Cytochrome c-mediated Apoptosis in Cells Lacking Mitochondrial DNA

Mitochondria serve as a pivotal component of the apoptotic cell death machinery. However, cells that lack mitochondrial DNA (ρ0 cells) retain apparently normal apoptotic signaling. In the present study, we examined mitochondrial mechanisms of apoptosis in ρ0 osteosarcoma cells treated with staurosporine. Immunohistochemistry revealed that ρ0cells maintained a normal cytochrome c distribution in mitochondria even though these cells were deficient in respiration. Upon staurosporine treatment, cytochrome c was released concomitantly with activation of caspase 3 and loss of mitochondrial membrane potential (Δψ m ). After mitochondrial loss of cytochrome c, ρ0 cells underwent little change in glutathione (GSH) redox potential whereas a dramatic oxidation in GSH/glutathione disulfide (GSSG) pool occurred in parental ρ+ cells. These results show that mitochondrial signaling of apoptosis via cytochrome c release was preserved in cells lacking mtDNA. However, intracellular oxidation that normally accompanies apoptosis was lost, indicating that the mitochondrial respiratory chain provides the major source of redox signaling in apoptosis.

Mitochondria serve as a pivotal component of the apoptotic cell death machinery. However, cells that lack mitochondrial DNA ( 0 cells) retain apparently normal apoptotic signaling. In the present study, we examined mitochondrial mechanisms of apoptosis in 0 osteosarcoma cells treated with staurosporine. Immunohistochemistry revealed that 0 cells maintained a normal cytochrome c distribution in mitochondria even though these cells were deficient in respiration. Upon staurosporine treatment, cytochrome c was released concomitantly with activation of caspase 3 and loss of mitochondrial membrane potential (⌬ m ). After mitochondrial loss of cytochrome c, 0 cells underwent little change in glutathione (GSH) redox potential whereas a dramatic oxidation in GSH/glutathione disulfide (GSSG) pool occurred in parental ؉ cells. These results show that mitochondrial signaling of apoptosis via cytochrome c release was preserved in cells lacking mtDNA. However, intracellular oxidation that normally accompanies apoptosis was lost, indicating that the mitochondrial respiratory chain provides the major source of redox signaling in apoptosis.
Mitochondria have recently become a focus of research in apoptosis, a cell death process in which caspase activation and DNA cleavage are characteristic features (1,2). The involvement of mitochondria was first suggested by the discovery that Bcl-2, a potent anti-apoptotic protein, is present in the mitochondrial outer membrane (3,4). However, an argument against the role of mitochondria was raised when Jacobson et al. (5) showed that cells without mitochondrial DNA (mtDNA), 1 and therefore lacking respiratory chain activity, are still able to undergo apoptosis after deprivation of survival factors or exposure to staurosporine. Overexpression of Bcl-2 was found to protect these cells from apoptosis.
Significant data strengthening the interpretation that mitochondria have a critical role in apoptosis have come from the study by Newmeyer et al. (6) who demonstrated that cell-free apoptosis in Xenopus egg extract required an organelle fraction enriched in mitochondria. Meanwhile, other researchers found that cells induced to undergo apoptosis show a fall in the mitochondrial membrane potential (⌬ m ) (7). The ⌬ m disruption occurs before exposure of phosphatidylserine on the outer cell membrane leaflet and before cells exhibit nuclear DNA fragmentation. Functional experiments indicate that the mechanism of the ⌬ m loss during apoptosis involves the opening of a large conductance channel known as permeability transition (PT) pore (8,9). Mitochondria undergoing PT liberate an apoptogenic protein, apoptosis inducing factor (AIF), which is capable of inducing nuclear apoptosis (10). AIF has recently been cloned and characterized (11); it is a protein that is translated with a mitochondrial targeting sequence that is cleaved upon import into mitochondria. The cleavage exposes a nuclear targeting sequence that allows the protein to target to the nucleus and induces nuclear changes upon release from the mitochondria. This mechanism has been suggested to be independent of a caspase-mediated pathway for apoptosis.
