Irradiation-induced Translocation of p53 to Mitochondria in the Absence of Apoptosis*

The tumor suppressor protein p53 promotes apoptosis in response to death stimuli by transactivation of target genes and by transcription-independent mechanisms. Recently, it was shown that during apoptosis p53 can specifically translocate to mitochondria, where it physically interacts with and inactivates prosurvival Bcl-2 proteins. In the present study, we therefore investigated the role of mitochondrial translocation of p53 for the stress response of tumor cells. In various cell lines, DNA damage induced by either ionizing irradiation or topoisomerase inhibitors triggered a robust translocation of a fraction of p53 to mitochondria to a similar extent. Nevertheless, the cells succumbed to apoptosis only in response to topoisomerase inhibitors, but remained resistant to apoptosis induced by ionizing radiation. Irradiated cells became senescent, although irradiation triggered a functional p53 response and induced expression of p21, Bax, and Puma. Interestingly, even the targeted expression of p53 to mitochondria was insufficient to launch apoptosis, whereas overexpression of wild-type p53 induced Bax activation and apoptotic alterations. Together, these results suggest that, in contrast to previous reports, mitochondrial translocation of p53 does not per se lead to cell death and that this might constitute a mechanism that contributes to the resistance of tumor cells to ionizing radiation-induced apoptosis.

The tumor suppressor protein p53 is an important cellular integrator of stresses that triggers cell cycle arrest, senescence, DNA repair, or apoptosis. Upon activation by DNA damage or other stress-induced signaling pathways, p53 promotes the expression of a huge number of genes that coordinate the cellular stress response (1)(2)(3)(4). The importance of p53 for the maintenance of genomic integrity and prevention of tumorigenesis becomes evident by the fact that p53 is mutated and thereby inactivated in nearly 50% of all malignant human tumors (5,6). These mutations are primarily point mutations, and more than 90% of them are located within the central DNA-binding domain, underscoring the crucial role of p53 as a nuclear transcription factor (7).
As various apoptosis-relevant genes are under the control of p53, induction of apoptosis by p53 was initially believed to be mainly because of its function as a transcription factor. Furthermore, especially in view of the fact that p53 controls expression of death receptors such as CD95/Fas/APO1 or tumor necrosis factor-related apoptosis-inducing ligand receptor 2 (TRAIL-R2/KILLER/DR5), apoptosis induced by p53 was attributed to the extrinsic death receptor pathway (8 -10). Soon it became evident, however, that p53 mediates cell death predominantly via the intrinsic death pathway, because cells deficient for crucial components of this pathway such as Apaf-1 or caspase-9 were shown to be highly resistant to p53-dependent apoptosis (11). The intrinsic pathway is initiated through the release of pro-apoptotic factors from the mitochondria, including cytochrome c and Smac/DIABLO (12). Released cytochrome c then allows the formation of a high-molecular weight complex, the apoptosome, containing Apaf-1 and caspase-9, that initiates the activation of the caspase cascade. Apoptotic signaling events at the level of the mitochondria are tightly controlled by pro-and antiapoptotic Bcl-2 proteins including Bax, Bak, Bcl-2, Bcl-x L , and others. These Bcl-2 proteins interact with another group of pro-apoptotic proteins that share at least one common motif with Bcl-2, namely the BH3domain. BH3-only proteins have been proposed to be allosteric regulators of the Bcl-2 proteins and serve as sensors of apoptotic signaling. Upon interaction of anti-apoptotic Bcl-2 proteins with BH3-only proteins, both Bax and Bak undergo a conformational change in response to apoptotic stimuli. This event, which mediates exposure of an occluded N terminus (13,14), results in the oligomerization of Bax and Bak and in the formation of a mitochondrial outer membrane pore with the subsequent release of cytochrome c. Interestingly, many pro-apoptotic Bcl-2 members, including, for instance, Bax and the BH3-only proteins Puma, Noxa, and Bid, are under the transcriptional control of p53 (15)(16)(17)(18)(19).
