¹Unidad Mixta de Investigaciones Médicas, Hospital Universitario San Cecilio, and Departamento de Radiología, Facultad de Medicina, Universidad de Granada, 18012-Granada, Spain

²Instituto de Parasitología y Biomedicina, CSIC, 18001-Granada, Spain

³UPR 9003 du CNRS, Ecole Supérieure de Biotechnologie, Université Louis Pasteur de Strasbourg, Boulevard Sébastien Brant, 67400 Illkirch, France

Correspondence to: F J Oliver, Unidad Mixta de Investigaciones Médicas, Hospital Universitario San Cecilio, Avda. Dr. Olóriz 16, 18012 Granada, Spain; E-mail: joliver@hsc.sas.cica.es

^aM T Valenzuela and R Guerrero contributed equally to this study

The tumour supressor protein p53 plays a key role in the cell's decision to arrest the cell cycle or undergo apoptosis following a genotoxic insult. p53 is stabilized and activated after DNA damage, however the cascade of events signalling from DNA lesions to p53 stabilization and activation is still controversial. Poly (ADP-ribosylation) of different nuclear acceptors by PARP-1 is an early event when a single strand DNA lesion is produced. We present here evidences that interplay between PARP-1 and p53 is dependent on the type of damage induced to DNA. Primary mouse embryonic fibroblasts derived from parp-1-/- mice exhibited decreased p53 accumulation and activation following -irradiation compared to parp-1 proficient cells. On the other hand, treatment with the single alkylating agent 2'-methyl-2'-nitrose-urea (MNU), resulted in the rapid and sustained accumulation and activation of p53 in parp-1-deficient cells, while very little accumulation was observed in parp-1+/+ cells. After IR, the turnover of the p53 inhibitory protein MDM-2 is perturbed and the level of phosphorylation of p53 at serine-15 is blunted in parp-1-/- cells. PARP-1 is determinant in the cytotoxic response to alkylating agents but only partially contributes to radiation-induced cell killing, as determined by colony forming assay. Altogether, these results suggest that PARP-1 participates in the p53 response following irradiation, resides upstream of p53 and indirectly modulates the level of phosphorylation of key substrates in this pathway while treatment with MNU results in an enhanced p53-mediated response in parp-1-null cells.

Oncogene (2002) 21, 1108-1116 DOI: 10.1038/sj/onc/1205169

PARP-1; p53-mediated DNA damage response; ionising radiation; alkylating agent

Introduction

The presence of damaged DNA in the cell activates repair mechanisms as well as signal transduction pathways leading to cell cycle arrest and programmed cell death. Several of these responses are mediated by the stabilization and activation of the tumour-suppressor protein p53 and failure of this activation leads to genetic instability and a predisposition to cancer (Lakin and Jackson, 1999). It is therefore crucial to understand the mechanisms that signal for the presence of lesions in DNA with the activation of p53. DNA-strand breaks generated by different genotoxic agents activate the nuclear/DNA binding protein poly (ADP-ribose) polymerase-1 (PARP-1, formerly PARP) (de Murcia and Ménissier-de-Murcia, 1994; D'Amours et al., 1999), allowing the post-translational modification by poly (ADP-ribosylation) of different nuclear protein substrates. A number of studies have suggested a role for PARP-1 and/or the poly (ADP-ribose) in p53-mediated DNA damage, although the nature and the consequence of this interaction are very controversial (reviewed in Oliver et al., 1999a).

PARP-1 is associated in vivo with XRCC1, a DNA repair protein involved, together with DNA polymerase and DNA ligase III, in the base excision repair of DNA (Masson et al., 1998). Treatment of parp-/- mice with either alkylating agents or -irradiation (IR) reveals an extreme sensitivity and a high genomic instability to both agents. Following whole body -irradiation, mutant mice died rapidly from acute radiation toxicity of the small intestine (Ménissier-de-Murcia et al., 1997). Cells derived from these mice display retarded kinetics of DNA end-rejoining following damage with an alkylating agent (Beneke et al., 2000; Trucco et al., 1998), indicating that PARP is involved in base excision repair, and more precisely in the long patch pathway, probably by recruiting DNA repair enzymes to the vicinity of a DNA lesion (de Murcia and Ménissier-de-Murcia, 1994; Dantzer et al., 1999). Genetic deletion of parp-1 attenuates tissue injury after transient cerebral ischemia (Eliasson et al., 1997), streptozotocin-induced diabetes (Burkart et al., 1999; Masutani et al., 1999), and improves the adverse clinical effects in different pathologies associated with inflammation (Oliver et al., 1999b; Szabo and Dawson, 1998).

