Oncogene (2001) 20, 430 ± 439 ã 2001 Nature Publishing Group All rights reserved 0950 ± 9232/01 $15.00 www.nature.com/onc Reciprocal down-regulation of and SOD2 expression ± implication in p53 mediated apoptosis

Pascal Drane1, Anne Bravard2,Ve ronique Bouvard1 and Evelyne May*,1

1Commissariat aÁ l'Energie Atomique (CEA), Laboratoire de CanceÂrogeneÁse MoleÂculaire, UMR217 CEA-CNRS, DRR, DSV, BP6 92265 Fontenay-aux-Roses Cedex, France; 2Commissariat aÁ l'Energie Atomique (CEA), Laboratoire de Radiobiologie Cellulaire, DRR, DSV, BP6 92265 Fontenay-aux-Roses Cedex, France p53 regulates the transcription of a number of cells containing damaged DNA. These functions are among which are di€erent redox-related genes. It has controlled, at least in part, by the property of p53 to been proposed that these genes can induce a cellular act as a transcriptional activator. In addition to leading to p53-dependent apoptosis stimulate the expression of cellular genes involved in (Polyak et al., 1997). MnSOD, the product of super- p53-dependent G1 arrest and apoptosis, p53 promotes oxide dismutase 2 (SOD2) gene, is one of the major the transcription of genes predicted to generate or cellular defences against oxidative stress. We demon- respond to oxidative stress suggesting that p53 strate here that p53 is able to repress SOD2 gene activation can result in an oxidative stress leading to expression and that this repression takes place at apoptosis (Polyak et al., 1997). level. We show the importance of this On the other hand, p53 has been reported to repress regulation for the p53 function, by demonstrating that the expression of a number of genes. Several mechan- an overexpression of MnSOD decreases p53-mediated isms have been proposed including interference with induction of apoptosis. Moreover, we demonstrate that initiation of transcription (Farmer et al., 1996; Wang MnSOD overexpression decreases p53- and Beck, 1998), binding to p53 responsive element at the promoter level. These ®ndings raise the hypothesis (p53RE) (Ori et al., 1998; Lee et al., 1999) or complex that p53 and SOD2 genes are mutually regulated leading formation with histone deacetylases (Murphy et al., to the modulation of various cellular processes including 1999). apoptosis. Oncogene (2001) 20, 430±439. The mitochondrial dismutase (MnSOD), involved in ROS detoxi®cation, .7 Keywords: p53; ; apoptosis; gene catalyses the dismutation of superoxide radical (O2 ) regulation; NF-kB into (H2O2) and (O2). There are two other SOD expressed in human cells, a cytosolic CuZnSOD (McCord and Fridovich, 1969) Introduction and an extracellular CuZnSOD (Marklund et al., 1982). Interestingly, while mice with disruption of one The p53 plays a pivotal role in of either form of CuZnSOD gene are viable, MnSOD maintaining the genomic integrity of cells. In response knockout mice have been shown to develop cardio- to various stresses, p53 has been shown to be activated myopathy and neonatal lethality (Huang et al., 1999). and stabilized. Activation of p53 results from post- This argues for MnSOD being one of the major translational events such as acetylation, phosphoryla- cellular defences against oxidative stress. tion and/or interaction with cellular proteins (May and MnSOD is encoded by the nuclear SOD2 gene May, 1999). The stabilisation of p53 is caused localized in the human 6q25 (Church et primarily by an increase in its half-life after post- al., 1992). Two main observations place SOD2 as a translational modi®cations that might be distinct from candidate tumor suppressor gene: the loss of hetero- those involved in p53 activation (Chernov et al., 1998). zygosity of chromosome region 6q found in about 40 Accumulation of p53 could also result from an up- percent of human malignant melanomas (Oberley and regulation of its gene transcription rate involving, in Oberley, 1997) and the deletion of the long arm of particular, the NF-kB transcription factor (Hellin et identi®ed in SV40 transformed human al., 1998). ®broblast (Bravard et al., 1992). In addition, MnSOD In response to genotoxic stress and depending on the overexpression suppresses the tumorigenicity of human cell-type, p53 response causes two major cellular melanoma cells (Church et al., 1993), breast cells events, either cell cycle arrest or programmed cell (Li et al., 1995) and glioma cells (Zhong et al., 1997). death (apoptosis), both preventing the propagation of Since both p53 and MnSOD appear to be involved in oxidative stress response, we have been interested in investigating a possible functional interaction between these two proteins. We present evidence that wt-p53 *Correspondence: E May Received 2 October 2000; revised 14 November 2000; accepted 14 down-regulates SOD2 gene expression at the promoter November 2000 level and that, in turn, an overexpression of MnSOD Cross-regulation of p53 and SOD2 gene expression P Drane et al 431 decreases the transcription level from p53 promoter and inhibits the p53-mediated induction of apoptosis.

Results

Constitutive SOD2 gene expression in MCF-7 and MCF-7/R-A1 cells MCF-7 and MCF-7/R-A1 are two related cell lines expressing wt-p53 and the mutant p53R280K, respec- tively (Cai et al., 1997). The wt status of p53 protein expressed in the parental MCF-7 cells used in this study was con®rmed by transient expression experi- ments using the luciferase reporter gene placed under the control of either wt (pE1B-hWAF1) or mutated (pE1B-hWAF1-mut) waf1-p53RE. Results presented in Figure 1a show in MCF-7 cells, a 700-fold increase of luciferase activity from wt waf1-p53RE compared to that obtained with the mutated one. By contrast and in agreement with the mutated status of p53 no such e€ect is observed with MCF-7/R-A1. By analysing similar cellular models, it has been reported that wt-p53 might participate to the constitu- tive expression of p53-target genes (Tang et al., 1998). We then took advantage of this cellular model to analyse a possible e€ect of wt-p53 on the constitutive expression of SOD2 gene. Representative Northern blots for waf1 and MnSOD2 mRNA are presented in Figure 1b. In agreement with previous study (Tang et al., 1998), the expression of waf1 is signi®cantly higher in MCF-7 than in MCF-7/R-A1, indicating that endogenous p53 might regulate the cellular waf1 gene expression. Inverse results are obtained for MnSOD (Figure 1b). The 4 kb-MnSOD mRNA species is Figure 1 Characterization of MCF-7 and the derivative MCF-7/ detectable in both cell lines, however, its intensity is R-A1 cell lines for: (a) their ability to activate transcription from the waf1 p53RE, (b) SOD2 mRNA level and (c) MnSOD activity. signi®cantly lower in MCF-7 than in MCF-7/R-A1. (a) MCF-7 and MCF-7/R-A1 cell lines were transfected with the Furthermore, in agreement with mRNA levels, the luciferase expression plasmid pE1B-hWAF1 (wt-p53RE) or pE1B- MnSOD activity is at least ®vefold lower in MCF-7 hWAF1-mut (having a mutated p53RE). Luciferase activity was than in MCF-7/R-A1 cellular extracts (Figure 1c). It is measured 24 h after transfection and normalized relative to the interesting to note that the high level of MnSOD in Renilla luciferase activity. (b) Twenty mg of total RNA extracted from con¯uent culture of MCF-7 and MCF-7/R-A1 cells were MCF-7/R-A1 correlates with an increase in their cell electrophoresed and transferred to Hybond N membrane doubling time compared to the parental MCF-7 cell (Amersham). Blot was successively hybridized with waf1 and line (data not shown). This is in agreement with the SOD2 cDNA probes. The loading was controlled by ethidium fact that MnSOD overexpression alters cell growth (Li bromide staining. (c) Activity of MnSOD was determined in both et al., 1995). We then hypothesised that SOD2 gene cell lines as described in Materials and methods expression might be down-regulated by wt p53.

