Bcl-2 Inhibits a Fas-induced Conformational Change in the Bax N Terminus and Bax Mitochondrial Translocation*

Members of the Bcl-2 family of proteins control the cellular commitment to apoptosis, although their role in Fas-induced apoptosis is ill-defined. In this report we demonstrate that activation of the Fas receptor present on a human breast epithelial cell line resulted in a conformational change in the N terminus of the pro-apoptotic protein Bax. This conformational change appeared to occur in the cytosol and precede Bax translocation to the mitochondria. Overexpression of the anti-apoptotic protein Bcl-2 inhibited both the conformational change of Bax as well as its relocalization to the mitochondria. Bcl-2 overexpression did not, however, inhibit Fas-induced cleavage of both procaspase-8 and the pro-apoptotic protein Bid, indicating that Bcl-2 functions downstream of these events. These results suggest that the mechanism by which Bcl-2 inhibits Bax mitochondrial translocation and subsequent amplification of the apoptotic cascade is not by providing a physical barrier to Bax, but rather by inhibiting an upstream event necessary for Bax conformational change.

dysfunction occurs, evidenced by dissipation of the electrochemical gradient, uncoupling of the respiratory chain, and production of reactive oxygen species (5)(6)(7). However, recent evidence suggests that the Fas-induced cell death pathway may bifurcate with predominantly mitochondrial-dependent or -independent mechanisms (8,9).
Members of the Bcl-2 family of proteins interact to set a threshold for cell death, although they may also act independently (10,11). Recent data indicate that the pro-apoptotic molecule Bax is a monomeric and predominantly cytosolic protein in unstressed cells despite the presence of a C-terminal hydrophobic domain that is required for membrane insertion and Bax-induced cell death (12,13). During apoptosis Bax translocates to the mitochondrial membrane (12,14). Insertion of Bax into the mitochondrial membrane appears to be sufficient to induce cytochrome c release, an event that may act to amplify the apoptotic signal (8,15,16). Mutations in the C terminus can also result in both constitutively mitochondrial or cytoplasmic Bax protein (17). Deletion of the N terminus, or enforced dimerization of Bax, results in direct targeting of Bax to the mitochondria and enhanced cytotoxicity (18,19).
Bid is a pro-apoptotic molecule that belongs to a class of the Bcl-2 family that has little homology to other members, containing only a single BH domain (BH3) (20). Bid is unique among the BH3-only class of proteins in that it has been reported to be able to complex with either pro-or anti-apoptotic members of the Bcl-2 family, potentially enhancing or antagonizing their function (21). Bid, which lacks a C-terminal transmembrane domain, is predominantly cytosolic in living cells. It is cleaved and activated by caspase-8, and the truncated protein translocates to the mitochondria apparently triggering cytochrome c release (22)(23)(24). A recent model of Fas-induced apoptosis suggests that Bid causes a conformational change in the N terminus of Bax, resulting in mitochondrial cytochrome c release and amplification of the apoptotic cascade (25). However, it remains unclear whether the conformational change in Bax occurs prior or subsequent to mitochondrial membrane insertion. Hsu and Youle (13,26) have reported a detergentinduced conformational change in the N terminus of Bax (amino acids [12][13][14][15][16][17][18][19][20][21][22][23][24], which appears to be required for both homoand heterodimerization with Bcl-x L . However, the physiological significance and relevance to its role in apoptosis is unknown. We have previously demonstrated that activation of the Fas receptor leads to Bax translocation from the cytosol to the mitochondria that can be inhibited by Bcl-2 overexpression (27). In the present report we demonstrate that activation of the Fas receptor results in a conformational change in the N terminus of Bax, which appears to precede its translocation to the mitochondria. Our findings that overexpression of Bcl-2 does not inhibit Fas-induced cleavage of procaspase-8 or Bid, but does inhibit both the Bax conformational change and mitochondrial translocation, provide further insight into the Fas apoptotic pathway.

EXPERIMENTAL PROCEDURES
Cell Lines and Fas Treatment-Mid-passage MCF10A1 cells, a spontaneously immortalized non-tumorigenic human breast epithelial cell line, were obtained from the Barbara Ann Karmanos Cancer Institute. The transfection, selection, and growth conditions used for both control vector transfected (N10) and stable Bcl-2-expressing (B30) clones have been described previously (27). Anti-Fas monoclonal antibody (clone CH11, Upstate Biotechnology) was stored at Ϫ20°C and diluted directly into tissue culture medium. Cells were exposed to 100 ng/ml anti-Fas antibody plus 1 g/ml cycloheximide (Calbiochem).