In addition to the transcriptional role of p53, there is now accumulating evidence for transcription-independent, p53-mediated apoptotic pathways (20 -23). In some cell types, for example, p53-dependent apoptosis occurs in the absence of any gene transcription or translation or even upon expression of transcriptionally inactive p53 mutants (24 -26). The precise molecular mechanisms, however, of how p53 induces apoptosis independently of its transcriptional activity are largely unknown. Intriguingly, p53 has been reported to directly induce mitochondrial cytochrome c release in a Bax-dependent manner in vivo and in vitro, providing evidence for a direct signaling pathway from the p53 protein to the downstream caspase activation cascade (27). These observations were further supported by recent findings demonstrating the appearance of p53 in the cytoplasm or even its translocation to the mitochondria during DNA damage-and stress-induced apoptosis (28 -30). Moll and co-workers also found that bypassing the nucleus by targeting p53 expression directly to the mitochondria was sufficient to induce cell death in p53-deficient tumor cells (30). As a mechanistic explanation, it was proposed that p53, like BH3-only proteins, directly induces the permeabilization of the outer mitochondrial membrane by forming inhibitory complexes with anti-apoptotic Bcl-2 and/or Bcl-x L proteins. Green and colleagues (28) have recently shown that the proapoptotic Bcl-2 protein Bax is one of the principal cytoplasmic targets of p53 (28). In a comparable scenario, a direct interaction of p53 with mitochondria-localized Bak rather than Bax was proposed that results in the disruption of the cyto-protective Mcl-1⅐Bak complex and hence in the oligomerization and activation of Bak (31). Thus, compelling evidence suggests that p53 directly impacts on mitochondria, although it is still a matter of debate by which route the transcription-independent apoptosis is achieved.
We have previously reported that MCF-7 breast carcinoma cells are sensitive to anti-cancer drug-induced apoptosis, but are remarkably resistant to apoptosis induced by ionizing radiation (IR) 2 regardless of whether or not they express caspase-3 (32). This finding was especially paradox because MCF-7 cells express a transcriptionally active wildtype p53 protein. Thus, we wondered whether the radio-resistant phenotype of these cells might be caused by an improper transcriptionindependent function of p53. We found that, although IR induced an efficient translocation of p53 to the mitochondria, this process was insufficient to induce apoptosis in MCF-7 and HCT116 carcinoma cells. In addition, in several cell lines that display different apoptosis susceptibilities toward IR or chemotherapeutics, we did not find a correlation between the amount of mitochondrial p53 and apoptosis induction. Interestingly, Bax oligomerization and apoptotic alterations were only observed when cells were transfected with wild-type p53, whereas the expression of p53 targeted directly to the mitochondria did not evoke an apoptotic response. Our results therefore suggest that the resistance to apoptosis induction by mitochondrial p53 might contribute to the radio-resistant phenotype of tumor cells.

MATERIALS AND METHODS
Cells, Reagents, and Antibodies-MCF-7 breast carcinoma cells and its subclones MCF-7/casp-3 (33), MCF-7 cells stably expressing a GFP-Bax fusion protein, as well as MCF-7/p53siRNA (34) and MCF-7/casp-3/p53siRNA cells (35), were cultured in RPMI 1640. The human B-lymphoblastoid cell line SKW6.4 and RKO colon carcinoma cells were cultured in Iscove's modified Dulbecco's medium and Dulbecco's modified Eagle's medium, respectively. All media were supplemented with 10% heat-inactivated fetal calf serum, 10 mM glutamine, and 100 g/ml of each streptomycin and 100 units/ml penicillin (PAA Laboratories GmbH, Cölbe, Germany). The HCT116 and HCT116/p53 Ϫ/Ϫ cells were maintained in McCoy's 5A medium supplemented as above. The monoclonal Chk1 antibody was obtained from Transduction Laboratories (Heidelberg, Germany), and the monoclonal p53 antibody (Ab-6) from Calbiochem (Bad Soden, Germany). The p21 and poly(ADP-ribose) polymerase monoclonal antibodies were from BD Pharmingen, and the polyclonal antibody to high mobility group 1 (HMG1) and the monoclonal antibody to Tom20 from BD Bioscience. The monoclonal antibodies to Bak and Bcl-2 were purchased from Oncogene Research Products (San Diego, CA) and Novocastra Laboratories (Newcastle, UK), respectively. Bcl-x L and Bax antibodies were from Trevigen (Gaithersburg, MD), the conformation-specific anti-Bax-NT antibody was from Upstate Biotechnology. The polyclonal rabbit antibody recognizing Puma as well as the monoclonal antibody against proliferating cell nuclear antigen (PCNA) were purchased from Santa Cruz (Heidelberg, Germany) and the antibodies directed against caspase-3 and actin were from R&D Systems (Wiesbaden, Germany) and Sigma, respectively. Secondary antibodies for the immunofluorescence studies including chicken anti-mouse and anti-rabbit antibodies coupled to Alexa Fluor 488 and Alexa Fluor 594, respectively, or goat anti-mouse antibody coupled to Alexa Fluor 350 as well as Mitotracker Red were from Molecular Probes (Leiden, The Netherlands). Peroxidase-labeled antibodies to rabbit and mouse IgG were from Promega GmbH (Mannheim, Germany). The nuclear stain 4Ј,6-diamidino-2-phenylindole was from Sigma (Deisenhofen, Germany). For protease inhibition, buffers were supplemented with a protease inhibitor mixture (Roche Diagnostics).
