Mitochondrial respiratory control is lost during growth factor deprivation

Abstract

The ability of cells to maintain a bioenergetically favorable ATP/ADP ratio confers a tight balance between cellular events that consume ATP and the rate of ATP production. However, after growth factor withdrawal, the cellular ATP/ADP ratio declines. To investigate these changes, mitochondria from growth factor-deprived cells isolated before the onset of apoptosis were characterized in vitro. Mitochondria from growth factor-deprived cells have lost their ability to undergo matrix condensation in response to ADP, which is accompanied by a failure to perform ADP-coupled respiration. At the time of analysis, mitochondria from growth factor-deprived cells were not depleted of cytochrome c and cytochrome c-dependent respiration was unaffected, demonstrating that the inhibition of the respiratory rate is not due to loss of cytochrome c. Agents that disrupt the mitochondrial outer membrane, such as digitonin, or maintain outer membrane exchange of adenine nucleotide, such as Bcl-x_L, restored ADP-dependent control of mitochondrial respiration. Together, these data suggest that the regulation of mitochondrial outer membrane permeability contributes to respiratory control.

Many biochemical reactions in cells depend on a tightly controlled ratio of ATP to ADP for their execution (1). This ratio is preserved by regulatory mechanisms that couple the rate of ATP consumption in the cell to the rate of ATP production by oxidative phosphorylation in mitochondria (2). Mitochondria use oxidizable substrates to produce NADH/FADH₂ that is used by the electron transport chain to produce an electrochemical potential of protons. This potential is used to produce ATP from ADP and P_i by the ATP synthase and to regulate ion exchange across the mitochondrial inner membrane.

Mitochondria can also be a source of signals that lead to apoptosis (3). Several apoptogenic factors that reside in the mitochondrial intermembrane space are released into the cytosol on apoptotic insults. These factors include cytochrome c, Smac/Diablo, Omi, AIF, and endonuclease G (3). The release of cytochrome c from mitochondria induces the assembly of the apoptosome, a cytosolic complex comprising cytochrome c, Apaf-1, and caspase-9 that activates processes leading to cell death (4, 5). In addition, changes in mitochondrial physiology have been demonstrated during many apoptotic processes (6–8), yet the connection between mitochondrial physiology and the release of apoptogenic factors is still unclear.

After growth factor withdrawal, mitochondria exhibit a decline in both the rate of respiration and mitochondrial membrane potential (9). Although such alterations in mitochondrial physiology have been documented, the relation of these changes to apoptosis, as well as the nature of their biochemical effectors, has not been determined (10). Moreover, a need still exists to establish whether the changes in mitochondrial physiology during apoptosis are simply byproducts of the apoptotic cascade or whether these changes represent essential events in the ultimate release of apoptogenic factors from the mitochondrial intermembrane space. Whether and how the role of mitochondria in regulating energy metabolism relates to a cell's commitment to life or death remains to be defined.

In this work we present evidence for a connection between extracellular signals and the regulation of oxidative phosphorylation. Mitochondria of the IL-3-dependent FL5.12 cell line lost their ability to regulate respiration in response to ADP early after IL-3 withdrawal. These mitochondrial changes occurred while cytochrome c was still available to the respiratory chain and at a time when resupplying IL-3 rescued the cells from apoptosis. The loss of ADP-coupled respiration seemed to result from a decline in the permeability of the mitochondrial outer membrane. In mitochondria isolated from growth factor-deprived cells, ADP-coupled respiration could be restored by either treatment with doses of digitonin sufficient to permeabilize the mitochondrial outer membrane or as a result of Bcl-x_L expression.

Materials and Methods

Materials.

