Cytochrome c Release upon Fas Receptor Activation Depends on Translocation of Full-length Bid and the Induction of the Mitochondrial Permeability Transition*

In Jurkat cells Bid was cleaved upon activation of the Fas receptor with an anti-Fas antibody. The caspase-8 inhibitor benzyloxycarbonyl-Ile-Glu(OMe)-Thr-Asp(OMe)-CH2F (IETD) prevented the cleavage of Bid and the loss of viability. The nuclear enzyme poly(ADP-ribose)polymerase (PARP) was also cleaved upon the activation of caspases, and IETD similarly prevented PARP cleavage. The PARP inhibitor 3-aminobenzamide (3-AB) restored the cell killing in the presence of IETD, an effect that occurred without restoration of the cleavage of Bid or PARP. In the presence of 3-AB and IETD, translocation occurred of full-length Bid to the mitochondria. The induction of the mitochondrial permeability transition (MPT) was documented by the cyclosporin A (CyA) sensitivity of the release of cytochrome c, the release of malate dehydrogenase from the mitochondrial matrix, the loss of the mitochondrial membrane potential, and the pronounced swelling of these organelles, as assessed by electron microscopy. In addition to preventing all evidence of the MPT, CyA prevented the loss of cell viability, without effect on the cleavage of either Bid or PARP. The prevention of PARP cleavage by inhibition of caspase-3 resulted in a 10-fold activation of the enzyme and a resultant depletion of NAD and ATP. The PARP inhibitor 3-AB prevented the loss of NAD and ATP. Depletion of ATP by metabolic inhibitors similarly prevented the cell killing. It is concluded that the cleaving of PARP in Fas-mediated apoptosis allowed expression of an energy-dependent cell death program that included the translocation of full-length Bid to the mitochondria with induction of the MPT.

The participation of mitochondria in the pathogenesis of apoptosis is generally held to be initiated by the interaction of these organelles with one or more of the proapoptotic members of the Bcl-2 family of proteins. A substantial recent effort has addressed both the interdependence of these proteins and the nature of their action on the mitochondria, whether individually or as components of a proapoptotic cascade.
Bax and Bak are members of the Bcl-2 family and share 3 homology domains that are designated BH1 to BH3 (1,2). By contrast, Bid and Bik share sequence homology with the other Bcl-2 proteins only at the BH3 or "death" domain (BH3-only proteins) (1,2). This feature suggested that Bid might function to receive death signals in the cytosol from upstream events and then, upon translocation to the mitochondrion, interact with membrane-bound downstream effector pathways (3).
Several different scenarios have been proposed that involve Bid acting alone or in association with other proapopotic proteins, specifically Bax or Bak. In staurosporine-induced apoptosis in HeLa cells, it is argued that Bid interacts with Bax to trigger a change in Bax conformation that leads, in turn, to the insertion of Bax into the outer mitochondrial membrane (4,5). In this model, Bid-induced Bax membrane insertion is independent of caspase activation (4). An association of Bid and Bax on the surface of the mitochondrial outer membrane is proposed to bring about changes in the conformation of Bax that are required for membrane insertion and activation. It deserves note, however, that embryonic fibroblasts prepared from mice deficient in Bid are not resistant to the cell killing by staurosporine (6).
Bid has been implicated in the TNF 1 and Fas death signal pathways. It is held that cytosolic p22 Bid represents an inactive conformation of the molecule that is cleaved by proteolysis to generate an active p15 or truncated Bid (tBid) (7)(8)(9). Engagement of either the TNF or Fas receptor recruits first FADD and then the proenzyme form of caspase-8. Activated caspase-8 then cleaves Bid to generate tBid that translocates to the mitochondria. tBid induces release of cytochrome c by a mechanism that is independent of the induction of the mitochondrial permeability transition (4,5,10,11). The ability of caspase-8 inhibitors to block Fas and TNF signaling is explained by the disruption of Bid processing. Hepatocytes, fibroblasts, or thymocytes derived from Bid-deficient mice activate caspase-8 but survive longer upon treatment with anti-Fas antibody or TNF (6). Other inducers of apoptosis, including staurosporine, dexamethasone, and ␥-radiation, which presumably do not act by activating either the TNF or Fas receptors, kill Bid-deficient cells to the same extent as wild type ones (6).
An interaction of Bid with Bak has also been proposed. tBid could not cause cytochrome c release from isolated Bak-deficient mitochondria (12). Resident in mitochondrial membranes, Bak is envisioned to undergo allosteric activation as a result of a conformational change induced by its interaction with tBid.
In the present study, we have utilized Jurkat cells, a human T-cell leukemia/lymphoma cell line, to re-examine the role of Bid in the cell killing that occurs subsequent to activation of the Fas receptor. Jurkat cells are very sensitive to activation of the Fas receptor. Caspase inhibitors readily prevent the resulting apoptosis. In addition, Jurkat cells do not contain Bax. The data presented below argue that upon activation of the Fas receptor full-length Bid translocates to the mitochondria, where it produces the release of cytochrome c as a result of induction of the MPT. The ability of caspase inhibitors to prevent the killing of Jurkat cells cannot be attributed to an inhibition of Bid processing. Rather, the likely basis of their anti-apoptotic effect is the prevention of PARP cleavage and the resultant activation of the enzyme.