clearly showed that PMA treatment leads to a signi®cant Activation of endogenous wt-p53 down-regulates the increase in MnSOD mRNA level in both MCF-7 SOD2 gene expression in MCF-7 (compare lanes 1 and 2) and MCF-7/R-A1 cells To further analyse the interaction between SOD2 (compare lanes 5 and 6). This is noticeable for both 1 expression and p53, we asked whether physiological and 4 kb mRNA species, generated by alternate activation of p53 by g-irradiation leads to a decrease of polyadenylation (Wispe et al., 1989). Interestingly, the steady-state level of MnSOD mRNA in MCF-7, SOD2 mRNA level is further increased in PMA- using MCF-7/R-A1 cells as negative control. To over- pretreated MCF-7/RA-1 cells, following g-irradiation come the diculty due to the very low level of MnSOD at a dose of 6 Gy (Figure 2a, compare lanes 6 and 8). In mRNA (Figure 1b), the cells were pretreated with PMA, contrast, a comparable treatment of MCF-7 leads to a a phorbol ester, that causes a consistent increase in signi®cant decrease of SOD2 mRNA level (Figure 2a, MnSOD mRNA level (Li et al., 1998). According to compare lanes 2 and 4). Both decrease (MCF-7) and these published data, results presented in Figure 2a increase (MCF-7/RA-1) in MnSOD mRNA level

Oncogene Cross-regulation of p53 and SOD2 gene expression P Drane et al 432

Figure 2 E€ect of g-radiation on basal and PMA-induced SOD2 mRNA level in MCF-7 and MCF-7/R-A1 (a) and e€ect of PMA pre-treatment on g-radiation-induced accumulation (b) and activation (c) of MCF-7 endogenous wt-p53. (a) Sub-con¯uent MCF-7 (lanes 1 ± 4) or MCF-7/R-A1 (lanes 5 ± 8) cells were mock treated (lanes 1, 3, 5, 7) or treated with 100 nM PMA (lanes 2, 4, 6, 8). Four hours later the medium was changed and the cells were irradiated at a dose of 6 Gy (lanes 3, 4, 7, 8) or not (lanes 1, 2, 5, 6). For each treatment, 20 mg of total RNA were electrophoresed and transferred to Hybond N membrane (Amersham). Blot was successively hybridized with SOD2, waf1 and GAPDH cDNA probes. (b) MCF-7 cells were incubated with PMA (lanes 2, 4) or DMSO (lanes 1, 3) for 4 h and exposed to g radiation at a dose of 6 Gy (lanes 3, 4) or not (lanes 1, 2). Two hours after irradiation, total cellular extracts were analysed by Western blot using anti-p53 and anti-PCNA (to control loading) antibodies. (c) Nuclear extracts prepared from cells treated as described in b were incubated with labeled oligonucleotide corresponding to the waf1-p53RE to be analysed by band-shift assay as described in Materials and methods

following g-irradiation are independent of PMA treat- could be the result of p53 repressing the SOD2 ment since similar irradiation-dependent variations are promoter activity. The plasmid pluc-SOD2 was con- observed with non-treated cells, although to a lower structed in which the luciferase reporter gene was extent (Figure 2a, compare lane 1 and 3 for MCF-7, and placed under the control of the rat SOD2 promoter. 5 and 7 for MCF-7/R-A1). PMA has been reported to The luciferase activity from this pluc-SOD2 construct attenuate p53-DNA binding activity after g-irradiation was then analysed following transient transfection of of NIH3T3 cells by decreasing the half-life of p53 protein MCF-7 and MCF-7/R-A1 cells. To take into account (Price and Calderwood, 1993). To make sure that the di€erences in the eciency of transfection from one decrease of MnSOD mRNA level after irradiation of the experiment to another and between both cell lines, the PMA-treated MCF-7 cells was not related to a decrease Fire¯y luciferase activity was normalised with Renilla of p53 level, the e€ect of g-irradiation on p53 protein luciferase activity and the results were expressed accumulation and activation was measured from un- relatively to normalized luciferase activity measured treated and PMA-treated MCF-7 cells. Results pre- from MCF-7 transfected cells. Results presented in sented in Figure 2b,c showed that, in our experimental Figure 3 give an average of three independent conditions, PMA treatment has no e€ect on p53 experiments, each performed in duplicate. The lucifer- accumulation and activation following irradiation ase activity is at least three times lower in MCF-7 than (compare lanes 3 and 4). Moreover, the expression of in MCF-7/R-A1 transfected cell extracts. This result waf1 gene is signi®cantly induced following g irradiation indicates that transcription from a reintroduced SOD2 of PMA-treated (Figure 2a, lane 4 compared to lane 2) as promoter is signi®cantly less ecient in MCF-7 than in well as non-treated (Figure 2a, lane 3 compared to lane MCF-7/R-A1, supporting the idea that in these cells 1) MCF-7 cells, indicating that PMA has no inhibiting endogenous wt-p53 could down-regulate gene expres- e€ect on p53 activation in our cellular model. We can sion from SOD2 promoter. To further support this then conclude that the SOD2 gene expression is down- assumption, pluc-SOD2 was cotransfected with regulated in response to physiologically activated pCMVDD that encodes the truncated protein endogenous wt-p53 in MCF-7 cells. p53DD, a dominant negative p53 mutant. This truncated protein has been shown to abrogate the transcriptional activity of endogenous wt-p53 protein SOD2 promoter activity is down-regulated by wt-p53 from various cell lines (Smart et al., 1999). Cotransfec- To further analyse the ability of p53 to down-regulate tion of pluc-SOD2 with pCMVDD increased signi®- SOD2 gene expression, we asked whether this e€ect cantly the relative luciferase activity of MCF-7, but not