Anti-cytochrome c oxidase subunit IV antibody, clone 20E8-C12, was purchased from Molecular Probes. Anti-cytochrome c antibody, clone 7H8.2C12, and recombinant IL-3 were purchased from PharMingen. Digitonin (high purity) and rotenone were purchased from Calbiochem. N,N,N′,N′-tetramethylphenylenediamine (TMPD), EGTA, mannitol, sucrose, ascorbate, glutamate, malate, succinate, antimycin A, BSA, cytochrome c, duroquinone, and ADP were purchased from Sigma. Sodium cacodylate and glutaraldehyde were purchased from Electron Microscopy Sciences (Fort Washington, PA).

Cell Culture.

The murine IL-3-dependent cell line FL5.12 and derived clones (FL5/neo and FL5/x_L) were cultured at 37°C and 5% CO₂ in RPMI medium 1640 supplemented with 10% fetal bovine serum, penicillin/streptomycin, 10 mM Hepes, 50 μM 2-mercaptoethanol, and 300 pg/ml IL-3. For apoptosis induction, cells were washed three times in serum-free RPMI medium 1640 and resuspended in complete medium with or without IL-3 for 12 h.

Mitochondrial Isolation.

Mitochondria were isolated by using a differential centrifugation method that retains mitochondrial structure and respiratory functions (11, 12). Cells (2.5–5 × 10⁸) were harvested and placed on ice for 15 min, centrifuged at 500 × g for 5 min at 4°C, washed one time with ice-cold PBS, and subsequently washed with ice-cold mitochondrial isolation buffer (MIB) (200 mM mannitol/70 mM sucrose/1 mM EGTA/10 mM Hepes; pH 7.4). Cells were resuspended in ice-cold MIB + 0.5 mg/ml BSA (MIB/BSA) and then homogenized in a syringe-driven cell disruptor. The lysate was spun at 800 × g for 10 min at 4°C. Supernatants were removed and spun at 10,000 × g for 10 min at 4°C. Pellets were resuspended in MIB/BSA and normalized for protein concentration by using the Bradford Reagent (Bio-Rad). Typically mitochondrial protein was adjusted to 2–3 mg/ml.

For the outer membrane permeabilization experiments, mitochondria were split into two samples in MIB (1 mg/ml) and 0.15 mg digitonin per mg of mitochondrial proteins was added to one of the tubes. Samples were kept on ice for 10 min with frequent shaking, and mitochondria were spun at 11,200 × g for 10 min at 4°C. Pellets were resuspended in MIB/BSA and normalized again to protein concentration.

Electron Microscopy Preparation.

Mitochondria (200 μg of protein) were added to 1 ml of reaction buffer (250 mM sucrose/20 mM Hepes/10 mM KCl/3 mM KH₂PO₄/1.5 mM MgCl₂/1 mM EGTA/0.5 mg/ml BSA; pH 7.4) with or without oxidizable substrates (7 mM malate/glutamate or 7 mM succinate) and ADP (500 μM). Mitochondria were incubated for 5 min at 30°C and then spun at 10,000 × g for 10 min at 4°C. Buffer was replaced with 1 ml of fixative (2% glutaraldehyde in 0.1 M sodium cacodylate; pH 7.4) and samples were kept at 4°C over night and analyzed by standard electron microscopy (EM) procedures.

Oxygen Consumption Assays.

Mitochondrial suspensions (500 μg of protein) were diluted to a total volume of 3.5 ml in reaction buffer. The oxygen consumption rate was measured in a water-jacketed respirometer chamber by using a polarographic oxygen electrode (Cameron Instrument Co., Port Aransas, TX) at 30°C. Reagents were added to the chamber during the assay by using Hamilton Microliter fixed-needle syringes. For respiratory control ratio (RCR) analysis, mitochondria were diluted in reaction buffer with 5 μM rotenone and 7 mM succinate (except when state 1 respiration was analyzed before the addition of succinate). Alternatively, mitochondria were analyzed without rotenone in the presence of malate and glutamate (7 mM each). After recording state 4 respiration, 150 μM ADP was added to induce state 3 respiration. Where indicated, mitochondria were incubated with 6.7 nM recombinant truncated Bid (tBid) and/or 10 μM cytochrome c for 5 min before the respiration analysis. For cytochrome c-dependent respiration, mitochondria were diluted in reaction buffer with 5 μg/ml antimycin A and placed in the respirometer chamber. After recording baseline respiration, 0.25 mM TMPD and 2.5 mM ascorbate were added to the chamber. For the analysis of the rate of uncoupled respiration with duroquinol, mitochondria were diluted in reaction buffer with 15 μM carbonylcyanide m-chlorophenlyhydrazone and 5 μM rotenone. Duroquinol (100 μM) was added after recording a baseline respiratory rate. Duroquinol was prepared by reducing duroquinone with potassium borohydride and hydrochloric acid as described (13).