Sample Preparation and Immunoblotting-For separation of subcellular cytosolic and nuclear/membrane fractions, 10 7 cells were washed twice with PBS at 4°C and resuspended in 1 ml of extraction buffer (50 mM PIPES, pH 7.0, 50 mM KCl, 5 mM MgCl 2 , 5 mM EGTA, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, 10 g/ml pepstatin A). Cells were lysed by five cycles of freezing in liquid nitrogen and thawing at 37°C. After microscopic examination with trypan blue verified that Ͼ95% cells were lysed, the crude lysate was centrifuged at 100,000 ϫ g for 1 h at 4°C. The resulting supernatant, which consisted of the cytosol, was separated from the pellet that contained the cellular membrane and organelles, and the pellet was resuspended in 1 ml of extraction buffer. Immunoblotting for the nuclear protein topoisomerase I (clone C-21, kindly provided by Dr. Y.-C. Cheng, Yale University School of Medicine, New Haven, CT) and the mitochondrial protein COX VIc (clone 3G5-F7-G3, Molecular Probes) revealed no cross-contamination of these organelles in the cytosolic fraction by this method of fractionation. As a cytosolic marker, a lactate dehydrogenase assay kit (Sigma) was used according to the manufacturer's instructions, and experiments were only deemed valid if Ͼ95% of the total lactate dehydrogenase activity was detected in the cytosolic fraction.
Whole cell extracts were prepared by lysing 10 7 cells in 1 ml of either Nonidet P-40 lysis buffer (50 mM Tris, pH 7.4, 150 mM NaCl, 0.2% Nonidet P-40, 50 mM NaF, 5 mM EDTA, 0.1 mM orthovanadate, plus protease inhibitor mixture (Sigma)) or CHAPS lysis buffer (10 mM HEPES, pH 7.4, 150 mM NaCl, 1% CHAPS, plus protease inhibitor mixture) on ice for 30 min. The crude lysate was then centrifuged at 10,000 ϫ g for 10 min at 4°C and the supernatant stored at Ϫ80°C. Protein concentration was estimated by the bicinchoninic acid assay method (Pierce) using a BSA standard. Equal amounts of protein, or equal volumes of cytosol and membrane fractions, were separated by SDS-PAGE and transferred to polyvinylidene difluoride membrane (Millipore). Primary antibodies used were rabbit polyclonal antisera raised against human Bcl-2, Bax (PharMingen), and actin (Sigma); a goat polyclonal antibody raised against a peptide corresponding to amino acids 176 -195 of human Bid (Santa Cruz); and a mouse monoclonal against COX VIc (Molecular Probes). Secondary antibodies used were horseradish peroxidase conjugates of anti-rabbit and anti-mouse IgG (Amersham Pharmacia Biotech) and anti-goat IgG (Santa Cruz). Chemiluminescent signals were generated by the addition of SuperSignal Ultra Chemiluminescent Substrate (Pierce). Proteins were visualized using radiographic film and quantified using a Bio-Rad GS-525 Molecular Imager System with Multi-Analyst version 1.0.2 software. Preliminary experiments had verified the linearity of protein quantification using the Molecular Imager System (data not shown).
Immunoprecipitation-Equal amounts of protein from whole cell lysates (1 mg) or cytosolic fractions (750 g) were used for immunoprecipitation. The KCl concentration of the cytosolic lysates was adjusted to 150 mM, and all samples were brought to a final volume of 450 l with 10 mM HEPES, pH 7.4, 150 mM NaCl. Samples were rotated for 2 h at 4°C with 6 g of monoclonal anti-Bax 6A7 (Trevigen) or 7 l of polyclonal anti-Bax (PharMingen). Antigen-antibody complexes were immobilized by rotation for 2 h at 4°C with protein G-agarose (Pierce) for monoclonal anti-Bax or protein A-Sepharose (Amersham Pharmacia Biotech) for polyclonal anti-Bax. The complexes were centrifuged and the supernatant removed with an aliquot being stored for subsequent analysis. The complexes were then washed five times with the same buffer used for the immunoprecipitation and subjected to SDS-PAGE and immunoblotted as described above.
