Cytochrome c release occurs via Ca2+-dependent and Ca2+-independent mechanisms that are regulated by Bax.

Release of cytochrome c from mitochondria is a key initiative step in the apoptotic process, although the mechanisms regulating this event remain elusive. In the present study, using isolated liver mitochondria, we demonstrate that cytochrome c release occurs via distinct mechanisms that are either Ca(2+)-dependent or Ca(2+)-independent. An increase in mitochondrial matrix Ca(2+) promotes the opening of the permeability transition (PT) pore and the release of cytochrome c, an effect that is significantly enhanced when these organelles are incubated in a reaction buffer that is based on a physiologically relevant concentration of K(+) (150 mm KCl) versus a buffer composed of mannitol/sucrose/Hepes. Moreover, low concentrations of Ca(2+) are sufficient to induce mitochondrial cytochrome c release without measurable manifestations of PT, though inhibitors of PT effectively prevent this release, indicating that the critical threshold for PT varies among mitochondria within a single population of these organelles. In contrast, Ca(2+)-independent cytochrome c release is induced by oligomeric Bax protein and occurs without mitochondrial swelling or the release of matrix proteins, although our data also indicate that Bax enhances permeability transition-induced cytochrome c release. Taken together, our results suggest that the intramitochondrial Ca(2+) concentration, as well as the reaction buffer composition, are key factors in determining the mode and amount of cytochrome c release. Finally, oligomeric Bax appears to be capable of stimulating cytochrome c release via both Ca(2+)-dependent and Ca(2+)-independent mechanisms.

Release of cytochrome c from mitochondria is a key initiative step in the apoptotic process, although the mechanisms regulating this event remain elusive. In the present study, using isolated liver mitochondria, we demonstrate that cytochrome c release occurs via distinct mechanisms that are either Ca 2؉ -dependent or Ca 2؉ -independent. An increase in mitochondrial matrix Ca 2؉ promotes the opening of the permeability transition (PT) pore and the release of cytochrome c, an effect that is significantly enhanced when these organelles are incubated in a reaction buffer that is based on a physiologically relevant concentration of K ؉ (150 mM KCl) versus a buffer composed of mannitol/sucrose/Hepes. Moreover, low concentrations of Ca 2؉ are sufficient to induce mitochondrial cytochrome c release without measurable manifestations of PT, though inhibitors of PT effectively prevent this release, indicating that the critical threshold for PT varies among mitochondria within a single population of these organelles. In contrast, Ca 2؉ -independent cytochrome c release is induced by oligomeric Bax protein and occurs without mitochondrial swelling or the release of matrix proteins, although our data also indicate that Bax enhances permeability transition-induced cytochrome c release. Taken together, our results suggest that the intramitochondrial Ca 2؉ concentration, as well as the reaction buffer composition, are key factors in determining the mode and amount of cytochrome c release. Finally, oligomeric Bax appears to be capable of stimulating cytochrome c release via both Ca 2؉ -dependent and Ca 2؉ -independent mechanisms.
Apoptosis, a gene-regulated form of cell death, is involved in cell deletion during organogenesis and in the control of cell proliferation and differentiation in adult tissues, as well as in various diseases (1,2). The biochemical machinery required for apoptotic cell death is constitutively present in virtually all mammalian cells and can be activated by a variety of extraand intracellular signals. Attempts to identify a common, uni-fying step in the apoptotic program, in response to various cytotoxic stimuli, have focused on the role of mitochondrial participation in this form of cell death (3)(4)(5)(6)(7).
Specifically, the release of several proteins, normally located in the intermembrane space of mitochondria, has been observed during the early stages of apoptotic cell death (8 -10). Among these proteins is cytochrome c, a component of the electron transport chain that shuttles electrons between complexes III (bc1) and IV (cytochrome oxidase). Only holocytochrome c, and not apocytochrome c, is able to stimulate procaspase-9 activation, which occurs when cytochrome c is released into the cytosol and interacts with Apaf-1 and procaspase-9 in the presence of dATP to form the apoptosome complex (11). Moreover, some observations indicate that both reduced and oxidized cytochrome c are able to activate procaspase-9 with similar efficiencies, an effect that may be related to the speed with which oxidized cytochrome c is reduced in vitro (12).