Results

Mitochondria of IL-3-Deprived Cells Failed to Conduct ADP-Regulated Structural Changes.

The release of apoptogenic factors from mitochondria has been established as a key determinant in the initiation of apoptosis in response to growth factor withdrawal. Whether the mitochondria simply act as storage organelles before the release of these apoptogenic factors by signals from outside the mitochondria, or whether changes in mitochondrial functions contribute to the loss of integrity of the outer membrane, remains a controversial issue. Changes in mitochondrial structure have been observed in cells during apoptosis, and these changes have been attributed to defects in ADP/ATP exchange between the cytosol and mitochondria (9, 14). To characterize the mitochondrial changes that take place early after growth factor deprivation, EM analysis was performed on mitochondria isolated from FL5.12 cells grown with IL-3 (control) and from cells 12 h after IL-3 withdrawal (IL-3-deprived). When mitochondria were incubated in vitro without oxidizable substrates, the matrix was condensed in both control and IL-3-deprived mitochondria (Fig. 1; no substrates). Matrix condensation of mitochondria (condensed configuration) has been previously observed in nonenergized isolated mitochondria and is characterized by dense mitochondrial matrix and increased intermembrane space (15). Addition of oxidizable substrates, such as succinate, an electron donor to complex II of the respiratory chain, resulted in matrix swelling and the formation of defined cristae [orthodox configuration (15)] (Fig. 1; +succinate), demonstrating that in vitro, substrate readdition can promote matrix swelling in mitochondria from both control and IL-3-deprived cells. When ADP was added to isolated mitochondria energized with substrates (in the orthodox configuration), the matrix of mitochondria isolated from control cells rapidly converted back to the condensed configuration. In contrast, when energized mitochondria isolated from IL-3-deprived cells were treated with ADP, they failed to undergo a reversion to a condensed state (Fig. 1; +succinate + ADP). The same results were observed when malate (together with glutamate) was used to generate mitochondrial NADH as an electron donor to complex I of the respiratory chain (data not shown). More than a hundred mitochondria from each treatment were analyzed and those shown in Fig. 1 represent more than 90% of the mitochondria of each group. These data suggest that mitochondria from IL-3-deprived cells have developed a physiological defect that is manifested by their inability to use ADP to regulate matrix volume.

Figure 1.

Mitochondria isolated from IL-3-deprived cells respond to reducing substrates but not to ADP. Cells were maintained in the presence or absence of IL-3 for 12 h. Mitochondria were isolated and placed in a reaction buffer either alone (no substrates = state 1) or together with succinate (+succinate = state 4), or succinate and ADP (+succinate +ADP = state 3) for 5 min followed by fixation and EM analysis. The dark regions (higher electron density) correspond to the matrix.

Mitochondria from IL-3-Deprived Cells Are Limited in Their Ability to Conduct ADP-Coupled Oxidative Phosphorylation.