The aim of the present study is to shed light on the interaction between PARP-1 and p53 in the cell's response to genetic damage produced by agents that differ (i) in the type of lesion induced to DNA and (ii) in the signal pathway connecting DNA lesion with the activation of p53. p53 is tightly controlled through a complex series of events including translational regulation, interaction with regulatory proteins such as the murine-double-minute-2 (MDM-2), and post-translational modifications including multisite phosphorylation and acethylation (reviewed in Lakin and Jackson, 1999). Two protein kinases have been implicated directly in p53 modifications induced by IR: ATM, required for the initial phase of p53 accumulation in response to this damage and ATR, involved in the later phase of this process (Canman et al., 1998). ATM is activated and mediates rapid phosphorylation of p53 on serine-15 (ser-15) (Canman et al., 1998), and also modifies the p53-inhibitory protein MDM-2 (Khosravi et al., 1999), interfering with the interaction between these two proteins and allowing the accumulation and subsequent initiation of p53-dependent processes such as trans-activation, cell cycle arrest and apoptosis.

In the present study we show that the role of PARP-1 in the p53-mediated response to DNA damage is dependent on the type of DNA lesion and on the upstream pathway leading to p53 activation. In summary, we show that treatment with IR results in an impaired p53 function in the absence of PARP-1; however, after treatment with a single alkylating agent, p53 accumulation and activation are enhanced in the absence of PARP-1.

Our results also demonstrate that PARP-1 is needed for optimal nuclear p53 accumulation, phosphorylation of ser-15 and stabilisation of p53, acting through the turnover of MDM-2, in response to -irradiation.

Results

Induction of p53 after DNA damage in parp-1 deficient cells

Nuclear accumulation of p53 protein in parp-1+/+ and -/- primary fibroblasts has been determined using a specific anti-p53 antibody after treatment with IR and with the alkylating agent MNU by immunoblot analysis. Figure 1a shows changes in p53 levels from 0 to 24 h in whole cell extracts from parp-1+/+ and parp-1 null cells. Following -irradiation, p53 accumulates very rapidly and remains at high levels even 24 h after irradiation in parp-1 parental cells (Figure 1a).

Parp-1 deficient cells fail to accumulate p53 in response to IR (Figure 1a), with only a discrete peak 4 h after IR-treatment. Co-incidently with this lack of p53 accumulation, a low increase in the level of p21 was observed (data not shown). On the other hand, following treatment with the single alkylating agent MNU, p53 was accumulated at much higher levels in parp-1 deficient cells than in their parp-1 counterparts (Figure 1b). The kinetic of p53 accumulation was similar between the two genotypes after UV treatment (Figure 1c). To further confirm the above results, indirect immunofluorescence to detect p53 was performed after treatment with the previously used DNA-damaging agents (Figure 1d). Clearly, a decrease in the nuclear accumulation of p53 was observed in parp-1-/- cells after IR, while MNU treatment was much more effective in promoting nuclear accumulation of p53 in parp-1 deficient cells.

We next tested whether PARP-1 affected the transactivation of p53 target genes. Semiquantitative analysis using the polymerase chain reaction with reverse transcription (RT-PCR) revealed that the p53 target genes p21 and MDM-2 were both upregulated in irradiated parp-1+/+ mouse embryonic fibroblasts (MEFs) while in irradiated parp-1-/- cells p21 was slightly increased and mdm-2 mRNA was not induced at the times tested (Figure 2). In contrast to irradiation, MNU exposure produced a strong up-regulation of p21 in parp-1 deficient cells, while in cells from both genotypes MNU was able to induce mdm-2 mRNA expression (Figure 2), although with different kinetic. Basal expression of both p21 and mdm-2 was slightly higher in parp-1 null cells. From all the above results, we conclude that the type of DNA damage is determinant of the role of PARP-1 in the p53-mediated response to genotoxic stress.

Regulation of MDM-2 protein levels and post-translational modification of p53

The p53 protein has a very short half-life. A key player in this regulation is the murine-double-minute-2 proto-oncogene. MDM-2 binds p53 within its trans-activation domain, blocks its transcriptional activity, and abrogates the ability of p53 to induce growth arrest and apoptosis. MDM-2 promotes p53 degradation by promoting the nuclear-cytoplasmic export of p53. Since MDM-2 is a direct target of p53, a negative autoregulatory feedback loop exists between these two proteins (Levine, 1997; Freedman and Levine, 1999).

A lag exists between the activation of p53 and the consequent induction of MDM-2, defining a time window within which p53 exerts its effects. We have determined the electrophoretic mobility pattern and turnover of MDM-2 following -irradiation or MNU treatments in immunoprecipitates obtained from cells of both genotypes. MDM-2 is found in several isoforms generated through alternative splicing and post-translational modifications (Olson et al., 1993). In parp-1+/+ cells a small increase shortly (15 min) after IR treatment was observed with a reproducible shift in the mobility of MDM-2 toward the faster migrating band (Figure 3). A remarkable increase in the levels of MDM2 was observed 4 h after IR treatment. Strikingly no such increase was observed in parp-1-/- cells, but rather a rapid and almost complete down-regulation of MDM-2 shortly after IR (Figure 3). Recovery of MDM-2 levels is also very rapid in the absence of PARP-1, that may account for the lack of nuclear accumulation and activation of p53 after irradiation in these cells (Figure 1). MNU treatment produced a rapid increase of MDM-2 protein in parp-1+/+ cells (in 30 min) and in a lesser extent in parp-1 deficient cells (Figure 3).

