Caspase-3-dependent Cleavage of Bcl-2 Promotes Release of Cytochrome c *

Caspases are cysteine proteases that mediate apoptosis by proteolysis of specific substrates. Although many caspase substrates have been identified, for most substrates the physiologic caspase(s) required for cleavage is unknown. The Bcl-2 protein, which inhibits apoptosis, is cleaved at Asp-34 by caspases during apoptosis and by recombinant caspase-3 in vitro. In the present study, we show that endogenous caspase-3 is a physiologic caspase for Bcl-2. Apoptotic extracts from 293 cells cleave Bcl-2 but not Bax, even though Bax is cleaved to an 18-kDa fragment in SK-NSH cells treated with ionizing radiation. In contrast to Bcl-2, cleavage of Bax was only partially blocked by caspase inhibitors. Inhibitor profiles indicate that Bax may be cleaved by more than one type of noncaspase protease. Immunodepletion of caspase-3 from 293 extracts abolished cleavage of Bcl-2 and caspase-7, whereas immunodepletion of caspase-7 had no effect on Bcl-2 cleavage. Furthermore, MCF-7 cells, which lack caspase-3 expression, do not cleave Bcl-2 following staurosporine-induced cell death. However, transient transfection of caspase-3 into MCF-7 cells restores Bcl-2 cleavage after staurosporine treatment. These results demonstrate that in these models of apoptosis, specific cleavage of Bcl-2 requires activation of caspase-3. When the pro-apoptotic caspase cleavage fragment of Bcl-2 is transfected into baby hamster kidney cells, it localizes to mitochondria and causes the release of cytochrome c into the cytosol. Therefore, caspase-3-dependent cleavage of Bcl-2 appears to promote further caspase activation as part of a positive feedback loop for executing the cell.

Caspases are a family of cysteine proteases that are activated during programmed cell death (reviewed in Ref. 1). These proteases are synthesized as proenzymes, which are proteolytically cleaved into active heterodimers. Caspases can be grouped according to their substrate specificities, which are largely determined by the amino acids preceding the cleavage site aspartic acid residue (2). The substrate specificity of one group of caspases that includes caspases-6, -8 (MACH/FLICE), and -9 is V/LEXD, a site similar to that found in caspase proenzymes (2). Therefore, these caspases may function as initiators of a proteolytic cascade by activating pro-caspases to amplify a death signal (2). During Fas (CD95/Apo-1)-induced apoptosis, pro-caspase-8 is recruited to the Fas receptor, where caspase-8 becomes activated (3,4). Once activated, caspase-8 cleaves and activates downstream caspases as well as the prodeath protein Bid (5)(6)(7). The Bid cleavage fragment is targeted to mitochondria, where it promotes release of cytochrome c. In HeLa cell extracts, cytochrome c and Apaf-1 can activate procaspase-9 (Mch6/ICE-LAP-6), which subsequently activates caspase-3 (8). The substrate specificity of a second group, consisting of caspases-2, 3, and 7 is DEXD, a cleavage site similar to that found in many target proteins that are cleaved during apoptosis, suggesting that these caspases function during the effector phase of cell death (2,9). However, the caspase cascade is not yet well defined, and some caspases may serve both upstream and downstream functions. For example, caspase-2 has some characteristics of an upstream caspase (10), whereas caspase-6 is presumably acting as a downstream caspase when it cleaves lamin (11).
Mice that lack caspase-3 have diminished cell death of the developing brain (12,13). Remarkably, cells from mice deficient in caspase-3 show stimulus-specific defects in apoptosis that include delayed kinetics, incomplete chromatin condensation, and the absence of DNA fragmentation (13). Similarly, MCF-7 cells that do not express caspase-3 also do not undergo DNA fragmentation during cell death (14). The absence of DNA fragmentation in cells lacking caspase-3 may be because of the failure to cleave the endonuclease inhibitor, ICAD/DFF45 (15,16). Cleavage of ICAD releases the nuclease CAD, which enters the nucleus to produce DNA fragmentation (16,17).