When mitochondria are maintained in a phosphate-containing sucrose buffer, without exogenous substrates, they consume oxygen slowly (state 1). When reducing substrates that can donate hydrogen atoms (electrons and protons) to the respiratory chain are added, mitochondria start to respire at a slightly higher rate (state 4) (16). However, because of the buildup of a proton gradient across the inner membrane, the state 4 rate of respiration remains low and is proportionate to the rate of proton leakage across the inner membrane (17). To induce an increased respiratory rate, the substrate-dependent proton gradient must be dissipated. One way to stimulate respiration of energized mitochondria is by adding ADP, which is then converted to ATP by the ATP synthase using the proton gradient established across the inner membrane by the electron transport chain. Consequently, an increase in mitochondrial oxygen consumption is achieved when ADP is supplied as a substrate for the ATP synthase (state 3). When limiting amounts of ADP are added to mitochondria, state 3 respiration persists until the added ADP is converted to ATP. Once this conversion reaches equilibrium, oxygen consumption returns to state 4 (17).

To determine whether mitochondria from IL-3-withdrawn cells fail to respond structurally to ADP because of a defect in ADP-coupled respiration, the respiratory properties of mitochondria were examined. Mitochondria were isolated from FL5.12 cells cultured either in the presence or absence of IL-3 for 12 h, and analyzed in a sealed chamber with an oxygen-sensitive electrode. Mitochondria were normalized to protein levels, and diluted into the reaction buffer containing succinate as an electron donor. State 4 respiration was recorded for 200 s before the addition of ADP to the chamber. No significant differences in the state 4 rate of oxygen consumption between mitochondria isolated from IL-3-deprived or control cells were observed (Fig. 2). This result correlated with the structural changes observed above in the configuration of the mitochondria after addition of succinate (Fig. 1). However, when ADP was added, mitochondria from FL5.12 cells deprived of IL-3 for 12 h showed a significantly decreased state 3 respiration as compared with mitochondria from control cells (Fig. 2).

Figure 2.

Mitochondria from IL-3-deprived cells display a reduction in ADP-coupled respiration. (A) Cells were maintained in the presence or absence of IL-3 for 12 h followed by mitochondria isolation. Mitochondria (500 μg) were diluted into reaction buffer containing rotenone and succinate, and oxygen levels were recorded. State 4 respiration was recorded for 200 s followed by the addition of ADP to induce state 3 respiration. (B) The rate of oxygen consumption, in states 4 and 3, of mitochondria isolated from control (+IL-3) or IL-3-deprived (-IL-3) cells. The results are the average and standard deviation of 3 independent experiments.

One important parameter that describes the mitochondrial response to ADP is the RCR. This is the ratio between the rates of oxygen consumption of mitochondria in state 3 to the rate of oxygen consumption in state 4. The RCR of mitochondria from cells cultured in the presence of IL-3 was 4.7 ± 0.4 (average ± SD from three independent experiments). In contrast, the RCR fell to 2.1 ± 0.2 in mitochondria from cells cultured in the absence of IL-3 for 12 h.

Cytochrome c Is Not Limiting for Respiration 12 h After IL-3 Withdrawal.

Because cytochrome c is a required component of the respiratory chain, a decrease in RCR could be a consequence of less efficient respiration due to cytochrome c release. However, as reported (18), cytochrome c was found to be fully contained within mitochondria in FL5.12 cells 12 h after IL-3 withdrawal (data not shown). This finding is in line with other reports in which respiratory disruptions were observed before cytochrome c release from mitochondria (8). Nevertheless, to rule out the possibility that cytochrome c is limiting for respiration in IL-3-deprived cells, TMPD was used as a substrate for mitochondrial respiration. TMPD is a membrane-permeable electron donor that directly reduces cytochrome c, allowing an assessment of functional cytochrome c. When TMPD was added to the respirometer chamber, together with excess amounts of ascorbate to recycle oxidized TMPD, the respiration rate of mitochondria from IL-3-deprived cells was equal to that of mitochondria from control cells (Fig. 3A). The TMPD-induced respiration was cytochrome c-dependent, because the removal of the outer membrane and the release of cytochrome c by mitoplast preparation through hypotonic shock and homogenization resulted in abrogation of TMPD-dependent respiration unless exogenous cytochrome c was added to the chamber (data not shown). The rate of nonenzymatically consumed oxygen by TMPD/ascorbate mixture in the reaction buffer without mitochondria was negligible (data not shown).