Oncogene Cross-regulation of p53 and SOD2 gene expression P Drane et al 433

Figure 3 SOD2 promoter activity in MCF-7 and MCF-7/R-A1 cell lines. Both cell lines (26105) were co-transfected with 500 ng of pluc-SOD2 and 1 mg of either the empty pcDNA3 plasmid (Invitrogen) or pCMVDD plasmid encoding the dominant negative p53-DD mutant. Luciferase activity was measured as described in Material and methods. Results are expressed relative to the activity of pluc-SOD2 in MCF-7 cell line

MCF-7/R-A1 cellular extract (Figure 3), leading to the conclusion that wt-p53 could inhibit SOD2 gene expression at the promoter level. This was further con®rmed in the p53-negative H1299 cells cotransfected with pluc-SOD2 and various amounts of pCMVhump53 coding for the human wt- Figure 4 E€ect of transient expression of wild-type (a) and mutant p53 (b) on the SOD2 promoter activity, in p53-null cell p53. Results presented in Figure 4a clearly show a p53 line H1299. H1299 cells (26105) were co-transfected with 500 ng dose-dependent decrease in luciferase activity. Indeed, of pluc-SOD2 and indicated amount of pCMVhump53 (a)or more than 80% inhibition of the SOD2 promoter 200 ng of empty vector (control), wild-type p53 (pCMVhump53), activity is observed in H1299 cells cotransfected with or di€erent p53 mutants (pCMVp53Ser249; p53-del11-69; 200 ng of p53 expression vector. In contrast to wt-p53, Trunc223NLS) (b) Results are expressed relative to the luciferase activity measured in the absence of p53 co-expression various mutants of p53 having either a single substitution in the DNA-binding domain (p53Ser249), or a deletion either in the N- (p53- del[11-69]) or C- (p53-223NLS) terminal domains, do p53 protein was performed and the cells were analysed not exhibit such inhibitory e€ect (Figure 4b). All by ¯ow cytometry. DNA content of p53-positive and taken together, these results indicate that p53 is able negative populations were examined to determine the to repress the gene expression from the SOD2 percentage of apoptotic cells as previously described promoter. (Yonish-Rouach et al., 1995). Results of four indepen- dent experiments are given in Table 1. We observed a MnSOD dose-dependent decrease in the percentage of Overexpression of MnSOD inhibits p53-dependent apoptotic cells in the p53-positive population, indicat- apoptosis ing that MnSOD might protect cells from p53- We then addressed the question whether an over- dependent induction of apoptosis. Since p53-dependent expression of MnSOD can interfere, in transient assay, apoptosis involves, at least partially, its transactivation with the p53-dependent induction of apoptosis. To test function (Yonish-Rouach et al., 1995; Polyak et al., this hypothesis, H1299 cells were co-transfected with 1997), we tested the e€ect of an overexpression of two recombinant plasmids, one encoding the human MnSOD on the p53-dependent transactivation of a wt-p53, the other the human MnSOD enzyme, both reporter gene. p53-mediated transactivation of the under the control of CMV promoter. Forty-eight hours waf1-p53RE was measured in H1299 with or without after transfection double staining of genomic DNA and coexpression of MnSOD (Figure 5). We ®rst checked,

Oncogene Cross-regulation of p53 and SOD2 gene expression P Drane et al 434 by Western blot, that transient overexpression of MnSOD had no e€ect on the level of the transiently expressed p53 protein, and, conversely, that transient overexpression of p53 did not a€ect the level of transiently expressed MnSOD protein (Figure 5, lower panel). But, unlike what was observed for apoptosis, the p53-mediated transactivation from waf1-p53RE is not a€ected by MnSOD overexpression (upper panel). These results suggest that the inhibition of p53- dependent apoptosis by MnSOD is independent of p53 transactivation function.