A number of other proteins have been identified as caspase substrates. Some of these regulate homeostatic functions like DNA repair (for example PARP 1 ) and the cell cycle (such as the cyclin-dependent kinase inhibitor p21) (18,19). Caspase cleavage inactivates these proteins and may prevent homeostatic mechanisms from impeding apoptosis. Cleavage of substrates such as PAK2 and gelsolin creates active proteolytic products, which appear to mediate some of the morphological changes of apoptosis (20,21). Other caspase substrates such as Bcl-2, Bcl-x L , and Akt/protein kinase B normally function to prevent apoptosis, and cleavage of these proteins appears to inactivate their survival function (22)(23)(24)(25). In addition the cleavage fragments of Bcl-2 and Bcl-x L are potently pro-apoptotic (22,24). Cell death induced by fragments of Bcl-2 or Bcl-x L is blocked by caspase inhibitors, indicating a potential positive feedback loop that ensures the death of the cell.
The lack of protease inhibitors that distinguish between individual caspases has made it difficult to identify the physiologic caspase(s) required to cleave specific substrates inside cells. To examine the caspase(s) required to cleave Bcl-2, we utilized apoptotic extracts from 293 cells that were depleted of specific caspases by immunoprecipitation. Furthermore, MCF-7 cells that lack caspase-3 were used to test whether caspase-3 is required for Bcl-2 cleavage in cells during staurosporine-induced cell death. Finally, the pro-death Bcl-2 cleavage product was shown to localize to mitochondria and cause release of cytochrome c, suggesting that Bcl-2 cleavage promotes caspase activation and cell death via a mitochondrial pathway.
Cell lines, Transfections, and Induction of Apoptosis-The human neuroblastoma cell line SK-NSH (30), the human leukemia cell line HL-60, and the human breast cancer cell line MCF-7 were each cultured in RPMI 1640 medium (with glutamine) containing 10% fetal bovine serum. Baby hamster kidney (BHK) cells were obtained from American Type Culture Collection and cultured in Dulbecco's modified Eagle's medium with glutamine containing 10% fetal bovine serum. MCF-7 cells were transfected with 2.5 g of the indicated plasmid in 14 l of LipofectAMINE according to the manufacturer's instructions (Life Technologies, Inc.), and cells were harvested 33 h post-transfection. Cell death was induced in SK-NSH cells with 4 gray ionizing radiation (1 gray ϭ 100 rads) from a 137 Cs source with or without a 1-h pretreatment with protease inhibitors (DEVD-CHO, YVAD-CHO, N-acetyl-Leu-Leu-norleucinal, aprotinin, or leupeptin obtained from Sigma). Cell death was induced in HL-60 cells and MCF-7 cells by treatment with 2 M staurosporine (Alexis) for 14 h.
Mitochondrial Isolation-7.5 ϫ 10 5 BHK cells were transiently transfected with 20 g of the indicated DNA with 80 l of Lipo-fectAMINE. After 3 h, the cells were placed in fresh medium in the presence or absence of 100 M z-VAD-fmk (Enzyme Systems Products, Dublin, CA). 8 h after adding the DNA-liposome complexes, the cells were rinsed twice with cold phosphate-buffered saline and scraped into 1 ml of cold MSHE buffer: 0.21 M mannitol, 70 mM sucrose, 10 mM Hepes-KOH (pH 7.4), 1 mM EGTA, 1 mM EDTA, 0.15 mM spermine, 0.75 mM spermidine, with 5 mM dithiothreitol, 2 g/ml leupeptin, 2 M benzamidine-HCl, and 1 g/ml pepstatin added just before the experiment. Cells were Dounce-homogenized with 25 strokes on ice. Nuclei and unlysed cells were cleared by centrifugation at 500 ϫ g for 12 min at 4°C. An aliquot of the supernatant was placed at Ϫ80°C, and the rest of the supernatant was centrifuged at 9500 ϫ g for 9 min at 4°C to pellet mitochondria. The supernatant containing the cytosol was frozen at Ϫ80°C. The mitochondrial pellet was washed once in MSHE buffer, and intact mitochondria were pelleted by centrifugation at 8500 ϫ g for 9 min at 4°C. The mitochondrial pellet was resuspended in MSHE buffer and frozen at Ϫ80°C before preparing samples for immunoblot analysis. 