Figure 3.

Cytochrome c loss does not account for the decrease in RCR after IL-3 withdrawal. (A) Cytochrome c-dependent respiration was measured using mitochondria from cells growing in the presence or absence of IL-3 for 12 h. After isolation, mitochondria (500 μg) were diluted into reaction buffer in the presence of antimycin A. Where indicated, TMPD and ascorbate, as electron donors for cytochrome c, were added to the chamber. (B) The RCR of succinate-mediated respiration of mitochondria isolated from control (+IL-3) or IL-3-deprived (−IL-3) cells in the presence or absence of exogenous cytochrome c. Mitochondria were prepared and analyzed as in Fig. 2. (C Upper) Mitochondria from cells growing in the presence of IL-3 were isolated and incubated with or without tBid for 5 min. (Lower) tBid-treated mitochondria were analyzed in the presence or absence of exogenous cytochrome c. After adding rotenone and succinate, samples were analyzed in the respirometer as in Fig. 2. (D) The RCR of mitochondria treated or untreated with tBid, with or without exogenous cytochrome c as shown in C.

To test whether mitochondrial outer membrane permeability allowed cytochrome c to escape from mitochondria, exogenous cytochrome c was added to the preparations. When succinate-mediated respiration was performed in the presence of 10 μM exogenously added cytochrome c, the drop in RCR observed in mitochondria from IL-3-deprived cells was not corrected (Fig. 3B), suggesting that cytochrome c had not been released from mitochondria in either preparation.

tBid is the active form of the proapoptotic Bcl-2 protein Bid and a potent cytochrome c-releasing agent from isolated mitochondria (19–22). When mitochondria isolated from control cells were incubated with recombinant tBid before respiration analysis, a very poor response to ADP was observed (Fig. 3C Upper). The addition of exogenous cytochrome c completely restored the ADP response in mitochondria that had been incubated with tBid before the respiration analysis (Fig. 3C Lower). The observation that exogenously added cytochrome c could fully correct the RCR of tBid-treated mitochondria (Fig. 3D), but not the RCR of mitochondria from IL-3-deprived cells (Fig. 3B), suggests that cytochrome c is not limiting for respiration in mitochondria from IL-3-deprived cells. Therefore, after IL-3 withdrawal, major physiological changes in the control of mitochondrial respiration have occurred before both the release of cytochrome c and the commitment to apoptosis. These changes precede those initiated by activation of proapoptotic Bcl-2 proteins by BH3-containing proteins such as tBid, and may directly regulate or contribute to the subsequent ability of mitochondria to release apoptogenic factors from the intermembrane space.

Permeabilization of the Mitochondrial Outer Membrane Restores the ADP Response of Mitochondria from IL-3-Deprived Cells.

As shown above, the decreased ability of mitochondria to perform ADP-coupled respiration early in apoptosis cannot be explained by the proposed effects of proapoptotic Bcl-2 proteins to create a membrane-permeant channel for cytochrome c (23–25). The defective response to ADP in terms of activating state 3 respiration in IL-3-deprived cells may be attributed either to a less active ATP synthase or to a decrease in ADP transport. To test the latter possibility, digitonin was used to permeabilize the cholesterol-rich mitochondrial outer membrane before respiration analysis. To assess the ability of digitonin to disrupt the mitochondrial outer membrane, digitonin-treated mitochondria were analyzed by EM. Disruptions of the mitochondrial outer membrane were observed in digitonin-treated mitochondria with no apparent damage to the inner membrane and matrix structure (Fig. 4A). The digitonin-induced disruptions of the mitochondrial outer membrane include membrane pinching, small gaps, and large gaps (Fig. 4A; arrows 1, 2, and 3, respectively). Permeabilizing the outer membrane with digitonin is not sufficient to induce substantial cytochrome c release, because the majority of the cytochrome c is retained in the mitochondria (Fig. 4B), which may be due to the fact that cytochrome c binds to the mitochondrial inner membrane through interactions with cardiolipins and/or cytochrome c oxidase (26).