A direct activation of caspase-mediated apoptosis was found to involve cytochrome c, an essential component of the mitochondrial respiratory chain (12). In many systems, release of cytochrome c from mitochondria to cytosol has been demonstrated to be a crucial step in the activation of apoptosis (13)(14)(15)(16). Once released from mitochondria, cytochrome c, in interaction with apoptotic protease activity factor-1 and caspase 9, initiates the cleavage and activation of caspase 3 (17). Active caspase 3 in turn activates other execution caspases and leads to the subsequent characteristic features of apoptosis, including cleavage of fodrin and lamin, chromatin condensation, and nuclear fragmentation (18). Bcl-2 and other related anti-apoptotic proteins, e.g. Bcl-X L , can prevent cytochrome c release and inhibit caspase activation and apoptosis (13,14,19). Although the mechanism(s) by which cytochrome c is released from mitochondria remain unknown, some observations suggest that the cytochrome c-mediated apoptosis is probably distinct from that mediated by AIF (2 and 16).
The role of mitochondria in apoptosis may be not restricted to releasing apoptogenic proteins. There is accumulating evidence that reactive oxygen species (ROS) may serve as signaling molecules in apoptosis. First, in some systems, addition of oxidants or depletion of endogenous antioxidants can promote apoptosis, while antioxidants can delay or inhibit apoptosis (20 -22). Second, increases in intracellular ROS are often associated with apoptosis (23,24). Finally, Bcl-2 has been found to have an antioxidant effect (25). Mitochondria are believed to be a major site of ROS production. Indeed, it has been found that both ROS accumulation and apoptosis require the presence of a * This work is supported by National Institutes of Health Grant ES09047. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Studies have indicated that disruption of mitochondrial electron transport and energy metabolism is involved in apoptosis. A disruption of the electron transport chain at the cytochrome b-c 1 /cytochrome c step was found to occur in cells and in isolated mitochondria treated with ceramide (27). In Fas-mediated apoptosis, an inhibition of electron transport was observed before nuclear changes of apoptosis, and results indicated that inactivation of cytochrome c at least in part contributes to the loss of mitochondrial function in this system (28). Overexpression of Bcl-2 can both prevent and reverse cytochrome c inactivation (29).
Taken together, the data indicate that mitochondria play an important role in apoptosis by releasing AIF and cytochrome c, by generating ROS, and by disrupting electron transport. These pathways may represent alternative and independent pathways or they may function together to provide an efficient and fail-safe cell death program. In the latter case, cells lacking mitochondrial oxidative phosphorylation may still undergo apoptosis if one or some of these pathways are still maintained.
In the present study, we report that pathways involving release of cytochrome c from mitochondria and disruption of ⌬ m are preserved in 0 osteosarcoma cells during staurosporine-induced apoptosis. However, the mitochondria-derived ROS pathway that causes oxidation in cellular thiol pool is largely attenuated, indicating that the mitochondrial release of cytochrome c and loss of ⌬ m are not due to the oxidation of the cellular thiols as reflected in the GSH pool. Thus, the mitochondrial redox signaling in apoptosis is a downstream effect not essential for caspase activation but probably functioning in late-stage events such as caspase inactivation or marking proteins and membrane-bound fragments for degradation.

EXPERIMENTAL PROCEDURES
Cells-The 0 cell line was derived from an osteosarcoma cell line 143BTK (ATCC CRL 8303) by growing cells in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, high glucose (4.5 mg/ml), 1 mM pyruvate, 50 g/ml uridine, and 50 ng/ml ethidium bromide for 13 weeks (30). Clones were screened for mtDNA content by Southern blot analysis and polymerase chain reaction and a cell line (143B-87) containing no mtDNA was selected and grown for several weeks in the medium without ethidium bromide (30). When oxygen consumption was measured with a Clark oxygen electrode, a rapid oxygen uptake was readily apparent in ϩ cells but undetectable in 0 cells.
Auxotrophy for uridine and pyruvate in 0 cells was demonstrated by the lack of growth in the absence of these nutrients (30). To eliminate the variance that could result from different culture conditions between 0 and ϩ , an identical medium was used to culture two cell lines in the present study, although the latter does not require uridine and pyruvate for growth.
Assessment of Apoptosis by TUNEL-TUNEL was performed using the in situ cell death detection kit following the standard protocol provided by manufacturer. Briefly, after treatment with 1 M staurosporine, both floating and adherent cells were collected and washed once with PBS. Cells were fixed with freshly prepared 4% paraformaldehyde in PBS, pH 7.4, for 30 min at room temperature and then permeabilized with 0.1% Triton X-100 in 0.1% sodium citrate for 2 min on ice. To label DNA strand breaks, cells were incubated with 50 l of TUNEL reaction mixture containing terminal deoxynucleotidyl transferase and fluorescein-dUTP in the binding buffer and incubated for 1 h at 37°C in a humidified atmosphere. Cells were washed twice with PBS and then analyzed by flow cytometry.