35 S-labeled, in vitro translated Bcl-2 was incubated with 293 cell extract in the presence of 1 mM ATP, cleavage of Bcl-2 occurred (⌬N34 Bcl-2 in Fig. 1a). Specificity of the cleavage reaction was demonstrated by the lack of cleavage of in vitro translated Bax (Fig.  1a). The Bcl-2 cleavage product comigrated with in vitro translated Bcl-2 lacking amino acids 2-33 (⌬N34 Bcl-2), and Bcl-2 cleavage by the 293 cell extract was inhibited by mutating the cleavage site Asp-34 to Ala (D34A Bcl-2) (Fig. 1b). Therefore, the 293 extract cleaves Bcl-2 at Asp-34, the same site cleaved in apoptotic cells and by recombinant caspase-3 in vitro (22). Cleavage of Bcl-2 by the 293 apoptotic extracts was inhibited by the addition of the caspase inhibitor DEVD-CHO, indicating that Bcl-2 cleavage by the 293 cell extract requires caspases (Fig. 1c). Thus, in vitro cleavage of Bcl-2 by the 293 apoptotic extract recapitulates the same Bcl-2 cleavage site and protease specificity previously observed in cells undergoing apoptosis (22).

Bcl-2 Cleavage by 293 Cell Extracts-When
Caspase-3-dependent Cleavage of Bcl-2 by 293 Extracts-The Bcl-2 cleavage site, D 31 AGD 34 , corresponds to the substrate specificity of caspases 2, 3, and 7. Because caspases 3 and 7 are activated in the 293 extract (Fig. 2), we focused our biochemical analysis on these caspases. 293 cell extracts were immunodepleted with monoclonal antibodies that immunoprecipitate either caspase-3 or caspase-7 to create 293 extracts that were depleted (⌬) for each caspase. Control extracts (Fig. 2, panels a-c, C) were immunodepleted with monoclonal antibodies that recognize caspase-3 or caspase-7 by immunoblot but fail to immunoprecipitate the respective caspase. Immunodepletion of caspase-3 completely blocked cleavage of Bcl-2 by 293 extracts in vitro (Fig. 2a). However, immunoprecipitation with control (immunoblotting only) antibodies had no effect on cleavage of Bcl-2 (Fig. 2a). Therefore, in vitro cleavage of Bcl-2 by 293 extracts requires caspase-3. In contrast, immunodepletion of caspase-7 had no effect on cleavage of Bcl-2 (Fig. 2a), demonstrating that caspase-7 is dispensable for Bcl-2 cleavage in this model system. Although caspase-7 is neither required for Bcl-2 cleavage nor appears to contribute to Bcl-2 cleavage activity in the 293 extracts, it remains possible that Bcl-2 has some susceptibility to caspase-7 in other contexts.
Immunoblot analysis of these cleavage reactions confirmed that caspase-3 was successfully immunodepleted from the extracts (Fig. 2b, lane 2). Similarly, immunoblot analysis of these reactions for caspase-7 showed that caspase-7 was successfully immunodepleted (Fig. 2c, lane 4). Interestingly, only unprocessed pro-caspase 7 (p35, Fig. 2c, top arrow) and caspase-7 lacking the short amino-terminal pro-domain (middle arrow) were present in extracts lacking caspase-3 (Fig. 2c, lane 2). However, fully processed caspase-7 (p19, bottom arrow) was observed in extracts containing caspase-3 (Fig. 2c, lanes 1 and  3). This suggests that caspase-3 is required to fully process and activate caspase-7 in this cell-free model of apoptosis. Thus, caspase-3 may not only cleave substrates during the effector phase of apoptosis, but it may also play a role in the initiation or amplification of a caspase cascade during the initiator phase of apoptosis before the activation of downstream caspases. Indeed, the decreased developmental apoptosis that occurs in brains from mice lacking caspase-3 (12, 13) has been interpreted as evidence that caspase-3 has a role in the initiator phase of apoptosis in the brain (1). An alternative explanation for the absence of caspase-7 activation after caspase-3 immunodepletion is that another factor required for caspase-7 processing was immunoprecipitated with the caspase-3 antibody. Consistent with this possibility is the observation that in other cell types, caspase-7 has been reported to be proteolytically processed in the presence of a caspase-3 inhibitor (33) or in the absence of caspase-3 (13).