Treatment Protocols-In all experiments anti-Fas/APO-1 human monoclonal antibody, clone CH-11 (Medical and Biological Laboratories, Nagoya, Japan) was added to a final concentration of 100 ng/ml. Ceramide (Biomol Research Laboratories, Plymouth Meeting, PA) was dissolved in Me 2 SO and added at a final concentration of 10 M. Where indicated the cells were pretreated for 30 min with one or more of the following reagents before addition of either anti-Fas antibody or ceramide. The caspase-8 inhibitor Z-Ile-Glu(OMe)-Thr-Asp(OMe)-CH 2 F (IETD) (Kamiya Biomedical Co., Seattle, WA) was dissolved in Me 2 SO and added at 50 M. Ac-Asp-Glu-Val-aspartic acid aldehyde, (DEVD) (Bachem Bioscience, King of Prussia, PA), a caspase-3 inhibitor, and 3-aminobenzamide (Sigma) were dissolved in Me 2 SO and added to the cells at 200 M. CyA (Biomol) was dissolved in Me 2 SO and added at a final concentration of 5 M. ArA (Biomol) was dissolved in phosphatebuffered saline and added at a final concentration of 50 M. In all cases the vehicles used to prepare stock solutions of the reagents had no effect (at 0.1% v/v) on the cells or the parameters measured.
Bad, Bid, Bik, Bak, and Bax Expression-Cells (1 ϫ 10 6 ) were pelleted at 700 ϫ g (5 min at 4°C) and lysed in 100 l of cell lysis buffer (20 mM Tris, pH 7.4, 100 mM NaCl, 1% Triton, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, 10 g/ml aprotinin). After 30 min on ice, the lysates were clarified by centrifugation (10 min at 4°C), and the supernatant was collected. Protein concentration was determined by the bicinchoninic acid assay (Sigma). Equivalent amounts of protein (40 g/lane) were electrophoresed on 15% SDS-polyacrylamide gels. Kaleidoscope Prestained Standards (Bio-Rad) were used to determine molecular weight. The gels were then electroblotted onto nitrocellulose membranes. Bid was detected with a rabbit polyclonal antibody (BIO-SOURCE, Camarillo, CA). Bax was detected with a rabbit polyclonal antibody (N-20) (Santa Cruz Biotechnology, Santa Cruz, CA). Bik was detected with a goat polyclonal antibody (N-19) (Santa Cruz Biotechnology). Bak was detected with a goat polyclonal antibody (N-20) (Santa Cruz Biotechnology). Bad was detected with a mouse monoclonal antibody (C-7) (Santa Cruz Biotechnology). In each case, the relevant protein was visualized by staining with the appropriate secondary horseradish peroxidase-labeled antibody followed by enhanced chemiluminescence.

FIG. 1. Proapoptotic proteins in Jurkat and HeLa cells.
A, whole lysates obtained from Jurkat and HeLa cells were analyzed for Bax, Bid, Bik, and Bak by SDS-PAGE and Western blotting as described under "Experimental Procedures." B, total RNA isolated from Jurkat and HeLa was used to obtain the cDNA for Bak, Bid, and Bax by RT-PCR as described under "Experimental Procedures." The PCR products were resolved on a 1.5% agarose gel. RT-PCR-To analyze Bid, Bax, and Bak RNA by RT-PCR, total RNA was isolated from Jurkat and HeLa cells. The following sense and antisense primers were designed and utilized for reverse transcription and polymerase chain reaction amplification of the cDNA for Bid, Bax, and Bak from the isolated total RNA: ATGGACTGTTGAGGTCAA-CAAC (sense) and TCAGTCCATCCCATTTCTGGCT (antisense) for Bid; ATGGACGGGTCCGGGGAGCAG (sense) and TCAGCCCATCT-TCTTCCAGAT (antisense) for Bax; and ATGGCTTCGGGGCAAGGC-CCAGGTCCTCCC (sense) and TCATGATTTGAAGAATCTTCGTAC-CACAAA (antisense) for Bak. The PCR products were resolved on a 1.5% agarose gel using a 100-bp marker (MBI Fermentas, Hanover, MD) as a standard.