Phosphorylation of p53 in response to DNA damage is one mechanism by which its activity may be modulated. More precisely, phosphorylation of p53 at ser-15 reportedly blocks nuclear export of p53 allowing efficient accumulation and maximal p53 activation (Zhang and Xiong, 2001). This stabilized p53 protein can then transactivate its target genes (Lakin and Jackson, 1999).

The ataxia-telangectasia gene product (ATM) (and in a lesser extent ATR) mediates phosphorylation of ser-15 of p53 and its kinase activity is activated in response to IR but not to UV or alkylation (Banin et al., 1998; Canman et al., 1998). The levels of phosphorylated p53 at ser-15, monitored by Western blot using a specific anti-phosphopeptide antibody, are much lower in parp-1 deficient cells compared to parental cells (Figure 4a) in response to -irradiation. 3-aminobenzamide (3-AB) is a well-known PARP inhibitor. 3-AB in a concentration of 5 mM after 2 h-exposure efficiently blocks PARP activity avoiding PARP-polymer formation (data not shown) in response to DNA damage. When PARP activity was inhibited in parental cells using 5 mM 3-aminobenzamide prior any treatment, no significant down-regulation in the levels of p53 phosphorylated at ser-15 were observed following neither IR nor MNU exposure (Figure 4b).

Cell survival after irradiation and MNU treatments

The clonogenic cell survival parameters and the survival curves for four embryonic fibroblasts cell lines derived from parp-1+/+ p53+/+, parp-1-/- p53+/+, parp-1+/+ p53-/- and parp-1-/- p53-/- mice, after irradiation are shown in Table 1 and Figure 5a, respectively. To better understand the role of PARP-1 in the cytotoxic effect of IR we have to consider that, according to the linear-quadratic model (LQ) (Curtis, 1986), -rays produce two types of cellular lesions: (i) lethal lesions, that are non-reparable, and represent the initial and direct damage to DNA and (ii) potentially lethal lesion, representing the probability of interaction between two breaks and whose fate depends on competing process of repair and fixation (Curtis, 1986). The efficiency of this single-track lethal damage (production of severe lesions) is represented by the coefficient (that is proportional to the dose). Lethal damage caused by interaction or fixation of potentially lethal lesions, is represented by the coefficient (that is proportional to the square of the dose), and fluctuates as consequence of the cellular recovery after damage repair.

Survival curves after irradiation for mammalian cells are usually presented with dose plotted on a linear scale and surviving fraction on a logarithmic scale (Figure 5a). At low doses, the survival curve starts out straight on the log-linear plot with a finite initial slope. At higher doses, the curve bends and this bending region extends over a narrow dose range. At very high doses the survival curve often tends to straight again.

SF2 (survival fraction at 2 Gy) values were very similar for the four cell lines although slightly lower for parp-1-/- p53+/+ and parp-1-/- p53-/-. The presence of PARP-1 increases the component of the curve, even when p53 has been genetically deleted, meaning an enhancement in the severity and irreversibility of the initial lesions produced by IR (Table 1). However, at doses higher than 4-6 Gy, the absence of PARP-1 does not contribute to IR-induced cytotoxicity. This is reflected in statistically significant changes (not shown) in the parameter (Table 1) for parp-1+/+ p53-/- and parp-1-/- p53-/- in respect to parp-1+/+ p53+/+; and also for parp-1+/+ p53-/- and parp-1-/- p53-/- in respect to parp-1-/- p53+/+.

The steepness and curvature of these survival curves is described by the / (Table 1). The absence of p53 produced a flatter survival curve in respect to the cells with a functional p53. In contrast, the absence of PARP-1 plays a radiosensitizing effect. The / ratio changed 4.2-fold for parp-1-/- p53+/+ in respect to parp-1+/+ p53+/+ and 2.5 fold for parp-1-/- p53-/- in respect to parp-1+/+ p53-/-.

The colony forming ability after exposure to MNU showed that parp-1+/+ p53-/- was the most resistant cell line while the most sensitive was parp-1-/- p53+/+. The high sensitivity of parp-1-/- p53+/+ was lost in MEFs lacking the p53 gene (Figure 5b). In this case PARP-1 clearly determines the cytotoxic response to the single alkylating agent and the increased sensitivity seems to be p53 dependent (Table 2).