Treatment and Transfection of Cells-Cells were exposed to ionizing radiation at the indicated dose using a Gammacell 1000 Elite (Nordion International Inc., Fleurus, Belgium) in the presence or absence of caffeine (1 mM, Sigma). Alternatively, cells were treated with the topoisomerase inhibitors etoposide (10 M) or camptothecin (5 M; both Sigma). The term "radio-resistant" that is used throughout the entire manuscript refers to "resistant to IR-induced apoptosis." Transfections of MCF-7 cells were performed with FuGENE reagent according to the instructions of the manufacturer (Roche Diagnostics). Mammalian expression vectors encoding for wild-type p53 (p53wt) or a p53 fused to the transmembrane domain of Bcl-2 (p53-CTM) were published before (30) and kindly provided by U. Moll. Stably transfected MCF-7 cells expressing a GFP-tagged Bax construct (36) were selected in G418 (1 mg/ml).
Immunofluorescence Microscopy-For immunofluorescence staining, cells were seeded on coverslides. At the indicated time points, cells were fixed for 20 min in ice-cold PBS containing 2% formaldehyde. Blocking and all following procedures were performed in PBS supplemented with 4% BSA and 0.05% saponin. Primary antibodies were applied at 4°C overnight at the appropriate concentration. Then, cells were washed three times for 20 min and incubated with the secondary antibody (1:500). Finally, samples were washed extensively and mounted in fluorescent mounting medium (Dako Corporation, Carpintera, CA) with or without 10 ng/ml 4Ј,6-diamidino-2-phenylindole. Pictures were taken on an Axiovert 135 Microscope (Zeiss, Germany) with an Apochromat ϫ63 oil immersion lens using OpenLab software (Improvision, Tübingen, Germany).
Flow Cytometric Analysis-For cell cycle analysis and detection of apoptosis, samples were prepared according to Nicoletti et al. (37). Briefly, cells were harvested by trypsinization, washed in cold PBS, and resuspended in lysis buffer containing 0.1% sodium citrate, 0.1% Triton X-100, and 50 g/ml propidium iodide. All flow cytometric analyses were performed on a FACScalibur (BD Biosciences) by using CellQuest analysis software. For each determination a minimum of 10,000 cells was analyzed.
Cell Extracts and Subcellular Fractionation-Total cell extracts were prepared in high-salt lysis buffer containing 1% Nonidet P-40, 20 mM HEPES (pH 7.9), 2 mM phenylmethylsulfonyl fluoride, 350 mM NaCl, 20% glycerol, 1 mM MgCl 2 , 0.5 mM EDTA, 0.1 mM EGTA, 0.5 mM dithiothreitol and protease inhibitors. For isolation of mitochondria, cells were harvested by scraping, centrifuged, washed, and resuspended in ice-cold buffer MB (400 mM sucrose, 50 mM Tris/HCl, pH 7.6, 1 mM EGTA, 5 mM 2-mercaptoethanol, 10 mM potassium phosphate, pH 7.6, and 0.2% BSA). After incubation on ice for 20 min, cells were Dounce homogenized with 40 strokes, and unbroken cells and nuclei were removed by centrifugation at 4,000 ϫ g for 1 min. The supernatant was centrifuged for 10 min at 15,000 ϫ g to pellet mitochondria. The resulting supernatant was removed and designated the cytoplasmic fraction, whereas the pellet enriched for mitochondria was resuspended in buffer MSM (300 mM mannitol, 10 mM potassium phosphate, pH 7.2, 0.1% BSA) and was referred to as the M1 fraction. To obtain highly purified mitochondria, the M1 fraction was loaded onto a discontinuous 1.2-1.6 M sucrose gradient buffered with 10 mM Tris/HCl (pH 7.5) and 0.1% BSA. After centrifugation at 70,000 ϫ g for 1 h, the interphase was harvested and diluted in MSM buffer at a ratio of 1:4. Mitochondria were centrifuged at 45,000 ϫ g for 15 min, and the resulting pellet was designated M2.