Immunofluorescence and Confocal Microscopy-Cells were grown on No. 1 glass coverslips under the same conditions described above. Prior to Fas activation, cells were labeled with Mitotracker Red CM-H 2 XRos (Molecular Probes). Preliminary experiments had determined the minimum concentration and time required for optimal Mitotracker Red staining of MCF10A1 cells to be 2 M for 45 min at 37°C. Treatments with Fas-activating antibody were carried out as described above. At the indicated time points, cells were washed twice in PBS, fixed for 30 min in 3% formaldehyde, and permeabilized for 2 min with 0.2% CHAPS/PBS. Cells were then incubated with mouse anti-Bax 6A7 monoclonal antibody (Trevigen) diluted 1:300 in 3% BSA/PBS for 1 h at 37°C in a humidified chamber. Excess antibody was removed by washing the coverslips six times with PBS. Cells were then incubated with fluorescein isothiocyanate-labeled goat anti-mouse IgG (Zymed Laboratories Inc.) diluted 1:20 in 3% BSA/PBS, for 1 h protected from light. After washing six times with PBS, coverslips were mounted onto microscope slides using ProLong antifade mounting reagent (Molecular Probes). Control slides were stained with secondary antibody alone. Immunofluorescence of total Bax was performed using a rabbit anti-Bax polyclonal antibody, as we have described previously (27). Cells were viewed under a Zeiss Axioskop 20 fluorescence microscope and imaged using the Bio-Rad MRC 1024 Laser Confocal Imaging System. For each time point, at least 300 cells were counted and the numbers reported represent the average and S.E. of two experiments.

RESULTS AND DISCUSSION
We have shown previously that treatment of MCF10A1 N10 cells with anti-Fas antibody results in a redistribution of Bax from the cytosol to the mitochondria, which is inhibited by overexpression of Bcl-2 (B30 cells) (27). To study the conformation of Bax during this translocation we used the 6A7 monoclonal antibody, which recognizes a detergent-induced conformation of Bax, but not native Bax (13,17,26). Fig. 1 shows immunofluorescence with anti-Bax 6A7 (green) either alone (A-C, G, and H) or merged with mitochondrial staining (red) (D-F, I, and J). Untreated N10 and B30 cells were negative for staining with 6A7 (A, D, G, I), consistent with previous results using COS-7 cells (17).
Treatment of N10 cells with anti-Fas induced a conformational change in Bax, resulting in a punctate staining pattern (B, C, E, F). At early time points (2 h, B and E) a proportion of conformationally changed Bax appeared not to co-localize to the mitochondria. In some cells there was extremely little colocalization as indicated by the green color observed in Fig. 1E, while in other cells there was a mixture of non-co-localized and co-localized Bax indicated by green and yellow, respectively (Fig. 1E). At a later treatment time point (4 h) a significant proportion of 6A7 staining was yellow (Fig. 1F), indicating its co-localization to the mitochondria, while apoptotic morphology was also observed in some cells. The remainder of the intracellular green staining observed in Fig. 1F is consistent with our previous report that only 35-40% of total Bax had translocated to mitochondria at 4 h following Fas activation (27).
Overall, these data raise the possibility that the conformational change of Bax precedes translocation to the mitochondria. Notably, the Bax conformational change was observed before morphological evidence of apoptosis, and once associated with the mitochondria, Bax maintained this altered conforma-FIG. 1. Bax immunofluorescence with the conformation-specific antibody, 6A7. N10 (A-F) and B30 (G-J) cells growing on glass coverslips were prelabeled with Mitotracker Red, exposed to anti-Fas for the indicated times, fixed, and stained for conformationally changed Bax (green). Co-localization of green and red fluorescence is indicated by yellow coloration. Arrows indicate cells with apoptotic morphology. K, graphic representation of the percentage of N10 (open circles and squares) and B30 (filled circles and squares) cells positive for 6A7 staining (circles) or exhibiting a punctate total Bax staining pattern using a polyclonal antibody that recognizes both native and conformationally active Bax (total Bax, squares) (27). Results are the mean and S.E. of two experiments. N10 cells could not be analyzed at a 6-h time point due to cell death and loss from the coverslip. tion as the cells executed apoptosis. Strikingly, overexpression of Bcl-2 inhibited the conformational change of Bax (Fig. 1, H  and J). Fig. 1K shows graphically the time-dependent increase in immunoreactivity with anti-Bax 6A7 in N10 cells, while the conformational change is inhibited in B30 cells.
