Oncogene (1997) 14, 801 ± 809  1997 Stockton Press All rights reserved 0950 ± 9232/97 $12.00

Modulation of (+)-anti-BPDE mediated p53 accumulation by inhibitors of protein kinase C and poly(ADP-ribose) polymerase

Sundaresan Venkatachalam2,3, Mikhail Denissenko1,4 and Altaf A Wani1,2

1Department of Radiology and 2Biochemistry Program, The Ohio State University, Columbus, Ohio 43210, USA

The rapid accumulation of the p53 gene product is of the cell cycle occur after DNA damage, most likely considered to be an important component of the cellular for the ecient and optimal removal of DNA lesions response to a variety of genotoxins. In order to gain from the cellular genome before cell division (Grana insights on the biochemical pathways leading to p53 and Reddy, 1995; Canman et al., 1994; Guengerich, stabilization, the e€ect of (+) 7,8-dihydroxy-anti-9, 1988; Kastan et al., 1991). It is suggested that the 10-epoxy-7,8,9,10-tetrahydrobenzo(a)-pyrene [(+)-anti- programmed death of cells may be yet another mode BPDE] induced DNA damage on p53 protein levels necessary for the elimination of potential precursor was investigated in various repair-pro®cient and repair- cells escaping DNA repair (Canman et al., 1994; de®cient human cells. Brief exposure of normal human Ho€man and Liebermann, 1994). Recent reports have ®broblasts to 0.05 ± 1 mM (+)-anti-BPDE resulted in demonstrated that the p53 tumor suppressor protein elevated p53 protein levels as compared to the plays an integral role in such cellular response constitutive levels of control cells. The rapid induction pathways to DNA damage (Nelson and Kastan, response, detectable within a few hours, was sustained up 1994; Kastan et al., 1991, 1992). The rapid accumula- to a period of at least 24 h. Repair-pro®cient and repair- tion of transcriptionally active p53 tumor suppressor de®cient (XPA) human lymphoblastoid cells showed a protein upon genotoxic damage has been shown to be similar response. The poly(ADP-ribose) polymerase essential for G1-S cell cycle arrest. This occurs via the inhibitor, 3-aminobenzamide (3-AB), diminished the p53 induction of p21, a potent inhibitor of cyclin dependent induction response by concomitantly decreasing the kinases regulating the cell cycle (El-Deiry et al., 1994; extent of (+)-anti-BPDE induced DNA damage in cells Dulic et al., 1994; Nelson and Kastan, 1994; Kastan et pretreated with the inhibitor. However, the direct al., 1991, 1992). While p53 is shown to play a de®nite involvement of poly ADP-ribosylation was also apparent role in triggering apoptosis following DNA damage as 3-AB was able to attenuate (*50%) the p53 response (Kastan et al., 1992; Smith et al., 1994; El-Deiry et al., by post-damage inhibitor treatment of the cells. 1994; Price and Park, 1994; Lowe et al., 1993), cells Inhibition of cellular DNA replication by hydroxyurea de®cient in p53, have been shown to undergo apoptotic and AraC, in the presence or absence of DNA damage, cell death indicating the presence of other p53 also resulted in rapid p53 accumulation in repair- independent pathways (Nelson and Kastan, 1994; de®cient cells. On the contrary, inhibition of protein Venkatachalam et al., 1993). Besides this function, kinase C (PKC) by calphostin-C led to an abrogation of p53 can act as a transcriptional activator of repair (+)-anti-BPDE mediated p53 induction. Analysis of the related genes like gadd45 and very recent studies have downstream e€ects of carcinogen treatment showed that implicated a direct role of p53 in the excision repair the lymphoblastoid cells undergo DNA fragmentation process (Ford and Hanawalt, 1995). Thus, the indicative of apoptosis while ®broblasts exhibit cell cycle intracellular accumulation of p53 after DNA damage arrest at the G1-S boundary. mediates a wide range of interrelated processes that ensure the faithful propagation of parental genomic Keywords: DNA damage; p53 induction; apoptosis; cell sequences and the avoidance of mutations. In cycle arrest concordance with its pivotal role in guarding the genome, the p53 gene has been reported to be mutated in a majority of human cancers (Hollstein et al., 1991; Greenblatt et al., 1994). Introduction The increase in p53 levels following DNA damage of cells has been shown to result from an increase in half- Continuous exposure of cells to exogenous and life of the protein (Liu and Pelling, 1995; Abrahams et endogenous DNA damage is implicated in the al., 1995). While the biochemical and molecular pathogenesis of various cancers. A host of complex modulators responsible for the stabilization of cellular cellular DNA repair mechanisms have evolved to p53 are yet to be discerned, some studies suggest that counteract the deleterious e€ects of DNA damage in excision repair coupled formation of transitory DNA humans (Sancar, 1995). In addition to the existence of strand breaks is an important initial factor in the DNA several detoxi®cation systems, delay at speci®c stages damage response pathway (Nelson and Kastan, 1994). Furthermore, DNA strand breaks that induce p53 protein levels also are known to activate various damage dependent cellular enzymes like poly(ADP- Correspondence: AA Wani ribose) polymerase and DNA dependent protein Present Addresses: 3Division of Molecular Virology, Baylor College kinases (Nelson and Kastan, 1994; Carson et al., of Medicine, Houston, TX 77030; 4Department of Biology, Beckman Research Institute, City of Hope, Duarte, CA 1986; Stierum et al., 1994; Anderson, 1993). Indirect Received 11 June 1996; revised 18 October 1996; accepted 18 October evidence from recent studies indicates that post- 1996 translational protein modi®cations, namely poly Modulation of DNA damage induced cellular p53 protein S Venkatachalam et al 802 ADP-ribosylation and phosphorylation may be in- indicated that the response was due to the interaction volved in the increased half-life of the p53 protein of the reactive epoxide metabolite with the genome (Whitacre et al., 1995; Khanna and Lavin, 1993). leading to DNA damage. The extent of p53 accumula- Benzo(a)pyrene is a major environmental pollutant tion increased proportionally with the dose of the present in automobile exhaust, cigarette smoke, carcinogen. However the p53 induction response was various foods and industrial wastes (Phillips, 1983). saturable at all the time points from treatment with The diol-epoxide metabolite, (+)-anti-BPDE reacts at doses higher than 0.5 mM (Figure 1, top and center). several nucleophilic sites in DNA and the covalent N2- Time course experiments revealed that the tumor dG-anti-BPDE adduct comprises more than 80% of suppressor protein begins to accumulate as early as the total adduct population (Osborne et al., 1981). 2.5 h and remains at elevated levels for at least 24 h The DNA base alterations arising from exposure to after exposure (Figure 1, center). Control cells treated (+)-anti-BPDE have been implicated in mutagenesis, with vehicle alone did not indicate any change in the carcinogenesis and cellular transformation (Harris, intracellular levels of the p53 protein. Cell lysates from 1991; Conney, 1982). In the present report we have transformed HeLa cells, containing high constitutive used (+)-anti-BPDE as a model DNA damaging levels of p53, served as a positive control for speci®c agent to study the modulation of the p53 induction antibody based detection by Western blotting. The p53 response by inhibitors of poly(ADP-ribose) polymer- protein speci®c binding of antibodies was further ase and PKC. The results indicate that the down established by using extracts of p53 null HL-60 cells regulation of PKC activity leads to an abrogation of which demonstrated absence of any immuno-reactive p53 protein elevation following exposure to (+)-anti- protein bands at 53 Kd (data not shown). BPDE. Inhibition of poly(ADP-ribose) polymerase activity by 3-aminobenzamide results in an attenua- tion of p53 accumulation in various cultured human