MnSOD overexpression down-regulates p53 promoter activity independently of NF-kB. Considering the fact that MnSOD has been shown to modulate NF-kB activity (Li et al., 1998), and that NF-kB regulates p53 promoter activity (Wu and Lozano, 1994; Hellin et al., 1998; Pei et al., 1999), we were led to analyse the e€ect of an overexpression of Figure 5 E€ect of MnSOD overexpression on p53 transactiva- MnSOD on p53 promoter activity. We then cotrans- tion function. H1299 cells (26105) were co-transfected with fected pluc-mp53, a plasmid encoding the luciferase 500 ng of pE1B-hWAF1 and, when indicated, 200 ng of reporter gene under the control of the mouse p53 pCMVhump53 and/or the indicated amount of pCMVhumMn- SOD. Twenty-four hours after transfection, cells were lysed and promoter, with increasing amounts of MnSOD expres- analysed for luciferase activity (upper panel) and for p53 and sion vector, in H1299 cells. The luciferase reporter gene MnSOD expression by Western blot (lower panel). Luciferase driven by the SOD2 promoter was used as control. No activity was expressed as fold stimulation relative to the luciferase signi®cant di€erence in luciferase activity was observed activity of cells transfected with pE1B-hWAF1, alone (®rst lane) when MnSOD expression plasmid was cotransfected with pluc-SOD2 (Figure 6a). In contrast, the activity of the p53 promoter was signi®cantly down-regulated by activity than MCF-7 cells, we found that the p53 MnSOD, in a dose-dependent manner (Figure 6b). promoter activity measured from the transfected pluc- Indeed, the transfection of 500 ng of pCMVhumMn- mp53 plasmid is lower in MCF-7/R-A1 than in MCF-7 SOD leads to a 50% decrease of luciferase activity. In cells (Figure 6c). Finally, we examined a possible support of these results, and in agreement with the fact connection between the NF-kB and MnSOD in the that MCF-7/R-A1 cells exhibit a higher MnSOD regulation of p53 promoter. For this purpose, pluc- p53NFkBmut was constructed from pluc-mp53 recom- binant by inactivating the p53-promoter NF-kB site. In agreement with already published data (Pei et al., Table 1 Dose-dependent inhibition of p53-dependent apoptosis in 1999), inactivation of NF-kB site signi®cantly di- H1299 cells by MnSOD overexpression minishes the level in the luciferase gene expression pCMVhumMnSOD/ from the p53-promoter (Figure 6d, control). Further- a pCMVhump53 more, the co-expression of IkBa inhibits the luciferase 0/1 1/1 4/1 pCMVhump53 Percentage of expression from pluc-mp53 but not from pluc- Experiments (mg) Population apoptotic bodiesb p53NFkBmut (Figure 6d,+IkBa). However, co-expres- sion of MnSOD further decreases the luciferase activity 1 0.2 p53+ 32.8 19.4 10.1 p537 3.5 2.6 1.7 of cells transfected with pluc-p53NFkBmut plasmid 1 p53+ 23.5 14.7 8.5 (Figure 6d, +MnSOD). The fact that NF-kB site p537 1.5 2.1 1.7 mutation shows an additive repressive e€ect over the 2 0.2 p53+ 20.0 nd 10.4 one of MnSOD overexpression, strongly suggests that p537 0.7 nd 0.5 MnSOD and NF-kB act via separate ways on p53 gene 1 p53+ 41.0 nd 29.9 p537 2.4 nd 2.3 expression. 3 1 p53+ 17.6 nd 8.3 p537 0.5 nd 0.8 4 0.2 p53+ 23.3 17.6 14.0 Discussion p537 2.5 2.4 2.4 1 p53+ 27.3 20.0 14.8 p537 3.2 2.9 2.2 We report herein data showing a reciprocal negative regulation of SOD2 and p53 gene expression by wt p53 aAmount pCMVhumMnSOD plasmid given by the ratio and MnSOD proteins, respectively. These bring b pCMVhumMnSOD/pCMVhump53. Percentage of apoptotic bodies argument to propose a cross-talk between p53 and (per cent of cells in sub-G1) was estimated by ¯ow cytometry for the p53-positive (p53+) and p53-negative (p537) populations. nd: not MnSOD that could participate to a ®ne regulation of determined p53 activity.

Oncogene Cross-regulation of p53 and SOD2 gene expression P Drane et al 435

Figure 6 E€ect of MnSOD overexpression on p53 promoter activity. H1299 cells (26105) were co-transfected with 500 ng of pluc- SOD2 (a) or pluc-mp53 (b) and 0, 200 or 500 ng of pCMVhumMnSOD. Luciferase activity measured 24 h after transfection was expressed relative to the activity of cells co-transfected with CMV empty vector taken as 100. (c) MCF-7 and MCF-7/R-A1 cells were transfected with 500 ng of pluc-mp53 and 10 ng of pSVE-RenLuc. Luciferase activity was normalised relative to Renilla luciferase activity. Results are expressed taking as 100 the normalized activity of MCF-7 cell extracts. (d)26105 MCF-7 cells were co-transfected with 500 ng of pluc-mp53 (&) or pluc-p53NFkBmut (&) and 500 ng of either empty (control), or MnSOD or IkBa expression plasmids. Luciferase activity measured 24 h after transfection, was expressed relative to activity of cotransfected cells with pluc-mp53 cells and the empty vector

DNA synthesis. Similar conclusion of p53 being Down-regulation of SOD2 gene by wt p53 slightly activated by cell culture conditions was In this study we took advantage of a cellular model reached by Mendrysa and Perry (2000) who showed consisting of two related cell lines: the wt p53 that the constitutive expression of MDM2 is higher in expressing human breast carcinoma MCF-7 cells and p53+/+ than in p537/7 mouse embryonic cell lines, the derived MCF7-RA1 line that expresses a mutant whereas no such p53-dependent e€ect could be p53, both lines expressing functional MnSOD protein observed in vivo. These observations tone down the (Li et al., 1995 and results herein). It has been reported largely accepted postulate that, in the absence of stress, that in MCF-7 cells, endogenous wt p53 could cells in culture express an inactive form of wt p53 and positively regulate the constitutive expression of at justify the use of wt p53 expressing cells to further least some p53-target genes, such as waf1 (Tang et al., characterize p53 positively or negatively regulated 1998) and c-Ha-Ras (Deguin-Chambon, unpublished candidate genes. Using this model, we show that the results). To account for the p53-dependent regulation constitutive level of both SOD2 mRNA and MnSOD of waf1 expression in these cells, Tang et al. (1998) activity is lower in MCF-7 than in MCF-7/R-A1, suggested that cells growing in vitro undergo low levels suggesting that p53 may be involved in down- of oxidative DNA damage resulting from errors during regulation of SOD2 gene expression. Recent published