Figure 4.

Digitonin causes disruption of the outer mitochondrial membrane but does not cause a significant release of cytochrome c. (A) Mitochondria were isolated from control (+IL-3) or IL-3-deprived (−IL-3) cells and maintained in the isolation buffer (MIB) without substrates. Where indicated, digitonin was added and mitochondria were treated as described in Materials and Methods and subjected to EM analysis. Arrows indicate different types of lesions to the outer membrane including 1) membrane pinching, 2) small gaps, 3) large gaps. (B) Cytochrome c (cyt c) levels and cytochrome c oxidase (COX IV) levels, as a loading control, were assessed by Western blot analysis of mitochondria isolated from cells growing in the presence or absence of IL-3 for 12 h with or without digitonin treatment.

When digitonin-treated mitochondria from cells growing in the presence of IL-3 were analyzed in the respirometer, they showed a reproducibly lower rate of state 3 respiration when compared with nondigitonized mitochondria (Fig. 5A). In contrast, mitochondria from IL-3-deprived cells showed a reproducible and statistically significant increase in the rate of state 3 respiration when treated with digitonin (Fig. 5B). This increase in ADP-stimulated respiration did not seem to be a result of inner membrane permeation to protons by digitonin. If uncoupling of mitochondria is the mechanism for digitonin-enhanced respiration, a fall in RCR should be observed, because mitochondria would not require ADP to dissipate the proton gradient across the inner membrane. Analysis of three independent experiments comparing digitonized versus nondigitonized mitochondria showed an increase in RCR of digitonized mitochondria from IL-3-deprived cells (Fig. 5C), whereas a decrease in RCR of mitochondria from control cells was observed.

Figure 5.

Digitonin treatment restores ADP response to mitochondria isolated from IL-3-deprived cells. Mitochondria were isolated from cells growing in the presence or absence of IL-3 for 12 h. Mitochondria were either treated or untreated with digitonin and analyzed in the respirometer as described in Fig. 2. Respiration was performed in the absence of exogenously added cytochrome c. (A) Mitochondria from control (+IL-3) cells. (B) Mitochondria from cells deprived of IL-3 for 12 h. (C) RCR of mitochondria isolated from cells growing in the presence or absence of IL-3 and incubated with or without digitonin. The results are the average and standard deviation of 3 independent experiments.

The increased RCR due to digitonin treatment suggests that the loss of respiratory control in mitochondria from IL-3-deprived cells is due to limitations in ADP transport across the outer membrane. Therefore, if the proton gradient across the mitochondrial inner membrane would be dissipated by an uncoupler, the mitochondrial respiration rate should be limited only by the accessibility of oxidizable substrates. Indeed, when the uncoupler carbonylcyanide m-chlorophenlyhydrazone was added to mitochondria in the presence of succinate, a significant increase in the rate of respiration was observed in both mitochondria from control and from IL-3-deprived cells (data not shown). Still, the uncoupled respiration rate was higher in the mitochondria from control cells, suggesting that either substrate access or electron transport were compromised in mitochondria from IL-3-deprived cells. To distinguish between these two possibilities, succinate was replaced with duroquinol as an electron donor. Unlike succinate, which requires facilitated transport across the mitochondrial membranes, duroquinol is membrane-permeant. Using duroquinol as a substrate, the respiration rate of mitochondria in the presence of carbonylcyanide m-chlorophenlyhydrazone was equal in both mitochondrial populations (Fig. 6A). The difference in the response of uncoupled mitochondria from IL-3-deprived cells to succinate or duroquinol seems to result from differences in substrate accessibility to mitochondria.

Figure 6.