Visualization of Cytochrome c, Nuclei, and Mitochondria by Confocal Microscopy-Cells grown on chambered coverglass (Lab-Tek 136439) were treated with 1 M staurosporine. After 6 h, cells still remaining on the coverglass were washed twice with PBS and fixed with 2% formaldehyde for 5 min at room temperature. Following a further fixation and permeabilization in ice-cold 100% methanol for 5 min, cells were gently washed three times, and nonspecific binding sites were blocked with a blocking buffer (2% horse serum, 0.5% Triton X-100, and 0.02% sodium azide in PBS) for 20 min at room temperature. Cells were then stained with anti-cytochrome c antibody (diluted 1:50 in PBS) or anti-HSP60 antibody (diluted 1:60 in PBS) for 45 min at room temperature. Excess antibody was removed by washing cells 3 times with PBS. The secondary antibody, FITC-conjugated goat anti-mouse IgG, diluted 1:60 in PBS, was added and the cells were incubated for 30 min at room temperature. To double-label nuclei, cells were further stained with 50 g/ml propidium iodide in PBS for 5 min at room temperature. To double-label mitochondria, cells were first incubated with 10 M Mito-Tracker Red for 30 min in a 5% CO 2 humidified incubator and then immunostained for cytochrome c or HSP60 as described above. The coverslips were analyzed using a Bio-Rad 1024 laser scanning confocal microscopy.
Cytosolic Extracts for Western Blot Analysis of Cytochrome c-A digitonin-permeabilization technique was used (31) to release cytosol from untreated or staurosporine-treated cells. Briefly, cells were washed once with PBS and then resuspended in a solution containing 70 mM Tris and 250 mM sucrose, pH 7.0. As a control for Fig. 3, 10 mM Tris and 300 mM sucrose, pH 7.0, was also used. Digitonin was added to provide a condition that resulted in 95% cells staining with 0.2% trypan blue (approximately 0.05 mg/ml final concentration). The cells were immediately centrifuged at 3000 rpm for 10 min at 4°C and the supernatant was collected as cytosol. The amount of protein in the cytosol was measured by the Bradford method, and 50 g of protein was loaded onto each lane of a 15% SDS-polyacrylamide gel. The separated proteins were blotted to 0.45 M Hybond-polyvinylidene difluoride membrane (Amersham Pharmacia Biotech). Nonspecific binding was blocked by incubation in 5% non-fat milk, 0.1% Tween 20, in PBS overnight at 4°C. The membrane was then stained with anti-cytochrome c (diluted 1:1600 in PBS containing 5% non-fat milk and 0.1% Tween 20) for 1 h at room temperature with gentle agitation. After being washed 3 times with PBS containing 0.1% Tween 20, the membrane was incubated with 1:3000 diluted horseradish peroxidase-coupled anti-mouse IgG for 1 h at room temperature. The specific protein was then identified using NEN TM Renaissance Western blot Chemiluminescence Reagent.
Measurement of DEVD-AMC Cleavage Activity-Caspase activity that cleaves DEVD-AMC was measured by a method modified from Nicholson et al. (32). Briefly, cells adherent to the culture plate after staurosporine treatment were washed once with PBS and collected in 50 l of PBS and frozen on a microtiter plate. Fifty microliters of reaction buffer containing 100 mM HEPES, 10% dithiothreitol, 10 Ϫ6 % Nonidet P-40, and 0.1% CHAPS at pH 7.25 were added to each well along with 50 M DEVD-AMC. Substrate cleavage to release free AMC (excitation 355 nm, emission 460 nm) was monitored at 37°C. Fluorescence was converted to picomoles of AMC using a standard curve generated with free AMC.

Measurement of ⌬ m by Flow
Cytometry-To measure ⌬, cells were incubated with 1.25 g/ml JC-1 for 30 min in culture medium at 37°C (33). The cells were washed once with PBS and immediately analyzed by flow cytometry.