Discussion

Interplay between PARP-1 and p53 in the early response to genotoxic stress

The activities of both PARP-1 and p53 increase during the cell's response to DNA damage. Nonetheless, the interplay between these two proteins remains to be clarified. Different groups have reported that inhibition or genetic disruption of PARP-1 result in increased p53 accumulation due to the persistence of unrepaired DNA. In a pioneer study, p53 accumulation induced by X-ray, was increased upon treatment of cells with the PARP-1 inhibitor 3-aminobenzamide (Lu and Lane, 1993). Results using splenocytes, bone marrow cells and MEF issued from either parp-1 deficient mice (Oliver et al., 1999a), or thymocytes derived from transgenic mice expressing the dominant negative mutant of PARP-1 in the thymus (Beneke et al., 2000) show an increased accumulation of p53 following treatment with an alkylating agent.

On the contrary, studies from other groups identify PARP-1 as a necessary step for p53 expression and activation. A recent study with A-172 cells (wild type-p53) and T98G cells (mutant-p53), suggested that poly (ADP-ribosylation) is required for rapid accumulation of p53, activation of p53 sequence-specific DNA binding and its transcriptional activity after DNA damage (Wang et al., 1998). Results obtained using parp-1-/- MEF issued from the Wagner's laboratory, suggest that activation of p53 is largely PARP-1-independent (Agarwal et al., 1997). At the molecular level, it has been proposed that PARP-1 could cause p53 induction either by direct protein-protein interactions or by poly (ADP-ribosylation) of p53 (Malanga et al., 1998).

The diversity of the p53 activating factors covering genotoxic stress such as IR, UV light and DNA-damaging drugs, and non-genotoxic stress such as heat shock, or depletion of growth factors or nucleotides, implicates p53 as an integrator for both intracellular and extracellular signals. The variety of stress signals able for inducing PARP-1 activation is restricted to agents that produce single strand breaks in DNA (D'Amours et al., 1999). The cross-talk between these activities is thought to be dependent on the type of insult arriving to the cell. In the present study we demonstrate that the type of DNA damage determines the p53 response in the absence of PARP-1. The -ray-induced accumulation of p53 is greatly reduced in PARP-1 knockout cells while MNU-induced p53 accumulation is increased in the absence of PARP-1.

In addition to the perturbation in the levels and nuclear accumulation of p53, the absence of PARP-1 has consequences on the expression of the gene products downstream of p53, p21 and MDM-2, which parallel the accumulation of p53 in cells following -ray or MNU treatment. Of particular interest is the almost complete absence of MDM-2 mRNA upregulation after -irradiation in cells lacking PARP-1, reflecting a severe deficiency of p53-dependent transactivation of a key modulator of p53 stability. When the levels of the mdm-2 gene product were monitored by immunoprecipitation, an early down-regulation was observed followed by rapid and complete recovery of MDM-2 levels in PARP-1 deficient cells after IR. This result is particularly intriguing due to the absence of MDM-2 mRNA upregulation (Figure 2). One can only speculate that this rise in MDM-2 protein levels could be probably due to enhanced translation of pre-existing MDM-2 mRNA that has been previously reported as a regulatory mechanism of MDM-2 protein expression (Capoulade et al., 1998).

This rapid elevation in the levels MDM-2, which is an inhibitor of p53, may account for deficient p53 accumulation and activation in -ray treated parp-1-/- cells. On the other hand, an important reduction of IR-induced phosphorylation on ser-15 of p53 in PARP-1 deficient cells has been found. These results suggest a defective activation of the kinase(s) responsible for this postranslational modification in p53 in these cells, most likely ATM which is responsible for the early phosphorylation of p53 following IR (Khanna et al., 1998), while ATR is involved in later events following this treatment (Tibbetts et al., 1999).

A sharply different situation has been found after treatment with the alkylating agent MNU. The absence of PARP-1 enhances both p53 accumulation and trans-activation of p21. Again we find that MDM-2 levels inversely correlates with p53 accumulation (Figure 3): 15 min after MNU treatment, MDM-2 is down-regulated in both cell types, however a rapid recovery is observed 15 min later in parp-1 +/+ cells while in parp-1-/- cells MDM-2 remains at low levels at all times tested afterwards. Low levels of phospho-ser-15 p53 have been observed after MNU treatment in parp-1 knockout cells, suggesting that this post-translational modification not only is not essential for the signal transduction to p53 of the DNA lesions produced by MNU, but might indeed inhibit p53 activation upon treatment with the alkylating agent. The insight of the mechanism involved in this damage-specific activation of p53 is being investigated in our laboratory.