Western Blot Analysis-Protein concentrations were determined using the BCA assay kit (Pierce Biotechnology), and 15 g of protein per lane were loaded onto standard SDS-PAGE gels and separated at 200 V. The proteins were transferred onto a 0.2-m pore size polyvinylidene difluoride membrane (Amersham Bioscience) by semi-dry blotting at a constant current of 40 mA for 90 min in a buffer containing 10 mM CAPS (pH 11) and 10% MeOH. The membranes were blocked in PBS containing 4% BSA and 0.05% Tween 20 for 1 h, followed by an overnight incubation with the primary antibody at 4°C. After washing the membranes extensively, the appropriate secondary antibodies (1:5000) were applied for 1 h. Membranes were washed for an additional 2 h in PBS, 0.05% Tween 20, and proteins were visualized using ECL reagents (Amersham Biosciences).

Irradiation Induces a Functional p53 Response in the Absence of
Apoptosis-The tumor suppressor p53 induces apoptosis in response to DNA damage in diverse cellular systems. Nevertheless, MCF-7 breast carcinoma cells are resistant to apoptosis induced by IR regardless of whether or not they express caspase-3, although they harbor a functional p53 gene (32,35). Hence, we wondered whether the resistance of MCF-7 and MCF-7/casp-3 cells to IR-induced apoptosis could be caused by a mechanism rendering wild-type p53 unresponsive toward IR in these cells. However, exposure to IR induced significant p21 expression in both MCF-7 cell lines, a process that is most likely mediated via stabilization and transcriptional activation of p53 (Fig. 1A). In addition, irradiated MCF-7 and MCF-7/casp-3 cells displayed a persistent G 2 /M arrest (Fig. 1B) that correlated with the induction of cellular senescence as evidenced by staining of the senescence-associated ␤-galactosidase activity and a dramatic increase in the cell volume (Fig. 1C). IR-induced senescence was substantially reduced in MCF-7 and MCF-7/casp-3 cells stably transfected with a p53siRNA construct (35), imply-  NOVEMBER 4, 2005 • VOLUME 280 • NUMBER 44 ing that senescence induction requires p53 expression in both cell lines (Fig. 1D). Together, these results demonstrate that the p53 protein expressed in MCF-7 cells is functionally active, at least with regard to its transcriptional activity.

Resistance to Mitochondrial p53-induced Apoptosis
Mitochondrial Translocation of p53 in Response to Irradiation-Accumulating evidence suggests that p53 can also induce apoptosis via transcription-independent events that are only poorly understood. One of these events might be the DNA damage-induced translocation of p53 to the mitochondria that was previously described (29,30). To examine whether this process was a critical event for the apoptosis resistance of irradiated MCF-7 cells, we initially analyzed the subcellular localization of p53 by immunofluorescence microscopy. Compared with untreated cells, irradiation triggered an increase in nuclear size associated with the formation of micronuclei ( Fig. 2A). Unlike expression of the nuclear HMG1 that remained unaffected by IR, p53 expression was barely detectable in untreated MCF-7 cells, but was highly induced in irradiated cells ( Fig. 2A). However, IR-induced p53 was predominantly found in the nucleus, and no mitochondrial-associated p53 could be detected with this technique. Also the use of different antibodies recognizing either N-or C-terminal p53 epitopes did not result in the detection of endogenous mitochondria-associated p53 following IR (data not shown). Although this result would support our initial hypothesis that the radio-resistant phenotype of MCF-7 cells might be caused by an improper p53 function, our failure to detect mitochondrial translocation of endogenous p53 following exposure of the cells to IR might simply reflect the sensitivity restrictions of the immunofluorescence technique applied. Therefore, we investigated the possible IR-induced p53 translocation by Western blot analyses of mitochondria-enriched fractions. Using this technique, we found indeed a significant amount of p53 protein that was associated with mitochondria (M1 fraction) only in irradiated, but not in untreated MCF-7 or MCF-7/casp-3 cells (Fig. 2B). However, the mitochondria-enriched M1 fractions prepared by such a method might be contaminated with diverse membrane particles and organelles including endoplasmic reticulum and other subcellular compartments. In addition, we had to rule out the possibility that the M1 fractions contained also nuclear or cytoplasmic p53 that was not specifically translocated to mitochondria during IR treatment in vivo, but was simply co-sedimented during the purification procedure. Western blot analyses revealed that the mitochondrial M1 fractions of both cell lines were indeed contaminated with nuclear proteins such as the PCNA but, for unknown reasons, not with poly(ADP-ribose) polymerase (Fig. 2B). Thus, the observed p53 signals in the M1 fractions of irradiated cells might only reflect contaminations with nuclear p53. To avoid such nuclear contaminations, and to undoubtedly establish whether IR induces the mitochondrial translocation of p53, we further purified the mitochondria-enriched M1 fractions by discontinuous sucrose gradient centrifugation. Western blot analyses of samples obtained with such a purification procedure (M2 fractions) now clearly revealed an irradiation-induced translocation of p53 to the mitochondria in both MCF-7 lines (Fig. 2B). Only the mitochondrial marker proteins Tom20 and cytochrome c (data not shown), but not the nuclear poly(ADP-ribose) polymerase or PCNA proteins were detected together with p53 in the M2 fraction of irradiated cells, confirming the composition and purity of these samples. In addition, these data suggest that the mitochondrial p53 translocation must have occurred in vivo as a response of the cell to IR exposure, but not because of an in vitro co-sedimentation artifact caused by the preparation procedure. It was reported previously that a fraction of p53 only localizes to mitochondria under conditions that provoke p53-dependent apoptosis, but not in apoptosis-resistant cells (28,30). However, we now show that IR induces also the mitochondrial translocation of p53 in MCF-7 and MCF-7/casp-3 cells that are highly resistant to apoptosis induced by IR.
