Synergistic activation of p53-dependent transcription by two ... - Nature

6 downloads 75 Views 271KB Size Report
1998; Waterman et al., 1998; Sakaguchi et al., 1998) ..... 5 Gy g-irradiation followed by increasing concentrations of (0 to ..... and Wynford-Thomas D. (1999).
Oncogene (2000) 19, 3829 ± 3839 ã 2000 Macmillan Publishers Ltd All rights reserved 0950 ± 9232/00 $15.00 www.nature.com/onc

Synergistic activation of p53-dependent transcription by two cooperating damage recognition pathways Jeremy P Blaydes1, Ashley L Craig1, Maura Wallace2, H Mei-Ling Ball1, Nicola J Traynor3, Neil K Gibbs3 and Ted R Hupp*,1 1

Department of Molecular and Cellular Pathology, The Cancer Research Campaign Laboratories, Dundee Cancer Research Centre, Ninewells Medical School, University of Dundee, Dundee, Scotland DD1 9SY; 2Department of Surgery and Molecular Oncology, The Cancer Research Campaign Laboratories, Dundee Cancer Research Centre, Ninewells Medical School, University of Dundee, Dundee, Scotland DD1 9SY; 3Department of Photobiology, The Cancer Research Campaign Laboratories, Dundee Cancer Research Centre, Ninewells Medical School, University of Dundee, Dundee, Scotland DD1 9SY

High level activation of p53-dependent transcription occurs following cellular exposure to genotoxic damaging agents such as UV-C, while ionizing radiation damage does not induce a similarly potent induction of p53dependent gene expression. Reasoning that one of the major di€erences between UV-C and ionizing radiation damage is that the latter does not inhibit general transcription, we attempted to reconstitute p53-dependent gene expression in ionizing irradiated cells by co-treatment with selected transcription inhibitors that alone do not activate p53. p53-dependent transcription can be dramatically enhanced by the treatment of ionizing irradiated cells with low doses of DRB, which on its own does not induce p53 activity. The mechanism of ionizing radiationdependent activation of p53-dependent transcription using DRB is more likely due to inhibition of gene transcription rather than prolonged DNA damage, as the non-genotoxic and general transcription inhibitor Roscovitine also synergistically activates p53 function in ionizing irradiated cells. These results identify two distinct signal transduction pathways that cooperate to fully activate p53-dependent gene expression: one responding to lesions induced by ionizing radiation and the second being a kinase pathway that regulates general RNA Polymerase II activity. Oncogene (2000) 19, 3829 ± 3839. Keywords: p53; cdk; ionizing radiation; DRB; Roscovitine Introduction The tumour suppressor p53 is a stress-inducible transcription factor which regulates the expression of genes implicated in growth control or apoptosis (Levine, 1997). The types of agents that activate p53 function include ionizing and non-ionizing radiation, ribonucleotide depletion and certain chemotherapeutic drugs (Giaccia and Kastan, 1998). The damageinducible nature of the p53 pathway is thought to be one mechanism whereby anti-cancer drugs or therapeutic levels of radiation induce death in tumour cells with a wild-type p53 pathway, whereas tumours null for p53 or harbouring mutant alleles may resist damage induced apoptosis (Wallace-Brodeur and

*Correspondence: TR Hupp Received 15 December 1999; revised 3 July 2000; accepted 4 July 2000

Lowe, 1999). The signalling pathways which regulate p53 protein in vivo are beginning to be de®ned and regulate both its anity for DNA consensus elements in the promoters of target genes (Blaydes and Hupp, 1998; Waterman et al., 1998; Sakaguchi et al., 1998) and also its interactions with other cellular proteins including the p14/19ARF/mdm2-dependent ubiquitination machinery (Tao and Levine, 1999; Kurokawa et al., 1999; Honda and Yasuda, 1999), the p300/CBP family of transcriptional activators (Sakaguchi et al., 1998; Yuan et al., 1999a,b) and TFIID and related components of the transcriptional machinery (Thut et al., 1995; Pise-Masison et al., 1998). These fundamental interactions of p53 are thought to be modulated, at least in part, by post-translational covalent modi®cations of p53 itself, including acetylation and phosphorylation (Giaccia and Kastan, 1998). Current models hold that DNA damage induces a concerted mechanism of p53-activation as a transcription factor, in part, by the observations that phosphorylation at two key highly conserved regulatory domains increases after cell damage: (i) DNA-PK and/or ATM/ATR-dependent phosphorylation of p53 at Ser15 (Shieh et al., 1997) may simultaneously reduce mdm2 protein binding and stimulate p300/CBP binding (Lambert et al., 1998) and (ii) increased phosphorylation at Ser392 (CK2/PKR site) may convert p53 protein from a latent to an activated state for sequence-speci®c DNA binding (Hupp et al., 1995) and p53-dependent transcription (Blaydes and Hupp, 1998; Kapoor and Lozano, 1998; Lu et al., 1998). These models provide a framework for further testing, and for example, independent evidence for posttranslational regulation of p53 is compelling based on the following data: (1) mutation of Ser15 on full-length p53 can alter p53 protein levels or its apoptotic activity under speci®c cellular conditions (Unger et al., 1999) and (2) mutation of the C-terminal CK2/PKR site a€ects p53 function when cells are grown to high density (Hao et al., 1996). Many other sites of posttranslational modi®cation are now known to be clustered within these regulatory domains, with a similar paradigm being supported: phosphorylation of p53 within the N-terminal BOXI domain at Thr18 or Ser20 can a€ect heterologous protein ± protein interactions with p53 (Craig et al., 1999b; Shieh et al., 2000) and C-terminal acetylation sites (Lys382/Lys320) (Sakaguchi et al., 1998; Abraham et al., 2000), phosphorylation sites (Ser371/Ser376/Ser378) (Waterman et al., 1998), and a sumolation site (Lys386) can

Modification of the p53 response JP Blaydes et al

3830

Oncogene

regulate p53 activity as a DNA binding protein (Gotissa et al., 1999; Rodriguez et al., 1999). The phosphorylation sites within the N-terminal regulatory domain of p53 exhibit distinct types of regulation, depending upon the context or damaging agent. Normal human ®broblasts constitutively modify p53 at the Ser20 site and oxidant stresses can result in hypo-phosphorylation at this site (Bond et al., 1999; Craig et al., 1999a), while the ionizing radiationinduced form of p53 protein is phosphorylated at Ser20 by a Chk2-dependent pathway (Chehab et al., 2000; Shieh et al., 2000; Hirao et al., 2000). The Thr18 site is modi®ed in human breast cancers (Craig et al., 1999a), induced during senescence (Webley et al., 2000) or transiently following ionizing radiation (Sakaguchi et al., 2000). More strikingly, Ser15 phosphorylation increases during quiescence, senescence, UV irradiation or ionizing irradiation, thus identifying a diverse signal recognition pathway that responds to stress in normal human cells (Webley et al., 2000). Presumably, this cluster of phosphorylation sites that di€erentially a€ects MDM2 and/or p300 binding will modulate the speci®c activity of p53 as a transcription factor and will be important in regulating the rates of p53-dependent tumour suppression in a cell-speci®c manner. Cterminal modi®cation of p53 in vivo by monoclonal antibodies that bind near the Ser392 site stimulates p53-dependent transcription and/or apoptosis (Hupp and Lane, 1995; Abarzua et al., 1995; Caron de Fromentel et al., 1999). These latter data suggest that C-terminal modi®cation of p53 can be a rate-limiting step in modulating p53 function and highlights the need to determine whether increased C-terminal phosphorylation of p53 at Ser392 does in fact occur after damage and if it correlates with p53 activation as a transcription factor. One enzyme known to target Ser392 is CK2 which is a ubiquitous, and apparently constitutively active, protein kinase responsible for the phosphorylation of Ser and Thr residues speci®ed by acidic side chains in many proteins, including several key enzymes, growth factor receptors, transcription factors and cytoskeletal proteins (Allende and Allende, 1995). A second enzyme, reported recently to phosphorylate p53 at Ser392, is double-stranded RNAactivated protein kinase PKR (Cuddihy et al., 1999). This enzyme is a putative tumour suppressor and is a key component of the interferon regulatory pathway that promotes inhibition of protein synthesis via phosphorylation of eIF-2. The nature of the cellular `lesion' that activates the p53 response to damage is not well-de®ned, mainly due to the diversity in the types of agents that can activate p53. These upstream signal transduction pathways that target p53 within the N-terminal and C-terminal regulatory domains have been mapped out mostly using cells exposed to `DNA damaging' agents like ionizing radiation. The models invoking DNA damage as a signalling lesion to p53 activation have been driven by the observations that p53 is a component of the ATM-radiation-a€ected signal transduction pathway, whose genetic defects in double-stranded DNA repair provide lesions that can be sensed by the upstream components of the p53 pathway (Lavin et al., 1999). However, genetic defects in p53 induction in cell lines containing the XP-A, but not the XP-C, complementation group mutation are defective in