The mechanisms regulating cytochrome c release remain obscure. However, with the exception of disturbances in osmotic balance accounting for a non-selective release of intramitochondrial proteins (13,14), two distinct models for cytochrome c release have emerged that can be distinguished on the basis of whether Ca 2ϩ is required for the event. In one instance, mitochondrial Ca 2ϩ overload results in swelling and rupture of the outer membrane followed by the release of cytochrome c and other intermembrane space proteins (15,16). The Ca 2ϩ -independent model asserts that a more selective protein release occurs without changes in mitochondrial volume. This mechanism involves specific channels/pores in the outer mitochondrial membrane that may be opened and regulated by certain proapoptotic members of the Bcl-2 family of proteins, including Bax (17,18). The precise way Bax modulates cytochrome c release is controversial. In particular, there are reports asserting that this protein functions directly on mitochondria to stimulate the release of cytochrome c by forming a selective pore in the outer membrane, whereas others argue that Bax interacts with the voltage-dependent anion channel and/or the adenine nucleotide translocator (ANT) 1 to facilitate opening of the permeability transition (PT) pore (19 -21). More recent evidence indicates that truncated Bid induces a conformational change in Bax that allows this protein to insert in the outer membrane and stimulate cytochrome c release (22). A different study demonstrated the ability of Bax to induce a complete loss of cytochrome c from mitochondria isolated from Xenopus eggs without otherwise affecting the functional integrity of these organelles (23).
The present study was undertaken to provide a more definitive understanding of the mechanisms regulating cytochrome c release in response to Ca 2ϩ -dependent and Ca 2ϩ -independent effects on isolated liver mitochondria. The results indicate that cytochrome c release triggered by Ca 2ϩ is significantly enhanced when mitochondria are incubated in KCl-versus MSHbased reaction buffers. Moreover, this release of cytochrome c can be observed prior to any measurable manifestations of MPT, such as swelling or the release of accumulated Ca 2ϩ . In comparison, oligomeric, but not monomeric, Bax triggers a Ca 2ϩ -independent release of cytochrome c that is not mitigated by inhibitors of MPT. Importantly, oligomeric Bax was also able to enhance Ca 2ϩ -induced cytochrome c release, pointing to a possible dual function for this protein. These data clearly demonstrate that cytochrome c release occurs by distinct Ca 2ϩdependent and Ca 2ϩ -independent mechanisms that are each modulated by oligomeric Bax.

Isolation of Rat Liver Mitochondria-Male
Harlan Sprague-Dawley rats (6 to 8 weeks old) were killed by CO 2 inhalation in accordance with the European directive of protection of vertebrate animals for scientific research. The liver was minced on ice, resuspended in 50 ml of MSH buffer (210 mM mannitol, 70 mM sucrose, 5 mM Hepes, pH 7.5) supplemented with 1 mM EDTA and homogenized with a glass Dounce homogenizer and Teflon pestle. Homogenates were centrifuged at 600 ϫ g for 8 min at 4°C. The supernatant was decanted and recentrifuged at 5,500 ϫ g for 15 min to form a mitochondrial pellet that was resuspended in MSH buffer without EDTA and centrifuged again at 5,500 ϫ g for 15 min. The final mitochondrial pellet was resuspended in MSH buffer at a protein concentration of 80 -100 mg/ml. The purity of the mitochondrial fraction was verified by the absence of both cytosolic (glyceraldehyde-3-phosphate dehydrogenase) and nuclear (poly(ADPribose) polymerase) proteins as determined by Western blot analysis.
