Mak. 2000. DNA damage-induced activation ofp53 by the checkpoint kinase chk2. Science 287:1824–1827.

30. Hiyama, H., A. Iavarone, and S. A. Reeves. 1998. Regulation of the cdkinhibitor p21 gene during cell cycle progression is under the control of thetranscription factor E2F. Oncogene 16:1513–1523.

31. Hollstein, M., D. Sidransky, B. Vogelstein, and C. C. Harris. 1991. p53mutations in human cancers. Science 253:49–53.

32. Jones, D. L., R. M. Alani, and K. Munger. 1997. The human papillomavirusE7 oncoprotein can uncouple cellular differentiation and proliferation inhuman keratinocytes by abrogating p21Cip1-mediated inhibition of cdk2.Genes Dev. 11:2101–2111.

33. Kapoor, M., and G. Lozano. 1998. Functional activation of p53 via phos-phorylation following DNA damage by UV but not gamma radiation. Proc.Natl. Acad. Sci. USA 95:2834–2837.

34. Ko, L. J., and C. Prives. 1996. p53: puzzle and paradigm. Genes Dev.10:1054–1072.

35. Krude, T. 1999. Mimosine arrests proliferating human cells before onset ofDNA replication in a dose-dependent manner. Exp. Cell Res. 247:148–159.

36. LaBaer, J., M. D. Garrett, L. F. Stevenson, J. M. Slingerland, C. Sandhu,H. S. Chou, A. Fattaey, and E. Harlow. 1997. New functional activities for thep21 family of CDK inhibitors. Genes Dev. 11:847–862.

37. Lees-Miller, S. P., K. Sakaguchi, S. J. Ullrich, E. Appella, and C. W. Ander-son. 1992. Human DNA-activated protein kinase phosphorylates serines 15and 37 in the amino-terminal transactivation domain of human p53. Mol.Cell. Biol. 12:5041–5049.

38. Li, Y., C. W. Jenkins, M. A. Nichols, and Y. Xiong. 1994. Cell cycle expres-sion and p53 regulation of the cyclin-dependent kinase inhibitor p21. Onco-gene 9:2261–2268.

39. Lomax, M., and M. Fried. 2001. Polyoma virus disrupts ARF signaling top53. Oncogene 20:4951–4960.

40. Lu, H., Y. Taya, M. Ikeda, and A. J. Levine. 1998. Ultraviolet radiation, butnot gamma radiation or etoposide-induced DNA damage, results in thephosphorylation of the murine p53 protein at serine-389. Proc. Natl. Acad.Sci. USA 95:6399–6402.

41. Meili, R., P. Cron, B. A. Hemmings, and K. Ballmer-Hofer. 1998. Proteinkinase B/Akt is activated by polyomavirus middle-T antigen via a phospha-tidylinositol 3-kinase-dependent mechanism. Oncogene 16:903–907.

42. Michel, M. R., B. Hirt, and R. Weil. 1967. Mouse cellular DNA enclosed inpolyoma viral capsids (pseudovirions). Proc. Natl. Acad. Sci. USA 58:1381–1388.

42a.Miller, S. D., G. Farmer, and C. Prives. 1995. p53 inhibits DNA replicationin vitro in a DNA-binding-dependent manner. Mol. Cell. Biol. 15:6554–6560.

43. Mor, O., M. Read, and M. Fried. 1997. p53 in polyoma virus transformedREF52 cells. Oncogene 15:3113–3119.

43a.Nemethova, M., and E. Wintersberger. 1999. Polyomavirus large T antigenbinds the transcriptional coactivator protein p300. J. Virol. 73:1734–1739.

44. Pipas, J. M., and A. J. Levine. 2001. Role of T antigen interactions with p53in tumorigenesis. Semin. Cancer Biol. 11:23–30.

45. Prives, C., and P. A. Hall. 1999. The p53 pathway. J. Pathol. 187:112–126.46. Qian, W., and K. G. Wiman. 2000. Polyoma virus middle T and small t

antigens cooperate to antagonize p53-induced cell cycle arrest and apoptosis.Cell Growth Differ. 11:31–39.