Analysis of GSH and GSSG-GSH and GSSG were measured as the S-carboxymethyl-N-dansyl-derivatives by high performance liquid chromatography (34). Cells were treated with ice-cold 5% perchloric acid containing 0.2 M boric acid and 5 M ␥-glutamylglutamate (internal standard). Following centrifugation to remove protein, 300 l is treated with 60 l of iodoacetic acid (7.4 mg/ml) and adjusted to pH 9.0 Ϯ 0.2 with 1 M KOH in saturated solution of potassium tetraborate. After 20 min, 300 l of a dansylchloride solution (20 mg/ml in acetone) was added, and extracts were incubated in the dark at room temperature overnight. Unreacted dansyl chloride was extracted with 500 l of chloroform and aliquots of the aqueous phase were separated by high performance liquid chromatography using a propylamine column (Custom LC, Houston). S-Carboxymethyl-N-dansyl-GSH and N,NЈ-bis-dansyl-GSSG derivatives were eluted at 1 ml/min with a linear gradient formed from an initial 80% solvent A (80% methanol in water) to 100% solvent B (64% methanol, 4 M sodium acetate buffer, pH 4.6) over 60 min. Dansyl derivatives were detected by fluorescence with 305-395 nm band-pass filter for excitation and a 520-nm cut off filter for emission. Peaks are quantified by integration relative to standards. For calculation of redox potential, intracellular volume was measured based upon 3 H 2 O uptake with [ 14 C]polyethylene glycol used for correction for extracellular volume. Redox potential (E h ) was calculated from the cellular GSH and GSSG concentrations using the Nernst equation and E 0 ϭ Ϫ0.258 V at pH 7.3.

RESULTS
Cells Lacking mtDNA Preserved Sensitivity to Staurosporineinduced Apoptosis-0 cells lacking functional mitochondria were treated with various concentrations of staurosporine for 17 h. Staining with Annexin-V revealed a dose-dependent exposure of phosphatidylserine on the surface of these cells (Fig.  1). A time course study with 1 M staurosporine showed that DNA cleavage could be detected as early as 1 h by TUNEL (Fig.  2). On agarose gels, the cleaved DNA had a ladder-like appearance typical of apoptosis (data not shown). Confocal microscopy of cells stained with propidium iodide illustrated a well dispersed chromatin structure in untreated cells (Fig. 3A). In contrast, cells treated with staurosporine showed chromatin that was condensed and fragmented. Therefore, despite complete loss of mtDNA and deficiency in respiratory function, these 0 cells maintained normal apoptotic machinery that could be activated by staurosporine.
Cytochrome c Had a Normal Mitochondrial Distribution in 0 Cells and Was Released during Apoptosis-Cytochrome c is a component of mitochondrial electron-transfer chain that is present in the intermembrane space. In cells containing respir-ing mitochondria, it can be released into the cytosol during the early phase of apoptosis (12)(13)(14)(15)(16). Like most mitochondrial proteins, cytochrome c is encoded by nuclear DNA and then translocated into the mitochondria. To determine whether cyto- chrome c is translocated into and retained within mitochondria in respiration-deficient cells, we immunostained 0 cells with an antibody that specifically recognizes cytochrome c. As illustrated in Fig. 3B, a superimposed punctuate perinuclear component was stained. Double staining with a mitochondria-selective vital dye MitoTracker Red revealed that mitochondria of 0 cells were also distributed in a punctuate perinuclear pattern and the punctuate fraction of cytochrome c colocalized with mitochondria (overlapping red and green pixels seen as yellow). When viewed with confocal microscopy, mitochondrial appearance and cytochrome c distribution in 0 cells were indistinguishable from those in parental ϩ cells (data not shown).
To determine whether cytochrome c in mitochondria of 0 cells is released during apoptosis, we examined cells stained with anti-cytochrome c antibody and then double-stained nuclei with propidium iodide. As illustrated in Fig. 3A, untreated cells showed the mitochondrial pattern of cytochrome c staining with diffuse nuclear staining by propidium iodide. However, in staurosporine-treated cells that have apoptotic morphology (condensed and fragmented nuclei), cytochrome c staining was diffuse with no mitochondrial staining pattern visible. The release of cytochrome c was further confirmed with Western blot analysis of cytosolic extracts prepared under a condition (70 mM Tris, 250 mM sucrose, pH 7) that keeps mitochondria intact. As shown in Fig. 3C, cytosol from untreated cells contained little cytochrome c. In contrast, cytosol from staurosporine-treated cells had a significant amount of cytochrome c as early as 1 h and a marked increase in cytochrome c level was seen in cytosol from cells treated for 6 h, a time point when numerous apoptotic nuclei were also observed (see Fig.  3A). Control experiments to verify that cytochrome c release was not due to the high ionic strength were performed with 10 mM Tris, 300 mM sucrose, pH 7. Comparable results were obtained for staurosporine treatment but there was somewhat more cytochrome c release in the untreated cells (results not shown).