Measurement of Functional Activity of Isolated Mitochondria-Mitochondria (1 mg/ml) were incubated in MSH buffer or a buffer containing 150 mM KCl, 1 mM KH 2 PO 4 , 5 mM succinate, and 5 mM Tris-HCl, pH 7.4, at 25°C. Rotenone (2 M) was added to maintain pyridine nucleotides in a reduced form. Estimation of ⌬ was performed using an electrode sensitive to the lipophilic cation tetraphenylphosphonium (TPP ϩ ). Energized mitochondria rapidly accumulate TPP ϩ from the incubation buffer and release this cation as ⌬ decays. Ca 2ϩ fluxes across the inner mitochondrial membrane were monitored using a Ca 2ϩ -sensitive electrode (model 97-20; Orion Research, Inc., Beverly, MA). Mitochondrial swelling was monitored continuously as changes in A 540 . Oxygen consumption by isolated rat liver mitochondria was measured using a Clark-type oxygen electrode (Yellow Spring Instrument Co., Yellow Springs, OH) at 25°C. Mitochondria with a respiratory control ratio (defined as the rate of respiration in the presence of ADP divided by the rate obtained following the expenditure of ADP) above 4 were used for all experiments. In some instances, mitochondrial respiration was uncoupled by the addition of 1 M carbonyl cyanide m-chlorophenylhydrazone (CCCP). Fresh mitochondria were prepared for each experiment and used within 4 h. At the end of the incubation period, mitochondrial suspensions were centrifuged at 10,000 ϫ g for 5 min, and the resulting supernatants and/or pellets were used for Western blot analysis.
Expression and Purification of Bax-Expression and purification of full-length Bax protein was performed as described previously (24). Briefly, the full-length human Bax cDNA sequence was amplified by standard polymerase chain reaction techniques. The polymerase chain reaction DNA fragment was isolated using the QIAquick kit (Qiagen) and subcloned into the NcoI and HindIII sites of the plasmid pBAD. The plasmid was transformed into Escherichia coli, and transformants were isolated by selection for ampicillin resistance. Cultures of the resistant colony were grown to an A 650 of ϳ7. After induction, the culture was further incubated for several hours, and cells were harvested by centrifugation. The cells were resuspended in lysis buffer (100 mM Hepes-NaOH, pH 8.0, 100 mM NaCl, 1 mM MgCl 2 , 0.1% 2-mercaptoethanol, 1% Triton X-100, a mixture of protease inhibitors, 30 g/ml DNase I, and 50 g/ml lysozyme) and broken by sonication. After centrifugation, Bax was recovered in the supernatant. Because the protein was expressed with a His tag at the N terminus, the purification by affinity chromatography on nickel-nitrilotriacetic acid-agarose (Qiagen) followed by ion-exchange chromatography on Q-Sepharose (Amersham Pharmacia Biotech) was performed according to the manufacturer's instructions. The protein was at least 95.8% pure as determined by SDS polyacrylamide gel electrophoresis. Multimerization of recombinant Bax to a stable complex, which did not dissociate into smaller species, was performed in the presence of 1% octyl glucoside (24).
Statistical Analysis-Data are presented as means Ϯ S.D., and significance was determined using a Student's t test. A value of p Ͻ 0.05 was considered to be significant.

Mitochondrial Permeability Transition and Cytochrome c Re-
lease-MPT is a consequence of Ca 2ϩ overload (15,16), and the sensitivity of mitochondria to permeability transition (PT) can be enhanced by different factors. Among these factors are an elevated level of P i , oxidative stress, the depletion of adenine nucleotides, and the oxidation of pyridine nucleotides. Fig. 1 is representative of a typical response of mitochondria to PTinducing agents. In particular, when P i or organic hydroperoxide is added to Ca 2ϩ -loaded isolated mitochondria ( Although the rate of Ca 2ϩ release during MPT is similar in KCl versus MSH incubation buffers ( Fig. 2A), the amount of cytochrome c released is significantly more pronounced when mitochondria are incubated in the more physiologically rele- vant KCl buffer as compared with MSH buffer (Fig. 2B). Further evidence of this is reflected in the assorted effects that cytochrome c release has on mitochondrial respiration in the two buffers (Fig. 2C). Specifically, the induction of MPT in MSH buffer resulted in a time-dependent acceleration of oxygen consumption that persisted until all oxygen was consumed. In contrast, incubation of mitochondria in KCl buffer resulted in an initial acceleration of respiration that was followed by its suppression, and no additional stimulation of oxygen consumption was observed upon the addition of the uncoupler CCCP. Moreover, Western blot analysis revealed that most cytochrome c was present in the supernatant fraction following MPT induction in KCl buffer, whereas only partial release of cytochrome c was observed in MSH buffer (Fig. 2D). Taken together, these data indicate that although PT-induced uncoupling of mitochondria occurs in both buffers, respiration is suppressed only in KCl buffer, an effect most likely because of a greater loss of cytochrome c.