Oncogene Cross-regulation of p53 and SOD2 gene expression P Drane et al 436 data led to the same conclusion in cell lines and in decreasing the transcription level from NF-kB respon- mouse liver (Pani et al., 2000; Shatrov et al., 2000). sive promoters (Li et al., 1998). We demonstrate This is further supported by our results obtained however, that in our experimental conditions, the following g irradiation of the MCF-7 and MCF-7/R- down-regulation of the p53 promoter activity by A1. We show that the SOD2 mRNA up-regulation MnSOD is not dependent of NF-kB inhibition by this induced by phorbol esters can be eciently repressed enzyme. Our data show indeed that, MnSOD over- by physiologically activated wt p53. Using transient expression is still able to down-regulate the activity of transfection assays we then show that the SOD2 the p53 promoter after mutation of the NF-kB site. promoter can confer transcriptional repression by This indicates that MnSOD and NF-kB act via endogenous or exogenous wt-p53 to a reporter gene, separate ways on p53 gene expression and suggests demonstrating that p53-dependent down-regulation of that MnSOD may alter the activity of another SOD2 transcription takes place at promoter level. transcription factor involved in this process. The SOD2 promoter is a GC-rich and TATA/ CAAT-less promoter (Kuo et al., 1999). Although Oxidative stress and apoptosis: a cross-talk between p53 p53 repression has been mainly reported on TATA and MnSOD containing promoters (Mack et al., 1993), few studies have shown that p53 is also able to negatively regulate Several reports indicate that oxidative stress is required some TATA-less promoters (Kaluzova et al., 2000; in many instances for execution of the apoptotic Dugimont et al., 1998; Iotsova et al., 1996). Several program (Buttke and Sandstrom, 1994). p53 acting as mechanisms have been proposed for the p53-dependent a sensor of the cellular response to stress and, MnSOD repression process. They mainly imply protein-protein being one of the main anti-oxidant enzymes, both are interactions of p53 with di€erent factors involved in involved in these processes. the basal transcriptional machinery such as the TBP p53 can induce apoptosis through a multistep (Lee et al., 2000) or Sp1 (Wang and Beck, 1998) or process involving transcriptional induction of a with histone deacetylases (Murphy et al., 1999). p53 number of genes such as redox-related genes playing binding to speci®c DNA sequences has also been a role in the production of reactive oxygene species reported to account for p53-dependent repression (Ori (ROS) (Polyak et al., 1997). This leads to oxidative et al., 1998; Lee and Rho, 2000). The analysis of the degradation of mitochondrial components and ®nally rat SOD2 promoter sequence revealed the existence of to apoptosis (Polyak et al., 1997). p53 has also been many Sp1 sites and also the presence of a half p53 shown to be capable of repressing the expression of consensus site that was demonstrated to participate in many genes among which several ones encoding for the p53-dependent repression of the HBV proteins with anti-apoptotic activities such as bcl-2 (Ori et al., 1998). Interestingly the same recognition (Miyashita et al., 1994), MAP4 (Murphy et al., 1999), sites can be found within the human SOD2 promoter. Presenilin 1 (Roperch et al., 1998) and others. The Experiments are under way to characterise for both present study allows us to add SOD2 to the list of SOD2 genes the mechanism responsible for p53- the anti-apoptotic p53 down-regulated genes. In dependent repression. addition, p53 was also reported to down-regulate the two other superoxide dismutases, SOD1 (Cu/Zn cytosolic SOD) (Cho et al., 1997) and SOD3 (Zhao et Down-regulation of p53 promoter activity by MnSOD al., 2000), arguing for a role of p53 in modulating the Although it is well recognised that activation of the oxidative stress via regulation of several anti-oxidant p53 protein by post-translational modi®cations is of enzymes. major importance for the p53-dependent cellular MnSOD protein was previously shown to negatively response to stress, Komarova et al. (1997) reported, interfere with apoptosis induced by various stimuli in vivo, a positive correlation between p53 mRNA level such as Tumour Factor (TNF), okadaic acid, and tissue radiosensitivity, suggesting that the regula- H2O2, taxol (Manna et al., 1998), paraquat (Wenk et tion of p53 mRNA expression could also be involved al., 1999) and antimycin (Kiningham et al., 1999). We in the modulation of p53 functional activity. present evidence herein that over-expression of In addition to the repressing e€ect of p53 on the MnSOD is also capable of protecting H1299 cells SOD2 transcription, our results also show that an from p53-dependent apoptosis. Although it was over-expression of the MnSOD protein leads to a reported that p53-dependent apoptosis could take down-regulation of the p53 promoter activity. The p53 place in the absence of p53-dependent activation of promoter sequence is highly conserved across the cellular gene expression in some cells (Caelles et al., di€erent species (Bienz-Tadmor et al., 1985). A number 1994), the p53 transcriptional activity was shown to be of transcription factors have been shown to regulate its required for the induction of apoptosis in our cellular activity (Pei et al., 1999; Furlong et al., 1996; Noda et model (Yonish-Rouach et al., 1995). Here, we demon- al., 2000), but among them only NF-kB has been strate nonetheless that the inhibition of p53-dependent reported to be redox-sensitive (reviewed in Sun and apoptosis by MnSOD overexpression is independent of Oberley, 1996). Recent published data indicated that the p53 transactivation function, suggesting that the MnSOD over-expression led to inhibition of both MnSOD overexpression acts downstream of p53 NF-kB and AP1 activity and was also capable of induced cellular gene expression.