47. Scheffner, M., B. A. Werness, J. M. Huibregtse, A. J. Levine, and P. M.Howley. 1990. The E6 oncoprotein encoded by human papillomavirus types16 and 18 promotes the degradation of p53. Cell 63:1129–1136.

48. Scherneck, S., R. Ulrich, and J. Feunteun. 2001. The hamster polyomavi-rus—a brief review of recent knowledge. Virus Genes 22:93–101.

49. Schuchner, S., and E. Wintersberger. 1999. Binding of polyomavirus small Tantigen to protein phosphatase 2A is required for elimination of p27 andsupport of S-phase induction in concert with large T antigen. J. Virol.73:9266–9273.

50. She, Q.-B., N. Chen, and Z. Dong. 2000. ERKs and p38 kinase phosphorylatep53 protein at serine 15 in response to UV radiation. J. Biol. Chem. 275:20444–20449.

51. Shieh, S. Y., J. Ahn, K. Tamai, Y. Taya, and C. Prives. 2000. The humanhomologs of checkpoint kinases Chk1 and Cds1 (Chk2) phosphorylate p53 atmultiple DNA damage-inducible sites. Genes Dev. 14:289–300.

52. Shieh, S. Y., M. Ikeda, Y. Taya, and C. Prives. 1997. DNA damage-inducedphosphorylation of p53 alleviates inhibition by MDM2. Cell 91:325–334.

53. Shieh, S. Y., Y. Taya, and C. Prives. 1999. DNA damage-inducible phos-phorylation of p53 at N-terminal sites including a novel site, Ser20, requirestetramerization. EMBO J. 18:1815–1823.

54. Siliciano, J. D., C. E. Canman, Y. Taya, K. Sakaguchi, E. Appella, and M. B.Kastan. 1997. DNA damage induces phosphorylation of the amino terminusof p53. Genes Dev. 11:3471–3481.

55. Silver, J. B., B. S. Schaffhausen, and T. L. Benjamin. 1978. Tumor antigensinduced by nontransforming mutants of polyoma virus. Cell 15:485–496.

56. Simmons, D. T. 2000. SV40 large T antigen functions in DNA replicationand transformation. Adv. Virus Res. 55:75–134.

57. Stephen, C. W., P. Helminen, and D. P. Lane. 1995. Characterization ofepitopes on human p53 using phase-displayed peptide libraries: insights intoantibody-peptide interactions. J. Mol. Biol. 248:58–78.

58. Tan, T.-H., J. Wallis, and A. J. Levine. 1986. Identification of the p53 proteindomain involved in formation of the simian virus 40 large T-antigen–p53protein complex. J. Virol. 59:574–583.

59. Tibbetts, R. S., K. M. Brumbaugh, J. M. Williams, J. N. Sarkaria, W. A.Cliby, S. Y. Shieh, Y. Taya, C. Prives, and R. T. Abraham. 1999. A role forATR in the DNA damage-induced phosphorylation of p53. Genes Dev.13:153–157.

60. Unger, T., R. V. Sionov, E. Moallem, C. L. Yee, P. M. Howley, M. Oren, andY. Haupt. 1999. Mutations in serines 15 and 20 of human p53 impair itsapoptotic activity. Oncogene 18:3205–3212.

61. Wang, E. H., P. N. Friedman, and C. Prives. 1989. The murine p53 proteinblocks replication of SV40 DNA in vitro by inhibiting the initiation functionsof SV40 large T antigen. Cell 57:379–392.

62. Weber, J. D., L. J. Taylor, M. F. Roussel, C. J. Sherr, and D. Bar-Sagi. 1999.Nucleolar Arf sequesters Mdm2 and activates p53. Nat. Cell Biol. 1:20–26.

63. Winocour, E. 1963. Purification of polyoma virus. Virology 19:158–168.64. Zhang, H., G. J. Hannon, and D. Beach. 1994. p21-containing cyclin kinases

exist in both active and inactive states. Genes Dev. 8:1750–1758.65. Zhou, B. P., Y. Liao, W. Xia, B. Spohn, M. H. Lee, and M. C. Hung. 2001.

Cytoplasmic localization of p21Cip1/WAF1 by Akt-induced phosphorylationin HER-2/neu-overexpressing cells. Nat. Cell Biol. 3:245–252.

9532 DEY ET AL. J. VIROL.