Cleavage of Bax during Apoptosis Is Inhibited by Leupeptin-The absence of Bax cleavage by 293 apoptotic extracts was somewhat surprising given that during chemotherapy-induced apoptosis of chronic lymphocytic leukemia cells, a smaller protein of approximately 18 kDa is produced, which is detected with an antibody to Bax (36). Similarly, an 18-kDa protein, which was reported to be a Bax cleavage product, is produced following Alphavirus-induced apoptosis (23). Moreover, we have observed similar cleavage of endogenous Bax following ionizing radiation-induced apoptosis of SK-NSH neuroblastoma cells, which produces an 18-kDa doublet (Fig. 3). It is possible that the protease that cleaves Bax in cells is not present or activated in the 293 extract. Alternatively, in vitro translated Bax may lack specific post-translational modifications that occur in apoptotic cells, rendering Bax a poor cleavage substrate in vitro. Finally, it is possible that one or more of the protease inhibitors used in the preparation of the 293 Proteins were analyzed as in Fig. 1. b, immunoblot analysis of the reactions in panel a using a caspase-3 monoclonal antibody (3-1-11). c, immunoblot analysis of the reactions in panel a with a caspase-7 monoclonal antibody (7-1-11). extracts, including phenylmethylsulfonyl fluoride, chymostatin, pepstatin, leupeptin, and antipain inhibits Bax cleavage.
To investigate the protease specificity of Bax cleavage during apoptosis, SK-NSH cells were treated with a variety of protease inhibitors before ionizing radiation. Twenty-four h after ionizing radiation, Bax was cleaved into a doublet, which migrated at approximately 18 kDa (Fig. 3). Pretreatment with the caspase inhibitor DEVD-CHO (Fig. 3, D) diminished Bax cleavage following ionizing radiation, suggesting that Bax cleavage may be downstream of caspase activation (Fig. 3). In contrast, YVAD-CHO (Fig. 3, Y) had no effect. However, when cells were pretreated with leupeptin, an inhibitor of serine and (noncaspase) cysteine proteases, Bax cleavage was abolished (Fig. 3 and Ref. 35), whereas PARP cleavage was not affected (data not shown). Furthermore, when cells were preincubated with the serine protease inhibitor, aprotinin, the upper band of the doublet was not detected, whereas preincubation with the (noncaspase) cysteine protease inhibitor N-acetyl-Leu-Leu-norleucinal prevented the appearance of the bottom half of the doublet (Fig. 3). Thus, the presence of protease inhibitors, such as 2 g/ml leupeptin present in the 293 extract, may prevent Bax cleavage in vitro. Taken together, these results suggest that serine and/or cysteine proteases other than caspases are involved in Bax cleavage.

Staurosporine-induced Cleavage of Bcl-2 in MCF-7 Cells
Requires Caspase-3-The results described above indicate that caspase-3 is required for cleavage of Bcl-2 in 293 apoptotic extracts. Although the 293 apoptotic extract cleaves Bcl-2 at the same site and with the same protease specificity as observed in intact cells, some cell-free systems do not always mimic all of the apoptotic events in cells (37). Therefore, to examine if caspase-3 is also required for Bcl-2 cleavage in intact cells, we used the breast cancer cell line MCF-7, which lacks endogenous caspase-3 (14, 38), because of a 47-base pair deletion within exon 3 of the caspase-3 gene (14). MCF-7 cells were treated with the protein kinase inhibitor staurosporine at a dose (2 M) that induces cell death in MCF-7 and HL-60 cells (Ref. 14 and data not shown) and also induces cleavage of endogenous Bcl-2 in HL-60 cells (Fig. 4a). Immunoblot analysis of MCF-7 cells treated with staurosporine showed a small amount of PARP cleavage but no Bcl-2 cleavage (Fig. 4b). When caspase-3 was transiently transfected into MCF-7 cells, a small amount of endogenous Bcl-2 was cleaved (Fig. 4b). However, treatment of caspase-3-transfected MCF-7 cells with staurosporine induced increased amounts of PARP and Bcl-2 cleavage (Fig. 4b). In contrast, treatment of caspase-1-transfected MCF-7 cells with staurosporine caused increased PARP cleavage compared with staurosporine alone but did not produce Bcl-2 cleavage (Fig. 4b).