Isolation of Cytosol and Mitochondrial Fractions-Cells (25 ϫ 10 6 ) were plated in 75-cm 2 polystyrene flasks. Following treatment the cells were harvested by centrifugation at 750 ϫ g for 10 min at 4°C. The cell pellets were resuspended in 1 ml of 20 mM HEPES-KOH, pH 7.5, 10 mM KCl, 1.5 mM MgCl 2 , 1 mM EDTA, 1 mM EGTA, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, 10 g/ml aprotinin, and 250 mM sucrose. The cells were broken open with 6 passages through a 26-gauge needle applied to a 1-ml syringe. The homogenate was centrifuged at 750 ϫ g for 10 min at 4°C to remove nuclei and unbroken cells. The supernatant was transferred to a high speed centrifuge tube. Centrifugation was conducted at 10,000 ϫ g for 15 min at 4°C. The resulting mitochondrial pellet was lysed in 50 l of 20 mM Tris, pH 7.4, 100 mM NaCl, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, 10 g/ml aprotinin. The supernatant from the 10,000 ϫ g spin was centrifuged at 100,000 ϫ g (60 min at 4°C), and the supernatant was used for preparation of cytosol. The cytosolic fraction was concentrated through a Microcon YM-10 Centrifugal Filter Device (Millipore, Bedford, MA). Protein content of the fractions was determined by the bicinchoninic acid assay (Sigma).
Western Blot of Cytochrome c, Malate Dehydrogenase, Cox4, and ␤-Actin-Mitochondrial and cytosolic fractions (equal amounts from 20 to 40 g of protein) were analyzed by electrophoresis and electroblotting as described above. Cytochrome c was detected by a monoclonal antibody (PharMingen, San Diego, CA). MDH was detected with a sheep polyclonal horseradish peroxidase-labeled antibody (Rockland, Gilbertsville, PA). To ensure that the same content of mitochondrial and cytosolic protein was loaded in each case, the same nitrocellulose blots used for the determination of cytochrome c or MDH were also stained for either both cytochrome c oxidase subunit IV (Cox4) or ␤-actin (antihuman mouse monoclonal antibodies against either Cox4 (CLONTECH, Palo Alto, CA) or ␤-actin (Santa Cruz Biotechnology).
Measurement of the Mitochondrial Membrane Potential-The induction of the mitochondrial permeability transition was determined in intact cells as the CyA-sensitive release of 3,3Ј-dihexyloxacarbocyanine (Molecular Probes Inc., Eugene, OR) by modification of the method of Krippner et al. (13). Briefly, 5 ϫ 10 5 cells in 500 l of serum-free RPMI medium were plated in 24-well plates. After overnight incubation, the cells were treated as described above. During the last 30 min of treatment, the cells were loaded with a final concentration of 100 nM of DiOC 6 (3). Afterward the cells were transferred to microcentrifuge tubes and centrifuged at 8,000 ϫ g for 10 min. The supernatant was removed, and the pellet was washed by resuspension in 500 l of phosphate-buffered saline. The washed pellet was lysed by the addition of 600 l of distilled, deionized water and was then homogenized. The fluorescence of DiOC 6 (3) in the homogenate was then read in a PerkinElmer Life Sciences LS-5 fluorescence spectrophotometer at 488 nm (excitation) and 500 (emission). Values for deenergized mitochondria were determined by simultaneous treatment of cells with 10 M CCCP and DiOC 6 (3). CCCP was dissolved in ethanol and added at a volume of 0.5%.
Measurement of PARP Activity-PARP activity was determined in Jurkat cells by a modification of the methods of Zingarelli et al. (14) and Wielckens and Delfs (15). Briefly, 5 ϫ 10 5 cells in 500 l of serum-free RPMI medium were plated in 24-well plates. After overnight incubation, the cells were treated as described above. Following treatment, the cells were transferred to microcentrifuge tubes and centrifuged at 8,000 ϫ g for 10 min. The medium was removed, and 500 l of a buffer containing 56 mM HEPES, 28 mM KCl, 28 mM NaCl, 2 mM MgCl 2 , 1 mM dithiothreitol, 50 g/ml digitonin, with 1.0 Ci of [ 3 H]NAD ϩ (4.5 Ci/ mmol; PerkinElmer Life Sciences) was added. The cells were then incubated for 15 min at 37°C. The reaction was stopped by addition of 200 l of 50% trichloroacetic acid. After 30 min at 4°C, the samples were centrifuged at 16,000 ϫ g for 20 min. The supernatant was removed and the pellet washed with 500 l of 3% trichloroacetic acid. The pellet was then solubilized with 1 ml of 2% SDS in 0.1 N NaOH. The samples were transferred to scintillation vials, and following neutralization with sufficient glacial acetic acid, the samples were counted in 10 ml of scintillation fluid on a Beckman LS 3801 scintillation counter.
Measurement of ATP Content-The cellular ATP content was determined by modification of the firefly luciferin-luciferase method as described previously (16). Briefly, cells in RPMI medium from 24-well plates were transferred into 1.5-ml microcentrifuge tubes, and 20 l of trichloroacetic acid was added to give a final concentration of 3%. After 30 min at 4°C, the cells were centrifuged at 14,000 ϫ g. An aliquot of the supernatant (250 l) was removed and neutralized with KOH. The volume was brought to 1 ml with 10 mM Tris-HCl, pH 7.4. One hundred l of a firefly extract (2 mg/ml in 50 mM K 2 HAsO 4 and 20 mM MgSO 4 , pH 7.4, and then filtered through 0.45-micron filter) was injected into a tube containing 20 l of sample. The resulting light emission was measured in an LB 9501/16 Lumat (EG & G Berthold Co.). ATP standards were placed in acidified medium and processed in identical fashion as the samples. The protein content of the pellet was determined by using bicinchoninic acid with bovine serum albumin as the standard (17).