Oncogene Cross-regulation of p53 and SOD2 gene expression P Drane et al 437 It has been recently described that p53 ®rst induces The pluc-mp53 construction was obtained by cloning up- mitochondrial cytochrome c release by a pathway stream the luciferase gene (between XmaI/BglII restriction requiring the Bax p53-target (Schuler et al., 2000). sites of pGL3-basic plasmid), the 0.7 kbp mouse p53 This is followed by the activation of the caspase promoter fragment isolated from p0.7CAT construction cascade. Since production of ROS can be both a cause (Ronen et al., 1991). The pluc-p53NFkBmut was constructed by PCR-directed mutagenesis from pluc-mp53 plasmid. The and a consequence of the mitochondrial permeability EcoRI-BsaAI 414 bp fragment of pluc-mp53 was replaced by transition (Sheng-Tanner et al., 1998), MnSOD acting the corresponding mutated fragment obtained by PCR- via ROS detoxi®cation may inhibit p53-dependent ampli®cation using the primers (TTGAATTCCAGGC- apoptosis, either before the permeability process CAGCCTTG) and (GGTGAGCACGTGGGAGCTTAAA- preventing the cytochrome c release or after, leading GTGATAATCC). Based on published data (Wu et al., 1994), to an inhibition of the caspase's activation, or both. In the nucleotide changes (nucleotides underlined) inactivate the agreement with this hypothesis, inhibition of both NF-kB binding site. To obtain pE1B-hWAF1 plasmid, the caspase 3 activation and cytochrome c release by two oligonucleotides TCGAGAACATGTCCCAACATGTTG MnSOD was previously reported (Manna et al., 1998; and CTAGCAACATGTTGGGACATGTTC, containing the Kiningham et al., 1999). Alternatively and in addition, waf1-p53RE (underlined) were annealed and cloned between the XhoI/NheI restriction sites of pGL3-E1bTATA as the ability of MnSOD to downregulate p53 gene previously described (Munsch et al., 2000). The pE1B- expression as shown here, could also play a role in a hWAF1-mut was constructed following the same strategy, negative modulation of the apoptotic p53 function. using the oligonucleotide pair TCGAGAAGATCTCCCAA- In conclusion, our results strongly suggest a cross- CATGTTG and CTAGCAACATGTTGGGAGATCTTC in talk between MnSOD and p53 that may result in a ®ne which two nucleotide changes (nucleotides underlined) were control of p53-dependent response to stress. It is then introduced, leading to the inactivation of the waf1-p53RE. tempting to speculate that the interplay between p53 and MnSOD could play a role in several human Cells and treatment chronic , such as atherosclerosis and cardio- vascular diseases, mutagenesis and cancer, several MCF-7 cell line expressing wt-p53 was derived from a human neurogenerative disorders, and even the aging process breast carcinoma H1299 cell line, a non-expressing p53, was per se, for which oxidative stress is one of the major derived from a human non small cell lung carcinoma. MCF- contributing factors (Frei, 1999). 7/R-A1 cells, a gift from Dr S Chouaib, was obtained from MCF-7 by continuous exposure to increasing doses of TNF-a (Cai et al., 1997). This cell line expresses a p53 mutated at amino acid residue 280 (R?K). Material and methods Cells were maintained at 378C in DMEM (MCF-7) or RPMI (H1299, MCF-7/R-A1) containing 10% fetal calf Plasmids and constructions serum (FCS). For PMA treatment, 48 h after plating (36106 per 10 cm Petri dish) 5 ml of a 200 mM solution of PMA All the p53- and MnSOD-expression plasmids are under the (Sigma) in DMSO (or 5 ml of DMSO for the mock treated control of the cytomegalovirus immediate-early (CMV) cells) were added per 10 ml of medium. After 4 h at 378C, the promoter. pCMVhump53 is an expression plasmid encoding cells were washed, refed with fresh medium and exposed to g the human wt-p53. The pCMV-p53Ser249, derived from irradiation at a dose of 6 Gy and a dose rate of 2.1 Gy/min, pCMVhump53, encodes the human protein mutated at codon using a 137Cs source (IBL 637 apparatus, Cisbio). Cells were 249. The mutant del(11 ± 69), deleted for the N-terminal part harvested 16 h following irradiation. of the p53 sequence from codon 11 ± 69, is the human counterpart of the mouse deletion mutant dl162 (Jenkins et al., 1988). The Trunc223NLS, derived from pCMVhump53, Antibodies encodes a truncated human p53 comprising the ®rst 223 residues fused to the C-terminal fragment (residues 282 ± 340) The monoclonal antibody DO-7, directed against p53, was a where the ®rst Nuclear Localization Domain is located gift from Dr D Lane. The monoclonal antibodies directed (NLS1, residue 316 ± 325). Plasmid pCMVDD expresses a against p53 (Ab1) and PCNA (Ab1) were from Oncogene truncated mouse p53 including residues 1 ± 14 linked to the Research Products (Calbiochem). The polyclonal antibody C-terminal fragment (residues 302 ± 390) (Shaulian et al., anti-MnSOD was from Valbiotech. 1992). pCMVhumMnSOD contains the human MnSOD cDNA cloned into the pcDNA3 expression vector (Invitro- gen). MnSOD cDNA, obtained from Dr Ho (Ho and Crapo, Quantification of apoptotic cells by flow cytometric analysis 1988), encodes the Threonine-58 MnSOD polymorphic form 16106 H1299 cells were seeded in a 10 cm Petri dish. Twenty (Borgstahl et al., 1996). The IkBa expression vector was a gift four hours later the cells were transfected with 1 ml of from Dr A At®. The pluc-SOD2 plasmid contains the rat calcium phosphate precipitate containing 0.2 or 1 mgof SOD2 sequence from position 7311 to +57, taking as +1 pCMVhump53, and the indicated amount of pCMVhumMn- the transcription initiation site according to Genbank SOD. In each experiment, total DNA was balanced to 30 mg/ (Accession number X56600). According to Huang et al. ml with pBS vector (Stratagene). Forty-eight hours after (1997) this region is suitable for initiating gene transcription. incubation, cells were trypsinised, ®xed with 70% ethanol and This fragment was generated by PCR ampli®cation from rat double stained with DO-7 hybridoma cell supernatant DNA using the oligonucleotide pair: (GCGGTACCACAGG- (dilution 1/10) for p53 and propidium iodide (sigma) for CAGAGGTGGCCAAGGC) and (GCAGATCTCCACGA- DNA, as previously described (Yonish-Rouach et al., 1995). CCGCTGCTCTCCTCA). It was inserted between the KpnI/ Flow cytometric analyses were carried out using a FacScan BglII restriction sites of the pGL3-basic plasmid (Promega). (Becton Dickinson).