All these results have a common point of convergence: PARP-1 deficient cells have a reduced ability to modulate MDM-2 levels following DNA damage and to phosphorylate p53 at ser-15. The ATM gene product is responsible for phosphorylation of both p53 at ser-15 and MDM-2 (Banin et al., 1998; Canman et al., 1998; Maya et al., 2001) in response to IR. According to our present data we can hypothesize that ATM is probably either modified by or (somehow) interact with PARP-1 and this interaction has consequences on ATM-kinase activation after IR (Figure 6a). In this case we should expect a positive role of PARP-1 in the signal transduction of DNA damage to p53 in the pathway where ATM is involved. When damage to DNA do not involve ATM-kinase activation, such as that induced by MNU, the absence of PARP-1 leads to a delay in the DNA repair ability of the cells and consequently to a longer persistence of DNA lesions with a passive enhancement of p53 accumulation and activation (Figure 6b).

Role of PARP-1 in the intrinsic response to IR

This study also demonstrates that the role of PARP-1 in the response to IR is complex. PARP-1 loss is clearly radiosensitizing for gut (Menissier-de-Murcia et al., 1997) and thymocyte apoptosis in culture (Beneke et al., 2000), while has a limited effect on the long-term viability of fibroblasts (Figure 5b). Cell type-specific interactions of PARP-1 and p53 may reflect differences in the molecular determinants of the apoptotic threshold (Westphal et al., 1997). Our observations support the hypothesis that PARP-1 resides upstream of p53 in mediating IR-induced stabilisation and, in part, trans-activation ability of p53 (Figure 6b). However, the remaining cytotoxic response observed in parp-1-/- fibroblasts suggest the existence of an alternative radiation-induced pathway involving p53 that functions independent of PARP-1. On the other hand, it has been recently shown that PARP-1 plays a key role in the earliest steps of cell response to IR (Fernet et al., 2000) and it was hypothesized that rapid poly (ADP-ribosylation) of target proteins, or recruitment of repair proteins by activated PARP-1 at the sites of DNA damage, bring about rapid chromatin remodelling that may affect the incidence of chromosomal damage upon re-irradiation.

The differences in the direct damage to DNA induced by IR, that are non-reparable, (represented by the value, Table 1) and in the / ratio upon genetic deletion of PARP-1, suggest that the clinically unresponsive tumours (which are treated with fractionated doses in the range from 1-2 Gy) can be made more sensitive to the production of initial irreparable damage by therapeutic inhibition of PARP-1 (Ruiz de Almodóvar et al., 1994; Schlicker et al., 1999). Moreover, PARP-1 clearly determines the cytotoxic effect of single alkylating agents, which are also commonly used in the treatment of tumours. In this case, pharmacological inhibition of PARP-1 may represent a clear therapeutic advantage.

Materials and methods

Cell culture and treatments

Primary mouse embryonic fibroblast derived from both parp-1+/+ and parp-1-/- were used in this study and all experiments with MEFs were performed at passage 5 or less. We have also used immortalized murine embryonic fibroblasts expressing or lacking PARP and p53: parp-1+/+ p53+/+, parp-1-/- p53+/+, parp-1+/+ p53-/- and parp-1-/- p53-/- for clonogenic assays. All of them were grown in 10% foetal bovine serum-supplemented Dulbecco's modified Eagle's medium (FBS-DMEM, Sigma, St. Louis, MO, USA) with gentamicin (10 g ml^-1). All cell lines were incubated at 37°C in a humidified atmosphere of 5% O₂, 5% CO₂ and 90% N₂. MEFs were treated with the monofunctional alkylating agent MNU at 2 mM, a single 6 Gy dose of irradiation or UV treatment (40 J/m²). IR was delivered using a linear accelerator delivering 18 MeV photons (Siemens).

Immunoprecipitation and Western blotting assay

Cells were collected by scraping and cellular pellets were washed with cold PBS and lysed for 20 min at 4°C in 150 mM NaCl, 10 mM HEPES, pH 7.2, 0.5% NP-40, 2 mM EDTA, 2 mM EGTA and protease inhibitors. After centrifugation at 14 000 g 15 min at 4°C, supernatant was collected for further experiments. After protein quantification, Western analysis was carried out using standard procedures. In summary, 20 g of protein were loaded into each well of a 12% SDS-polyacrylamide gel. The gels were run at 150 V for 90 min in a Bio-Rad mini gel system. Proteins in the gel were transferred to a nitrocellulose membrane (100 V, 1 h) and then blocked for 15 min in 3% non-fat milk at room temperature. A polyclonal antibody to p53 (CM-1, Novocastra, Newcastle upon Tyne, UK), a monoclonal antibody to p21 (F-5, Santa Cruz Biotechnology) and a monoclonal antibody to -tubulin (Sigma, St. Louis, MO, USA) were used for detection of the respective proteins. An antibody against phospho-p53 (ser-15) was obtained from Oncogene. Antibody reaction was revealed with chemiluminescence detection procedures according to the manufacturer's recommendations (ECL kit, Amershan Corp., Buckinghamshire, UK). For immunoprecipitation, cell lysates were precleared by constant mixing for 2 h with protein A-Sepahrose (Pharmacia). The beads were removed by centrifugation, and the supernatant was mixed constantly overnight with a monoclonal antibody against MDM-2 (Biosource International). Immune complex were adsorbed onto protein A-sepharose, boiled and electrophoresed on polyacrylamide gels and revealed with the same antibody.