NAD ϩ Determinations-The cellular NAD ϩ content was determined by modification of the cycling assay (18,19) as described previously (20). After treatment cells were transferred to microcentrifuge tubes and centrifuged at 8,000 ϫ g to pellet the cells. The supernatants were removed, and the pellets were resuspended in 500 l of 100 mM potassium phosphate buffer containing 3% trichloroacetic acid. After 30 min on ice, the samples were centrifuged at 12,000 ϫ g. Electron Microscopy-After treatment the cells were trypsinized and pelleted. The cells were then suspended and fixed overnight at 4°C in 2% glutaraldehyde with 1% tannic acid in 0.1 M sodium cacodylate, pH 7.3. The cells were rinsed 3 times in the sodium cacodylate buffer and then incubated in 2% osmium tetroxide in the same buffer for 2 h at room temperature. The cells were then rinsed 3 times in distilled water and exposed to 1% uranyl acetate in water for 15 min at room temperature. The cells were rinsed twice in distilled water, spun down into 3% agarose at 45°C, and cooled to form blocks. The agarose blocks were dehydrated in graded steps of acetone and embedded in Spurr's low viscosity media. Following polymerization overnight at 65°C, 80-nm sections were cut on a Reighert-Jung Ultra cut E ultramicrotome and picked up on copper grids. The grids were post-stained in uranyl acetate and bismuth subnitrate. The sections were observed in a Hitachi 7000 STEM and micrographs recorded on Kodak 4489 sheet film.
All experiments using Western blotting were repeated three times (cytochrome c, MDH, Bid, Bik, Bak, Bad, Cox4, ␤-actin, and PARP as above). The immunoblots from each experiment were scanned with an Image Station-440CS densitometer (Eastman Kodak Co.). The relative densities with respect to the appropriate controls were calculated. In all cases below, statements regarding changes in the intensity of Western blot staining are based on comparisons of the calculated relative densities rather than simply from visual inspection of the gels.

Expression of Pro-apoptotic Proteins in Jurkat Cells-The
Jurkat cell is a human T-cell leukemia/lymphoma that is frequently used to study the mechanisms mediating Fas-induced apoptosis. Accordingly, the content of proapoptotic proteins in these cells is of significant concern. Fig. 1A contrasts the Jurkat cell with a malignant human epithelial cell, the HeLa cell, with respect to the content of Bax, Bid, Bik, and Bak. Jurkat cells contain Bid, Bak, and Bik but not Bax. By contrast, HeLa cells contain predominantly Bax and Bak. In the HeLa cell, there is no Bik and much less Bid than in the Jurkat cell.
The absence of the Bax protein in Jurkat cells was reported (21) to be a consequence of single base deletions and additions in a polyguanine tract within the Bax open reading frame. RT-PCR was used to detect the mRNA for Bax, Bid, and Bak in both the Jurkat and HeLa cells. A 579-bp species that corre-sponds to Bax ␣ RNA was present in HeLa cells but not in Jurkat cells (Fig. 1B). By contrast, Jurkat cells contain a 433-bp species that likely represents the RNA of the Bax ␦ isoform. Bid and Bak RNA were readily detectable in both Jurkat and HeLa cells (Fig. 1B). There was no detectable Bax ␦ protein in the Jurkat cells (data not shown). Interestingly, the Bax ␦ protein was not detectable in DG75 cells that overexpressed Bax ␦ RNA (21), a finding suggesting that the mRNA was unstable. Alternatively, at least in the Jurkat cells, the Bax ␦ gene may be mutated.
Bid Cleavage Upon Activation of the Fas Receptor-Bid was cleaved by activated caspase-8 upon the initiation of Fas signaling (7)(8)(9). Cleavage of Bid was held to release a potent proapoptotic activity (tBid) that acts on the mitochondria to release cytochrome c (10). We report here a dissociation of the cleavage of Bid from the killing of Jurkat cells upon Fas receptor activation. anti-Fas antibody, the content of full-length Bid was reduced, and the cleavage fragment was present. By 6 h both these changes were more prominent. The caspase-8 inhibitor IETD (22) prevented the accumulation of the p15 cleavage fragment and maintained the content of full-length Bid at both 3 and 6 h. The PARP inhibitor 3-aminobenzamide had no effect on the cleavage of Bid. With both IETD and 3-AB, no evidence of Bid cleavage was seen at either 3 or 6 h.
The effect of these same manipulations on the extent of cell killing is shown in Fig. 3. Within 6 h of activation of the Fas receptor, 60% of the Jurkat cells were dead. IETD completely prevented the cell killing. In the presence, however, of both IETD and 3-AB, the cell killing was restored. In other words, the addition of 3-AB reversed the ability of the caspase inhibitor to prevent the cell killing. Importantly, 3-AB alone (in the absence of IETD) had no effect on the rate or extent of the cell killing (data not shown).