transcription-coupled repair in response to UV irradiation (Yamaizumi and Sugano, 1994; Ljungman and Zhang, 1996). These data suggest that inhibition of RNA Polymerase II-gene expression, rather than DNA damage per se, can signal to the p53 pathway. In addition, apoptotic cell death in post-mitotic neurons induced by exposure to either kainic acid or N-methylD-aspartate occurs where p53 protein induction precedes DNA fragmentation, suggesting that DNA damage is a consequence rather than a cause of p53 induction (Sakhi et al., 1994, 1997; Hughes et al., 1997). Cellular damaging agents as diverse as heat shock (Nitta et al., 1997), hypoxia (Graeber et al., 1996), perturbation of nucleotide metabolism (Linke et al., 1996), low extracellular pH (Williams et al., 1999), and TGF-b (Bellamy et al., 1997a) can also activate p53 function, thus suggesting either that the signalling lesions that target p53 protein are quite diverse and/or that strikingly di€erent damaging agents function through identical signalling lesions. This study addresses the types of lesions generated by a genotoxic agent leading to the observation that two distinct damage recognition pathways cooperate to fully-activate p53. The genotoxic agents UV-C radiation or cisplatin typically activate a relatively high level of p53-dependent gene expression, while ionizing radiation induces a form of p53 with a much lower speci®c activity (Bond et al., 1996; Bellamy et al., 1997b). These data suggest that ionizing radiation fails to induce a lesion which is more potently promoted by UV-C radiation. The mechanism of UV-C-induced activation of p53 is dissected by attempting to reconstitute p53-dependent gene expression in cells damaged with ionizing radiation, which on its own does not strongly activate p53 transactivation function. We describe that the p53-dependent transcriptional response to ionizing radiation can be reconstituted with RNA Polymerase II transcriptional inhibitors DRB or Roscovitine, which on their own do not activate p53 function. These results identify two distinct lesions which signal to the p53 pathway by synergistically promoting activation of p53 as a transcription factor in vivo. The ®rst being a pathway that responds to the oxidative damage induced by ionizing radiation and the second compromising DRB-sensitive or Roscovitine-sensitive protein kinases whose inhibition blocks RNA Polymerase II transcription thus facilitating full activation of p53-dependent gene expression.

Results Differential activation of p53-dependent gene expression by non-ionizing and ionizing radiation The microinjection or intracellular synthesis of a monoclonal antibody to the C-terminal domain of p53 (Hupp and Lane, 1995; Abarzua et al., 1995; Caron de Fromentel et al., 1999) or microinjection of antibodies to MDM2 protein can alone induce p53dependent gene expression in the absence of exogenous DNA damaging agents (Blaydes et al., 1997; Blaydes and Wynford-Thomas, 1998; Bottger et al., 1997). These data indicate that C-terminal modi®cation of p53 or inhibition of MDM2 activity are two ratelimiting steps for activating the p53 pathway and that

Modification of the p53 response JP Blaydes et al

DNA damage is not absolutely necessary to elicit a p53 response. However, it is generally held that DNA damage is a trigger for activating the p53 response, based on the observations that: (1) the ionizing radiation-responsive ATM pathway is required for p53 protein induction (Kastan et al., 1992; Siliciano et al., 1997); (2) the microinjection of damaged DNA templates (Huang et al., 1996) or (3) the introduction of restriction enzymes into cells can induce p53 activity (Lu and Lane, 1993). Nevertheless, these latter data do not address whether the speci®c activity of p53 as a transcription factor is relatively high and therefore whether DNA damage alone can completely assemble the factors needed to maximally coordinate a p53 response to cell injury. UV-C and X-irradiation are two distinct agents known to promote DNA damage to a cell and to activate p53-dependent gene expression. UV-C is a type of non-ionizing radiation that induces oxidative stress as well as bulky-adduct DNA damage along with activation of an XP-dependent transcription-coupled nucleotide excision DNA repair pathway. Ionizing radiation induces the generation of reactive oxygen intermediates, protein damage, lipid peroxidation, the production of single-stranded and double-stranded DNA breaks and activation of an ATM-dependent checkpoint-repair pathway. The speci®c activity of p53 as a transcription factor from a stably-integrated bgalactosidase gene containing a consensus p53 binding site in its promoter region was analysed after cellular exposure to UV-C and X-irradiation to begin to dissect quantitatively the signal recognition pathways that can modulate p53 function. We have been using primarily the A375 human melanoma cell line as a model system to dissect upstream signalling events that activate p53 protein post-translationally in response to UV-C irradiation. These studies have shown that: (1) decreased Ser20-site phosphorylation (unmasking of the DO-1 epitope); (2) decreased Ser371/376-site phosphorylation (unmasking of the PAb421 epitope) and (3) increased Ser392-site phosphorylation of p53 (Craig et al., 1999a) correlates with increased p53dependent gene expression after low doses of UV-C irradiation (Blaydes and Hupp, 1998). As normal human diploid ®broblasts similarly promote Ser20-site dephosphorylation and elevated Ser392 site phosphorylation without changes in p53 protein levels after oxidant stresses including UV-C radiation (Bond et al., 1999; Craig et al., 1999a; Webley et al., 2000), the A375 cells are a good model to dissect the signalling to p53 that occurs in normal human ®broblasts. A striking di€erence in the level of p53-dependent gene expression was observed 10 and 24 h postirradiation, depending upon the type of damaging agent utilized. UV-C irradiation gave rise to relatively high levels of p53-dependent gene expression 10 or 24 h post-irradiation at 10, 20 or 40 Jm72 (Figure 1a). In contrast, ionizing radiation gave rise to a relatively weak induction of p53-dependent gene expression (Figure 1a). Previous reports had also indicated that the ionizing radiation or the ionizing radiation-mimetic bleomycin gave rise to lowered induction of p53 activity compared to genotoxic agents including UVC radiation in normal ®broblasts and hepatocytes (Bond et al., 1996; Bellamy et al., 1997b), consistent with the data shown in Figure 1.

a

3831

b

Figure 1 Di€erential a€ects of UV-C and ionizing radiation on p53-dependent activity. (a) p53 activity. Arn8 melanoma cells [A375 parental cell line containing a stably integrated p53dependent reporter construct] were exposed to: (i) altering concentrations of UV-C radiation (40, 20 and 10 J/m2; ®lled circle, open triangle, and ®lled square, respectively) and (ii) altering concentrations of X-irradiation (5 Gy and 2.5 Gy; open circle and open square, respectively). Ten or 24 h post-radiation, the cells were pelleted and examined for changes in p53 dependent transcription of the b-galactosidase reporter gene by enzyme assay. The data represent the b-galactosidase in O.D. as indicated in the Materials and methods. (b) Levels of p53 protein. Arn8 cells were exposed to (i) altering concentrations of UV-C radiation (40 to 0 J/m2, from left, lanes 1 ± 4) and (ii) altering concentrations of X-irradiation (5 Gy, 2.5 Gy, 0, from left, lanes 5 ± 7). Ten hours (top panel) or 24 h (bottom panel) postradiation, the cells were pelleted, lysed, and examined for changes in p53 protein levels by immunoblotting with DO-12. The * indicates the position of p53 protein