cells exposed to carcinogenic (+)-anti-BPDE via, (a) a „—˜le I g—r™inogen dose dependent indu™tion of @CAE—ntiEf€hi

concomitant decrease in the absolute adduct levels hxe —ddu™ts in hum—n ™ells

upon pre-incubation of the cells with 3-AB and (b) a T wA eddu™tsGIH nu™leotides @m

direct decrease in the protein accumulation upon post- hose

QFUTCHFRS

damage incubation of the cells with the PARP HFHS VFUTCIFPV

inhibitor. The study also demonstrates that the HFI HFS RWFSCHFQT

IHUCIHFWP formation of DNA stand breaks might not be the IFH

only necessary factor in the p53 accumulation xorm—l lympho˜l—stoid ™ells were tre—ted with the indi™—ted doses

response triggered by (+)-anti-BPDE. Lastly the data of the ™—r™inogen for QH minD lysed —nd pro™essed for hxe

are presented to show that the downstream e€ects of isol—tionF „he represent—tive —ddu™t levels of s—mples ™orresponding

DNA damage i.e. cell cycle arrest and apoptosis occur to the experiment shown in ®gure I @˜ottomAD were qu—ntit—ted ˜y

in a cell type dependent fashion. the nonE™ompetitive immunoEslot ˜lot —ss—y —s des™ri˜ed in m—teri—ls —nd methodsF „he levels represent the me—nCsFdF of repli™—te me—surements

Results Time 0 2.5 10 Dose dependence and time course of p53 accumulation [hr] BPDE after (+)-anti-BPDE treatment µ 0.00 0.05 0.50

M 0.10 0.50 1.00 0.00 0.01 0.05 [ ] HeLa 0.00 0.01 0.05 0.10 0.50 1.00 DNA damage serves as an important trigger to activate p53 the cellular responses such as cell cycle arrest, DNA repair and apoptosis (Nelson and Kastan, 1994; BPDE Kastan et al., 1991). To establish a dose-response µ relationship between (+)-anti-BPDE induced DNA [ M] 0.0 0.05 0.5 Time damage and p53 induction, excision repair-pro®cient 0 2.5 5 10

[hr] 24 0 2.5 5 0 24 human ®broblast and lymphoblastoid cells were treated 10 24 with increasing concentrations of the carcinogen (0.05 p53 to 1.0 mM). The DNA damage levels as detected by non-competitive immuno-slot blot assays, for varying does used in these experiments, ranged from 3.76 to Time 6

107 adducts/10 nucleotides (Table 1). To correlate the [hr] 10 0 2.5 5 10 24 formation of DNA damage with p53 protein accumu- BPDE lation, whole cell extracts from treated and untreated µ 0.00 0.05 0.10 0.50 1.00 [ M] 0.50 control cells were analysed using antibodies speci®c for p53 di€erent epitopes of the tumor suppressor protein. Results with low passage normal human diploid Figure 1 Carcinogen dose and time-dependent accumulation of ®broblasts, showed a rapid and sustained accumula- p53 protein in ®broblasts (top and center) and lymphoblastoid tion of the p53 protein in response to (+)-anti-BPDE cells (bottom) after (+)-anti-BPDE treatment. Exponentially treatment. A brief treatment of the cells with the growing cells were treated with the indicated doses of the genotoxin resulted in a p53 induction response that carcinogen for 30 min and lysed at various time points. Cellular extracts (equivalents of 1.26105 cells/lane) were separated by was apparent with doses as low as 0.05 mM. Failure of SDS-polyacrylamide gel electrophoresis, transferred to nitrocellu- p53 accumulation upon treatment of cells with the lose, and visualized by enhanced chemiluminescence as described inactivated carcinogen (hydrolyzed in aqueous media) in Materials and methods Modulation of DNA damage induced cellular p53 protein S Venkatachalam et al 803 For subsequent studies to relate the e€ects of DNA a damage on cell cycle arrest and apoptosis, lympho- 3–AB Post None Pre Post blastoid cells grown in suspension cultures provided an BPDE useful cellular model. In order to establish that the p53 [µM] 0.0 0.5 0.5 0.5 induction response was similar to that of ®broblasts, Time 048048048048 lymphoblastoid cells were evaluated under similar [hr] conditons. Genotoxin treated normal and repair p53 de®cient (XPA) lymphoblastoid cells displayed an identical response in which the p53 protein levels were increased in a dose and time dependent manner (Figure 1, bottom panel and data not shown). p53