PARP is another well documented substrate for activated caspases, and PARP cleavage occurred upon activation of the Fas receptor (23). Fig. 4 illustrates that the details of PARP cleavage were similar to those of the cleavage of Bid in Fig. 2. PARP cleavage was readily evident within 2 h and complete within 4 h. IETD prevented PARP cleavage, whereas 3-AB alone did not. As with the cleavage of Bid (Fig. 2), PARP cleavage was not seen in the presence of both IETD and 3-AB.
The data in Figs. 2 and 3 document a condition under which the Jurkat cells were killed in the absence of Bid cleavage upon activation of the Fas receptor. Rather, we show below that the cell killing correlated more closely with the translocation of full-length Bid from the cytosol to the mitochondria.
Translocation of Bid to the Mitochondria- Fig. 5 details the mitochondrial content of both full-length Bid (p22) and its cleavage fragment (p15). Within 2 h of activation of the Fas receptor, the content of full-length Bid had increased, as well as that of the cleavage product. The caspase-8 inhibitor IETD prevented both the accumulation of full-length Bid and its cleavage product. In the presence of both IETD and 3-AB, there was more full-length Bid associated with the mitochondria than occurred with IETD alone. In other words, the ability of 3-AB to reverse the protection afforded by the caspase inhibitor IETD correlated with the translocation of full-length Bid to the mitochondria. Fig. 5 also shows that the content of the mitochondrial marker Cox4 did not vary under the conditions studied, a result indicating that the data cannot be interpreted as reflecting differences in the number of mitochondria compared under the various conditions illustrated.  was primarily found in the mitochondrial fraction. The content of mitochondrial Bik increased within 3 h of addition of the anti-Fas antibody, an effect that was not prevented by the caspase inhibitor IETD alone or in the presence of 3-AB. Although it is not readily apparent, in the control cells Bik migrated as a doublet of 24 -25 kDa. The heavier more prominent band corresponded to the previously identified phosphorylated form of Bik (24), and the increased content of Bik likely represents an accumulation of the phosphoprotein. As reported previously (25), Bak was similarly localized primarily in the mitochondria in Jurkat cells (Fig. 6B). There was no change in mitochondrial content of Bak at either 1 or 3 h after activation of the Fas receptor.
The Role of Bad in the Translocation of Bid-Bad is a proapoptotic protein that is capable of forming heterodimers with the antiapoptotic proteins Bcl-X L and Bcl-2 (26). It was postulated that the binding of Bad to Bcl-X L or Bcl-2 antagonized the antiapoptotic activity of these proteins. Phosphorylation of Bad at serine 136 prevented its binding to either Bcl-X L or Bcl-2 (27). Thus, the phosphorylation of Bad promoted cell survival.
Bad was constitutively phosphorylated in Jurkat cells (Fig.  7). The activation of the Fas receptor resulted in the dephosphorylation of Bad (Fig. 7). Thus, upon activation of the Fas receptor, Bad became competent to bind Bcl-X L or Bcl-2, an effect that prevented the interaction of these antiapoptotic proteins with Bid (28). The protection afforded by the caspase-8 inhibitor IETD, however, cannot be attributed to an inhibition of the de-phosphorylation of Bad. Fig. 7 shows that no protection occurred against the decrease in P-Bad in the presence of IETD.
The Release of Cytochrome c upon the Translocation of Fulllength Bid-The content of cytochrome c in the mitochondria decreased within 2 h of activation of the Fas receptor (Fig. 8). IETD prevented this loss of cytochrome c from the mitochondria. In the presence of both the caspase inhibitor and the PARP inhibitor 3-AB, there was again loss of cytochrome c from the mitochondria. In Fig. 8 the content of Cox4 documents that the number of mitochondria examined was constant.
The loss of cytochrome c from the mitochondria was accompanied by its accumulation in the cytosol. Within 2 h of activation of the Fas receptor, there was an increase in the cyto-chrome c content of the cytosol, an effect that was inhibited by IETD (Fig. 8). In the presence of both IETD and 3-AB, the content of cytochrome c in the cytosol again increased. In Fig.  8 the content of ␤-actin assessed the relative content of cytosolic proteins in the various fractions studied. Despite the significant differences in the content of cytochrome c among the various cytosol preparations, the content of ␤-actin did not vary.
Fas Receptor Activation Induced the Mitochondrial Permeability Transition-Several reports (4, 5, 10, 11) document the ability of the main cleavage fragment of Bid to release cytochrome c from isolated mitochondria in vitro by a mechanism that is independent of the MPT. By contrast, Bid-induced mitochondrial membrane permeabilization, both in vitro and in intact cells, could be blocked by inhibitors of the permeability transition pore complex (29). We report the ability of the inhibitor CyA to prevent in intact Jurkat cells both the release of cytochrome c and cell killing upon activation of the Fas receptor.