The lowered speci®c activity of p53 protein in ionizing irradiated cells cannot be explained based on reduced levels of p53 protein induced by this agent. Immunoblotting of lysates from damaged cells harvested at 10 h post-damage (Figure 1b, top panel) or 24 h post-damage (Figure 1b, bottom panel) indicated that p53 protein levels at the highest level of damaging agent are similar (2.5 ± 5 Gy of ionizing radiation, 10 ± 40 Jm72 of UV-C). Together these data suggest the possibility that ionizing radiation is a weak inducer of p53 activity from some promoters because it fails to impinge upon a signalling pathway a€ected by the genotoxic agent UV-C and not because a defect in p53 protein accumulation. In addition, these data demonstrate that increases in p53 protein levels are not sucient to induce p53-dependent gene expression, consistent with the hypotheses that post-translational modi®cation is an important stage in p53 activation. We therefore took advantage of these quantitative di€erences by attempting to reconstitute full p53 Oncogene

Modification of the p53 response JP Blaydes et al

3832

activity in an ionizing irradiated cell and thereby begin to dissect signal recognition pathways that give rise to complete p53 activation. One of the major di€erences between the damages induced by UV-C and ionizing radiation is that the latter fails to inhibit general RNA polymerase II dependent transcription (Ljungman et al., 1999) and ionizing radiation fails to induce Ser392site phosphorylation (PKR/CK2-site phosphorylation) in the C-terminal regulatory domain of p53 (Kapoor and Lozano, 1998; Lu et al., 1998). We therefore investigated whether Ser392-site phosphorylation of p53 was defective in ionizing irradiated cells and/or whether inhibition of general RNA Polymerase II transcription could reconstitute p53 activity in an ionizing irradiated cell.

Table 1

De®nition of the FPS392 epitope on p53 by peptide-library phage-display and species speci®city

Phage display ± clone A ± clone B ± clone C ± consensus P53 ± human ± rat ± mouse

Sequence

No. of times selected

FHENWPS FHESPSPAGGR FHSDWPGQTFTW FH ± WP KLMFKTEGPDSD KPMIKKVGPDSD KTMVKKVGPDSD

5 2 1 N/A N/A N/A N/A

Ionizing radiation can induce Ser392-phosphorylation of human p53 The phosphorylation site at Ser392 lies within the Cterminal negative regulatory domain of p53, modi®cation of which is a key rate limiting step for the activation of p53 in vivo (Hupp and Lane, 1995; Abarzua et al., 1995; Caron de Fromentel et al., 1999). Although classic mutagenesis-coupled transfection approaches have often failed to prove a role for Ser392 phosphorylation in activating p53, mutation of the Ser residue to Asp constitutively activates p53 in contact inhibited ®broblasts (Hao et al., 1996). Phosphorylation at Ser392 in vitro enhances the sequence-speci®c DNA binding activity of p53 directly and recent studies have determined that steady-state levels of Ser392 phosphorylation are elevated when p53 function is activated in vivo in response to UV-C irradiation (Blaydes and Hupp, 1998), which is consistent with an activating role for this phosphorylation. Recent reports using a phospho-speci®c polyclonal antibody to the PKR/CK2-site have shown that p53 is phosphorylated in response to UV-C, but not ionizing radiation (Kapoor and Lozano, 1998; Lu et al., 1998). We thus generated a phospho-speci®c monoclonal antibody as a reagent to determine whether the lowered speci®c activity of p53 in ionizing irradiated cells (Figure 1) was due to the fact that the signal recognition pathway that targets Ser392 does not respond to ionizing radiation. It is the experience of this laboratory that polyclonal phospho-speci®c reagents are not reliable due to the natural variability of phospho-speci®c IgG within such a polyclonal antibody preparation (Craig et al., 1999b). As a result, we have placed e€orts in making phospho-speci®c monoclonal antibodies to phosphorylation sites including Ser15, Thr18, Ser20, Ser315, and Ser392 of human p53 that are well-characterized and monospeci®c (Blaydes et al., 2000). Phospho-Ser392 containing peptides were used as antigens to develop hybridomas producing IgG with speci®city for the Phospho-Ser392 epitope of fulllength human p53 and the monoclonal antibody developed, named FPS392, was characterized by six distinct assays: peptide-phage display (Table 1), denaturing immunoblot (Figure 2), ELISA, immunoprecipitation, EMSA, and immunohistochemical cell staining (data not shown). Phage-peptide display was ®rst used to de®ne the FPS392 consensus epitope. ELISA wells were coated Oncogene

Figure 2 Monoclonal antibody FPS392 is speci®c for Ser392phosphorylated p53 protein. Full-length human p53 protein puri®ed from a bacterial expression system was incubated in the absence or presence of recombinant human CK2 to phosphorylate p53 at Ser392. Following phosphorylation, the speci®city of the monoclonal antibody FPS392 was analysed by denaturing immunoblots. Unmodi®ed p53 (left lane; 20 ng) and Ser392phosphorylated p53 (right lane; 20 ng) were immunoblotted and membranes were probed with the monoclonal antibody DO-12 speci®c for the core domain of p53 (top panel) or with the phospho-speci®c monoclonal antibody FPS392 (bottom panel)

with puri®ed FPS392 and used to select bacteriophage from libraries containing random peptides inserted within the phage III coat protein. After two cycles of ampli®cation, antibody-speci®c peptide-phage were identi®ed by ELISA and the inserts sequenced. Although two di€erent libraries were screened (a cyclic 7-mer and a linear 12-mer), only the 12-mer peptide library gave rise to peptide-phage populations which express epitopes with detectable anity to FPS392. A general peptide consensus of Phe-His-x-x-Trp-Pro was observed (Table 1) which exhibited homology to the Cterminal region of human p53 protein that was used as the original antigen (Phe385-x-x-x-x-Pro390). Since this C-terminal region of p53 protein is partially divergent amongst di€erent species (Table 1) the anity of FPS392 for murine or rat p53 protein was examined. Recombinant p53 protein was expressed in insect cells to ensure relatively extensive phosphorylation at the PKR/CK2-site and denaturing immunoblots were used to demonstrate that FPS392 recognizes p53 protein from both these species (data not shown). This demonstrates that FPS392 can also be used to study PKR/CK2-site signalling to p53 protein in species other than human and suggests that the conserved amino acids F-x-x-x-GPDSPO3D are critical in the recognition of denatured antigen by this monoclonal antibody. Immunohistochemical blotting was also used to demonstrate that FPS392 binds speci®cally to p53 protein in denaturing immunoblots only after in vitro Ser392-phosphorylation of p53 by CK2 (Figure 2, right lane vs left lane). As a control to demonstrate that the same levels of p53 protein were processed in each