Matrix Heat Shock Protein Was Largely Maintained in Mitochondria of Cells Undergoing Apoptosis-
The release of cytochrome c in 0 cells could result from overall damage to mitochondrial structure in apoptotic cells or it could be a specific event that causes little changes in mitochondria except outer membrane disruption as seen in other cell types. We distinguished these two possibilities by examining whether the mitochondrial matrix protein, HSP60, was also released during apoptosis. Staining the cells with an anti-HSP60 antibody revealed a similar pattern obtained with anti-cytochrome c antibody, i.e. a superimposed punctuate perinuclear staining that is colocalized with mitochondria (Fig. 4A). In apoptotic 0 cells, a change in the staining pattern of heat shock protein due to the shrinkage of cytosol was observed (Fig. 4B). However, the diminished and diffused immunostaining pattern like that of cytochrome c was not seen. Therefore, matrix heat shock protein was largely retained in mitochondria during apoptosis, indicating that cytochrome c was released from mitochondria by mechanism(s) that do not cause rupture of overall mitochondrial structure.
Appearance of Caspase-3-like Activity in Apoptotic 0 Cells-In many apoptosis systems, release of cytochrome c into cytosol results in activation of a proteolytic cascade in which procaspase 3 is cleaved and activated (12)(13)(14)(15). Activated caspase 3 then cleaves specific substrates and leads to apoptotic cytosolic and nuclear changes. To determine whether a similar process is also preserved in 0 cells, caspase 3-like activities were analyzed in staurosporine-treated cells by measuring cleavage of the peptide substrate DEVD-AMC. As shown in Fig. 5, in untreated cells, a background level of DEVD-AMC cleavage activity was detected. However, an increase in DEVD-AMC cleavage activity was observed 1 h after treatment with 1 M staurosporine and continued to increase at subsequent times. Thus, caspase 3-like activity is detectable early in apoptosis with a time course similar to that of cytochrome c release from mitochondria.
A Reduction of ⌬ m in 0 Cells during Apoptosis-Loss of ⌬ m is another early mitochondrial change in cells undergoing apoptosis but its relation to cytochrome c release is still unclear (2,3). To investigate whether ⌬ m is lost in 0 cells during apoptosis, a time course study of ⌬ m was performed in 0 cells treated with staurosporine. ⌬ m was measured with the potential sensitive dye, JC-1, which undergoes a molecular aggregation and shift in fluorescence from green to red at high membrane potentials (33). As shown in Fig. 6, after exposure to 1 M staurosporine a decrease of JC-1 dye uptake as measured by a decrease in JC-1 red fluorescence was observed at 1 h, at the time when cytochrome c release and caspase activation were also detectable. In order to better understand the temporal sequence of these events, we repeated this experiment with 0.1 M staurosporine and shorter incubation times. As shown in Fig. 7, although a slight decrease in fluorescence intensity of JC-1 was observed after 0.1 h of treatment, a significant loss in ⌬ m can be detected only after 1 h. Similarly, the activity that cleaves DEVD-AMC was undetectable until 1 h, indicating that the loss of ⌬ m occurs over a same time course with activation of caspase 3.
Redox Status in 0 and ϩ Cells before and after Staurosporine Treatment-Cytochrome c is an electron carrier in the mitochondrial respiratory chain and one of the consequences of loss of cytochrome c from mitochondria is generation of superoxide due to 1e Ϫ reduction of oxygen, which results in oxidation of GSH pool (36). We therefore examined whether intracellular oxidation also occurred during apoptosis in respiration-deficient 0 cells. As shown in Fig. 8, in the parental ϩ cells that have normal mitochondrial electron transport, a decrease in the intracellular GSH level was observed immediately after treatment with staurosporine. The intracellular GSSG content remained unchanged during 10 h of treatment but slightly increased thereafter. The GSH/GSSG ratio was initially increased but dropped to the control level by 6 h and decreased further to the level lower than control after 15 h. When redox potential was calculated, the results show that the cellular GSH/GSSG pool became oxidized after 10 h with a 40-mV oxidation in redox potential being detected by 24 h.