The increased cleavage of endogenous PARP and Bcl-2 in the presence of both caspase-3 and staurosporine is explained by caspase-3 immunoblot analysis of these cells (Fig. 4c). As previously reported (14,38), MCF-7 cells lack caspase-3 expression (Fig. 4c). In MCF-7 cells transiently transfected with caspase-3 in the absence of staurosporine, most of the caspase-3 is unprocessed (p32 in Fig. 4c). However, with the addition of a death stimulus, most of the caspase-3 is now processed to the active heterodimer (p17 and p10 in Fig. 4c).
These results demonstrate that staurosporine is capable of initiating a death stimulus in MCF-7 cells that can promote caspase activation. Although the caspases activated in MCF-7 cells have some ability to cleave PARP, they fail to cleave Bcl-2. After transient transfection of caspase-3, staurosporine can activate caspase-3 and initiate cleavage of Bcl-2. What role does caspase-3 play in Bcl-2 cleavage in MCF-7 cells and in 293 apoptotic extracts? Since purified, recombinant caspase-3 can cleave Bcl-2 directly (22) at the same site cleaved by the 293 apoptotic extracts and in MCF-7 cells following staurosporine treatment, it is possible that caspase-3 is the protease responsible for cleavage of Bcl-2 in vivo. Although caspase-3 is required for Bcl-2 cleavage in 293 extracts and in MCF-7 cells treated with staurosporine, it is also possible that caspase-3 may activate a more downstream protease that directly cleaves Bcl-2.
Studies of caspase-3 knockout mice have identified stimulusspecific and tissue-specific requirements for caspase-3 in apoptosis (12,13). An important conclusion from these studies is that although distinct death stimuli in different tissues cause apoptosis with similar morphological and biochemical characteristics, individual caspases (such as caspase-3) may play very different roles depending on the specific context. Therefore, it remains possible that other death stimuli may cause caspase-3-independent cleavage of Bcl-2 or even that staurosporine in other cellular contexts may initiate Bcl-2 cleavage in the absence of caspase-3.
The lack of caspase-3 in MCF-7 cells prevents DNA fragmentation following staurosporine treatment (14) and, as reported here, also prevents cleavage of Bcl-2. Cleavage at Asp-34 of Bcl-2 not only inactivates Bcl-2 but also can produce a proapoptotic Bcl-2 proteolytic product (22). Therefore, failure to cleave Bcl-2 could delay cell death and contribute to tumorigenesis. Remarkably, in a cell line derived from a follicular lymphoma, the cleavage site Asp was mutated to His (34). Because mutation of Asp-34 inhibits caspase cleavage of Bcl-2 (22), mutation of the Bcl-2 cleavage site is an alternative mechanism to prevent Bcl-2 cleavage.