In an attempt to explain this discrepancy, we prepared sucrose gradient fractions from MCF-7/casp-3 cells that were exposed to IR in combination with the methylxanthine and ATM inhibitor caffeine, a treatment that efficiently induces apoptosis in these cells (35). When compared with purified M2 fractions of cells that were only irradiated, co-treatment with caffeine resulted in even further increased p53 levels at the mitochondria (Fig. 2C). At first glance, this result would therefore support the idea that mitochondrial translocation of p53 correlates with apoptosis induction. However, IR-induced apoptosis in the presence of caffeine proceeds in a p53-independent manner in MCF-7/casp-3 cells (35).
Mitochondrial p53 Translocation in HCT116 Cells-In light of the surprising finding that IR induces the mitochondrial translocation of p53 in MCF-7 and MCF-7/casp-3 cells that are both resistant to irradi-ation-induced apoptosis, we extended our investigation to another experimental system. As demonstrated by FACS analyses, HCT116 colon carcinoma cells were also resistant to apoptosis induced by high dose IR regardless of whether or not they express p53 (Fig. 3A). Similar to MCF-7 cells (32), HCT116 cells displayed a persistent irradiationinduced G 2 /M arrest, but were efficiently killed by the anti-cancer drug etoposide in a p53-dependent manner verifying their p53 status (Fig.  3A). Western blot analyses confirmed these observations and demonstrated the disappearance (and hence, activation) of the pro-caspase-3 protein predominantly in etoposide-treated HCT116 cells expressing wt-p53 (Fig. 3B, lower panel). Expression of p21 was clearly dependent on p53, as it was not observed in p53-deficient HCT116 cells (Fig. 3B). More importantly, p53 was found together with the mitochondrial marker protein Tom20 in the highly purified M2 fractions of both irradiated and etoposide-treated wild-type HCT116 cells, regardless of whether or not the cells were killed by this treatment (Fig. 3C). Again, the mitochondrial accumulation was specific for p53 because other nuclear proteins such as PCNA (Fig. 3C) and HMG1 (data not shown) were not detected in the M2 fraction. Thus, the specific mitochondrial translocation of p53 in the absence of apoptosis is not only restricted to the MCF-7 cell system, but occurs also in HCT116 colon carcinoma cells that are largely resistant to IR-induced apoptosis.
Mitochondrial Translocation of p53 Does Not Correlate with Apoptosis Susceptibility-It was recently reported that mitochondrial translocation of p53 occurs specifically during p53-dependent apoptosis, but not during p53-mediated cell cycle arrest or in p53-independent apoptosis (20). Our data, however, demonstrate p53 accumulation at the mitochondria following a treatment (IR) that does not induce apoptosis in two different cell lines. To investigate this discrepancy in more detail, we compared apoptosis susceptibility and mitochondrial p53 translocation also in the RKO colon carcinoma cell line. In agreement with a previous report (29), we found that these two events correlated well when the cells were treated for 6 h with camptothecin, which efficiently induced apoptosis (Fig. 4A) and p53 accumulation at the mitochondria (Fig. 4B). We detected, however, comparable amounts of mitochondrial p53 when the cells were exposed to IR, a treatment that does not induce apoptosis in these cells (Fig. 4). Similar results were obtained when the effects of both treatments were compared in HCT116 and MCF-7/ casp-3 cells (Fig. 4). In addition, we did not observe a correlation with regard to apoptosis sensitivity and mitochondrial p53 accumulation when we performed the experiments with the human B-lymphoblastoid cell line SKW6.4. Although these cells were equally susceptible to apoptosis induction by both treatments, a significantly larger portion of p53 accumulated at the mitochondria following incubation with camptothecin (Fig. 4). Interestingly, neither cell death nor p53 translocation appears to correlate with expression of the anti-apoptotic Bcl-2 protein at the mitochondria. Whereas the radio-resistant cell lines HCT116, MCF-7/casp-3, and RKO express extremely low and moderate levels of Bcl-2, respectively, we found very high Bcl-2 expression in IR-sensitive SKW6.4 cells (Fig. 4B).