Modification of the p53 response JP Blaydes et al

reaction the products of the kinase reactions were also immunoblotted with the monoclonal antibody speci®c for the core domain of p53 (DO-12) (Figure 2, right lane vs left lane). In addition, FPS392 can be used in clinical material to de®ne the levels of Ser392-site phosphorylation in formalin-®xed sections when p53 is denatured (data not shown) and will be discussed elsewhere. Lysates from normal cells were also analysed for changes in steady-state phosphorylation at Ser392 of endogenous p53 after exposure to ionizing radiation. The cell types used are those which have previously been shown to activate p53-dependent gene expression in response to ionizing radiation, namely primary human ®broblasts (Kastan et al., 1991) and cells from the spleen and thymus (Clarke et al., 1993). Normal human diploid ®broblasts exposed to ionizing radiation showed increased steady-state phosphorylation at Ser392 (probed with FPS392) relative to p53 protein levels (probed with DO-12) 1 h post-irradiation (Figure 3a, top panel vs bottom panel), with extended times post-irradiation showing higher p53 protein levels and elevated Ser392 phosphorylation. In addition, animals exposed to whole body ionizing radiation where p53 protein levels are elevated 40 min after exposure in both spleen (Figure 3b, top CM5 panel) and thymus (Figure 3b, bottom CM5 panel), also show an increase in CK2-site phosphorylation of p53 (Figure 3b, ap53Pser392 panel), with more pronounced increases in steady-state phosphorylation, relative to p53 protein levels, between 40 and 120 min in the thymus. Induction of Ser392 phosphorylation of p53 in normal lymphocytes and ®broblasts established that this signalling pathway can respond to ionizing radiation in a physiologically relevant cell type and further prompted examining whether two human tumour cell models we commonly use to dissect p53 activation mechanisms responded to ionizing radiation by inducing Ser392 phosphorylation. MCF7 breast cancer cells have been widely used to study regulation of wild-type p53 protein stabilization and activity. Exposure of MCF7 cells to ionizing radiation (Figure 3c; MCF7 panel; 3 h time point) or non-ionizing UV-C radiation (Figure 3c; MCF7 panel; 3 and 5 h time points) induced p53 protein accumulation to similar extents (as de®ned with the monoclonal antibody DO-12; Figure 3c, MCF7 panel, top row). Under these conditions, the p53 protein induced both ionizing and non-ionizing radiation was phosphorylated at Ser392 to similar extents (as de®ned with the monoclonal antibody FPS392; Figure 3c, MCF7 panel, bottom row), indicating that both types of irradiation damage signal to the Ser392-site in this cell line. The damage induced by ionizing radiation can trigger a general increase in p53 protein levels in the A375 cell line (de®ned with the antibody DO-12; Figure 3c, A375 panel, top row) and the induced p53 protein remains phosphorylated at Ser392 (de®ned with the antibody FPS392; Figure 3c, A375 panel, bottom row). Together, these data indicate that the Ser392 signalling pathway functions in ionizing irradiated cells and indicate that a global defect in the signal recognition pathway that a€ects Ser392-site phosphorylation cannot explain the reduced speci®c activity of p53 in an ionizing irradiated cell.

a

3833

b

c

Figure 3 Ionizing radiation induces phosphorylation of p53 at Ser392 in human cell lines. (a) Ser392 phosphorylation in human ®broblasts. Normal human diploid ®broblasts were left unirradiated (from left, lane 1), exposed to 5 Gy ionizing radiation and harvested 1 or 3 h post irradiation (from left, lanes 2 and 3, respectively), or exposed to 20 Jm72 UV-C radiation (from left, lane 4) and cell lysates were blotted with the indicated monoclonal antibody: top row (DO-12) and bottom row (FPS392). (b) CK2-site phosphorylation in vivo. Wistar rats were exposed to whole body ionizing irradiation (5 Gy) and sacri®ced at the indicated time points of 40, 80 or 120 min after irradiation. Lysates from the spleen and thymus were processed for immunoblotting and were probed to determine: (1) total p53 protein levels (left panels; using the CM5 polyclonal serum) and (2) the levels of CK2-phosphorylated p53 protein (right panel; ap53Pser392). Rabbit polyclonal antibodies speci®c for CK2phosphorylated p53 protein were used in this experiment, instead of FPS392 monoclonal antibody, to avoid problems with secondary antibody reactivity with rodent immunoglobulins. (c) Ser392 phosphorylation in human tumour cells. The human tumour cell line A375 was unirradiated (from left, lane 1) or harvested 1 and 3 h after exposure to 5 Gy of ionizing radiation (from left, lanes 2 and 3, respectively). MCF-7 cells were left unirradiated (from left, lanes 4 and 7), harvested 3 and 5 h after exposure to 20 J/m2 of UV-C radiation (from left, lanes 5 and 6, respectively), or harvested 1 and 3 h after exposure to 5 Gy of ionizing radiation (from left, lanes 8 and 9, respectively). The total p53 protein levels were measured with DO-12 (top panel) and Ser392-phosphorylation was measured with FPS392 (bottom panel)

DRB activates p53's trans-activation function after ionizing radiation damage One of the major di€erences between the damages induced by UV-C irradiation and by ionizing radiation is that the latter fails to inhibit general RNA Polymerase II-dependent transcription (Ljungman et al., 1999). Previous reports have suggested that inhibition of RNA Polymerase II transcription by treatment of cells with high doses of the transcription Oncogene

Modification of the p53 response JP Blaydes et al

3834

Oncogene

inhibitor DRB can be an important factor in the activation of p53 (Ljungman and Zhang, 1996). A close correlation also exists between the inhibition of RNA Polymerase II transcription by either DRB or aamanatin and the induction of apoptosis in tumour cell lines (Koumenis and Giaccia, 1997). We thus reasoned that inhibition of RNA Polymerase IIdependent transcription by a non-genotoxic kinase inhibitor like DRB that alone does not a€ect p53 function, may re-activate p53-dependent gene expression in ionizing irradiated cells. To determine whether the transcription inhibitor DRB could enhance p53 activity as a transcription factor, b-galactosidase reporter gene activity was monitored before and after ionizing irradiation without and with increasing concentrations of DRB (Figure 4a). Quantitation of the p53-dependent gene expression shows that ionizing radiation induces 0.008 Units of p53 activity in this assay (Table 2). Additionally, quantitation of the p53-dependent gene expression induced shows that DRB alone induces 0.007 Units of p53 activity in this assay (Table 2). However cotreatment with DRB at concentrations from 10 ± 20 mM maximally enhanced p53-dependent reporter activity in ionizing irradiated cells (Figure 4a). Quantitation of the b-galactosidase shows that ionizing radiation induces 0.149 Units of p53 activity in this assay, or approximately 17 ± 19-fold above controls (Table 2). These data suggest that inhibition of RNA Polymerase II transcription may in fact be an important signal recognition pathway that fully activates p53 in cooperation with a pathway that responds to the lesions induced by ionizing radiation. p53 protein induction and Ser392 phosphorylation were also followed in parallel to p53 activity. As reported previously (Blaydes and Hupp, 1998), DRB at concentrations of up to 20 mM had no e€ects on p53 protein levels in unirradiated cells (Figure 4b, DO-12 panel; from left, lanes 1 ± 5). Phosphorylation at Ser392 was also una€ected by DRB concentrations up to 10 mM, with a small degree of induction being detected by the FPS392 antibody at 20 mM DRB (Figure 4b, FPS392 panel; from left, lanes 1 ± 5). It is notable that higher DRB concentrations (40 mM) that resulted in extensive phosphorylation at Ser392, as well as p53 protein induction (Figure 4b; from left, lane 6), did not give rise to signi®cant p53 activity. These data indicate that even substantial induction of p53 protein levels and phosphorylation by a non-genotoxic agent like DRB can produce a transcriptionally-incompetent form of p53 and that inhibition of RNA Polymerase II activity by DRB alone is not sucient to trigger a potent p53-response. In contrast, when g-irradiated cells were exposed to levels of DRB that alone has no detectable a€ect, the levels of p53 protein induction and Ser392 phosphorylation were signi®cantly enhanced in a concentration dependent manner (Figure 4b; from left, lane 7 vs lanes 8 ± 12). Low concentrations of DRB (2.5 ± 5 mM) in irradiated cells resulted in an enhancement of Ser392 phosphorylation without a€ecting p53 protein stabilization (Figure 4b; from left, lane 7 vs lanes 8 and 9). Higher doses of DRB (10 mM to 20 mM), in the presence of ionizing radiation, resulted in an increase in p53 protein levels and concomitant elevated Ser392 phosphorylation (Figure 4b; from left, lane 7 vs lanes 10 and 11). It is notable

that DRB alone (at 40 mM) or DRB treatment (10 mM) of irradiated cells gave rise to similar levels of p53 protein and Ser392-site phosphorylation (Figure 4b, a