The Bcl-2 Cleavage Product Localizes to Mitochondria and Promotes Release of Cytochrome c-We have demonstrated that the Bcl-2 cleavage product, ⌬N34Bcl-2, has pro-death activity (22). Deletion of the carboxyl-terminal transmembrane domain abolishes the pro-apoptotic activity, suggesting that proper subcellular localization is required for this killing activity (22). To determine the subcellular localization of ⌬N34Bcl-2, a Flag epitope was added to ⌬N34Bcl-2. Flag-⌬N43Bcl-2 was as potent at killing BHK cells in a luciferase assay as untagged ⌬N34Bcl-2 (data not shown). Flag-⌬N43Bcl-2 was cotransfected with HA-tagged, full-length Bcl-2 and the caspase inhibitor P35 into BHK cells. Indirect immunofluorescence detected HA-Bcl-2 in a pattern consistent with previously published reports that Bcl-2 localizes to the outer nuclear envelope, endoplasmic reticulum, and mitochondria (Fig. 5a) (reviewed in Ref. 39). Indirect immunofluorescence detected the Flag-⌬N34Bcl-2 in a similar pattern (Fig. 5b), as expected since the two proteins share the same membrane-targeting transmembrane domain. Confocal microscopy demonstrated that the two proteins co-localized to several intracellular membrane com-partments (yellow in Fig. 5c) including organelles that appeared to be mitochondria (arrows). Similar patterns were observed when the proteins were transfected alone into BHK cells or when the ⌬N34Bcl-2 protein was HA-tagged (data not shown), demonstrating that the location of ⌬N34Bcl-2 did not require cotransfection of full-length Bcl-2 or the Flag epitope.
To confirm that Flag-⌬N34Bcl-2 localizes to mitochondria, mitochondria from BHK cells transfected with Flag-⌬N34Bcl-2 were isolated by differential centrifugation. Mitochondria were successfully isolated as measured by the Western blot of the outer mitochondrial membrane protein porin/VDAC. Porin was concentrated from the whole cytoplasm into the mitochondrial pellet fraction but was absent from the cytosolic supernatant (Fig. 6). The Flag-⌬N34Bcl-2 was similarly concentrated in the mitochondrial fraction, demonstrating that Flag-⌬N34Bcl-2 is targeted to the mitochondria. Some Flag-⌬N34Bcl-2 was also present in the cytosolic supernatant (Fig. 6), consistent with the ER localization observed by confocal microscopy (Fig. 5).
Mitochondria appear to play an important role in apoptosis by releasing cytochrome c to promote caspase activation (40 -42). Transient transfection of the pro-death protein Bax was reported to localize to mitochondria and cause release of cytochrome c as measured by indirect immunofluorescence and subcellular fractionation of mitochondria (32). Furthermore, the caspase-8 cleavage product of Bid, a pro-death Bcl-2 family member, localizes to mitochondria, where it triggers cytochrome c release (6,7). Because Flag-⌬N34Bcl-2 also localizes to mitochondria, Flag-⌬N34Bcl-2 may cause cell death by promoting cytochrome c release. To test this hypothesis, BHK cells were cultured on slides and transfected with control pSG5 plasmid, Bax, Bcl-x L , or Flag-⌬N34Bcl-2. At 10 h post-transfection, cells were co-stained for the transfected protein (red, Fig. 7) and for cytochrome c (green, Fig. 7). The cytochrome c staining of control cells (panel b) was punctate and in a mitochondrial distribution around a well defined nucleus (Fig. 7, N). In contrast, in a cell transfected with Bax (Fig. 7, arrow, panel  c), the cytochrome c was diffuse throughout the cytosol, and the nucleus was no longer clearly defined (panel d). However, when cells were transfected with the anti-death protein Bcl-x L (Fig.  7, panel e), cytochrome c retained its mitochondrial pattern (arrow, panel f) and the nucleus (N) was well defined. When cells were transfected with Flag-⌬N34Bcl-2 (Fig. 7, arrows,  panels g and i), cytochrome c was diffusely localized throughout the cells (arrows, panels h and j). Although neighboring, untransfected cells retained cytochrome c staining in a mitochondrial pattern around prominent nuclei, similar to cells transfected with Bax, the diffuse cytochrome c staining in cells expressing Flag-⌬N34Bcl-2 concealed the nuclei (Fig. 7, ar- rows, panels h and j). Remarkably, diffuse staining of cytochrome c was observed in the absence (data not shown) and presence of 100 M caspase inhibitor z-VAD-fmk (panels h and j) or with cotransfection of the caspase inhibitor P35 (data not shown), suggesting that cytochrome c release following Flag-⌬N34Bcl-2 transfection is independent of caspase activity.