Oncogene Cross-regulation of p53 and SOD2 gene expression P Drane et al 438 20 min at 48C in bu€er A (20 mM HEPES, pH 7.6, 10 mM RNA extraction and Northern blot NaCl, 1.5 mM MgCl2, 0.2 mM EDTA, 1 mM DTT, 20% Total RNA was extracted using RNAplus reagent (Biop- glycerol, 4 mM Pefabloc) containing 0.1% NP-40. Nuclei robe). Twenty mg of total RNA were separated on a 1% were collected by centrifugation at 2000 g,at48C, and lysed agarose/6% formaldehyde gel. Northern-blot analysis were 30 min at 48C in bu€er A containing 0.5 M NaCl and 0.1% performed as previously described (Deguin-Chambon et al., NP-40. The supernatant was collected by centrifugation at 2000). The 500 bp EcoRI/BamHI-fragment isolated from the 15 000 g,at48C, for 15 min. The complementary oligonu- pCMVhumMnSOD plasmid, the 2.1 kbp EcoRI-fragment cleotides TCGAGAACATGTCCCAACATGTTG and isolated from pZL-WAF1 (El-Deiry et al., 1993) and the CTAGCAACATGTTGGGACATGTTC (containing the 1.3 kbp GAPDH-cDNA PstI-fragment were used as probes waf1-p53 binding site) were annealed and labeled by ®ll-in. for Northern-blot analysis. Signals were quanti®ed using the Then, nuclear extracts (10 mg of total protein) were incubated STORM apparatus (Molecular Dynamics). with 32P-labeled oligonucleotide (0.4 ng), salmon sperm DNA (1 mg) in 30 ml ®nal volume of bu€er A. Fifty ng of PAb421 (Ab1) monoclonal antibody were added to the reaction Western-blot mixture containing the waf1-p53 binding site before incuba- Cells from one Petri dish (10-cm diameter) were scraped, tion at room temperature, for 30 min. Then, 4 ml of reaction washed twice with ice-cold PBS and resuspended at 48Cin mixture was loaded on a Phast Gradient 4 ± 15% native bu€er containing 50 mM Tris-base (pH 8.0), 150 mM NaCl, acrylamide gel and electrophoresed for 75 min at 400 V in 5mM EDTA, 1% NP-40 and 1 mM Pefabloc (Boehringer the PhastSystem apparatus (Pharmacia LKB). Mannhein). After 30 min, at 48C, the extracts were centrifuged for 15 min at 10 000 g. Total protein concentra- Assays for SOD activity tion of the supernatant was determined by Bradford assay (Biorad Protein Assay). A volume of supernatant correspond- Cells were harvested by trypsination in exponential growth ing to 20 mg of total protein was separated on a 10% (for p53 phase, washed twice in NaCl 0.9% and stored as dried pellets and PCNA) or 12% (for MnSOD) SDS ± PAGE. After at 7808C until extraction. Cells disruption was performed by electrophoresis, the proteins were electrotransferred to successive freezing and thawing in 10 mM Tris-HCl (pH 7.5), nitrocellulose ®lters and the ®lters were probed with either 0.1% Triton X-100 and 200 mM sucrose. The homogenate DO-7 hybridoma cell supernatant (diluted 1/5000) to reveal was centrifuged at 20 000 g for 30 min and the supernatant p53 or Ab-1 monoclonal antibody directed against PCNA was used for enzyme assays. Total SOD activity was (diluted 1/2000) to reveal PCNA or with the polyclonal determined by measuring the inhibition of nitro-blue- antibody anti-MnSOD (diluted 1/2000). The immuno-com- tetrazolium reduction by xanthine- as plexes were detected by chemiluminescence (ECL, Amersham previously described (Bravard et al., 1992). MnSOD activity Pharmacia Biotech). was determined as the remaining SOD activity after addition of 5 mM KCN. One unit of SOD activity corresponds to 50% inhibition of nitro-blue-tetrazolium reduction. The protein Luciferase assay content was measured using Biorad kit. Cells (26105) were plated in 6-well plates (Falcon) and transfected with 200 ml of DNA-calcium phosphate precipitate. To each DNA-precipitate, 50 ng/ml of pSVE-RenLuc (plasmid from Promega coding for the Renilla luciferase) was added as internal control and the total DNA concentration was balanced to 30 mg/ml with pBS (Stratagene). Each transfection was done in duplicate and repeated four to ®ve times. Six hours after adding the DNA precipitate, the cells were washed twice, re-fed with fresh medium and incubated at 378C for 18 h. The Acknowledgments cells were then proceeded as previously described (Deguin- We are grateful to A At®, J Jenkins, M Oren and V Rotter, Chambon et al., 2000), using the `Dual Luciferase Reporter for providing some of the plasmids used in this study, to assay' kit (Promega). Results are expressed as ®re¯y luciferase YS Ho for providing the rat SOD2 cDNA, to S Chouaib activity normalised to the Renilla luciferase activity. for the MCF-7/R-A1 cell line and to D Lane for providing DO-7 hybridoma cells. We wish to thank JC Lelouf for kindly providing his help in EMSA experiments and P May Preparation of nuclear extracts and electrophoretic mobility for critical reading of the manuscript. This work was shift assay (EMSA) supported by a grant from the `Association pour la Cells (70% con¯uency) from one 10-cm diameter Petri dish Recherche sur le Cancer' (ARC), and by a grant from were scraped, washed twice with ice-cold PBS, and lysed for `Electricite de France' (EDF). VB was supported by EDF.

References

Bienz-Tadmor B, Zakut-Houri R, Libresco S, Givol D and Buttke TM and Sandstrom PA. (1994). Immunol. Today, 15, Oren M. (1985). EMBO J., 4, 3209 ± 3213. 7±10. Borgstahl GE, Parge HE, Hickey MJ, Johnson MJ, Boissinot Caelles C, Helmberg A and Karin M. (1994). Nature, 370, M, Hallewell RA, Lepock JR, Cabelli DE and Tainer JA. 220 ± 223. (1996). Biochemistry, 35, 4287 ± 4297. Cai ZZ, Capoulade C, Moyretlalle C, AmorGueret M, Bravard A, Sabatier L, Ho€schir F, Ricoul M, Luccioni C Feunteun J, Larsen AK, Bressacdepaillerets B and and Dutrillaux B. (1992). Int. J. Cancer, 51, 476 ± 480. Chouaib S. (1997). Oncogene, 15, 2817 ± 2826.