Bax and Puma Expression Do Not Correlate with Apoptosis Susceptibility-Although we found in several cell lines no correlation between apoptosis susceptibility and mitochondrial p53 accumulation, we wondered whether the resistant phenotype of these cells to IR-induced apoptosis might be caused by defects in events downstream of mitochondrial p53. To address this possibility, we analyzed in MCF-7 cells expression of the p53-inducible pro-apoptotic proteins Bax and Puma that play key roles in apoptotic pathways mediated by mitochondria (17,23). Western blot analyses of whole cell extracts of MCF-7/ casp-3 and MCF-7/casp-3/p53siRNA cells showed that IR clearly induced up-regulation of Bax and Puma in a p53-dependent manner, whereas expression of other Bcl-2 family members such as Bak, Bcl-2, or Bcl-x L was neither affected by this treatment nor by the p53 status of the cells (Fig. 5A). Interestingly, when we analyzed the mitochondrial M2 fractions of several cell lines exposed to either IR or the chemotherapeutic drugs etoposide or camptothecin, we did not observe a correlation between Puma expression and apoptosis susceptibility. Puma was induced in all cell lines by either treatment regardless of whether or not the cells succumbed to apoptosis (Fig. 4B). In fact, at both time points analyzed (6 h and 24 h), IR induced an even higher Puma expression in the mitochondrial M2 fraction of IR-resistant HCT116 cells than etoposide or in IR-sensitive SKW6.4 cells. Furthermore, we found similar Puma levels at the mitochondria of MCF-7/casp-3 and RKO cells exposed to either IR or to chemotherapeutic drugs although only the latter treatment induces efficient apoptosis in these cells. Thus, neither the mitochondrial translocation of p53 nor expression of Bcl-2 or Puma appear to correlate with apoptosis sensitivity in the cell lines investigated. Although these data do not rule out a possible contribution of these and other Bcl-2 proteins toward the apoptosis-resistant phenotype of the cell lines investigated, the observation that Bax and Puma can be induced by IR in these cells indicates the functionality of the p53 pathway.
Bax Activation in MCF-7 Cells-Next, we tested the functionality of Bax that is known to be located upstream of caspase activation in the apoptosis signaling pathway. Bax exerts pro-apoptotic activity upon its conformational change and the exposure of an occluded N terminus, resulting in Bax oligomerization and cytochrome c release. Using an antibody that specifically detects the N terminus of such a conformationally changed, i.e. active, Bax protein, we could demonstrate activation of Bax only in etoposide-treated MCF-7 cells, but not in cells that were exposed to IR (Fig. 5C). Analysis of an overexpressed GFP-Bax protein and N terminus-specific Bax staining in these cells confirmed Bax activation exclusively in etoposide-treated but not in irradiated MCF-7 cells (Fig. 5C). This is in agreement with our previous study (35) demonstrating that exposure of MCF-7 or MCF-7/casp-3 cells to IR does not result in a decrease of the mitochondrial membrane potential, which usually coincides with Bax clustering. Together with the observed p53-dependent expression of Bax and Puma, these results not only verify the functional wild-type status of p53 in these cells, but also suggest that Bax or Puma are not the limiting factors responsible for the apoptosis-resistant phenotype of MCF-7 cells toward IR.