Low Doses of Ca 2ϩ Stimulate MPT and Cytochrome c Release in a Subpopulation of Mitochondria-In the aforementioned experiments, cytochrome c release was a consequence of PT induction, accompanied by swelling of mitochondria and rupture of the outer membrane. The next step was to determine whether cytochrome c release could occur when Ca 2ϩ loading was insufficient to induce observable manifestations of MPT. As seen in Fig. 3A, Ca 2ϩ loading of mitochondria alone did not induce changes characteristic of MPT (cf. Fig. 1C), and this cation was retained by mitochondria unless CCCP, an uncoupler of oxidative phosphorylation, was added. In addition, Ca 2ϩ -loaded mitochondria exhibited controlled respiration (Fig. 3B), although increasing the Ca 2ϩ retention time from 1 to 5 min prominently diminished the rate of CCCP-directed uncoupled respiration (Panel B, trace a versus trace b). Taking into consideration that the observed decrease in the rate of uncoupled respiration might be because of damage of the respiratory chain and/or the release of cytochrome c, the presence of this protein in the incubation buffer was analyzed. As seen in Fig. 3C, the accumulation of Ca 2ϩ by mitochondria in the absence of observable PT was sufficient to stimulate a release of cytochrome c, an effect that was considerably more pronounced the longer mitochondria retained accumulated Ca 2ϩ . Inhibitors of MPT, such as ADP and Mg 2ϩ , suppressed this release (Fig. 3C) and restored the rate of uncoupled respiration (Fig. 3D). Additional evidence of cytochrome c release in the absence of observable MPT is presented in Fig. 4 where the sequential addition of Ca 2ϩ pulses to isolated mitochondria led to a stepwise decrease in the overall optical density (Fig. 4A, traces 1-5) without inducing large amplitude swelling characteristic of PT. Meanwhile, the amount of cytochrome c released from mitochondria was enhanced as Ca 2ϩ loading increased (Fig. 4B). This release reflected PT induction in a subpopulation of mitochondria, which was further supported by the fact that co-treatment of isolated mitochondria with 1 M CsA eliminated cytochrome c release induced by Ca 2ϩ additions ranging between 20 and 80 nmol/mg protein (data not shown). Cytochrome c release ultimately reached a pinnacle at 80 nmol Ca 2ϩ /mg protein (Fig. 4B, lanes 6 and 7) as all mitochondria underwent PT (Fig. 4A).

Bax-induced Cytochrome c Release Can Occur via Both MPTindependent and MPT-dependent
Mechanisms-Next, we tested the effect of recombinant Bax protein on cytochrome c release. Bax has been reported to facilitate cytochrome c release in response to different stimuli, although the precise mechanism responsible for this event remains unclear. Both monomeric and oligomeric forms of Bax (Fig. 5A) were generated for our study because of reports indicating that oligomerization of Bax is a critical event for integration of this protein into membranes (22). It should be noted that our monomeric form of Bax also contained the oligomeric form (Fig. 5A, lane 2), although it is clear from the data that these different pools of recombinant Bax had strikingly dissimilar effects on cytochrome c release.
The addition of oligomeric Bax to isolated mitochondria prominently induced the release of cytochrome c (Fig. 5B). This release did not depend on MPT, because the incubation buffer used for these experiments contained 1 mM EGTA, a concentration sufficient to chelate available Ca 2ϩ and hence prevent MPT. In contrast to the results obtained with oligomeric Bax, cytochrome c release was marginal in the presence of the mo-nomeric form of Bax (Fig. 5B), which is in agreement with other data indicating that only oligomeric Bax is able to induce cytochrome c release (25). The slight release observed in the presence of monomeric Bax was most likely because of the presence of some of the oligomeric form in this preparation (Fig. 5A, lane 2).
In addition to stimulating an MPT-independent release of cytochrome c, recombinant oligomeric Bax also facilitated Ca 2ϩ -induced MPT as assessed by both the release of accumulated Ca 2ϩ (Fig. 5C) and a drop in mitochondrial membrane potential (Fig. 5D). In contrast, the monomeric form of this protein had no effect on either parameter as compared with control mitochondria (Fig. 5, C and D). Western blot analysis of cytochrome c release in samples taken 6 min after the addition of Ca 2ϩ revealed that oligomeric Bax enhanced MPT-dependent cytochrome c release (Fig. 5E). Finally, an attempt to compare the ability of oligomeric Bax to enhance MPT-independent and MPT-dependent release of cytochrome c revealed, as we expected, that cytochrome c release was significantly more pronounced in samples taken from Bax-treated mitochondria that were stimulated to undergo PT (Fig. 5F).