b

Figure 4 E€ects of the protein kinase inhibitor, DRB, on the activation of p53 protein by ionizing irradiation. (a) p53 activity. Arn8 melanoma cells [A375 parental cell line containing a stably integrated p53-dependent reporter construct] were exposed to: increasing concentrations of DRB (®lled squares; from 0 to 40 mM) or 5 Gy g-irradiation followed by increasing concentrations of DRB (open diamonds; from 0 to 40 mM). Nine hours later, the cells were pelleted and examined for changes in p53 dependent transcription of the b-galactosidase reporter gene by enzyme assay. The data represent the b-galactosidase as indicated in the Materials and methods. The raw data for p53 activity after exposure to ionizing radiation and/or DRB (20 mM) are summarized in Table 2. (b) Levels of p53 protein and Ser392site phosphorylation. Arn8 cells were exposed to: increasing concentrations of DRB alone (0 to 40 mM, from left, lanes 1 ± 6); 5 Gy g-irradiation followed by increasing concentrations of (0 to 40 mM, from left, lanes 7 ± 12); or 10 Jm72 UV-C irradiation (from left, lanes 13 and 14, without or with DRB, respectively). Nine hours later, the cells were pelleted and examined for changes in p53 protein levels with DO-12 (top panel) and Ser392-site phosphorylation with FPS392 (bottom panel)

Table 2 Quantitation of p53-dependent transcription by combined treatment of cells with ionizing radiation and DRB Baseline b gal level in cells b gal level after X-ray Induction by X-rays b gal level after 20 mM DRB Induction by DRB b gal level after DRB+X-rays Induction by DRB+X-rays Relative increase of DRB+X compared to X alone Relative induction of DRB+X compared to DRB alone

(A) (B) (C=B7A) (D) (E=D7A) (F) (G=F7A)

0.044 0.053 0.008 0.052 0.007 0.193 0.149

(H=G/C)

17.5

(I=G/E)

19.9

The 17.5-fold increase in p53-dependent activity refers to the e€ect of DRB+ionizing irradiation compared to the e€ects of either treatment alone. The numbers are the mean of duplicate values for b-galactosidase activity (A=450 nm) with background (enzyme assay w/o cell extract ± performed in duplicate) subtracted

Modification of the p53 response JP Blaydes et al

3835

a

Figure 5 The e€ect of DRB on the extent of repair of girradiation induced DNA damage. Cells were irradiated (5 Gy) and either placed on ice immediately, or refed witih medium with or without 20 mM DRB and incubated for a further 30 min at 378C as indicated in the Materials and methods. Cell pellets were then assayed by the COMET assay to quantify the extent of single and double strands breaks. The data are represented as %Tail DNA (% DNA breaks) in: (a) untreated cells; (b) irradiated cells incubated for 30 min at 08C; (c) irradiated cells incubated for 30 min at 378C; and (d) irradiated cells incubated for 30 min at 378C in media containing DRB

b

from left, lane 6 vs lane 10), whereas only the cell subjected to combined treatment gave rise to signi®cant p53 activity (Figure 4a, Table 2). Thus, the combined treatment appears to a€ect two distinct signalling pathways that cooperate to activate p53 function. DRB does not prolong global DNA damage Whilst the p53-activating e€ects of DRB have been attributed to its e€ects on RNA polymerase IIdependent transcription, DRB-sensitive enzymes are involved in many cellular processes and we thus considered alternative mechanisms which could account for the observed hyperstimulation of the radiation-induced p53 response. The inhibition of DNA repair can also enhance the radiation-induced induction of p53 by prolonging the presence of DNA strand breaks (Lu and Lane, 1993) and we considered the possibility that DRB may be acting through this mechanism. Cyclin H-cdk7, in addition to its role in transcription, is also a component of TFIIH which plays a role in transcription-coupled DNA repair, and the DRB-sensitive enzyme CK2 has been implicated in activating the function of DNA Ligase (Prigent et al., 1992) which is required for re-joining of repaired DNA. Furthermore, a previous report has shown that DRB increases mutagenesis in hamster cells exposed to UV irradiation (Stone-Wol€ and Rossman, 1981). Thus, the inhibition of cyclin H-cdk7 and/or CK2 by DRB may reduce the rate of DNA repair and/or DNA joining in irradiated cells, leading to the `stabilization' of an ionizing radiation-dependent damage signal that potentiates and prolongs p53 protein activation. To determine whether general DNA repair after ionizing radiation is prolonged in A375 cells treated with DRB, a COMET assay was used to measure relative rates of DNA re-ligation after ionizing radiation induced DNA strand breaks. Compared to irradiated control cells, the presence of DRB at levels that hyperinduce bgalactosidase activity (Figure 4), there is little detectable di€erence in the extent of DNA re-ligation 30 min after DNA damage (Figure 5). These data suggest that

Figure 6 E€ects of the protein kinase inhibitor, Roscovitine, on the activation of p53 protein by ionizing irradiation. (a) Roscovitine can sensitize cells to activation of p53 by ionizing radiation. Arn8 melanoma cells [A375 parental cell line containing a stably integrated p53-dependent reporter construct] were, as indicated; (i) left untreated; (ii) exposed to X-rays (5 Gy); (iii) treated with either DRB or Roscovitine (ROS) alone; and (iv) treated with a combination X-rays and either DRB or Roscovitine (ROS). Nine hours later, the cells were pelleted and examined for changes in p53 dependent transcription of the bgalactosidase reporter gene by enzyme assay. (b) Immunoblot of p53 protein levels and the extent of Ser392 phosphorylation of p53 with increasing concentrations of Roscovitine in the presence of the indicated damaging agent. Arn8 cells were exposed to: increasing concentrations of Roscovitine alone (0 to 20 mM, from left, lanes 1 ± 4); 5 Gy X-irradiation followed by increasing concentrations of (0 to 20 mM, from left, lanes 5 ± 8); or 10 Jm72 UV-C irradiation (0 to 20 mM; from left, lanes 9 ± 12). Nine hours later, the cells were pelleted and examined for changes in p53 protein levels with DO-12 (top panel), changes in Ser392site phosphorylation with FPS392 (middle panel), and changes in p21WAF1 protein levels (bottom panel)

it is not the attenuation in the repair of a damagedDNA intermediate, but it may be the low-level inhibition of the RNA Polymerase-II dependent transcription via inhibition of the DRB-sensitive protein kinases that leads to hyper-activation of p53 function in the presence of ionizing radiation. Inhibition of cyclin-dependent kinases is sufficient to sensitize p53-dependent transcription to ionizing radiation We wished to more precisely de®ne the DRB-sensitive enzymes which account for its e€ects on p53-dependent Oncogene