The proportion of N10 and B30 cells exhibiting a punctate Bax staining pattern, indicative of mitochondrial translocation, was assessed as we have described (27) using a polyclonal antibody that recognizes both native and conformationally active Bax (total Bax, Fig. 1K). The increase in the proportion of cells with punctate total Bax staining showed a similar time course, albeit slightly delayed, as that for conformationally changed Bax (6A7).
In an attempt to confirm the immunofluorescence results, immunoprecipitation with 6A7 was performed on the cytosolic fractions of N10 and B30 cells that had been lysed in the absence of detergents. As controls, immunoprecipitations were performed on cells lysed with 0.2% Nonidet P-40 or 1% CHAPS. Consistent with the results of Hsu and Youle (13), lysis with Nonidet P-40 induced a conformational change in Bax, allowing its immunoprecipitation with 6A7 ( Fig. 2A, lane a). In contrast, CHAPS lysis did not induce a Bax conformational change, prohibiting immunoprecipitation with 6A7 (lane b). In the cytosolic fraction of untreated N10 or B30 cells lysed in the absence of detergents, Bax was in its native conformation and therefore was not immunoprecipitated by anti-Bax 6A7 ( Fig.  2A, lanes c and f), in agreement with a previous study using mouse L929 cells (17). After 3-and 6-h treatment with anti-Fas, the conformationally changed Bax was detected in the immunoprecipitate of the cytosolic fractions of N10 cells ( Fig.  2A, lanes d and e). Fig. 2B shows Bax relocalization from the cytosol (lanes a-c) to the membrane compartment (lanes d-f), and subsequent proteolysis (lane f) during Fas activation, demonstrating that the decreased amount of Bax protein in the 6-h supernatant from the immunoprecipitation (Fig. 2A, lane e) reflects its relocalization to the membrane compartment.
Thus, we have demonstrated by two different methods that the conformational change in Bax can be detected in the cytosol of anti-Fas-treated N10 cells. These results differ from those of Desagher et al. (25) who described a conformational change in Bax but concluded that Bax is constitutively associated with the mitochondria and undergoes a conformational change within that location. The apparent differences in Bax localiza-tion between this study and that by Desagher et al. (25) may be due to the methods used to prepare subcellular fractions or to cell type-specific variations. Notably, our studies have not made the distinction between Bax that is merely associated with the mitochondrial membrane and that which is inserted (i.e. resistant to alkali extraction).
Overexpression of Bcl-2 significantly delayed Bax conformational change and membrane translocation. After 6 h of treatment of N10 cells, greater than 90% of Bax was membranebound, and the majority of the cells were apoptotic (Fig. 2B,  lane f, and data not shown). In contrast, there was no detectable conformational change or translocation of Bax within the first 6 h of treatment of B30 cells (Fig. 2A, lane g, Fig. 2B, lanes  k-m, and Fig. 1, H and J). At 6-h treatment and beyond there was a loss of Bcl-2 expression presumably due to proteolysis (Fig. 2B, lanes m and n) (28). At 12 h of treatment the Bax conformational change was detected in B30 cells (Fig. 2A, lane  h), and its membrane translocation and proteolysis were observed (Fig. 2B, lane n). Clearly, however, the Bax conformational change and membrane translocation in B30 cells occurred subsequent to the loss of Bcl-2 expression.
These results suggest that Bcl-2, a membrane-bound protein, inhibits the conformational change of Bax that occurs in the cytosol. A potential upstream event for Bcl-2 regulation of Bax conformation change is the cleavage of Bid. Overexpression of Bcl-2 did not alter the subcellular localization of Bid, which was cytosolic in both N10 and B30 cells (Fig. 3A). Fig. 3B demonstrates that overexpression of Bcl-2 did not inhibit processing of procaspase-8, which supports previously published reports that Bcl-x L functions downstream of caspase-8 activation (29,30). Fig. 3B also demonstrates that overexpression of Bcl-2 did not inhibit Fas-induced cleavage of Bid. We were, however, unable to detect a cleaved product in the membrane compartment of N10 or B30 cells. The time course of Bid cleavage does suggest that it preceded insertion of Bax since after 3 h of treatment, greater than 70% of Bid was cleaved (Fig. 3B), while there was only a slight (approximately 20%) relocalization of Bax to the membrane compartment in N10 cells (Fig. 2B, lanes  b and e).