To determine whether the combined or individual effects of Bax and Ca 2ϩ on cytochrome c release were unique for this apoptogenic protein, we also evaluated the release of other mitochondrial proteins. Results indicated that the intermembrane space protein adenylate kinase-2 was released into the cytosol during MPT-independent and MPT-dependent modes of protein release, an effect that was consistent with our cytochrome c release data (Fig. 6, A and B). In contrast, the mitochondrial matrix protein adenylate kinase-3 was released only via PT, although oligomeric Bax significantly enhanced this result by potentiating the onset of PT (Fig. 6C). Taken together, our data indicate that the effect of Bax alone is not specific for cytochrome c release insofar as adenylate kinase-2, a different intermembrane space protein, is also released. However, treatment of isolated mitochondria with Bax alone does not appear to compromise matrix integrity, because this protein did not initiate or enhance the release of adenylate kinase-3 unless PT was induced.

DISCUSSION
Although it was originally believed that MPT induction was the root mechanism responsible for cytochrome c release in response to different cytotoxic stimuli, more recently this notion has been challenged, and the precise mechanisms regulating the release of this protein are unclear. In fact, many of the early results on mechanisms of cytochrome c release were generated using cell-free systems wherein isolated mitochondria and nuclei were treated with different PT pore activators, which, in turn, led to mitochondrial swelling, the release of cytochrome c (and other proteins), and subsequent changes in nuclear morphology that were characteristic of apoptosis. However, ample evidence from more recent studies suggests that although MPT is likely to be a mechanism responsible for cytochrome c release, it is no longer regarded as the mechanism (26). In particular, Martinou et al. (27) demonstrated that mitochondria of neurons undergoing apoptosis, induced by neuronal growth factor depletion, actually reduced in size and resumed their normal function when reincubated with neuronal growth factor. A more recent study from our laboratory demonstrated that etoposide stimulated cytochrome c release from isolated mitochondria, despite the presence of 1 mM EGTA (a known inhibitor of MPT) in the reaction buffer (28). Thus, the current study was undertaken to examine more comprehensively the mechanisms regulating the release of this important apoptogenic protein and the potential effect recombinant Bax protein has on this event. We clearly demonstrate here, using isolated liver mitochondria, that cytochrome c release occurs via distinct mechanisms that can be divided into two groups, Ca 2ϩ -dependent and Ca 2ϩ -independent.
As mentioned previously, it is well known that Ca 2ϩ is the basic PT pore activator (16). Coincidentally, Ca 2ϩ -dependent cytochrome c release occurs when mitochondria experience Ca 2ϩ overload, resulting in the induction of PT. Yet a number of reports in the literature make claims of MPT-dependent cytochrome c release in the absence of Ca 2ϩ ; in fact, many of these studies base this claim on the fact that CsA, a pharmacological inhibitor of MPT, effectively blocks the release of this apoptogenic protein (29,30). For example, Kroemer and coworkers (29) recently reported the ability of Bid (also a proapoptotic Bcl-2 family protein) to induce cytochrome c release by opening the permeability transition pore, an effect they base solely on the fact that it was preventable by CsA and other inhibitors of PT. Although this may be the case, a far more accurate and convincing determination of MPT would have included the presence of Ca 2ϩ -mediated large amplitude swelling. In short, claims of cytochrome c release via MPT in the absence of Ca 2ϩ are at best improbable. This was our rationale for employing a multiparameter mitochondrial functional analysis, including measurements of swelling, decreases in ⌬, changes in respiration, and the release of accumulated Ca 2ϩ , to more accurately link changes in mitochondrial integrity and the release of cytochrome c.