Modification of the p53 response JP Blaydes et al

3836

Oncogene

transcription. The highly selective cdk inhibitor, Roscovitine, has recently been shown to inhibit RNA polymerase II-dependent transcription at a similar concentration to DRB (Sankrithi and Eskin, 1999) and share in common the ability to inhibit cyclin Hcdk7 activity (Yankulov et al., 1995; Hajduch et al., 1999). However, Roscovitine does not inhibit the DRB-sensitive enzymes CK2 or CK1. Additionally, a recent report indicates that Roscovitine at high levels is a non-genotoxic inducer of p53 protein levels and induces nucleolar fragmentation (David-Pfeuty, 1999). Treatment of the p53-responsive reporter-cell line with Roscovitine had very similar e€ects on ionizing radiation dependent-transcription as DRB, causing hyperstimulation of p53 activity in an irradiated cell at a drug concentration of 10 mM (Figure 6a). As controls, Roscovitine alone (ROS), DRB alone, or Xrays alone gave rise to very low levels of p53-dependent reporter activity (Figure 6a). Analysis of Ser392 phosphorylation and p53 protein levels, however, shows similarities and di€erences between the e€ects of Roscovitine (10 mM) and DRB (10 ± 20 mM) on ionizing-radiation-induced perturbations of p53. First, similar to DRB, Roscovitine alone at concentrations up to 10 mM did not a€ect p53 protein levels or the extent of Ser392 phosphorylation (Figure 6b, from left, lanes 2 and 3 vs lane 1; both panels). However, the phosphorylation at Ser392 induced by combined treatment of ionizing radiation and Roscovitine (10 mM) is no di€erent than that observed in irradiated cells only (Figure 6b, from left, lane 7 vs lane 5; bottom panel). These data indicate that the stimulation of ionizing radiation-induced p53-dependent transcription by 10 mM Roscovitine is not accompanied by a signi®cant increase in Ser392 phosphorylation or changes in p53 protein levels (Figure 6b, from left, lane 6 and 7 vs lane 5, top panel), providing further evidence that Ser392 phosphorylation and p53 activation can be uncoupled. However, at the highest level of Roscovitine (20 mM) that alone induces p53 protein Figure 6b, from left, lane 4 vs lane 1, top panel) which is phosphorylated at Ser392 (Figure 6b, from left, lane 4 vs lane 1, bottom panel), a dramatic hyperstimulation of Ser392-site phosphorylation is observed in irradiated cells (Figure 6b, from left, lane 8 vs lane 5; bottom panel). These data together indicate that the extent of Ser392 phosphorylation can be synergistically stimulated by combined treatment of cells with DRB/ ionizing radiation (Figure 4) or Roscovitine/ionizing radiation (Figure 6), providing a correlation between p53 phosphorylation and p53 activity. However, the drug-titrations clearly show an uncoupling of phosphorylation from p53 activity. In the case of DRB, the drug alone (40 mM) can induce high levels of p53 protein which is phosphorylated at Ser392, but this form of p53 is transcriptionally inactive. In the case of Roscovitine, concentrations of the drug that hyperstimulate p53 activity in irradiated cells (10 mM) do not alter p53 protein levels or Ser392 phosphorylation. Thus, our conclusions regarding the synergistic e€ects of ionizing radiation and kinase inhibitors on p53 activation exclude a mechanism whereby protein levels or its modi®cation at Ser392 play a direct role.

Discussion The activation of a p53-dependent transcriptional response plays a de®nitive role in determining the fate of cells subjected to a wide range of stressinducing stimuli and abnormal growth states. Amongst the most clearly de®ned of these stimuli are the cellular changes associated with malignant transformation and immortalization (Bond et al., 1996), as well as responses to the genotoxic damage induced by the majority of currently available anticancer modalities (Bunz et al., 1999). Therefore a thorough understanding of the mechanisms which regulate the p53 response will both increase our understanding of the processes of malignant transformation and facilitate the development of more e€ective methods of inducing tumour cell death. Ionizing radiation remains one of the most widespread treatments used for cancer and, at least in certain tumour types, its ecacy is due, in part, to the induction of a p53-dependent apoptotic response (Wallace-Brodeur and Lowe, 1999). In this study we have investigated the signal transduction pathways whereby ionizing radiation induces a p53-dependent transcriptional response and have subsequently identi®ed a means of enhancing this response using the protein kinase and general transcription inhibitors, DRB and Roscovitine. The use of ionizing radiation as a therapeutic agent has been recognized for almost a century and is widely used in the treatment of human cancers, but a critical balance must be achieved when using radiation as form of anti-cancer treatment to ensure tumour cell death with minimal side-e€ects to normal tissue and organ function. Biochemical pathways which govern the di€erential survival or repair of normal cells and tissues exposed to ionizing radiation in vivo are only beginning to be de®ned. p53 protein can activate an ionizing radiation-dependent apoptotic pathway in tissue known to be acutely sensitive to radiation in vivo (Macleod et al., 1995). In addition to p53 and other components of the DNA-damage repair pathway including scid and PARP (Biedermann et al., 1991; Wesierska-Gadek et al., 1999), two factors implicated in oxidative stress-responses and which regulate the rates of irradiation injury include Superoxide Dismutase and Catalase (Hardmeier et al., 1997). Ionizing radiation induces damage not only to DNA but to lipids and proteins in part from the generation of highly reactive oxygen intermediates. Cells respond to the production of reactive oxygen molecules by inducing anti-oxidant enzymes that scavenge reactive oxygen species and minimize chemical damage. That radiation-resistant strains of mice rapidly induce after ionizing irradiation both Superoxide Dismutase and Catalase activities are consistent with these models. Thus, it remains unclear whether the primary lesions from ionizing radiation exposure that result in complete p53 protein stabilization (Figure 1b), but which fail to fully activate p53dependent transcription (Figure 1a) are from reactive oxygen species/oxidant stress, from direct DNA damage, or from combinations of both. What is more clear is that ionizing radiation does induce one component of the p53 pathway that can lead to full p53 activation in cells exposed to kinase inhibitors

Modification of the p53 response JP Blaydes et al

that a€ect RNA Polymerase II-dependent transcription. One of the major di€erences between the damages induced by UV-C and ionizing radiation is that the latter strikingly fails to inhibit general RNA polymerase II dependent transcription (Ljungman et al., 1999). Kinase inhibitors that can inhibit RNA Polymerase II-dependent gene expression (Hajduch et al., 1999) can cooperate with ionizing radiation to activate the p53 response (Figures 4 and 6). An increasing body of evidence identi®es this transcriptional inhibition as an important factor in the activation of p53 by UV-C. The UV-C induced p53 response is enhanced in XP-A and Cockaynes Syndrome cells in which transcription-coupled repair is speci®cally defective and the recovery of RNA synthesis following UV-C exposure is delayed (Yamaizumi and Sugano, 1994) (Ljungman and Zhang, 1996), and, when used at high concentrations, inhibitors of RNA polymerase II-dependent transcription such as DRB, H7 and a-amanitin are e€ective inducers of p53 protein accumulation (Andera and Wasylyk, 1997; Chang et al., 1999; Ljungman et al., 1999) (Figure 4). We therefore reasoned that combined treatment of cells with ionizing radiation and low concentrations of transcriptional inhibitors would synergize to induce e€ective stimulation of p53dependent transcription and indeed this proved to be the case. Thus, the partial inhibition of RNA synthesis by 20 mM DRB, which produces a similar level of RNA synthesis inhibition as 10 Jm72 UV-C (Ljungman et al., 1999), or Roscovitine, a recently identi®ed inhibitor of transcription (Sankrithi and Eskin, 1999) and a more selective inhibitor of cdks including cyclin H-cdk7 (Hajduch et al., 1999), synergizes with the ionizing radiation-induced ATM-dependent signal transduction pathway to hyper-induce the p53response. Relatively high concentrations of the RNA polymerase II inhibitor, DRB, has been demonstrated to induce apoptosis in cell lines by p53 dependent (te Poele et al., 1999) and p53-independent (Koumenis and Giaccia, 1997) mechanisms, and has thus been proposed as a potential therapeutic agent. Our studies of the signal transduction pathways which regulate p53 have identi®ed an additional potential use for signi®cantly lower concentrations of DRB, or more speci®c inhibitors such as Roscovitine, as extremely e€ective sensitizers of the p53 response to therapeutic doses of ionizing radiation. One aspect of our current work is centred on studying the DRB-sensitive and Roscovitin-sensitive kinases whose inactivation promotes ionizing radiation-hyperactivation of p53 protein, as these enzymes may provide good candidates for modi®ers of the ionizing radiation response through p53-dependent mechanisms. Further understanding the mechanisms whereby ionizing radiation activates p53 will not only help to de®ne upstream signalling cascades that play a role in the radiation response, but will hopefully assist in developing strategies to minimize normal cell injury while promoting maximal tumour-speci®c killing upon irradiation. These data highlight the therapeutic potential for the use of speci®c non-genotoxic kinase inhibitors as modi®ers of the p53 response to ionizing radiation in cancers with a wild-type p53 pathway (Figure 7).