Two previous reports suggest that Bid directly interacts with both Bax and Bcl-2 (20,25). To investigate interactions between Bid and Bax during Fas-induced apoptosis, N10 cells were treated, lysed in 1% CHAPS, and immunoprecipitated with a polyclonal anti-Bax antibody. Fig. 3C demonstrates that neither full-length nor cleaved Bid co-immunoprecipitated with Bax. We also re-probed the immunoblot of the anti-Bax 6A7 immunoprecipitation in Fig. 2A for Bid and were unable to detect any full-length or cleaved Bid that co-immunoprecipitated with conformationally changed Bax (data not shown). In addition, immunoprecipitations with anti-Bcl-2 antibody detected no Bid/Bcl-2 interactions (data not shown). Thus, using samples lysed with CHAPS we were unable to detect native Bid, full-length or cleaved, interaction with native Bax or Bcl-2. This does not rule out the possibility that Bid may be responsible for Bax conformational change, but our results indicate that if a Bid/Bax interaction occurs the interaction is not stable enough to survive immunoprecipitation.
To verify that Bax/Bid interactions do not occur under conditions that allow Bax/Bcl-2 dimerization (26), N10 and B30 cells were lysed in the presence of Nonidet P-40 and immunoprecipitated using polyclonal anti-Bax antibody. Fig. 3D shows that a significant amount (approximately 20%) of Bcl-2 was co-immunoprecipitated with Bax from the B30 cell lysate. Despite this observation, there was no indication that any of the Bid protein co-immunoprecipitated with Bax from either the N10 or B30 cell lysates. Interestingly, previous reports of Bid CHAPS are untreated N10 whole cell lysates prepared with buffer containing 0.2% Nonidet P-40 or 1% CHAPS, respectively. The immunoprecipitate was exposed to radiographic film three times longer than the Super. B, immunoblot of cytosolic (cyto) and membrane (mem) fractions of N10 and B30 cells, probed for Bcl-2 (p26) and Bax (p21). COX VIc (p11) demonstrates the presence of mitochondria in the membrane fractions only.
interacting with Bax or Bcl-2 have used recombinant proteins in vitro or cells expressing ectopic Bid that were lysed with Nonidet P-40 (20,25). Our data suggest that endogenous cellular Bid and Bax do not dimerize to any significant degree.
The structure of Bax is currently unknown; however, x-ray crystallography of the related family member Bcl-x L suggests that the N terminus and BH3 domain are adjacent to the C-terminal amino acid 197 (31). The Bax C terminus is known to be required for redistribution during apoptosis (12), while the N terminus negatively regulates Bax membrane insertion (18). Based on these and other studies it has been hypothesized that in its native conformation the N terminus of Bax masks domains at the C terminus that are required for mitochondrial translocation (17). Bax molecules containing C-terminal deletions or substitutions that cause its constitutive localization to mitochondria do not undergo N-terminal conformational change until the application of an apoptotic stress (17). In the same study, the temporal relationship between N-terminal conformational change and mitochondrial translocation of wildtype Bax was not delineated (17). Our data support a model in which, following application of an apoptotic stress such as Fas activation, Bax undergoes an N-terminal conformational change while residing in the cytosol, thereby unmasking the C-terminal hydrophobic domain allowing insertion into the mitochondrial membrane. In support of this model, high pH was recently shown to induce an N-terminal conformational change of cytosolic Bax in lymphoid cells (32).
It is unlikely that Bcl-2 can inhibit the Bax conformational change by direct interaction with Bax, since the two proteins appear to be in different subcellular locations when the conformational change occurs. Furthermore, Bcl-2/Bax interactions are not observed in immunoprecipitation experiments in the absence of nonionic detergents (13). 2 Our data demonstrate that overexpression of Bcl-2 does not sequester Bid to the membrane compartment, nor does it prevent Bid cleavage during Fas-induced apoptosis. Thus, Bcl-2 appears to function downstream of caspase-8 and Bid activation. It is possible, however, that Bcl-2 may inhibit activation of other caspases that are responsible for the Bax conformational change. These mechanisms are currently under investigation.