The 0 cells had a similar level of intracellular GSH as ϩ cells had but a 5-fold higher GSSG content and a calculated GSH/GSSG redox potential of Ϫ174 mV, which was more oxidized than that in untreated ϩ cells (Ϫ217 mV, Fig. 8). Upon treatment of 0 cells with staurosporine, both GSH and GSSG pool were rapidly decreased with little change in GSH/GSSG ratio. The decrease in cellular GSH is in agreement with studies showing an activation of GSH efflux during apoptosis (37).
However, the corresponding change in GSSG and little change in GSH/GSSG indicate that the loss of GSH, per se, does not necessarily reflect a change in the redox state of the GSH/ GSSG pool. This is further shown by the calculations of E h for the GSH/GSSG pool. Following staurosporine treatment, only a 15-mV oxidation in redox potential was detected after 24 h, compared with a 40-mV oxidation in ϩ cells. Thus, unlike ϩ cells, 0 cells underwent little change in GSH/GSSG redox potential during apoptosis. These data show that 0 cells have some mitochondria-derived apoptosis pathways, i.e. loss of ⌬ m and release of cytochrome c, but that the dramatic thiol oxidation was largely lost in the respiration-deficient cells. These results suggest that oxidative stress may not be essential for cytochrome c release and caspase activation but rather be a parallel or potentiating event associated with the mitochondrial changes. DISCUSSION Mitochondrial function depends on proteins that are encoded by nuclear DNA and mtDNA. mtDNA codes for 13 subunits of enzyme complexes of the electron-transport chain present in the inner mitochondrial membrane, including 7 of NADH-Q reductase, 1 of cytochrome c reductase, 3 of cytochrome c oxidase, and 2 of ATP synthase (38). Therefore, in 0 cells that lack mtDNA, the mitochondrial electron transport chain cannot function in normal respiration, and oxidative phosphorylation cannot take place. Because of this lack of respiration, Jacobson et al. (5) used 0 cells to investigate the role of mitochondria in apoptosis. They showed that apoptosis can still occur in 0 cells and that overexpression of Bcl-2 can protect these cells from apoptosis. Thus, they concluded that if mitochondria are required for apoptosis, then functions other than oxidative phosphorylation are involved.
Our present results show that respiration-deficient 0 cells had a normal distribution of cytochrome c in mitochondria. This finding is consistent with available knowledge on cytochrome c import into mitochondria. Cytochrome c is a nuclear DNA-encoded protein that is imported into mitochondria via a pathway that does not require a signal sequence, electrochemical potential, and general protein translocation machinery (39). After being imported to mitochondria, cytochrome c resides in the space between the outer and inner membrane of mitochondria, where it can transfer electrons from Complex III to the cytochrome c oxidase, both of which are located in the inner membrane. Nijtmans et al. (40) have reported that in 0 cells, nuclear-encoded subunits of cytochrome c oxidase can still be translated, imported into mitochondria and processed. However, cytochrome c oxidase is not assembled due to the lack of the mitochondrial encoded subunits. Accordingly, cytochrome c imported to mitochondria in 0 cells cannot exert its electron transfer function because of the absence of functional cytochrome c oxidase.
However, its second role in activation of apoptosis is preserved. By confocal microscopy and Western blot analysis, we were able to detect release of cytochrome c from mitochondria into cytosol in staurosporine-treated 0 cells. The release of cytochrome c was accompanied by activation of DEVD-AMC cleavage activity. Therefore, cytochrome c-mediated caspase activation is preserved in cells lacking mitochondrial function.
The mechanisms by which cytochrome c is released from mitochondria are largely unknown. In some systems, a dissipation of ⌬ m could be observed before cytochrome c is detected in cytosol and therefore the opening of PT pore and associated loss in ⌬ m has been proposed to be responsible for its release (2). However, in other systems, the loss in ⌬ m was not detectable until after release of cytochrome c (13,16). In those systems, opening of a specific channel or pore located in the mitochondrial outer membrane could allow specific release, but evidence for that is still not available. On the other hand, mitochondrial swelling without loss of ⌬ m could cause rupture of the outer membrane and release cytochrome c together with other mitochondrial intermembrane proteins (41,42). In the latter case, a transient opening of PT pore with only a transient mitochondrial depolarization may allow a repeated, respiration-driven re-establishment of ⌬ m so that cytochrome c release can occur without ⌬ m collapse (1).