Recovery of uncoupled respiration of mitochondria from IL-3-deprived cells by a membrane permeant electron donor and prevention of the decrease in coupled respiration by Bcl-x_L expression. (A) Respiration analysis of uncoupled mitochondria by using duroquinol as an electron donor. Mitochondria (500 μg) from control or IL-3-deprived cells were incubated in the reaction buffer in the presence of rotenone and carbonylcyanide m-chlorophenlyhydrazone. Where indicated, 100 μM duroquinol (DQH2) was added to the chamber. (B) FL5.12 cells transfected with control vector (Neo) or with vector encoding Bcl-x_L (XL) were incubated in the presence or absence of IL-3 for 12 h. Mitochondria (500 μg) were diluted into reaction buffer in the presence of rotenone and oxygen consumption was measured. State 1 respiration (no substrate) was recorded for 100 seconds followed by the addition of succinate. ADP was added 200 s later to induce state 3 respiration.

Bcl-x_L Prevents the Early Decrease in Oxidative Phosphorylation During Apoptosis.

The antiapoptotic, mitochondrial outer membrane protein Bcl-x_L has been reported to maintain metabolite exchange across the mitochondrial outer membrane (27). FL5.12 cells were transfected with Bcl-x_L-encoding plasmid (FL5/x_L), and compared with cells transfected with the neomycin-resistance gene only (FL5/neo). Unlike mitochondria from IL-3-deprived FL5/neo cells, mitochondria from IL-3-deprived FL5/x_L cells maintained their response to ADP as judged by the ability of exogenous ADP to induce state 3 respiration (Fig. 6B). Mitochondria from IL-3-deprived FL5/x_L cells maintained a state 3 respiration and RCR similar to that of mitochondria from FL5/neo cells growing in IL-3 (Fig. 6B). Because the drop in RCR occurs early in apoptosis, before the release of cytochrome c, these data point to a role for the anti-apoptotic protein Bcl-x_L in maintaining ADP-coupled respiration in growth factor-deprived cells.

Discussion

The fact that mitochondria participate in the apoptotic response to a variety of stimuli is well established (3, 5, 24). Mitochondria are repositories of apoptogenic factors that on apoptotic signals are released into the cytosol to induce the apoptotic cascade. The mechanism by which these factors are released from mitochondria, however, is still obscure. A significant body of evidence supports the hypothesis that large pores are formed in the mitochondrial outer membrane that mediates the release of apoptogenic factors from the mitochondrial intermembrane space (25, 28, 29). The loss of outer membrane integrity has been proposed to be directly induced by the pore forming properties of the pro-apoptotic Bcl-2 proteins. In this scenario, the role of the anti-apoptotic Bcl-2 proteins is to antagonize the pore-forming activity of the pro-apoptotic members of the Bcl-2 family.

It was recently shown that disruption of the mitochondrial outer membrane is not sufficient to achieve complete cytochrome c release. Studies of mitochondrial structure have revealed that most of the cytochrome c protein is sequestered in the cristae (30). Moreover, cytochrome c can be tightly bound to the inner membrane through cardiolipin (26). To achieve complete release of cytochrome c, as is generally observed during apoptosis (31), structural and functional changes of the inner membrane may be needed (22, 32). A possible way to achieve cytochrome c detachment from the inner membrane would be through lipid peroxidation of the mitochondrial inner membrane (32) by reactive oxygen species generated as a result of defective regulation of oxidative phosphorylation (33). Thus, persistent changes in the control of mitochondrial bioenergetics may be required if sufficient apoptogenic factors are to be released to lead to cell death as a result of loss of outer membrane integrity.