Modulation of p53 accumulation by inhibition of poly(ADP-ribose) polymerase FG It is now well documented that exposure to various physical and chemical carcinogenic DNA damaging agents cause an increase in poly(ADP-ribose) poly- b merase (PARP) activity within minutes of genotoxic insult (Whitacre et al., 1995). The synthesis of poly(ADP-ribose) polymers has been described as a necessary accompaniment for ecient DNA repair processes (Satoh and Lindahl, 1992). The periodic proximity and similarity of PARP activity induction and elevated p53 protein levels suggest that PARP may be a potential DNA damage inducible candidate involved in the post-translational modi®cation of the p53 protein conferring increased half-life (Whitacre et al., 1995). To evaluate the importance of PARP for the p53 induction response, cells with normal PARP activity but di€erent repair capacities were treated with the PARP speci®c inhibitor, 3-aminobenzamide (3-AB) before and after exposure of cells to (+)-anti- BPDE. As shown in Figure 2a (top), pre-incubation of normal lymphoblastoid cells with 3-AB led to a decrease in the overall induction of the p53 protein compared to cells treated with (+)-anti-BPDE alone. Densitometric quantitation of the ®lms, exposed to non-saturating densities, revealed a 50 ± 60% decrease in the intensity of the p53 protein bands at 4 and 8 h after DNA damage. The treatment of cells with 3-AB Figure 2 E€ect of 3-aminobenzamide (3-AB) on p53 accumula- tion. (a) Exponentially growing normal lymphoblastoid (top) and before carcinogen exposure also led to a concomitant XPA lymphoblastoid cells (center) were treated with 0.5 mm(+)- decrease in the (+)-anti-BPDE-DNA adduct levels by anti-BPDE in the presence (pre) or absence (post) of 3- *60% (Figure 2b), indicating that the p53 accumula- aminobenzamide (5 mM), resuspended in fresh media containing tion was clearly a consequence of the induced DNA 3-AB and p53 accumulation was monitored after 0, 4 and 8 h. damage. However, the treatment of the cells with the Control cells (lanes 1 ± 3) were treated with 1% ethanol and then incubated with 5 mM 3-AB (post). Fast green (FG) staining of the PARP inhibitor, after genotoxic insult also resulted in bottom portion of the protein blot from normal lymphoblastoid a decreased p53 induction response (Figure 2a). Thus cells to indicate equal protein loading is shown (bottom panel). poly ADP-ribosylation itself may be directly involved (b) E€ect of 3-aminobenzamide pre-treatment on the levels of in the pathway for the stabilization of the p53 protein. DNA damage in normal and XP lymphoblastoid cells. For each corresponding experiment, the (+)-anti-BPDE adduct levels were Since poly(ADPR) synthesis in XPA cells has been quantitated by immuno-slot blot assay as described in Materials shown to be defective due to the lowered amount of and methods. The adduct values, derived with varying amounts of excision repair mediated strand breaks after DNA each isolated DNA sample, represent the mean (+s.e.) of at least damage, repair-de®cient (XPA) lymphoblastoid cells four determinations from two separate experiments. The adduct were also treated under identical conditions (Berger et levels obtained with the cells treated with (+)-anti-BPDE alone (positive controls) were set as 100% al., 1980; Berger and Sikorski, 1981). An analogous decrease in p53 levels was noticed in these cells when 3- AB was present prior to and after (+)-anti-BPDE treatment (Figure 2a, center). In addition, pre- incubation of XP lymphoblastoid cells with 3-AB also Induction of p53 levels by inhibitors of DNA replication resulted in a concurrent decrease in the (+)-anti-BPDE DNA adduct levels (Figure 2b). The decrease in p53 While various types of DNA lesions are shown to levels was not due to a reduced amount of protein elevate the intracellular levels of p53, the initiating loaded in the gel as indicated by the fast green dye pathways responsible for this e€ect remain unclear. (FG) staining of the lower portion of a representative The role of DNA strand breaks in triggering p53 levels protein blot at *35 Kd range (Figure 2a, bottom). has been suggested but not proven unequivocally Modulation of DNA damage induced cellular p53 protein S Venkatachalam et al 804 (Nelson and Kastan, 1994; Hess et al., 1994). To observable as a result of cellular DNA damage (Figure investigate the role of replication mediated DNA 4, top and center). This failure of p53 accumulation strand breaks at the polymerase pause sites due to response by damage was observed at 4 and 8 h post- (+)-anti-BPDE-DNA adducts, repair-de®cient lympho- incubation after carcinogen treatment. Once again the blastoid cells were incubated with inhibitors of DNA decrease in the amount of the p53 protein was not due replication immediately after exposure to the carcino- to changes in the total amount of protein loaded as gen at doses of 0.125 mM and 0.5 mM. Since these cells indicated by the fast green staining of the protein blots are de®cient in the excision repair pathway, DNA (Figure 4, bottom). However, at the 8 h post treatment damage and the inhibition of replication would not time, there was a slight increase in the levels of the p53 lead to an increase in the formation of DNA strand protein in the control cells treated with calphostin-C breaks either through repair or replication (Nelson and alone (Figure 4). In parallel experiments, treatment of Kastan, 1994). The results shown in Figure 3 describe cells with a non-speci®c inhibitor of PKC, staurospor- the e€ect on p53 accumulation when the ribonucleotide ine (3 nM) did not appear to a€ect the DNA damage reductase inhibitor, hydroxyurea and the DNA induced accumulation of p53 levels (data not shown). replication inhibitor, AraC were included in the post incubation period after carcinogen exposure. Densito- Cell lineage dependent cellular response to DNA damage metric evaluation showed that the presence of ± apoptosis vs cell cycle arrest replication inhibitors in the culture medium following carcinogen exposure resulted in an enhancement (*2- Earlier reports have shown that various e€ects and the fold) of the p53 protein accumulation when compared subsequent fate of a cell in response to DNA damage, to cells treated with (+)-anti-BPDE alone, at 4 h i.e., to undergo cell cycle arrest and/or apoptosis that (Figure 3, lanes 2 and 3 vs lanes 8 and 9) as well as 8 h occur is dependent on factors like cellular histotype, (Figure 3, lanes 5 and 6 vs lanes 11 and 12). extent of genotoxic damage and funtional status of the Interestingly, in control untreated cells, block of relevant gene products i.e. p53 and bcl2 (Canman et replication with HU and AraC, stimulated a p53 al., 1994; Di Leonardo et al., 1994; Lowe et al., 1993; accumulation response that was detectable in 4 and 8 h Fan et al., 1994; Liebermann et al., 1995). To delineate as opposed to vehicle treated cells not blocked for the di€erences in the post-damage responses of distinct replication (Figure 3, lanes 1 and 7 vs lanes 4 and 10). cellular histotypes, exponentially growing ®broblast This indicates that the perturbation of DNA synthesis and lymphoblastoid cells were treated with (+)-anti- might itself be a trigger for p53 induction. In separate BPDE and analysed for any apoptosis related inter- experiments, treatment of cells with AraC alone caused nucleosomal DNA fragmentation. As shown in Figure an accumulation of p53 protein levels analogous to the 5a, agarose gel electrophoresis of cellular DNA from response seen in combination with HU (data not carcinogen treated lymphoblastoid cells showed a clear shown). DNA breakdown but with fragments ranging in size from large high molecular weight to small oligonucleo- tides (Figure 5a, lanes marked L-N). This type of E€ect of Calphostin-C on DNA damage induced p53 degradation is associated with early stages of levels programmed cell death. Time course experiments It has been recently suggested that the phosphorylation revealed that the lymphoblastoid cells undergo cell of the tumor suppressor protein might be necessary for its stabilization after DNA damage (Hupp and Lane, 1994). To investigate the role of phosphorylation on p53 stabilization/induction, ®broblasts and lympho- CalC –+ blastoid cells were pre-treated for 1.5 h with 300 nM Time 0804808048 PKC inhibitor, calphostin-C and the induction patterns [hr] of the p53 protein were monitored subsequent to (+)- BPDE anti-BPDE treatment. Pre-treatment of both the [µM] 0.0 0.0 0.5 0.5 0.5 0.0 0.0 0.5 0.5 0.5 ®broblast and lymphoblastoid cells with calphostin-C led to an abrogation of the elevated p53 response p53