p53 indirect immunofluorescence

Cells were grown on slides 24 h before -irradiation, MNU or UV exposure. After the treatment cells were washed three times in PBS, fixed in fresh cold methanol-acetone (1 : 1) for 10 min at 4°C and then washed again with PBS-Tween 0.1%. The primary antibody for p53 analysis was used diluted 1/100 in PBS-Tween 0.1% and bovine serum albumin (BSA) 1%. Cells were incubated overnight at 4°C and then washed three times in PBS-Tween 0.1%. The secondary antibody used was the FITC-conjugated goat anti-rabbit antibody (Sigma, St. Louis, MO, USA) diluted 1/400 in PBS-T 0.1%-BSA 1%. The cells were incubated 1 h at room temperature in the dark. Finally cells were washed three times in PBS-T 0.1% and stained with 4',6'-diamidino-2-phenylindole dihydrochloride (DAPI) 1/1000 10 min. Slides were prepared using the Dako mounting medium (Dako Corporation), coverslipped and stored in the dark at -20°C.

RT-PCR

The primary cells were treated with 6 Gy -irradiation or 2 mM MNU. At the end of each treatment the cells were incubated in complete culture medium for 2 and 4 h. Then cells were harvested and total cellular RNA was isolated using the Trizol method (Gibco) according to the directions for use provided by the manufacturer. The RNA extracted was quantified spectrophotometrically. The control cultures were handled under identical conditions. At least two different experiments were done for each cell line. First-strand cDNA was synthesized using oligo (dT)_12-18 (Perkin Elmer) and M-MLV Reverse Transcriptase (Gibco) for RT-PCR. cDNA was used as template in PCR reactions. Primer sets were previously described by Jimenez et al. (1999): p21 5'-CGGTCCCGTGGACAGTGAGC-3' and 5'-AAATCTGTCAGGCTGGTCTGCC-3' (amplified fragment size: 371 bp), MDM-2 5'-GGAGCGCAAAACGACACTTACA-3' and 5'-CTCGCTGCTGCTGCTGCTAC-3' (512 bp), -tubulin 5'-GACAGTGTGGCAACCAGATCG-3' and 5'-GTACGGAAGCAGATGTCGTAG-3' (612 bp). Reaction conditions were as follows: initial denaturation step of 95°C for 9 min followed by 35 cycles of 93°C for 60 s, 66°C for 60 s and 72°C for 60s, with a final extension step of 72°C for 10 min. For the normalization of quantitative PCR the amplification with the three primers set was performed in the same reaction for each cell line. PCR products were analysed by agarose gel electrophoresis and ethidium bromide staining.

Irradiation and colony formation

Cell survival following IR and MNU-treatment was measured by clonogenic assay in monolayer. Cells were harvested and suspended in full culture medium. Single-cell suspensions were plated out at appropriate densities in duplicate in 25 cm² plastic flask (Sarstedt). In all the experiments, cells in exponential growth were used. Three separate experiments were done for each cell type and for each treatment.

Irradiation (dose range: 0-12 Gy) was performed 16 h after plating when cells were attached. After the treatment cells were incubated in complete culture medium for 8-12 days. Colonies of at least 50 cells were scored as surviving cells. Survival data (SF) were fitted using the linear-quadratic model which has two components of cell killing: one is proportional to dose (D) and the other is proportional to the square of the dose (D²): lnSF=-(D+D²). This model offers a good description of radiation response in the low-dose region (0-3 Gy). The cell survival curve is continuously bending with no final straight portion at high radiation doses. Its shape is determined by the ratio / that is the dose at which the linear contribution to the damage (D) equals the quadratic contribution (D²). When the quadratic component dominates, the survival curve bends over and becomes curved. The LQ model has now taken over as the model of choice to describe survival curves and generally works well in describing responses to radiation in vitro and also in vivo (Hall, 1994; Steel, 1997). and -parameters were determined by non-linear regression analysis and the surviving fraction at 2 Gy from the experimental curve.

To perform the clonogenic assay after drug exposure, cells were treated during 30 min with MNU-dilutions made in DMEM, range 0-2 mM. Each dose was used in two replicate flasks. At the end of treatment the MNU solution was removed and cells were incubated in complete culture medium for 8-12 days. Colonies of at least 50 cells were scored as surviving cells. The lethal dose 50% (LD50) in each cell line was determined from the experimental curve.

Statistical analysis

The survival parameters obtained in the four cell genotypes after irradiation (SF2, , ) and MNU treatment (LD50) were compared for statistically significant differences by ANOVA. The software used was GraphPad Software Prism (GraphPad Software Inc., San Diego, CA, USA).