The effect of CyA and ArA, a phospholipase A 2 inhibitor, on the extent of cell killing is shown in Fig. 3. The ability of CyA to inhibit the MPT was enhanced and prolonged by phospholipase inhibitors, both in vitro with isolated mitochondria and in the intact cell (30,31). Whereas within 6 h after activation of the Fas receptor 60% of the cells are dead, the presence of CyA and ArA reduced the killing to less than 10% of the cells. Importantly, CyA and ArA also prevented the cell killing that occurred with IETD and 3-AB, a result indicating that the mechanism of cell killing in this case was not different from that with activation of the Fas receptor alone. In this regard, it deserves noting that CyA and ArA did not prevent the cleavage of Bid (data not shown) or PARP (Fig. 4). It should also be noted that CyA prevented the liver cell death that followed Fas receptor activation in intact mice (32,33).
CyA inhibited calcineurin, a calcium-dependent phosphatase that is involved in signal transduction in T-cells (34). However, the ability of CyA to prevent Fas-induced apoptosis was not a consequence of the inhibition of calcineurin. The immunosuppressive drug FK506 similarly inhibited calcineurin, but it was inactive against the MPT (35). FK506 (alone or in combination with ArA) was without effect on the cell killing by anti- Fas   FIG. 8. Cytochrome c and MDH release upon activation of Fas receptor and its prevention by IETD. Jurkat cells were treated with an anti-Fas antibody. Where indicated the cells were pretreated with IETD or IETD plus 3-AB. After 2 h the cells were processed to obtain a mitochondrial and cytosolic fraction as described under "Experimental Procedures." The content of cytochrome c (Cyt.c), MDH, Cox4, and ␤-actin was determined by SDS-PAGE and Western blotting.
antibody (data not shown). Cypermethrin, another potent inhibitor of calcineurin, was similarly unable to prevent the cell killing (data not shown). Finally, an inhibition of the de-phosphorylation of Bad cannot account for the protection afforded by CyA (see Fig. 7).
The MPT caused the loss of the mitochondrial membrane potential (⌬⌿ m ) (13,36,37). The fluorescent dye DiOC 6 (3) localized to mitochondria as a consequence of ⌬⌿ m , and the MPT reduced the accumulation of DiOC 6 (3) upon the loss of ⌬⌿ m . Fig. 9 shows that activation of the Fas receptor resulted in 40% less DiOC 6 (3) fluorescence within 1 h and 65% less within 3 h. CyA and Ara completely prevented this effect. There was only minimal, if any, loss of viability prior to 2 h (data not shown). Thus, prevention of the loss of DiOC 6 (3) fluorescence cannot be attributed to a nonspecific consequence of the preservation of cell viability. Treatment of the Jurkat cells with CCCP, a protonophore that dissipates ⌬⌿ m , similarly resulted in less DiOC 6 (3) fluorescence (data not shown). However, CyA and ArA did not affect this result. In this situation and in contrast to that occurring with activation of the Fas receptor, the loss of DiOC 6 (3) fluorescence did not document the MPT.
The MPT released cytochrome c from the intermembrane space (38) and other proteins from the mitochondrial matrix (39). CyA and ArA prevented the release of cytochrome c to the cytosol and a concomitant decrease in the content of cytochrome c in the mitochondria upon activation of the Fas receptor (Fig. 10). Malate dehydrogenase is a soluble protein present in the mitochondrial matrix (39). Figs. 8 and 10 show that the MDH content of the cytosol increased substantially upon treatment of the Jurkat cells with the anti-Fas antibody, an effect that was accompanied by a loss of the enzyme from the mitochondria. Activation of the Fas receptor in the presence of IETD (Fig. 8) or CyA and ArA (Fig. 10) prevented both the accumulation of MDH in the cytosol and its loss from the mitochondria. 3-AB restored the cytosolic accumulation of MDH (Fig. 8). The content of Cox4 was unchanged in any of the mitochondrial fractions (Figs. 8 and 10). Similarly, the content of ␤-actin in the cytosolic fractions was constant ( Fig. 8 and 10). Accordingly, changes in the relative content of mitochondria and cytosolic proteins in the respective subcellular fractions cannot explain the redistribution of cytochrome c and MDH illustrated in Figs. 8 and 10.
Examination of intact Jurkat cells by electron microscopy 3 h after treatment with anti-Fas antibody revealed prominent mitochondrial swelling (Fig. 11A). Clearing of the matrix space with loss of cristae accompanied an increase in the size and configuration of the mitochondria. Importantly, these mitochondrial alterations occurred at a time when the nuclear membrane was still intact, and the chromatin was not condensed or clumped (Fig. 11A). The plasma membrane was also intact (Fig. 11A). In the presence of CyA and ArA, mitochondrial ultrastructure was preserved upon activation of the Fas receptor (Fig. 11B).