3837

Figure 7 A model for the synergistic activation of p53dependent transcription by oxidative damage and inhibition of RNA polymerase II-dependent transcription. Ionizing radiation induces signals including Ser392 phosphorylation which on their own do not fully activate p53-dependent transcription but do stabilize p53 protein. In addition, DRB and Roscovitine at low concentrations do not fully-activate p53-dependent gene expression, indicating that blocking the kinases that a€ect RNA Polymerase II-dependent transcription does not alone maximally a€ect p53. When ionizing irradiated cells are co-treated with low concentrations of cdk inhibitors such as DRB and Roscovitine, signals induced by the inhibition of RNA polymerase IIdependent transcription cooperate with radiation-induced DNA damage to reconstitute the p53 response. The ionizing radiation signalling pathway presumably involves the ATM-Chk2 damage recognition pathway, while the RNA Polymerase II inhibition pathway presumably involves kinases such as cyclin H-cdk7 that can regulate RNA Polymerase II gene transcription

Materials and methods Reagents, enzymes and proteins Anti-p53 antibodies DO-1, DO-12, PAb421, CM5 and ap53-Pser392 have been described previously (Blaydes and Hupp, 1998). Expression of full-length human p53 in E. coli, puri®cation by Heparin-Sepharose chromatography, expression of human p53 in Sf9 insect cells and in vitro phosphorylation of E. coli expressed human p53 by human CK2 was performed as described previously (Blaydes and Hupp, 1998). For the generation of the monoclonal antibody FPS392, mice were immunized with a Keyhole Limpet Haemocyanin-conjugated phospho-peptide, SRHKKLMFKTEGPDSPO3D, corresponding to amino acids 378 ± 393 of human p53 (synthesized by Dr G Bloomberg, University of Bristol) and IgG were generated with speci®city for the phospho-C-terminal peptide according to previously established procedures (Craig et al., 1999b) and screened as described in Figure 2. Definition of the FPS392 epitope Peptide-library phage display was used to de®ne the FPS392 epitope. The monoclonal antibody FPS392 was anity puri®ed from hybridoma supernatant using a Protein ASepharose matrix (Pharmacia). ELISA wells were coated with puri®ed FPS392 (1 mg/ml in 0.1 M sodium borate bu€er [pH 9.0]) and used to select bacteriophage from libraries containing random 12-mer peptides inserted within the phage III coat protein (New England Biolabs) according to the manufacturer's protocol. Brie¯y, after two cycles of ampli®cation, antibody-speci®c phage were identi®ed by ELISA and the inserts sequenced. A general peptide consensus of PheHis-x-x-Trp-Pro was observed which exhibited homology to the C-terminal region of human p53 protein which was used as the original immunogen (Phe385-x-x-x-x-Pro390; see Table 1). Human cell lines, culture conditions, cell lysis, and COMET assay A375 malignant melanoma cells and the subclone Arn8 [containing a stably transfected p53-dependent reporter construct] have been described previously (Blaydes and Hupp, 1998). The MCF-7 breast cancer cell line was obtained Oncogene

Modification of the p53 response JP Blaydes et al

3838

from the ECACC. The HCA2 strain of normal human ®broblast were from Dr James Smith (Houston). All cells were cultured in Dulbecco's Modi®ed Eagle's Medium supplemented with 10% foetal calf serum (GIBCO/BRL). For UV-C irradiation cells were ®rst washed with Hank's balanced salt solution (GIBCO/BRL) and then irradiated in the absence of medium using a model 2400 Stratalinker (Stratagene), before being refed with fresh medium. Whole body g-irradiation of Wistar rats was 5 Gy at 1.25 Gy/minute and cell irradiation was performed by Dr B Vojtesek (Masaryk Institute for Cancer Research, Brno, Czech Republic) using a TORREX 150D source (Astrophysics research group, Long Beach, CA, USA) at a rate of 5 Gy/ minute and lysates harvested using denaturing urea bu€er (6.4 M urea, 0.1 M dithiothreitol, 0.1% Triton, 25 mM NaCl, 20 mM HEPES pH 7.6) were the kind gift of Dr B Vojtesek. DRB (Calbiochem/Novabiochem) was stored for up to 1 month at 7208C as a 10 mM stock in ethanol and added to the medium as indicated. ELISA analysis was performed as described (Blaydes and Hupp, 1998). For immunoblotting, cells were scraped o€ the dishes on ice-cold PBS, pelleted by centrifugation and snap frozen. Frozen pellets were lysed for 15 min at 48C in denaturing urea bu€er and lysates clari®ed

by centrifugation at 13 000 g for 10 min. Protein concentration was determined by the method of Bradford (Biorad) and aliquots stored at 7708C until required. For the determination of b-galactosidase activity, cells were lysed by repeated rounds of freeze ± thaw, and lysates assayed as previously described (Blaydes et al., 1997). The COMET assay for the detection of single- and double-stranded breaks was performed as described in (Reavy et al., 1997) except that the COMET tails were scored using image analysis software (COMET 2.2, kinetic imaging, Liverpool, UK).

Abbreviations DRB, 5,6-dichloro-1-beta-D-ribofuranosylbenzimidazole; UV, ultraviolet radiation.

Acknowledegments This work was supported by the UK Medical Research Council, the Cancer Research Campaign, Tenovus-Scotland, and by Moravian Biotechnologies.

References Abarzua P, LoSardo JE, Gubler ML and Neri A. (1995). Cancer Res., 55, 3490 ± 3494. Abraham J, Kelly J, Thibault P and Benchimol S. (2000). J. Mol. Biol., 295, 853 ± 864. Allende JE and Allende CC. (1995). FASEB J., 9, 313 ± 323. Andera L and Wasylyk B. (1997). Mol. Med., 3, 852 ± 863. Bellamy CO, Clarke AR, Wyllie AH and Harrison DJ. (1997a). FASEB J., 11, 591 ± 599. Bellamy CO, Prost S, Wyllie AH and Harrison DJ. (1997b). J. Pathol., 183, 177 ± 181. Biedermann KA, Sun JR, Giaccia AJ, Tosto LM and Brown JM. (1991). Proc. Natl. Acad. Sci. USA, 88, 1394 ± 1397. Blaydes J, Vojtesek B, Bloomberg G and Hupp T. (2000). Stress Responses: Methods and Protocols, Keyes S (ed). Humana Press. Blaydes JP, Gire V, Rowson JM and Wynford-Thomas D. (1997). Oncogene, 14, 1859 ± 1868. Blaydes JP and Hupp TR. (1998). Oncogene, 17, 1045 ± 1052. Blaydes JP and Wynford-Thomas D. (1998). Oncogene, 16, 3317 ± 3322. Bond J, Haughton M, Blaydes J, Gire V, Wynford-Thomas D and Wyllie F. (1996). Oncogene, 13, 2097 ± 2104. Bond JA, Webley K, Wyllie FS, Jones CJ, Craig A, Hupp T and Wynford-Thomas D. (1999). Oncogene, 18, 3788 ± 3792. Bottger A, Bottger V, Sparks A, Liu WL, Howard SF and Lane DP. (1997). Curr. Biol., 7, 860 ± 869. Bunz F, Hwang PM, Torrance C, Waldman T, Zhang Y, Dillehay L, Williams J, Lengauer C, Kinzler KW and Vogelstein B. (1999). J. Clin. Invest., 104, 263 ± 269. Caron de Fromentel C, Gruel N, Venot C, Debussche L, Conseiller E, Dureuil C, Teillaud JL, Tocque B and Bracco L. (1999). Oncogene, 18, 551 ± 557. Chang D, Chen F, Zhang F, McKay BC and Ljungman M. (1999). Cell Growth Di€er., 10, 155 ± 162. Chehab NH, Malikzay A, Appel M and Halazonetis TD. (2000). Genes Dev., 14, 278 ± 288. Clarke AR, Purdie CA, Harrison DJ, Morris RG, Bird CC, Hooper ML and Wyllie, AH. (1993). Nature, 362, 849 ± 852. Craig AL, Blaydes JP, Burch LR, Thompson AM and Hupp TR. (1999a). Oncogene, 18, 6305 ± 6312. Craig AL, Burch L, Vojtesek B, Mikutowska J, Thompson A and Hupp TR. (1999b). Biochem. J., 342, 133 ± 141.