Apoptosis Induction by Wild-type p53 but Not by Mitochondria-targeted p53-So far, we have shown that, although IR induces the translocation of p53 to the mitochondria, this event does not induce apoptosis of MCF-7 cells. This is most likely the case because IR does neither trigger Bax activation nor does it lead to a decrease of the mitochondrial membrane potential. We therefore hypothesized that the levels of endogenous p53 detected at the mitochondria following IR might simply not be sufficient to induce apoptotic events such as Bax oligomerization. To test this, we transiently transfected MCF-7 cells with either wild-type p53 or with a p53 construct fused to the Bcl-2 transmembrane domain (p53-CTM) that directly targets expression of the fusion protein to the mitochondria as shown previously (30). Targeted expression was verified using immunofluorescence analyses that clearly demonstrated an exclusive expression of the latter construct at the mitochondria (Fig. 6A). Interestingly, only cells expressing the wild-type p53 protein displayed the typical signs of apoptosis, such as shrinkage and rounding, and only those cells could be stained with the antibody recognizing conformationally changed, active Bax protein (Fig. 6, B and C). Importantly, during the whole course of these experiments, we could never detect wt-p53 localized at the mitochondria. In contrast, MCF-7 cells that were transfected with the mitochondria-targeted p53-CTM construct were still attached to the plastic surface and appeared healthy (Fig. 6, B and C). This finding was in contrast to the original reports demonstrating that transfection of this construct resulted in efficient apoptosis induction even in p53-deficient cells (29,30). In addition, and consistent with our results, we could not detect conformationally changed, active Bax protein in MCF-7 cells transfected with the p53-CTM construct (Fig. 6B). Although we cannot completely rule out the  possibility that the different apoptotic outcome upon transfection of these two constructs is caused by their differential expression, our results provide strong evidence that neither endogenous p53 upon IR treatment nor transiently expressed p53 targeted to mitochondria are able to induce Bax oligomerization and subsequent apoptotic events in MCF-7 cells.

DISCUSSION
For the last few years, substantial evidence has emerged suggesting the existence of p53-dependent apoptosis pathways that do not require its transcriptional activity. Recently, p53 was demonstrated to translocate into the cytoplasm and even onto mitochondria during DNA damage-and stress-induced apoptosis (28 -30). Although it is still controversially discussed by which mechanism and interaction partners, in particular of the Bcl-2 family, p53 exerts an extranuclear, pro-apoptotic function, p53 may directly impact on mitochondria. p53 was shown to interact either with anti-apoptotic Bcl-2 proteins leading to the inhibition of their cytoprotective function (30), whereas in other studies a direct or indirect interaction with the pro-apoptotic Bcl-2 members Bak or Bax has been proposed (28,31). A mitochondrial translocation of p53 was observed in a variety of human and mouse cell lines and occurred specifically only in p53-dependent apoptosis such as in drug-, hypoxia-, or radiation-induced apoptosis, but was not detected during p53-mediated cell cycle arrest or in p53-independent apoptosis (20). In addition, mitochondrial p53 appears to contribute also to the physiological p53 stress response in vivo, as it could be detected following IR or intravenous application of etoposide in several radio-sensitive murine organs including thymus, spleen, testis, and brain, but not in radio-resistant organs such as liver and kidney (40).
Based on these and other reports it was surprising to find that IR treatment induced the translocation of p53 to mitochondria also in MCF-7 and MCF-7/casp-3 cells that display a persistent G 2 /M arrest following IR, but are resistant to apoptosis induction by this treatment (32). Although the IR-induced mitochondrial p53 levels were below the immunofluorescence detection limit, the presence of p53 at the mitochondria following IR could be readily demonstrated by Western blot analyses of subcellular fractions (M1 and M2). Only the M1 fractions were partially contaminated with nuclear proteins, whereas composition and purity of the M2 fractions could be confirmed by the absence and presence of nuclear and mitochondrial marker proteins such as HMG1, PCNA, and Tom20, respectively. Irradiation-induced translocation of p53 to mitochondria without any signs of apoptosis was not restricted to the MCF-7 cell line, but was also a prominent feature of similarly treated HCT116 and RKO cells. In addition, we did not observe a correlation between mitochondrial p53 and the extent of apoptosis as exposure of various cell lines to IR, etoposide, or camptothecin resulted in comparable levels of mitochondrial p53, whereas their apoptotic responses toward these stimuli differed dramatically. Hence, our data that are in line with a very recent report showing that a transcriptional inactive p53 mutant is completely unable to induce apoptosis in vitro and in vivo (41), demonstrate that the physical translocation of p53 to mitochondria is not necessarily associated with the induction of apoptosis as it was proposed previously (29,30).
Although the exact reason for this discrepancy is presently unknown, one possible explanation might be that too little amounts of p53 protein are translocated to the mitochondria following IR treatment. Provided that the apoptotic machinery is only activated above a certain threshold level of mitochondrial p53, the IR-induced p53 levels at the mitochondria might simply be not sufficient to induce apoptosis. Several observations, however, argue against this explanation and clearly suggest that even elevated mitochondrial p53 levels do not trigger an apoptotic response at least in MCF-7 cells.