The Mechanism of Action of 3-Aminobenzamide- Table I indicates that in the presence of anti-Fas antibody alone, no change occurred in the activity of PARP. In the presence, however, of both the antibody and DEVD, a more specific caspase-3 inhibitor than IETD (22), a 10-fold increase in PARP activity occurred within 30 min. The presence of the PARP inhibitor 3-AB prevented the increase in the enzyme seen with anti-Fas antibody and caspase inhibition (Table I). Table I also shows that DEVD, like IETD (detailed in Fig. 3), prevented the cell killing upon activation of the Fas receptor. In turn, 3-AB reversed this protection. In other words, in the presence of 3-AB and the specific caspase-3 inhibitor, the extent of cell killing was identical to that with the anti-Fas antibody alone. 3-AB alone or in combination with DEVD was not toxic (data not shown).
PARP is a nuclear enzyme involved in the repair of DNA damage. PARP hydrolyzed NAD into an ADP-ribose moiety and nicotinamide (40,41). Repetition of this activity produced extended chains of ADP-ribose. The nicotinamide released from NAD by poly(ADP-ribosylation) was reconverted to NAD in a reaction that consumed ATP. In addition, depletion of NAD, an electron carrier required for mitochondrial respiration, can disrupt ATP production. Thus, the activation of PARP can substantially deplete NAD and ATP.
Depletion of both NAD and ATP (Table II) accompanied the increase in PARP activity detailed in Table I. Treatment with the anti-Fas antibody alone did not change the content of either NAD or ATP. In the presence of both the anti-Fas and DEVD, however, NAD decreased by 90% and ATP by more than 50% within 30 min. 3-AB prevented the depletion of both NAD and ATP. In other words, restoration of cell killing by 3-AB was reflected in the maintenance of the content of NAD and ATP. Treatment of Jurkat cells with 2-deoxyglucose plus oligomycin or rotenone reduced the ATP content of the cells by a mechanism unrelated to an effect on PARP. These inhibitors similarly protected against the cell killing (Table III).
Translocation of Bid to the Mitochondria with Ceramide-Upon activation the Fas receptor generated ceramide, as a consequence of the sequential involvement of phosphatidylcholine-specific phospholipase C and an acidic sphingomyelinase (42). Thus, the toxicity of ceramide can be used to model downstream events in the absence of other upstream consequences of Fas receptor activation, namely the activation of caspase-8.
In the presence of 10 M ceramide, 60% of the Jurkat cells were dead within 6 h. With ceramide full-length Bid translocated to the mitochondria (Fig. 12A), an effect that occurred in the absence of detectable Bid cleavage (Fig. 12B).

DISCUSSION
The above data document that, upon activation of the Fas receptor, full-length Bid translocates to the mitochondria with induction of the MPT, the release of cytochrome c, and then cell death. In addition, the ability of caspase inhibitors to prevent the killing of the Jurkat cells is not attributable to an inhibition of Bid processing. Rather, the data imply that the prevention of PARP cleavage is the likely basis of the anti-apoptotic effect of early caspase inhibition. Finally, the participation of full-length Bid in the mitochondrial phase of Fas-induced cell killing does not necessarily depend on its ability to interact with Bax.
Activation of the Fas receptor readily kills Jurkat cells (Fig.  3), an effect that occurs in the absence of Bax expression (Fig.  1). Two distinct pharmacologic interventions that prevent this  cell killing were then used to probe the responsible events. A caspase-3 inhibitor (DEVD), a caspase-8 inhibitor (IETD), and an inhibitor of the MPT (CyA) prevent Fas-induced apoptosis in the Jurkat cells ( Fig. 3 and Table I). Protection from cell killing upon caspase inhibition correlates with the inhibition of Bid cleavage, a result consistent with the hypothesis (7-9) that tBid formed upon caspase activation mediates the release of cytochrome c from the mitochondria. When the Jurkat cells, however, are treated with the anti-Fas antibody in the presence of both the caspase inhibitor and 3-AB (a PARP inhibitor), cell killing is fully restored. Cell death under these conditions occurs in the absence of Bid cleavage. Rather, the cell killing occurs upon the translocation of full-length Bid to the mitochondria and induction of the MPT. The Jurkat cells are also killed by ceramide. In this case full-length Bid translocates to the mitochondria in the absence of detectable, early Bid cleavage. Thus, in two distinct situations the killing of Jurkat cells is accompanied by the translocation of full-length Bid to the mitochondria. The interaction of full-length Bid with this organelle results in the release of cytochrome c and the initiation of an effector phase of caspase activity. The release by tBid of cytochrome c from isolated mitochondria in vitro is reported to be independent of the MPT (4, 5, 10, 11). By contrast, the release of cytochrome c in the intact cell by full-length Bid depends on the MPT. Our data document that the induction of the MPT is an essential feature of the mechanism of cell killing in Fas-induced apoptosis in the Jurkat cell. CyA inhibits the MPT and prevents the cell killing (Fig. 3). The MPT was documented in intact Jurkat cells by the sensitivity to CyA as follows: (a) the loss of cytochrome c from the intermembrane space and MDH from the mitochondrial matrix, (b) the loss of the mitochondrial membrane potential, and (c) the swelling of these organelles, as determined by electron microscopy.