Oncogene

Cuddihy AR, Wong AH, Tam NW, Li S and Koromilas AE. (1999). Oncogene, 18, 2690 ± 2702. David-Pfeuty T. (1999). Oncogene, 18, 7409 ± 7422. Giaccia AJ and Kastan MB. (1998). Genes Dev., 12, 2973 ± 2983. Gostissa M, Hengstermann A, Fogal V, Sandy P, Schwarz SE, Sche€ner M and Del Sal G. (1999). EMBO J., 18, 6462 ± 6471. Graeber TG, Osmanian C, Jacks T, Housman DE, Koch CJ, Lowe SW and Giaccia AJ. (1996). Nature, 379, 88 ± 91. Hajduch M, Havlieek L, Vesely J, Novotny R, Mihal V and Strnad M. (1999). Adv. Exp. Med. Biol., 457, 341 ± 353. Hao M, Lowy AM, Kapoor M, Dee A, Liu G and Lozano G. (1996). J. Biol. Chem., 271, 29380 ± 29385. Hardmeier R, Hoeger H, Fang-Kircher S, Khoschsorur A and Lubec G. (1997). Proc. Natl. Acad. Sci. USA, 94, 7572 ± 7576. Hirao A, Kong YY, Matsuoka S, Wakeham A, Ruland J, Yoshida H, Liu D, Elledge SJ and Mak TW. (2000). Science, 287, 1824 ± 1827. Honda R and Yasuda H. (1999). EMBO J., 18, 22 ± 27. Huang LC, Clarkin KC and Wahl GM. (1996). Proc. Natl. Acad. Sci. USA, 93, 4827 ± 4832. Hughes PE, Alexi T and Schreiber SS. (1997). Neuroreport, 8, v ± xii. Hupp TR and Lane DP. (1995). J. Biol. Chem., 270, 18165 ± 18174. Hupp TR, Sparks A and Lane DP. (1995). Cell, 83, 237 ± 245. Kapoor M and Lozano G. (1998). Proc. Natl. Acad. Sci. USA, 95, 2834 ± 2837. Kastan MB, Onyekwere O, Sidransky D, Vogelstein B and 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 Jr AJ. (1992). Cell, 71, 587 ± 597. Koumenis C and Giaccia A. (1997). Mol. Cell. Biol., 17, 7306 ± 7316. Kurokawa K, Tanaka T and Kato J. (1999). Oncogene, 18, 2718 ± 2727. Lambert PF, Kashanchi F, Radonovich MF, Shiekhattar R and Brady JN. (1998). J. Biol. Chem., 273, 33048 ± 33053. Lavin MF, Concannon P and Gatti RA. (1999). Cancer Res., 59, 3845 ± 3849. Levine AJ. (1997). Cell, 88, 323 ± 331.

Modification of the p53 response JP Blaydes et al

Linke SP, Clarkin KC, Di Leonardo A, Tsou A and Wahl GM. (1996). Genes Dev., 10, 934 ± 947. Ljungman M and Zhang F. (1996). Oncogene, 13, 823 ± 831. Ljungman M, Zhang F, Chen F, Rainbow AJ and McKay BC. (1999). Oncogene, 18, 583 ± 592. Lu H, Taya Y, Ikeda M and Levine AJ. (1998). Proc. Natl. Acad. Sci. USA, 95, 6399 ± 6402. Lu X and Lane DP. (1993). Cell, 75, 765 ± 778. Macleod KF, Sherry N, Hannon G, Beach D, Tokino T, Kinzler K, Vogelstein B and Jacks T. (1995). Genes Dev., 9, 935 ± 944. Nitta M, Okamura H, Aizawa S and Yamaizumi M. (1997). Oncogene, 15, 561 ± 568. Pise-Masison CA, Radonovich M, Sakaguchi K, Appella E and Brady JN. (1998). J. Virol., 72, 6348 ± 6355. Prigent C, Lasko DD, Kodama K, Woodgett JR and Lindahl T. (1992). EMBO J., 11, 2925 ± 2933. Reavy HJ, Traynor NJ and Gibbs NK. (1997). Photochem. Photobiol., 66, 368 ± 373. Rodriguez MS, Desterro JM, Lain S, Midgley CA, Lane DP and Hay RT. (1999). EMBO J., 18, 6455 ± 6461. Sakaguchi K, Herrera JE, Saito S, Miki T, Bustin M, Vassilev A, Anderson CW and Appella E. (1998). Genes Dev., 12, 2831 ± 2841. Sakaguchi K, Saito S, Higashimoto Y, Roy S, Anderson CW and Appella E. (2000). J. Biol. Chem., 275, 9278 ± 9283. Sakhi S, Bruce A, Sun N, Tocco G, Baudry M and Schreiber SS. (1994). Proc. Natl. Acad. Sci. USA, 91, 7525 ± 7529. Sakhi S, Bruce A, Sun N, Tocco G, Baudry M and Schreiber S. (1997). Exp. Neurol., 145, 81 ± 88. Sankrithi N and Eskin A. (1999). J. Neurochem., 72, 605 ± 613. Shieh SY, Ahn J, Tamai K, Taya Y and Prives C. (2000). Genes Dev., 14, 289 ± 300. Shieh SY, Ikeda M, Taya Y and Prives C. (1997). Cell, 91, 325 ± 334.

Siliciano JD, Canman CE, Taya Y, Sakaguchi K, Appella E and Kastan MB. (1997). Genes Dev., 11, 3471 ± 3481. Stone-Wol€ DS and Rossman TG. (1981). Mutat. Res., 82, 147 ± 157. Tao W and Levine AJ. (1999). Proc. Natl. Acad. Sci. USA, 96, 6937 ± 6941. te Poele RH, Okorokov AL and Joel SP. (1999). Oncogene, 18, 5765 ± 5772. Thut CJ, Chen JL, Klemm R and Tjian R. (1995). Science, 267, 100 ± 104. Unger T, Sionov RV, Moallem E, Yee CL, Howley PM, Oren M and Haupt Y. (1999). Oncogene, 18, 3205 ± 3212. Wallace-Brodeur RR and Lowe SW. (1999). Cell. Mol. Life Sci., 55, 64 ± 75. Waterman MJ, Stavridi ES, Waterman JL and Halazonetis TD. (1998). Nat. Genet., 19, 175 ± 178. Webley K, Bond JA, Jones CJ, Blaydes JP, Craig A, Hupp T and Wynford-Thomas D. (2000). Mol. Cell. Biol., 20, 2803 ± 2808. Wesierska-Gadek J, Wang ZQ and Schmid G. (1999). Cancer Res., 59, 28 ± 34. Williams AC, Collard TJ and Paraskeva C. (1999). Oncogene, 18, 3199 ± 3204. Yamaizumi M and Sugano T. (1994). Oncogene, 9, 2775 ± 2784. Yankulov K, Yamashita K, Roy R, Egly JM and Bentley DL. (1995). J. Biol. Chem., 270, 23922 ± 23925. Yuan ZM, Huang Y, Ishiko T, Nakada S, Utsugisawa T, Shioya H, Utsugisawa Y, Shi Y, Weichselbaum R and Kufe D. (1999a). Oncogene, 18, 5714 ± 5717. Yuan ZM, Huang Y, Ishiko T, Nakada S, Utsugisawa T, Shioya H, Utsugisawa Y, Yokoyama K, Weichselbaum R, Shi Y and Kufe D. (1999b). J. Biol. Chem., 274, 1883 ± 1886.

3839

Oncogene