First, as quantitated by densitometry of the immunoblots, mitochondrial p53 comprised ϳ2% of the total induced p53 in MCF-7 or MCF-7/casp-3 cells. This amount is very similar to the mitochondrial fraction of p53 in other cell types in which mitochondrial p53 was originally proposed to provoke apoptosis (29). Second, although we found even further increased p53 levels associated with mitochondria in M2 fractions of MCF-7/casp-3 cells that were exposed to IR in the presence of caffeine than compared with cells exposed to IR alone, apoptosis induced by the combined treatment of IR and caffeine proceeds in a p53-independent manner (35). Hence, these results suggest that elevation of mitochondrial p53 levels induced by this treatment does not contribute to the rate of apoptosis in these cells. In addition, these data demonstrate that mitochondrial translocation of p53 can be also observed in p53-independent apoptosis systems, a finding that is in sharp contrast to previous studies (30,40). Third, it is well accepted that the ratio of pro-and anti-apoptotic Bcl-2 proteins is an important factor determining the fate of a DNA-damaged cell. As p53 was shown to interact either with the pro-apoptotic Bcl-2 proteins Bax or Bak (28,31) or with the anti-apoptotic proteins Bcl-2 and Bcl-x L (30), it can be assumed that aberrant expression of either of these proteins would interfere with the pro-apoptotic activity of mitochondrial p53. However, MCF-7 cells do not express abnormal levels of either Bcl-2 or Bcl-x L and Bax or Bak, respectively. In fact, IR treatment of MCF-7/ casp-3 cells solely increased expression of the pro-apoptotic proteins Bax and Puma without affecting Bak, Bcl-2, or Bcl-x L levels. Hence, also these results do not favor the above mentioned threshold hypothesis. This view is also supported by our findings that Bcl-2 and Puma levels neither correlate with apoptosis sensitivity nor with the amount of p53 translocated to the mitochondria following exposure of the cells to IR, camptothecin, or etoposide.
Finally, and most importantly, we have shown that even MCF-7 cells that are deficient for caspase-3 displayed oligomerized Bax (Fig. 6) and Bak (data not shown) and succumbed to apoptosis following overexpression of a wt-p53 construct. In contrast, when MCF-7 cells were transfected with a mitochondria-targeted expression construct of wt-p53 (p53-CTM) that was previously postulated to efficiently induce apoptosis (30), no activation of Bax or Bak (data not shown) or any other apoptotic sign was observed in these cells. Therefore, our data demonstrate that not only the IR-induced mitochondrial translocation of endogenous p53, but even the targeted overexpression of exogenous mitochondrial p53 is insufficient to induce apoptosis of MCF-7 cells. As chemotherapeutic drugs or overexpression of wt-p53 were sufficient to induce Bax activation and apoptosis, our data suggest that the radioresistant phenotype of these cells is not caused by a general resistance or unresponsiveness to p53-dependent cell death. Nevertheless, one has to keep in mind that apoptosis induction by overexpression of wt-p53 is most likely achieved via its transcriptional activity as overexpressed wt-p53 was mainly located in the nucleus, but was never detected at the mitochondria.
It is still possible that the inability of mitochondrial p53 to induce Bax oligomerization and apoptosis is caused by the absence or presence of certain inhibitory or stimulatory factors. In addition, it might well be the case that the interaction of mitochondrial p53 with Bcl-2 or Bcl-xL is for unknown reasons hampered in these cells, a scenario that would easily explain their unresponsiveness toward mitochondrial p53. Also the lack of post-translational modifications of p53 itself that might be required for its pro-apoptotic activity at the mitochondria might account for the observed resistance. With regard to phosphorylation and acetylation, no major differences were observed in the modification patterns between the pro-apoptotic mitochondrial and nuclear p53 proteins (39). However, these data do not rule out that other post-translational events, such as ubiquitinylation or sumoylation, might play a role in apoptosis induction by mitochondrial p53. Finally, we cannot exclude the possibility that translocation of p53 to mitochondria is merely a consequence of DNA damage signaling without any implications for the fate of the cell.
In summary, we show that the IR-induced translocation of p53 to the mitochondria is not unequivocally associated with apoptosis and that mechanisms might exist that counteract the pro-apoptotic ability of mitochondrial p53. This could be convincingly demonstrated in several cell lines of different origin. The reason why most of these cell lines are resistant to apoptosis induction by mitochondrial p53 is completely unknown, but is especially intriguing as it might represent a potential mechanism that contributes to the occurrence of a radio-resistant tumor cell phenotype.