Consistent with data reported by other investigators, we demonstrate the ability of mitochondria undergoing PT to release cytochrome c (31,32). The fact that this effect was more pronounced in KCl versus MSH buffers is likely to be related to the understanding that cytochrome c binding to the inner mitochondrial membrane is weaker in buffers with higher ionic strength (33). In addition to cytochrome c release occurring during full-blown MPT, our results indicate that this protein is released even when Ca 2ϩ loading is below the threshold needed to elicit observable manifestations of PT, though different inhibitors of MPT, including ADP and Mg 2ϩ , were able to suppress cytochrome c release. Thus, it appears that low Ca 2ϩ loading is sufficient to induce PT in a subpopulation of mitochondria, and as recently described by Rizzuto et al. (34) the released Ca 2ϩ is subsequently taken up by neighboring polarized mitochondria. This effect can likely be explained by the inherent heterogeneity of mitochondria (35) that make up any particular population of these organelles in terms of their sensitivity to PT. In other words, all mitochondria are not created equally.
The mechanisms responsible for Ca 2ϩ -independent cytochrome c release are less clear, although increasing evidence indicates that certain proapoptotic members of the Bcl-2 family of proteins, notably Bax, are able to participate in this process. One possible, and increasingly popular, mechanism is that in the absence of MPT, or any changes in mitochondrial volume, oligomeric Bax inserts and forms a channel in the outer membrane large enough to allow the release of cytochrome c (20,36). As our data suggest, monomeric Bax is incapable of stimulating cytochrome c release, which is consistent with reports indicating that oligomerization of Bax is required prior to its insertion in the membrane. It should be noted that in addition to the ability of Bax to stimulate cytochrome c release in the absence  A and B) and matrix (C) proteins. After a 2-min stabilization period, mitochondria were treated with 20 g/ml oligomeric Bax and/or Ca 2ϩ (40 nmol/mg protein), and samples were taken for Western blot analysis after either 10 (control or Bax alone) or 6 min (Ca 2ϩ plus Bax or Ca 2ϩ alone) of incubation. Results are typical of three or more independent experiments. Cyt c, cytochrome c; AK, adenylate kinase. of any changes in mitochondrial volume, this protein also possesses the ability to enhance MPT-mediated release of cytochrome c. In this case, Bax may interact with the permeability transition pore proteins ANT or voltage-dependent anion channel and hasten the opening of the pore. In fact, it was previously reported that Bax interacts directly with ANT and facilitates ANT opening induced by atractyloside (an ANT inhibitor) and that ANT-deficient yeast are resistant to Baxinduced cell death, suggesting that ANT may also be a functional target of Bax (26,37). At the same time, it was demonstrated that Bax and Bak directly target and open voltagedependent anion channel in liposomes and induce changes in ⌬ (21). Interestingly, it should be noted that more recent results from our laboratory indicate that Ca 2ϩ -independent cytochrome c release can also occur via mitochondrial swelling stimulated by alterations in osmotic balance. 2 Recently, evidence was provided that cardiolipin, which is present exclusively in mitochondrial membranes, mediates the targeting of truncated Bid to mitochondria through a previously unknown three-helix domain in truncated Bid and that this interaction is critical for cytochrome c release (38). The involvement of cardiolipin in the release of cytochrome c was also documented by data showing that elevated levels of mitochondrial phospholipid hydroperoxide glutathione peroxidase in cells triggered to undergo apoptosis completely suppressed the release of cytochrome c (39). Taken together, these findings suggest that cytochrome c release may occur by a two-step process, wherein this protein is first liberated from cardiolipin and then released via specific or nonspecific pores/channels in the outer mitochondrial membrane.
In summary, we have demonstrated that mitochondria release cytochrome c via distinct Ca 2ϩ -dependent or Ca 2ϩ -independent mechanisms. In the first case, mitochondrial Ca 2ϩ overload causes swelling, rupture of the outer mitochondrial membrane, and the nonspecific release of cytochrome c. In contrast, Ca 2ϩ -independent protein release is regulated by the oligomeric form of the proapoptotic protein Bax, is specific for intermembrane space proteins, and occurs without MPT. Importantly, oligomeric Bax also enhances MPT-induced protein release. Which, if any, of these mechanisms predominates under physiological conditions remains unclear. However, it is tempting to speculate that under pathological conditions when the level of intracellular Ca 2ϩ increases, such as during ischemia-reperfusion injury, swelling, rupture, and the release of mitochondrial proteins occurs by PT, whereas during "authorized" cell death or receptor-mediated killing this release occurs without swelling and proceeds via a Bax-and/or truncated Bid-regulated pathway.