AraC –––+++–––+++ p53 + HU Time [hr] 48 BPDE FG [µM] 0.0 0.1 0.5 0.0 0.1 0.5 0.0 0.1 0.5 0.0 0.1 0.5 p53 Figure 4 Abrogation of the p53 induction response by treatment of human cells with the PKC inhibitor, calphostin-C. Exponen- tially growing normal ®broblast (top) and lymphoblastoid cells Figure 3 E€ect of inhibition of DNA replication on DNA (center) were pre-incubated with 300 nM of calphostin-C for 1.5 h damage induced p53 accumulation. Repair de®cient lymphoblas- and then treated with the indicated dose of the carcinogen. Cell toid cells (XPA) were treated with the indicated doses of (+)-anti- lysates were prepared at the indicated time points and p53 protein BPDE and incubated in the presence (+) or absence (7)of was detected using a mixture of antibodies as described in hydroxyurea (2 mM) and AraC (10 mM) for 4 and 8 h and Materials and methods. The bottom panel shows the protein blot processed for p53 protein levels as described in Figure 1. Control of the lymphoblastoid cellular extracts stained with fast green cells were treated with 1% ethanol (FC) to indicate equal protein loading a

M b 0.0 S– L–N OSU–2

0.5

1.0 0.0

0.5

1.0

M ouaino N aaeidcdclua 5 protein p53 cellular Venkatachalam induced S damage DNA of Modulation bols OU2 n ypolsodcls(-)wr treated normal were growing (L-N) Exponentially cells 1.0 lymphoblastoid culture. with Lanes and in (OSU-2) methods. cells ®broblast h. gel ( and 24 human ladder. agarose kbp for Materials treated by 1 the media in are analysed M of fresh described marked and doses in as separated indicated incubated electrophoresis was the and DNA with min Fragmented 30 treated for were carcinogen Cells lymphoblastoid normal cells. and (L-N) (OSU-2) ®broblasts in fragmentation ( treatment. 5 Figure ramn ceue.Fo-yoercaayi a efre as methods and performed Materials was in analysis described Flow-cytometric schedules. treatment tie ihpoiimidd o N otn nlss h G The analysis. content DNA for iodide propidium with stained 0.4 containing media -hs,G S-phase, m M 2 Apoptosis Madaottcclsaeidctdfrtevarious the for indicated are cells apoptotic and /M a tal et ( €c f( of E€ect ) + )- anti BD o 0mnadicbtdi fresh in incubated and min 30 for -BPDE m /lncdzl o 4h h el were cells The h. 24 for nocodazole g/ml vs rwhars fe ( after arrest growth + b )- .Cl yl nlsso carcinogen of analysis cycle Cell ). anti BD ramn nDNA on treatment -BPDE + )- anti -BPDE 1 , 805 Modulation of DNA damage induced cellular p53 protein S Venkatachalam et al 806 cycle arrest before the induction of apoptosis at DNA damage with repair-pro®cient and repair- approximately 24 h (data not shown). In contrast, de®cient cells (Lu and Lane, 1993; Abrahams et al., agarose gel analysis of DNA from ®broblasts did not 1995; Kaspin and Baird, 1996). Our studies were aimed indicate any detectable inter-nucleosomal DNA frag- at elucidating the biochemical mediators responsible mentation or release of any high molecular weight for the p53 damage response. However, in contrast to DNA fragments (Figure 5a, lanes marked OSU-2). the reported ®ndings showing the requirement of DNA Morphological analysis of the lymphoblastoid cells also strand breaks for the p53 induction response in XP revealed cellular changes characteristic of apoptosis cells (Nelson and Kastan, 1994), our data indicate that while the ®broblasts exhibited only an increase in the increase in p53 levels may not be mediated solely cellular size resembling large senescent cells but no via strand breaks. The induction and super-induction apoptosis associated changes (data not shown). of p53 in XP lymphoblastoid cells treated with To substantiate the di€erences in the above replication inhibitors in the absence and presence of mentioned responses, cells were treated with the (+)-anti-BPDE-DNA damage suggest that p53 eleva- carcinogen and their transition through the cell cycle tion may result from any perturbation to the DNA was analysed by ¯ow cytometry at various post- replication machinery and/or stuctural distortions to treatment times. The mitotic inhibitor nocodazole the DNA helical structure. However, these results do (0.4 mg/ml) was added immediately to cultures after not rule out the possibility of an increase in the the genotoxin treatment to avoid confounding results amount of strand breaks at the sites of the stalled