When BHK cells were transfected with Bax, immunoblot analysis of isolated mitochondria and cytosol confirmed that cytochrome c moved from mitochondria into the cytosol as expected (Fig. 8a, left panel, lane 2). In contrast, cytochrome c was retained in the mitochondrial pellet when cells were transfected with control pSG5 (Fig. 8a, left, lane 1). Similar to Bax, Flag-⌬N34Bcl-2 also induced release of cytochrome c (Fig. 8a,  left, lane 3). Cotransfection of Flag-⌬N34Bcl-2 with the caspase inhibitor P35 (Fig. 8a, left, lane 4) or incubation with the caspase inhibitor, z-VAD-fmk (Fig. 8a, left, lane 5), did not inhibit release of cytochrome c into the cytosol. The quality of the subcellular fractionation was monitored by Western blot analysis of the mitochondrial pellet and the cytosolic supernatant for the outer mitochondrial membrane protein porin/ VDAC. In all of the samples, porin/VDAC localized to the mitochondrial pellet and was not present in the cytosolic supernatant, demonstrating that there was no mitochondrial contamination in the cytosol (Fig. 8a, right). Furthermore, despite lower expression of ⌬N34Bcl-2 compared with wild type Bcl-2 (Fig. 8c), subcellular fractionation of cells transfected with ⌬N34Bcl-2 had significantly more cytosolic cytochrome c (Fig. 8b, lane 4) compared with cells transfected with wild type Bcl-2 (Fig. 8b, lane 3). In addition, mitochondrial cytochrome c was depleted in cells expressing ⌬N34Bcl-2 (Fig. 8b, lane 2) compared with cells expressing wild type Bcl-2 (Fig. 8b, lane 1).
These results suggest that caspase cleavage of Bcl-2 is part of a positive feedback loop for regulating apoptosis. Caspase-3-de- FIG. 7. Flag-⌬N34Bcl-2 causes cytochrome c release. BHK cells were transiently transfected in panels a-b with pSG5, panels c-d with Bax, and panels e-f with Bcl-x L or panels g-j with Flag-⌬N34Bcl-2. At 10 h post-transfection, cells were identified with their respective antibodies (red, left panels), and cytochrome c location in the same cells was identified with an anti-cytochrome c antibody (green, right panels). Cells transfected with Bax or Flag-⌬N34Bcl-2 exhibit diffusely staining cytochrome c that obscures the nucleus (arrows in panels d, h, and j), in contrast to neighboring untransfected cells where the cytochrome c staining was punctate around a well defined nucleus (N). Treatment with a caspase inhibitor z-VAD-fmk (100 M) had no effect on relocalization of cytochrome c by ⌬N34Bcl-2 (panels g-j and data not shown). Results are representative of multiple transfected cells in four independent experiments. The scale bars represent 10 m. pendent cleavage of Bcl-2 removes the amino-terminal BH4 domain, which is required to inhibit apoptosis (22,23,(43)(44)(45). Cleavage may facilitate exposure of the pro-apoptotic BH3 domain of Bcl-2 as recently reported for Bid (6,7). In addition, the Bcl-2 cleavage product localizes to mitochondria and can cause release of cytochrome c into the cytosol. Cytosolic cytochrome c can serve as a co-factor for caspase-9, which further activates downstream caspases (8). Thus, cleaved Bcl-2 may amplify the caspase cascade. This is supported by the findings that cleaved Bcl-2 requires caspases to induce cell death (22) but not to induce the release of cytochrome c (Figs. 7 and 8). Therefore, cleavage of Bcl-2 may not only diminish the anti-apoptotic activity of Bcl-2 but, by producing a protein that can further promote caspase activation, may also accelerate the apoptotic pathway.
In summary, we have shown that caspase-3 is required for Bcl-2 cleavage by 293 extracts. Furthermore, in the MCF-7 cell line that lacks endogenous caspase-3 expression, induction of cell death by staurosporine does not cause Bcl-2 cleavage unless the cells are transfected with caspase-3. Together these results suggest that in some models of apoptosis, caspase-3 is required for Bcl-2 cleavage. We have also demonstrated that the Bcl-2 cleavage product localizes to mitochondria and can cause release of cytochrome c. This suggests a model where caspase cleavage of Bcl-2 produces a product that can promote further caspase activation to ensure the execution of the cell.