The data presented demonstrate that mitochondria undergo major physiological changes early after growth factor deprivation. As shown in Figs. 1 and 2, the major mitochondrial flaw observed is a loss of ADP regulation of mitochondrial matrix volume and oxidative phosphorylation. Because, early after IL-3 withdrawal, a decrease in glycolytic products availability to mitochondria was observed together with a fall in mitochondrial membrane potential (18), such a change in ADP response may be adaptive, because it allows mitochondria to preserve their remaining membrane potential for the processes of ion exchange and organelle homeostasis. However, the subsequent metabolism of alternative oxidizable substrates through stimulation of enzymes involved in β-oxidation will restore the mitochondrial supply of electron transport substrates. If under these conditions ADP/ATP exchange is not reestablished, the production of substrates will lead to inner membrane hyperpolarization and increased regulation of oxidative phosphorylation generation by the electron transport chain, events that have been reported in the late stage of growth factor withdrawal (14, 34). In addition, matrix swelling and herniation of the matrix through the mitochondrial outer membrane have been described as late mitochondrial apoptotic events (14, 21, 22). We showed here that matrix volume is dynamically regulated both by the levels of oxidizable substrates and ADP available to the mitochondrial matrix. The development of the inability of mitochondria to import ADP would promote matrix swelling if additional oxidizable substrates subsequently become available as a result of up-regulation of β-oxidation and autophagy.

The ADP responsiveness of mitochondria from growth factor-deprived cells can be restored by permeabililizing the mitochondrial outer membrane, suggesting that the voltage-dependent anion channel, the major ATP/ADP channel in the mitochondrial outer membrane, is being induced to adopt a closed configuration in response to growth factor withdrawal. We also found that the expression of Bcl-x_L maintains the mitochondrial capacity to respond to ADP after IL-3 withdrawal. This effect of Bcl-x_L may be due to the reported ability of Bcl-x_L to maintain the open state of voltage-dependent anion channel (35). While these results suggest that outer membrane permeability contributes to the control of ADP-coupled oxidative phosphorylation, digitonin may also disrupt the formation of contact sites between the inner and outer mitochondrial membranes. These sites may have a role in regulating mitochondrial physiology in general and respiratory control in particular (36).

The fact that mitochondria from Bcl-x_L-expressing cells maintain coupled respiration in response to ADP addition with an RCR that is slightly higher than control cells (Fig. 6B) suggests that Bcl-x_L promotes ADP-coupled respiration even in cells maintained in IL-3. Nevertheless, it is not clear whether the effect of Bcl-x_L on mitochondrial physiology is a direct biochemical effect or whether Bcl-x_L expression changes mitochondrial physiology indirectly. It was described previously that extracts from Bcl-2-expressing cells can protect mitochondria in vitro from physiological changes that follow treatment with cytosolic extracts from apoptotic cells (37). Yet, the ultimate test for a direct biochemical effect of Bcl-2 on mitochondria is the employment of a recombinant Bcl-2 protein during the respiratory analysis. Using recombinant Bcl-x_L protein, we could not reverse the mitochondrial physiological changes that occur after IL-3 withdrawal (data not shown), suggesting that either, once they occur, these changes are irreversible, or that the Bcl-x_L effect on mitochondria in cells is indirect.

Judging from the response of isolated mitochondria to tBid (Fig. 3 C and D; see also refs. 21 and 38), in which cytochrome c became the limiting factor for respiration, the mitochondrial response to IL-3 withdrawal, documented here, is not mediated by tBid. Although there may be other proapoptotic Bcl-2 proteins with different activities, which do not directly induce cytochrome c release, the results presented here demonstrate that a loss of mitochondrial outer membrane exchange of ADP precedes cytochrome c release. If this block in mitochondrial outer membrane adenine nucleotide exchange persists, the changes in the regulation of mitochondrial respiration and volume homeostasis could eventually lead to cell death by necrosis. However, under most circumstances, the activities of proapoptotic Bcl-2 family members intervene, releasing intermembrane apoptogenic factors and initiating the apoptotic elimination of bioenergetically compromised cells.

Abbreviations

EM

electron microscopy

MIB

mitochondrial isolation buffer

tBid

truncated Bid

RCR

respiratory control ratio

TMPD

N,N,N′,N′-tetramethylphenylenediamine