Acknowledgements

This work was supported by a Grant from the Fondo de Investigación Sanitarias through the project FIS 00/0948 to FJ Oliver and by the Comisión Interministerial de Ciencia y Tecnología through the project PM97-0185 to JM Ruiz de Almodóvar. MT Valenzuela (009371) and R Guerrero are supported by fellowships from Fondo de Investigación Sanitaria (BEFI) and Instituto de Salud Carlos III, respectively. We are also indebted to Dr Josiane Ménissier- de-Murcia and Dr Abelardo López-Rivas for helpful discussion and reviewing of the manuscript and to Carmen Conde and Guadalupe de la Rubia for their excellent help in getting primary cells.

Agarwal ML, Agarwal A, Taylor WR, Wang ZQ, Wagner EF, Stark GR. (1997). Oncogene, 15: 1035-1341. MEDLINE

Banin S, Moyal L, Shieh S-H, Taya Y, Anderson CW, Chessa L, Smorodinsky NI, Prives C, Reiss Y, Shiloh Y, Ziv Y. (1998). Science, 281: 1674-1677. Article MEDLINE

Beneke R, Geisen C, Zevnik B, Bauch T, Muller WU, Kupper JH, Moroy T. (2000). Mol. Cell. Biol., 20: 6695-6703. MEDLINE

Burkart V, Wang ZQ, Radons JR, Heller B, Herceg Z, Stingl L, Wagner EF, Kolb H. (1999). Nat. Med., 5: 314-319. Article MEDLINE

Canman CE, Lim DS, Cimprich KA, Taya Y, Tamay K, Sakaguchi K, Appella E, Kastan MB, Siciliano JD. (1998). Science, 281: 1677-1679. Article MEDLINE

Capoulade C, Bressac-de Paillerets B, Lefrere I, Ronsin M, Feunteun J, Tursz T, Wiels J. (1998). Oncogene, 16: 1603-1610. MEDLINE

Curtis SB. (1986). Radiat. Res., 106: 252-270. MEDLINE

D'Amours D, Desnoyers S, D'Silva I, Poirier GG. (1999). Biochem. J., 342: 249-268. MEDLINE

Dantzer F, Niedergang SVC, Trucco C, Flatter E, de la Rubia G, Oliver FJ, Rolli V, Ménissier-de-Murcia J, de Murcia G. (1999). Biochimie, 81: 69-75. Article MEDLINE

de Murcia G, Ménissier-de-Murcia J. (1994). Trends Biochem. Sci., 19: 172-176. MEDLINE

Eliasson MJ, Sampei K, Mandir AS, Hurn PD, Traystman RJ, Bao J, Pieper A, Wang ZQ, Dawson TM, Snyder SH, Dawson VL. (1997). Nat. Med., 3: 1089-1095. MEDLINE

Fernet M, Ponette V, Deniaud-Alexandre E, Menissier-de-Murcia J, de Murcia G, Giocanti N, Megnin-Chanet F, Favaudon V. (2000). Int. J. Radiat. Biol., 76: 1621-1629. MEDLINE

Freedman DA, Levine AJ. (1999). Cancer Res., 59: 1-7. MEDLINE

Jimenez GS, Bryntessib F, Torres-Arzayus MI, Priestley A, Beeche M, Saito S, Sakaguchi K, Appella E, Jeggo PA, Taccioli GE, Wahl GM, Hubank M. (1999). Nature, 6739: 81-83.

Hall EJ. (1994). Radiobiologist for the radiologist Papadopoulos D, Cox K, Ryan JD. (eds) JB Lippincott Company: Pennsylvania, pp 29-44.

Khanna KK, Keating KE, Kozlov S, Scott S, Gatei M, Hobson K, Taya Y, Gabrielli B, Chan D, Lees-Miller SP, Lavin MF. (1998). Nat. Genet., 4: 398-400.

Khosravi R, Maya R, Gottlieb T, Oren M, Shiloh Y, Shkedy D. (1999). Proc. Natl. Acad. Sci. USA, 96: 14973-14977. Article MEDLINE

Lakin DL, Jackson SP. (1999). Oncogene, 18: 7644-7655. MEDLINE

Levine AJ. (1997). Cell, 88: 323-331. MEDLINE

Lu X, Lane DP. (1993). Cell., 75: 765-778. MEDLINE

Malanga M, Pleschke JM, Kleczkowska HE, Althaus FR. (1998). J. Biol. Chem., 273: 11839-11843. MEDLINE

Masson M, Niedergang C, Schreiber V, Muller S, Menissier-de-Murcia J, de Murcia G. (1998). Mol. Cell. Biol., 18: 3563-3571. MEDLINE

Masutani M, Suzuki H, Kamada N, Watanabe M, Ueda O, Nozaki T, Jishage K, Watanabe T, Sugimoto T, Nakagama H, Ochiya T, Sugimura T. (1999). Proc. Natl. Acad. Sci. USA, 96: 2301-2304. MEDLINE