due to cells escaping the G2/M phase (Bae et al., 1995). replication enzyme leading to an enhanced e€ect on It may be noted that, non-adherent ®broblasts were p53 accumulation. Nonetheless, previous studies have collected along with the trypsinized cells for the ¯ow shown that PARP is necessary for the ligation of DNA cytometric analysis. As shown in Figure 5b, normal strand breaks (Chatterjee and Berger, 1994). If DNA ®broblasts and lymphoblastoid cells treated with strand breaks are the only necessary trigger for the p53 ethanol vehicle alone showed a typical cell cycle elevation response, our experiments with the PARP pro®le of exponentially growing cultures with a higher inhibitor would have shown an increase in the above

fraction of the cells in the G1 phase analysed at 24 h. mentioned response in repair-pro®cient cells due to the Nocodazole treatment of these cells caused the lack of strand break ligation. These results, showing a expected shift and accumulation of control cells at decrease in the p53 accumulation response in the

the G2/M phase. However, treatment of ®broblasts to presence of the PARP inhibitor, argue against the role an initial dose of 1 mM of the carcinogen prevented the of DNA strand break formation as the only initiating

progression of cells from the G1 phase and caused cell factor in the p53 response pathway. Therefore, p53

cycle arrest at the G1/S boundary. This response was might serve as a common cellular regulatory mediator clearly demonstrated from (+)-anti-BPDE treated cells for intracellular events speci®cally related to DNA

that were prevented from escaping the G2/M phase by replication that may potentially result in heritable the presence of nocodazole in the cultures. About 82% genetic changes. This assumption is consistent with

of the original G1 population was arrested at the G1 earlier ®ndings which have indicated that DNA phase in ®broblast cells treated with 1 mM of the replication inhibitors can induce cellular p53 levels