Oncogene Cross-regulation of p53 and SOD2 gene expression P Drane et al 439 Chernov MV, Ramana CV, Adler VV and Stark GR. (1998). MiyashitaT,KrajewskiS,KrajewskaM,WangHG,Lin Proc. Natl. Acad. Sci. USA, 95, 2284 ± 2289. HK, Liebermann DA, Ho€man B and Reed JC. (1994). Cho G, Kang S, Seo SJ, Kim Y and Jung G. (1997). Biochem. Oncogene, 9, 1799 ± 1805. Mol. Biol. Int., 42, 949 ± 956. Munsch D, Watanabe-Fukunaga R, Bourdon JC, Nagata S, Church SL, Grant JW, Meese EU and Trent JM. (1992). May E, Yonish-Rouach E and Reisdorf P. (2000). J. Biol. Genomics, 14, 823 ± 825. Chem., 275, 3867 ± 3872. Church SL, Grant JW, Ridnour LA, Oberley LW, Swanson Murphy M, Ahn J, Walker KK, Ho€man WH, Evans RM, PE, Meltzer PS and Trent JM. (1993). Proc. Natl. Acad. Levine AJ and George DL. (1999). Genes Dev., 13, 2490 ± Sci. USA, 90, 3113 ± 3117. 2501. Deguin-Chambon V, Vacher M, Jullien M, May E and Noda A, Toma-Aiba Y and Fujiwara Y. (2000). Oncogene, Bourdon JC. (2000). Oncogene, 19, 5831 ± 5841. 19, 21 ± 31. Dugimont T, Montpellier C, Adriaenssens E, Lottin S, Oberley TD and Oberley LW. (1997). Histol. Histopathol., Dumont L, Iotsova V, Lagrou C, Stehelin D, Coll J and 12, 525 ± 535. Curgy JJ. (1998). Oncogene, 16, 2395 ± 2401. Ori A, Zauberman A, Doitsh G, Paran N, Oren M and Shaul El-Deiry WS, Tokino T, Velculescu VE, Trent JM, Lin D, Y. (1998). EMBO J., 17, 544 ± 553. Mercer WE, Kinzler KW and Vogelstein B. (1993). Cell, Pani G, Bedogni B, Anzevino R, Colavitti R, Palazzotti B, 75, 817 ± 825. Borrello S and Galeotti T. (2000). Cancer Res., 60, 4654 ± Farmer G, Friedlander P, Colgan J, Manley JL and Prives C. 4660. (1996). Nucleic Acids Res., 24, 4281 ± 4288. PeiXH,NakanishiY,TakayamaK,BaiFandHaraN. Frei B. (1999). FASEB J., 13, 963 ± 964. (1999). J. Biol. Chem., 274, 35240 ± 35246. Furlong EE, Rein T and Martin F. (1996). Mol. Cell Biol., Polyak K, Xia Y, Zweier JL, Kinzler KW and Vogelstein B. 16, 5933 ± 5945. (1997). Nature, 389, 300 ± 305. Hellin AC, Calmant P, Gielen J, Bours V and Merville MP. Price BD and Calderwood SK. (1993). Oncogene, 8, 3055 ± (1998). Oncogene, 16, 1187 ± 1195. 3062. Ho YS and Crapo JD. (1988). FEBS Lett., 229, 256 ± 260. Ronen D, Rotter V and Reisman D. (1991). Proc. Natl. Acad. Huang TT, Carlson EJ, Raineri I, Gillespie AM, Kozy H and Sci. USA, 88, 4128 ± 4132. Epstein CJ. (1999). Ann. NY Acad. Sci., 893, 95 ± 112. Roperch JP, Alvaro V, Prieur S, Tuynder M, Nemani M, Huang Y, Peng J, Oberley LW and Domann FE. (1997). Free Lethrosne F, Piou€re L, Gendron MC, Israeli D, Dausset Radic. Bio. Med., 23, 314 ± 320. J, Oren M, Amson R and Telerman A. (1998). Nat. Med., Iotsova V, Crepieux P, Montpellier C, Laudet V and Stehelin 4, 835 ± 838. D. (1996). Oncogene, 13, 2331 ± 2337. SchulerM,Bossy-WetzelE,GoldsteinJC,FitzgeraldPand Jenkins JR, Chumakov P, Addison C, Sturzbecher HW and Green DR. (2000). J. Biol. Chem., 275, 7337 ± 7342. Wade-Evans A. (1988). J. Virol., 62, 3903 ± 3906. Shatrov VA, Ameyar M, Bouquet C, Cai Z, Stancou R, Kaluzova M, Pastorekova S, Pastorek J and Kaluz S. (2000). Haddada H and Chouaib S. (2000). Int. J. Cancer, 85, 93 ± Biochimica et Biophysica Acta-Gene Structure and Expres- 97. sion, 1491, 20 ± 26. Shaulian E, Zauberman A, Ginsberg D and Oren M. (1992). Kiningham KK, Oberley TD, Lin S, Mattingly CA and St Mol. Cell. Biol., 12, 5581 ± 5592. Clair DK. (1999). FASEB J., 13, 1601 ± 1610. Sheng-Tanner X, Bump EA and Hedley DW. (1998). Radiat. KomarovaEA,ChernovMV,FranksR,WangK,ArminG, Res., 150, 636 ± 647. Zelnick CR, Chin DM, Bacus SS, Stark GR and Gudkov Smart P, Lane EB, Lane DP, Midgley C, Vojtesek B and Lain AV. (1997). EMBO J., 16, 1391 ± 1400. S. (1999). Oncogene, 18, 7378 ± 7386. KuoS,ChesrownSE,MellottJK,RogersRJ,HsuJLand Sun Y and Oberley LW. (1996). Free Radic. Biol. Med., 21, Nick HS. (1999). J. Biol. Chem., 274, 3345 ± 3354. 335 ± 348. Lee KC, Crowe AJ and Barton MC. (1999). Mol. Cell. Biol., Tang HY, Zhao K, Pizzolato JF, Fonarev M, Langer JC and 19, 1279 ± 1288. Manfredi JJ. (1998). J. Biol Chem., 273, 29156 ± 29163. Lee SG and Rho HM. (2000). Oncogene, 19, 468 ± 471. Wang QJ and Beck WT. (1998). Cancer Res., 58, 5762 ± 5769. Lee YI, Lee S, Das GC, Park US, Park SM and Lee YI. Wenk J, Brenneisen P, Wlaschek M, Poswig A, Briviba K, (2000). Oncogene, 19, 3717 ± 3726. Oberley TD and Schar€etter-Kochanek K. (1999). J. Biol. Li JJ, Oberley LW, Fan M and Colburn NH. (1998). FASEB Chem., 274, 25869 ± 25876. J., 12, 1713 ± 1723. Wispe JR, Clark JC, Burhans MS, Kropp KE, Korfhagen Li JJ, Oberley LW, St Clair DK, Ridnour LA and Oberley TR and Whitsett JA. (1989). Biochim. Biophys. Acta, 994, TD. (1995). Oncogene, 10, 1989 ± 2000. 30 ± 36. Mack DH, Vartikar J, Pipas JM and Laimins LA. (1993). Wu HY and Lozano G. (1994). J. Biol. Chem., 269, 20067 ± Nature, 363, 281 ± 283. 20074. Manna SK, Zhang HJ, Yan T, Oberley LW and Aggarwal Yonish-Rouach E, Deguin V, Zaitchouk T, Breugnot C, BB. (1998). J. Biol. Chem., 273, 13245 ± 13254. Mishal Z, Jenkins JR and May E. (1995). Oncogene, 11, Marklund SL, Holme E and Hellner L. (1982). Clin. Chim. 2197 ± 2205. Acta, 126, 41 ± 51. Zhao R, Gish K, Murphy M, Yin Y, Notterman D, Ho€man May P and May E. (1999). Oncogene, 18, 7621 ± 7636. WH, Tom E, Mack DH and Levine AJ. (2000). Genes McCord JM and Fridovich I. (1969). J. Biol. Chem., 244, Dev., 14, 981 ± 993. 6056 ± 6063. Zhong W, Oberley LW, Oberley TD and St Clair DK. (1997). Mendrysa SM and Perry ME. (2000). Mol. Cell. Biol., 20, Oncogene, 14, 481 ± 490. 2023 ± 2030.

Oncogene