Ménissier-de-Murcia J, Niedergang C, Trucco C, Ricoul M, Dutrillaux B, Mark M, Oliver FJ, Masson M, Dierich A, LeMeur M, Walztinger C, Chambon P, de Murcia G. (1997). Proc. Natl. Acad. Sci. USA, 94: 7303-7307. MEDLINE

Maya R, Balass M, Kim ST, Shkedy D, Leal JF, Shifman O, Moas M, Buschmann T, Ronai Z, Shiloh Y, Kastan MB, Katzir E, Oren M. (2001). Genes Dev., 15: 1067-1077. Article MEDLINE

Oliver FJ, de Murcia G, Ménissier-de-Murcia J. (1999a). Am. J. Human Genetics, 64: 1282-1288.

Oliver FJ, Ménissier-de-Murcia J, Nacci C, Decker P, Andriantsitohaina R, Müller S, de la Rubia G, Stoclet JC, de Murcia G. (1999b). EMBO J, 18: 4446-4454. Article MEDLINE

Olson DC, Marechal V, Momand J, Chen J, Romocki C, Levine AJ. (1993). Oncogene, 8: 2353-2360. MEDLINE

Ruiz de Almodóvar JM, Núñez MI, McMillan TJ, Olea N, Mort C, Villalobos M, Pedraza V, Steel GG. (1994). Br. J. Cancer, 69: 457-462. MEDLINE

Schlicker A, Peschke P, Bürkle A, Hahn EW, Kim JH. (1999). Int. J. Radiat. Biol., 75: 91-100. MEDLINE

Steel GG. (1997). Basic Clinical Radiobiology Steel GG. (ed) Arnold, London, pp 52-57.

Szabo C, Dawson VL. (1998). Trends. Pharmacol. Sci., 19: 287-298. MEDLINE

Tibbetts RS, Brumbaugh KM, Williams JM, Sarkaria JN, Cliby WA, Shieh S-Y, Taya Y, Prives C, Abraham RT. (1999). Genes Dev., 13: 152-157. MEDLINE

Trucco C, Oliver FJ, de Murcia G, Ménissier-de-Murcia J. (1998). Nucleic Acids Res., 26: 2644-2649. MEDLINE

Wang X, Onishi K, Takahashi A, Onishi T. (1998). Oncogene, 17: 289-2825.

Westphal C, Rowan S, Schmaltz C, Elson A, Fisher DE, Leder P. (1997). Nature Genetics, 16: 397-401. MEDLINE

Zhang Y, Xiong Y. (2001). Science, 292: 1910-1915. MEDLINE

Figure 1 Time course expression of p53 in whole cell extracts from parp-1+/+ and parp-1-/- MEFs by Western blot: (a) After -irradiation (6Gy). (b) After MNU-treatmen (2 mM). (c) After UV-exposure (40 J/m²). -Tubulin expression is shown as loading control. (d) Nuclear accumulation of p53 by indirect immunofluorescence in parp-1+/+ and parp-1-/- cells subjected to -irradiation (6 Gy), MNU (2 mM) during 5 h

Figure 2 RT-PCR of the p53 target genes p21 and mdm-2 in parp-1+/+ and parp-1-/- cells 2 and 4 h after -irradiation (6Gy) and MNU treatment (2 mM). The expression of -tubulin is shown as internal control

Figure 3 Kinetic of MDM-2 expression immunoprecipitated from parp-1+/+ and parp-1-/- cells after treatment with -irradiation (6 Gy) or MNU (2 mM)

Figure 4 Phosphorylation of ser-15 p53 monitored by western blot: (a) In whole cell extracts from parp-1+/+ and parp-1 null cells after -irradiation and MNU-treatment. -Tubulin expression is shown as control of protein loading. (b) In whole cell extracts from parp-1+/+ incubated 2 h with 5 mM 3-AB prior any treatment. Cells were harvested 30 min after 6 Gy irradiation or MNU exposure. -tubulin expression is shown as control of protein loading

Figure 5 Cell survival curves for four embryonic fibroblasts cell lines derived from parp-1+/+ p53+/+, parp-1-/- p53+/+, parp-1+/+ p53-/- and parp-1-/- p53-/- mice: (a) After -irradiation. (b) After MNU-treatment. Experiments were performed at least three times with each cell line, and pooled data after irradiation were fitted to a linear-quadratic equation to obtain these estimates of the surviving fraction at 2 Gy, - and -coefficients. Colonies of at least 50 cells were scored as surviving cells

Figure 6 Proposed role of PARP-1 in the pathways leading to p53 activation: (a) After ionizing radiation. (b) After treatment with an alkylating agent

Table 1 Clonogenic cell survival parameters after -irradiation

Table 2 Clonogenic cell survival parameters after MNU treatment