carcinogen while the G1 fraction had decreased to 25% (Khanna and Lavin, 1993; Hess et al., 1994). in the control cells incubated with nocadazole. In Interestingly, the ability of p53 to bind structural contrast to ®broblasts, lymphoblastoid cells were able distortions in DNA has recently been demonstrated to activate programmed cell death observed at 24 h (Lee et al., 1995). Though these results were obtained following carcinogen exposure (1.0 and 0.5 mM) and in vitro, it is likely that p53 might be capable of binding apoptotic cells appeared as a distinct population with a gross structural distortions in the DNA as encountered lower DNA content (Figure 5b). Furthermore, analysis with u.v. and bulky chemical lesions within the cell. of ®broblasts treated with higher doses (up to 4 mM)of In spite of a plethora of published reports the carcinogen and isolated after longer post-treatment concerning the role of p53 in the DNA damage intervals (48 h) did not indicate any evidence of response pathway (Nelson and Kastan, 1994; Kastan apoptosis (data not shown). et al., 1991, 1992), the early steps in the signal transduction mechanism remain mostly hypothetical. Our data, with PARP and PKC inhibitors demonstrate Discussion that more than one pathway may be in place to ensure the elevation of p53 levels in response to DNA Accumulation of cellular p53 protein in response to a damage. The 50 ± 60% reduction in p53 levels in cell variety of agents that damage DNA has been incubated in the presence of the PARP inhibitor 3-AB, demonstrated in several earlier reports (Kastan et al., indicate that poly(ADP-ribosylation) of p53 might be 1991; Hess et al., 1994; Bae et al., 1995; Bjelogrlic et an important factor in stabilizing p53 levels. However, al., 1994; Hall et al., 1993). While the kinetics of the an earlier study has shown an enhancement of (+)- p53 induction response has been shown to be similar in anti-BPDE induced cellular p53 levels by 3-aminoben- repair-pro®cient and repair-de®cient cells, the biochem- zamide (Stierum et al., 1995). The analytical methods ical and molecular events that lead to the increase in and the dosage of the carcinogen (2.5 mM) used in these p53 levels have not been established clearly (Lu and studies may be a reason for these contrasting e€ects. Lane, 1993; Abrahams et al., 1995). The enhanced p53 As shown in the results, the decrease in p53 protein levels, due of stabilization of the protein, persists was detectable only at lower exposure times of the longer in repair-de®cient than repair-pro®cient cells immuno-blots that did not lead to saturating densities (Abrahams et al., 1995). Consistent with earlier studies, of the signal. Furthermore, poly ADP-ribosylation of we observed an identical mode of p53 induction after the chromatin has been implicated in various nuclear Modulation of DNA damage induced cellular p53 protein S Venkatachalam et al 807 processes like transcriptional regulation and DNA DO-7) used in this study do not discriminate between repair (Satoh and Lindahl, 1992; Lautier et al., 1993). the phosphorylated and the unphosphorylated forms of The reduced levels of carcinogen-DNA adduction in p53 (Kumar and Spandau, 1995). While the stabiliza- cells pre-incubated with 3-AB show that the dynamic tion of p53 protein and its binding to consensus nature of the chromatin can also modulate the extent sequences (after DNA damage) might be mutually of DNA damage caused by bulky carcinogens and is exclusive, earlier reports have indicated that the consistent with an earlier report (Kurian et al., 1992). binding of p53 to its consensus sequences in the Nevertheless, the lowering of p53 levels upon post- DNA leads to a change in the protein conformation damage incubation with 3-AB suggest that p53 is a that is similar to the mutant form (Halazonetis et al., potential candidate for modi®cation by poly ADP- 1993) and also confers resistance to ubiquitin mediated ribosylation. The increased half-life of poly ADP- proteolysis (Molinari and Milner, 1995). Thus, the ribosylated p53 protein would provide an inequivocal inhibition of PKC activity might in turn perturb the proof for this assumption. Another explanation for the ability of p53 to bind damaged DNA thereby relative reduction in post-damage p53 levels may be the increasing the degradation of the tumor suppressor obstruction of the DNA binding capacity of the tumor protein which leads to a decrease in cellular p53 levels. suppressor protein due to the lack of poly ADP- The cellular response to DNA damage is a complex ribosylation of histones. A recent report has documen- phenomenon that is highly dependent upon various ted the reduced p53 induction response in cells de®cient factors such as extent of damage, di€erentiation status of in PARP synthesis thus con®rming the potential role of the cell and cellular type. Indeed, our results show that poly ADP-ribosylation in this process (Whitacre et al., cell cycle arrest and apoptosis occur in a cell type 1995). Further studies are needed to address the role of dependent fashion. The prolonged cell cycle arrest PARP in stabilizing p53 levels. exhibited by ®broblasts even at higher doses indicate Our results with the PKC inhibitor, calphostin-C that these cells might be resistant to apoptotic cell death appear to indicate that phosphorylation of p53 might induced by DNA damage. This is in agreement with be the primary mechanism by which the protein is earlier reports that have shown that ®broblasts do not stabilized in the nucleus following DNA damage. undergo apoptosis after DNA damage (Canman et al., However, the ®nal conclusion about the role of 1994; Di Leonardo et al., 1994). While the lymphoblas- protein kinase C in DNA damage induced signal toid cells and ®broblast cells appear to sustain a similar transduction pathway(s) that lead(s) to p53 stabiliza- post-damage p53 induction response, the outcome of tion cannot be elucidated from the present data. these events namely, cell cycle arrest and/or apoptosis Recent data also indicate that phosphorylation might seem to depend on cell type speci®c factors. It is of be involved in the regulation of the post-damage p53 interest that the lymphoblastoid cells do not shown a response and the ability of the tumor suppressor typical oligo-nucleosomal DNA fragmentation as protein to bind its consensus sequences in the DNA encountered with treatment of HL-60 cells (Venkata- (Zhang et al., 1994; Price and Calderwood, 1993). chalam et al., 1993). The DNA fragmentation patterns Consistent with the role of p53 phosphorylation, the suggest the presence of high molecular weight DNA AT (ataxia-telangiectasia) gene product, a PI-3 kinase, fragments ranging from several kilobases to several has been suggested to be involved in the signal hundred kilobases and are characteristic of early stages transduction pathway leading to the stabilization of of programmed cell death (Walker et al., 1994). p53 levels (Savitsky et al., 1995; Khanna et al., 1995; In summary, this study shows that DNA damage Khanna and Lavin, 1993). This has been further induced by the carcinogenic metabolite, (+)-anti- substantiated by the abrogation of the sequence BPDE leads to a rapid and sustained increase in speci®c DNA binding capacity of p53 by a speci®c cellular p53 levels. The primary mode of p53 protein inhibitor of PI-3 kinase (Price and Youmell, 1996). Our stabilization might occur via the phosphorylation of results are consistent with the report showing that the tumor suppressor protein and poly(ADP-ribosyla- calphostin-C can down-regulate p53 induction after tion) might serve as secondary means of regulation. In gamma and u.v. irradiation in normal and ataxia- addition, the results demonstrate that DNA strand telangiectasia cells (Khanna and Lavin, 1993). A direct breaks are not the only necessary alterations respon- role of PKC in regulating p53 function has been sible for the p53 elevation response. The immediate demonstrated by in vitro studies in which phosphoryla- response of p53 protein stabilization accompanies the tion of the carboxy terminus of p53 leads to the downstream processes of cell cycle arrest and/or activation of the DNA binding function of the protein apoptosis in a cell lineage dependent fashion. (Hupp and Lane, 1994). However, in spite of the evidence that p53 phosphorylation is involved in the DNA damage response, the sequence of events that Materials and methods link DNA damage and the down-stream stabilization of p53 remains unclear. Since the activity of kinases Cell culture and treatment can in turn be regulated by phosphorylation, the role of PKC and PI-3 kinase in the p53 induction response Normal human ®broblasts (OSU-2) were established in may either be a direct one involving p53 phosphoryla- culture as described (Gibson-D'Ambrosio et al., 1986) and tion, or an indirect one in which one of the proteins is grown in DMEM supplemented with 10% fetal calf serum. Normal (GM02184C) and repair-de®cient (GM02250E, regulated by the other. In addition, it is important to XPA) lymphoblastoid cells were obtained from the note that the observed decrease in p53 levels after NIGMS Human Genetic Cell Repository (Camden, NJ) calphostin-C treatment is not due to the decreased and grown in RPMI 1640 medium supplemented with 10% binding of antibodies to the unphosphorylated forms fetal calf serum and antibiotics at 378C in a humidi®ed of p53 as the mixture of antibodies (clones 1801 and atmosphere of 5% CO2. Exponentially growing cells were Modulation of DNA damage induced cellular p53 protein S Venkatachalam et al 808 treated for 30 min with (+)-anti-BPDE (Midwest Research pelleted by centrifugation at 13 000 r.p.m. The precipitates Institute, NCI Repository, Kansas City, MO) by the were brie¯y rinsed with 70% ice-cold ethanol, resuspended addition of 1/100 volume of freshly prepared carcinogen in TE and processed by microdiphenylamine assay as stock in 95% ethanol after replacing the growth medium described (Venkatachalam et al., 1993). Protein extracts with Hank's balanced salt solution. The monolayers of from equivalent number of cells (1.26105 cells i.e. 875 ng ®broblast cells were washed with PBS and maintained in DNA from TCA precipitates) were separated in 8% SDS- fresh culture medium for varying post-treatment times. The polyacrylamide gels and transferred to nitrocellulose lymphoblastoid cells, grown as suspension cultures were membranes by using a semi-dry electroblotter (Hoe€er, centrifuged and cell pellets were washed once with PBS San Francisco, CA). Equal protein loading was con®rmed before resuspending in fresh culture medium at a density of from visualization of the membranes stained for 15 to 0.36106 cell/ml. Cells were harvested at various post- 20 min with fast green (0.1% fast green, 5% glacial acetic treatment times, lysed by boiling for 10 min in sample acid and 20% methanol). For p53 protein detection, the bu€er (2% sodium dodecyl sulfate, 10% glycerol, 10 mM nitrocellulose membranes were blocked with 5% milk dithiothreitol in 62 mM Tris-HCl, pH 6.8, 10 mg/ml powder in TBST bu€er (100 mM Tris, 150 mM NaCl, pepstatin and 10 mg/ml leupeptin), snap frozen in liquid 0.05% Tween, pH 7.4) for 2 h and incubated overnight nitrogen and stored at 7208C. For treatment with 3- with a mixture of anti-p53 protein antibodies at 1 : 200 aminobenzamide, cells were either pre-incubated for 1 h dilution (p53 Ab-2 and p53 Ab-6 from hybridoma clones with 5 mM 3-AB before genotoxin exposure or mock 1801 and DO-1 respectively; Neomarkers, Freemont, CA) treated with PBS. Both the pre-incubated and mock at room temperature. The membranes were further treated cells were further resuspended in fresh medium incubated with the secondary antibody reagent (goat anti- containing 5 mM 3-AB. In experiments with calphostin-C, mouse horseradish peroxidase conjugate antibodies; Boeh- cells were pre-incubated for 1.5 h with 300 nM of the PKC ringer Mannheim, Indianapolis, IN) at 1 : 2000 dilution for inhibitor before carcinogen treatment. Post-treatment times 2 h at room temperature. Following each antibody were calculated after removal of (+)-anti-BPDE by incubation the membranes were thoroughly washed with washing and zero time coincides with the start of TBST bu€er. The p53 protein bands were detected from continuous culturing in the presence of appropriate the peroxidase activity using the enhanced chemilumines- chemical modulator. All experiments were repeated at cence substrate reaction (Pierce, Rockford, IL) essentially least twice and the results of representative experiments are according to the manufacturer's instructions using X- shown in various ®gures. OMAT AR ®lm (Kodak, Rochester, NY).