The sufficient conditions for induction of the MPT upon activation of the Fas receptor remain to be defined. The most obvious concern is the mechanism that causes Bid to move from the cytosol to the mitochondria. In principle, this translocation can result from a change in Bid itself, or a change in the mitochondria, or both. It is currently held that N-terminal protein processing of Bid upon caspase activation exposes the BH3 domain (43,44) and results in the translocation of tBid to the mitochondria. The data presented here argue against this action of Bid and require an alternative explanation of its activation. There is evidence that a conformational change in Bax is required for the translocation of this proapoptotic protein to the mitochondria (45)(46)(47). At present, there are no similar data with respect to Bid.
A change in the mitochondria could provide an environment that promotes Bid translocation. The proapoptotic proteins Bik and Bak are localized primarily to the mitochondria (24,25). Human Bik is a phosphoprotein, and evidence exists that phosphorylation plays a role in the apoptotic activity of Bik (24). An interaction of Bik with Bid, however, has not been reported. By contrast, conformational changes in mitochondrial Bak are induced by tBid, and Bak is required for Bid-mediated cytochrome c release from mitochondria (12). However, the cytochrome c release mechanism that resulted from tBid-Bak binding and the oligomerization of Bak was independent of the MPT (12). Finally, mention needs be made of the possibility that a change in the lipid composition of the mitochondria may promote Bid translocation. In particular, it is possible that TABLE III Inhibition of ATP synthesis and alkalinization of acidic compartments prevent the cell killing by the anti-Fas antibody Where indicated, the cells were pretreated for 30 min with either 25 M rotenone plus 10 mM 2DG or 0.1 g/ml oligomycin plus 2DG. The ATP content of cells pretreated with rotenone plus 2DG was 8.5 Ϯ 1.3 nmol/mg protein and with oligomycin plus 2DG 11.0 Ϯ 0.7. The ATP content of untreated cells was 26.8 Ϯ 2.5. After the pretreatment, the cells were treated with 100 ng/ml anti-Fas antibody, and cell killing was determined after 6 h. The data are the mean Ϯ S.D. from three independent experiments. FIG. 12. Ceramide causes the translocation of full-length Bid to the mitochondria. A, Jurkat cells were treated with ceramide for 1 or 2 h. The cells were then processed to obtain a mitochondrial fraction as described under "Experimental Procedures." The content of Bid and COX4 was determined by SDS-PAGE and Western blotting. B, Jurkat cells were treated with ceramide for 1 or 2 h. Whole cell lysates were obtained and analyzed for Bid by SDS-PAGE and Western blotting as described under "Experimental Procedures." ceramide may mediate a change in the mitochondrial membranes that promotes Bid translocation.
In addition to preventing the processing of Bid, caspase inhibitors also block the cleavage of PARP. In turn, this inhibition of PARP cleavage is accompanied by a 10-fold activation of this enzyme, an effect that was associated with a depletion of NAD and ATP. In the absence of a caspase inhibitor, no increase in PARP activity occurs upon activation of the Fas receptor, and no change in the content of NAD and ATP is seen. In other words, PARP cleavage during Fas-induced apoptosis prevents an otherwise marked activation of this enzyme.
The depletion of ATP is a likely mechanism whereby caspase inhibition prevents the MPT and, thereby, prevents cell killing. At least two downstream steps that follow activation of the Fas receptor require ATP, namely the internalization of the Fas receptor complex and the acidification of the resulting endosome. Treating the Jurkat cells with 2DG plus either oligomycin or rotenone demonstrates the requirement for ATP in Fasinduced apoptosis. The resulting loss of ATP prevents the cell killing (Table III). Alternatively, an accumulation of ADP, rather than the depletion of ATP per se, is the mechanism of inhibition of the MPT.
The prevention of Fas-induced apoptosis by either inhibition of caspase-3 or caspase-8 is reversed by 3-AB, a PARP inhibitor. In the presence of 3-AB and the caspase inhibitor, PARP activity does not increase (Table I), and no loss of either ATP or NAD occurs (Table II). Nevertheless, cell killing still ensues (Fig. 3). Importantly, in the presence of 3-AB and a caspase inhibitor, PARP is not cleaved. In other words, in the presence of 3-AB and DEVD, Fas-induced apoptosis ensues despite the inhibition of either caspase-3 or caspase-8. By inhibiting PARP and, thus, by preventing the depletion of ATP that occurs when this enzyme is not cleaved, 3-AB allows the energy-dependent induction of the MPT to proceed. Accordingly, it is our conclusion that the action of inhibition of caspase-8 or caspase-3 in Fas-mediated apoptosis permits expression of the cell death program that depends on the induction of the MPT.
We have not defined the mechanism by which PARP was activated upon treatment of the Jurkat cells with the anti-Fas antibody in the presence of a caspase inhibitor. We hypothesize that the early cleavage of DNA into large fragments prior to internucleosomal fragmentation (48,49) activates PARP. Such large fragmentation of DNA may not be dependent on early caspase activation, and thus would not be prevented by caspase inhibition.