In the 143B-87 0 cells, we found that despite complete absence of mitochondrial respiration, ⌬ m appeared normal. ⌬ m is thought to be generated as a consequence of the transport of protons out of the mitochondrial matrix. In normally respiring mitochondria, three electron-transfer complexes (I, III, and IV) function as proton pumps. Therefore, lack of electron transport in 0 cells should exclude establishment of ⌬ m by these systems. An alternative explanation for retention of ⌬ m in the absence of electron transport is by a reverse action of ATP synthase (F 0 ,F 1 -ATPase). This could occur if the mitochondria consume ATP and pump H ϩ out. The nuclear subunits of F 1 part of F 0 ,F 1 -ATPase have been found to be assembled in 0 cells even though 2 of the subunits are missing (40). Indeed, using oligomycin, an inhibitor of F 0 ,F 1 -ATPase, we were able to decrease, although not completely, ⌬ m in 0 cells (data not shown). Thus, these data are consistent with an ATP-driven maintenance of ⌬ m in 0 cell.
After staurosporine treatment, a reduction of ⌬ m was observed in 0 cells. As ⌬ m loss can be due to PT and PT causes release of cytochrome c from mitochondria, a PT may underlie the dissipation in ⌬ m and cytochrome c release in staurosporine-treated 0 cells. The proteins thought to participate in the formation of PT, i.e. the adenine nucleotide translocase, the peripheral benzodiazepine receptor, and the voltage-dependent anion channel, are all coded by nuclear genes and their production should not be affected by depletion of mtDNA. In contrast, increases in adenine nucleotide translocase 2 and 3 isoforms are detected in 0 cells (43). With the patch-clamp technique, the activity and functional expression level of multiple conductance channels, which appear to be identical to PT pores (44), are found also normal in 0 cells (45). Therefore, in respiration-deficient 0 cells, staurosporine could cause PT release of cytochrome c and loss of ⌬ m. However, unlike ϩ cells, some of which can retain ⌬ m during this process, 0 cells do not contain a functional electron transport chain and a respiration-driven re-establishment of ⌬ m cannot occur.
Other proapoptotic proteins, such as AIF (46) and procaspases (41,42) would also be released if the mitochondria swell and undergo the PT. Thus, the present data showing cytochrome c release and caspase activation in 0 cells do not address whether these other signaling pathways are present. However, the GSH/GSSG redox studies indicate that the dramatic oxidation that has been commonly observed in apoptotic cells is largely absent. These results show that mitochondrial respiration and ROS generation are not essential for the induction of apoptosis by staurosporine in 143B-87 0 cells. The death signals from cytochrome c release and ⌬ m collapse may be abundant in these cells and therefore induce apoptosis even when other signals are defective. However, in systems where release of caspase-activating proteins or loss of ⌬ m are insufficient or missing, disruption of mitochondrial electron transport and ROS generation may be required to make apoptosis complete. This may explain why functional mitochondria seem necessary for tumor necrosis factor-induced caspase activation and apoptosis (26) and why an interruption of electron transport is obligatory for a Fas-mediated apoptosis system where cytochrome c release is not detected (28). This may also explain why antioxidants can prevent apoptosis in some (oxidative stress acts as an activator) but not other (⌬ m collapse and cytochrome c release) systems (21).
Another important finding of the present study, probably unrelated to the activation of apoptosis, is that untreated 0 cells are more oxidized than parental cells. More ROS generation was also detected in 0 cell with dihydrorhodamine, a fluorescence indicator of ROS (data not shown). Cellular sources of ROS are numerous. Besides mitochondrial respiratory chain enzymes, cytochrome P-450 system, nitric oxide synthase, xanthine oxidase, NADPH oxidase, lipoxygenases, cyclooxygenase, etc. are also involved in generation of ROS. Therefore, the oxidation in respiration-deficient 0 cells may come from the increased activity of these enzymes.
In conclusion, our data show that in cells lacking mtDNA and defective in mitochondrial respiration, mitochondria-derived apoptotic pathways including release of cytochrome c and loss of ⌬ m are intact and can be activated by staurosporine treatment. The mitochondria-derived ROS generation is substantially diminished and has no effect on caspase activation or morphologic changes of apoptosis. This shows that redox changes are only a supplemental signal and that non-redox, caspase-mediated apoptosis in preserved in 0 cells.