Quantitation of DNA damage by immuno-slot blot assays Flow cytometric analysis of cells The (+)-anti-BPDE-DNA adduct levels were quantitated Carcinogen treated and mock treated human cells in using non-competitive immuno-slot blot assays essentially culture (1 ± 26106) were rinsed with phosphate bu€ered as described (Venkatachalam et al., 1995). Brie¯y, cells saline, resuspended in fresh cell culture media containing were lysed after treatment with the carcinogen for 30 min, 0.4 mg/ml nocodazole and incubated further for 24 h (Fan treated with RNAase-A and proteinase K and subjected to et al., 1994). Cells were then collected, ®xed in ice-cold multiple organic extractions. DNA concentration was ethanol (70%), centrifuged at 1500 r.p.m. washed with PBS estimated using the microdiphenylamine assay (Venkata- and resuspended in 0.8 ml PBS to give a ®nal concentra- chalam et al., 1993). Various amounts of modi®ed and tion of 1 ± 26106 cells per ml. To the resulting single cell unmodi®ed single-stranded cellular DNA samples were suspensions, RNase-A (0.1 ml of 0.1 mg/ml) and propi- blotted onto nitrocellulose membranes (Schleicher & dium iodide (0.1 ml of 0.4 mg/ml) were added and cells Schuell, Keene, NH) under low vacuum in BRL Con- were incubated at 378C for 30 min. Cell cycle analysis was vertible device (Life Technologies, Gaithersburg, MD). The performed using a Coulter EPICS Elite cytometer nitrocellulose membranes were baked at 808Cfor1hand (Hialeah, Fl). At least 15 000 gated cells were analysed blocked overnight at 378C with Blotto (20 mM Tris-HCl, for each time point with Multicycle software (Phoenix pH 8.0, 120 mM glycine, 5% dry milk powder, 0.5% Flow Systems, San Diego, CA) and the coecient of gelatin, 0.1% sodium azide, 0.01% antifoam-A and 0.1% variance (CV) values were below 5%. Tween-20). The ®lters were then incubated with primary antibodies (BP1, rabbit polyclonal antibodies speci®c for (+)-anti-BPDE-DNA adducts) (Venkatachalam and Wani, DNA fragmentation analysis 1994) followed by secondary antibodies (Goat anti-rabbit DNA fragmentation analysis was performed essentially as alkaline phosphatase conjugate, Boehringer Mannheim, described (Sellins and Cohen, 1987). Brie¯y, DNA from Indianapolis, IN) for 1 h at 378C. The bound enzymatic cells was isolated by cellular lysis and fragmented DNA activity was determined by color development with a was separated from native intact DNA by centrifugation. substrate solution of 0.33 mg/ml nitro blue tetrazolium The 13 000 g supernatant was treated with RNAase-A for and 0.17 mg/ml 5-bromo-4-chloro-3-indolyl phosphate 30 min at 378C followed by Proteinase K for 1 h at 458C. (Sigma) in Tris-HCl bu€er (100 mM Tris, 100 mM NaCl, Aliquots of DNA corresponding to *16106 cells were 5mMMgCl2, pH 9.5). The adduct levels were calculated by resolved by gel (0.8% agarose) electrophoresis and comparing the band intensities of the samples with (+)- visualized by ethidium bromide staining. anti-BPDE modi®ed DNA standard samples run in parallel as described (Venkatachalam et al., 1995).

Acknowledgements Immunoblot analysis of p53 protein levels We thank Mr Joseph Trask of The Ohio State University Western blot analysis was performed with aliquots from a Comprehensive Cancer Center-Analytical Cytometry La- constant number of cells obtained from the estimation of boratory (supported by P30 CA16058), for help with ¯ow DNA content in the cell lysates. For DNA content cytometry and Mr John Croyle (Department of Radiol- estimation, aliquots of the cell lysates were precipitated ogy), for help with photography. This work was supported with 5% ice cold TCA in the presence of 0.5 mgtRNAand by NIEHS grants ES2388 and ES6074. Modulation of DNA damage induced cellular p53 protein S Venkatachalam et al 809 References

Abrahams PJ, Schouten R, Van Laar T, Houweling A, Khanna KK, Beamish H, Yan J, Hobson K, Williams R, Terleth C and van der Eb AJ. (1995). Mutat. Res. DNA Dunn I and Lavin MF. (1995). Oncogene, 11, 609 ± 618. Repair, 336, 169 ± 180. Khanna KK and Lavin MF. (1993). Oncogene, 8, 3307 ± Anderson CW. (1993). Trends Biochem. Sci., 18, 433 ± 437. 3312. Bae I, Fan S, Bhatia K, Kohn KW, Fornace AJ Jr and Kumar M and Spandau DF. (1995). Biochem. Biophys. Res. O'Connor PM. (1995). Cancer Res., 55, 2387 ± 2393. Commun., 214, 744 ± 753. Berger NA, Sikorski GW, Petzold SJ and Kurohara KK. Kurian P, Kumari HL and Milo GE. (1992). Carcinogenesis, (1980). Biochemistry, 19, 289 ± 293. 13, 489 ± 491. Berger NA and Sikorski GW. (1981). Biochemstry, 20, Lautier D, Lagueux J, Thibodeau J, Menard L and Poirier 3610 ± 3614. GG. (1993). MCB, 122, 171 ± 193. Bjelogrlic NM, MaÈ kinen M, StenbaÈ ck F and VaÈ haÈ kangas K. Lee S, Elenbaas B and Levine A. (1995). Cell, 81, 1013 ± (1994). Carcinogenesis, 15, 771 ± 774. 1020. Canman CE, Chen C-Y, Lee M-H and Kastan MB. (1994). Liebermann DA, Ho€man B and Steinman RA. (1995). Cold Spring Harbor Symp. Quant. Biol., 59, 277 ± 286. Oncogene, 11, 199 ± 210. Carson DA, Seto S and Wasson DB. (1986). Exp. Cell Res., Liu M and Pelling JC. (1995). Oncogene, 10, 1955 ± 1960. 164, 273 ± 281. Lowe SW, Schmitt EM, Smith WS, Osborne BA and Jacks T. Chatterjee S and Berger NA. (1994). Mol. Cell. Biochem., (1993). Nature, 362, 847 ± 849. 138, 61 ± 69. Lu X and Lane DP. (1993). Cell, 75, 765 ± 778. Conney AH. (1982). Cancer Res., 42, 4875 ± 4917. Molinari M and Milner J. (1995). Oncogene, 10, 1849 ± 1854. Di Leonardo A, Linke SP, Clarkin K and Wahl GM. (1994). Nelson WG and Kastan MB. (1994). Mol. Cell. Biol., 14, Genes Dev., 8, 2540 ± 2551. 1815 ± 1823. Dulic V, Kaufmann WK, Wilson SJ, Tlsty TD, Lees E, Osborne MR, Jacobs S, Harvey RG and Brookes P. (1981). Harper JW, Elledge SJ and Reed SI. (1994). Cell, 76, Carcinogenesis, 2, 553 ± 558. 1013 ± 1023. Phillips DH. (1983). Nature, 303, 468 ± 472. El-Deiry WS, Harper JW, O'Connor PM, Velculescu VE, Price BD and Calderwood SK. (1993). Oncogene, 8, 3055 ± Canman CE, Jackman J, Pietenpol JA, Burrell M, Hill 3062. DE,WangY,WimanKG,et al. (1994). Cancer Res., 54, Price BD and Park SJ. (1994). Cancer Res., 54, 896 ± 899. 1169 ± 1174. Price BD and Youmell MB. (1996). Cancer Res., 56, 246 ± Fan S, El-Deiry WS, Bae I, Freeman J, Jondle D, Bhatia K, 250. Fornace AJ Jr, Magrath I, Kohn KW and O'Connor PM. Sancar A. (1995). Annu. Rev. Genet., 29, 69 ± 105. (1994). Cancer Res., 54, 5824 ± 5830. Satoh MS and Lindahl T. (1992). Nature, 356, 356 ± 358. Ford JM and Hanawalt PC. (1995). Proc. Natl. Acad. Sci. Savitsky K, Bar-Shira A, Gilad S, Rotman G, Vanagaite L, USA, 92, 8876 ± 8880. Tagle DA and Smith S. (1995). Science, 268, 1749 ± 1753. Gibson-D'Ambrosio RE, Samuel M and D'Ambrosio SM. Sellins KS and Cohen JJ. (1987). J. Immunol., 139, 3199 ± (1986). In Vitro Cell. Dev. Biol., 22, 529 ± 534. 3206. Grana X and Reddy EP. (1995). Oncogene, 11, 211 ± 219. Smith ML, Chen IT, Zhan Q, Bae I, Chen C-Y, Gilmer TM, Greenblatt MS, Bennett WP, Hollstein M and Harris CC. Kastan MB, O'Connor PM and Fornace AJ Jr. (1994). (1994). Cancer Res., 54, 4855 ± 4878. Science, 266, 1376 ± 1380. Guengerich FP. (1988). Cancer Res., 48, 2946 ± 2954. Stierum RH, Van Herwijnen MHM, Hageman GJ and Halazonetis TD, Davis LJ and Kandil AN. (1993). EMBO J., Kleinjans JCS. (1994). Carcinogenesis, 15, 745 ± 751. 12 (3), 1021 ± 1028. Stierum RH, Van Herwijnen MHM, Pasman PC, Hageman Hall PA, McKee PH, Menage HP, Dover R and Lane DP. GJ, Kleinjans JCS and Van Agen B. (1995). Carcinogen- (1993). Oncogene, 8, 203 ± 207. esis, 16, 2765 ± 2771. Harris CC. (1991). Cancer Res., 51, 5023s ± 5044s. Venkatachalam S, Denissenko MF, Alvi N and Wani AA. Hess R, Plaumann B, Lutum AS, Haessler C, Heinz B, (1993). Biochem. Biophys. Res. Commun., 197, 722 ± 729. Fritsche M and Brandner G. (1994). Toxicol. Lett., 72, Venkatachalam S, Denissenko MF and Wani AA. (1995). 43 ± 52. Carcinogenesis, 16, 2029 ± 2036. Ho€man B and Liebermann DA. (1994). Oncogene, 9, 1807 ± Venkatachalam S and Wani AA. (1994). Carcinogenesis, 15, 1812. 565 ± 572. Hollstein M, Sidransky D, Vogelstein B and Harris CC. Walker PR, Weaver VM, Lach B, Leblanc J and Sikorska M. (1991). Science, 253, 49 ± 53. (1994). Exp. Cell Res., 213, 100 ± 106. Hupp TR and Lane DP. (1994). Cold Spring Harbor Whitacre CM, Hashimoto H, Tsai ML, Chatterjee S, Berger Symposia on Quantitative Biology, LIX, 195 ± 206. SJ and Berger NA. (1995). Cancer Res., 55, 3697 ± 3701. Kaspin LC and Baird WM. (1996). PAH, (in press). Zhang W, McClain C, Gau J, Guo XD and Deisseroth AB. Kastan MB, Onyekwere O, Sidransky D, Vogelstein B and (1994). Cancer Res., 54, 4448 ± 4453. Craig RW. (1991). Cancer Res., 51, 6304 ± 6311. Kastan MB, Zhan Q, El-Deiry WS, Carrier F, Jacks T, Walsh WV, Plunkett BS, Vogelstein B and Fornace AJ. (1992). Cell, 71, 587 ± 597.