Oncogene (1998) 17, 2515 ± 2524 ã 1998 Stockton Press All rights reserved 0950 ± 9232/98 $12.00 http://www.stockton-press.co.uk/onc Inhibition of mitochondrial respiratory chain complex I by TNF results in cytochrome c release, membrane permeability transition, and apoptosis{

Masahiro Higuchi*,1, Rita J Proske1 and Edward TH Yeh1,2

1Research Center for Cardiovascular Diseases, Institute of Molecular Medicine for the Prevention of Human Diseases and 2Division of Molecular Medicine, The University of Texas-Houston Health Science Center, Houston, Texas 77030, USA

Mitochondria have been shown to play a key role in apoptosis-inducing protein apoptogeneic protease apoptosis induction. However, the sequence of changes (Susin et al., 1996) is released from mitochondria. that occur in the mitochondria in the initial step of Several reagents can open mitochondrial permeability apoptosis has not been clearly elucidated. Here, we transition pores, leading to a release of apoptogeneic showed that mitochondrial respiratory chain (MRC) protease and caspase 3 activation (Susin et al., 1997). complex I was inhibited during the early phase of Inducers of MPT can release cytochrome c from TNF- or serum withdrawal apoptosis. The importance mitochondria (Yang and Cortopassi, 1998). However, of complex I inhibition in apoptosis is also supported by the induction of MPT is not always necessary for the the observation that rotenone, an inhibitor of complex I release of cytochrome c (Yang et al., 1997). Several but not that of other complexes, could induce apoptosis reports have shown that reactive oxygen species (ROS), in a manner comparable to TNF. We hypothesized that phospholipases, and Ca2+ ions can promote the inhibition of complex I could a€ect electron ¯ow through induction of MPT (Skulachev, 1996); however, the other complexes leading to cytochrome c release by an link between the stimulus of MPT inducing reagents antioxidant-sensitive pathway and caspase 3 activation and the induction of MPT is not clear. Overexpression followed by the induction of membrane permeability of bcl-2 abrogated TNF-induced apoptosis, increased transition (MPT). This hypothesis is supported by the mitochondrial membrane potential (Marchetti et al., following observations: (1) TNF and rotenone induced 1996), and inhibited cytochrome c release (Kluck et al., MPT and cytochrome c release; (2) TNF-induced 1997; Yang et al., 1997). This suggests that these events complex I inhibition was observed prior to cytochrome c are closely related during apoptosis. However, the release and MPT induction; (3) MPT induction was relationship between MPT induction and cytochrome c inhibited by a caspase 3 inhibitor, z-DEVD-CH2F, and release in TNF-induced apoptosis has not yet been an antioxidant pyrrolidine dithiocarbamate (PDTC), elucidated. whereas cytochrome c release was only inhibited by Involvement of mitochondria has been implicated in PDTC. Thus, these results suggest that MRC complex I TNF-induced cell death. During the early stage of cell plays a key role in apoptosis signalings. death induced by TNF, mitochondria swell (Rutka et al., 1988) and mitochondrial respiration chains (MRC) Keywords: reactive oxygen species; caspases; apoptosis; are inhibited (Higuchi et al., 1992; Schulze-Ostho€ et necrosis; cytochrome al., 1992). Because mitochondria are the source of ATP, inhibition of MRC will ultimately decrease cellular ATP levels leading to cell death. Additional mechanisms may be involved in TNF-induced cyto- Introduction toxicity. Since mitochondria are the main source of ROS, which have been implicated in both apoptosis Changes in mitochondrial function are thought to be and necrosis (Lennon et al., 1991), mitochondria- involved in apoptois induction. In staurosporine- derived ROS may provide a link between mitochon- induced apoptosis, cytochrome c is released from dria and cell death. This hypothesis is also supported mitochondria which leads to the activation of by the observation that under low oxygen conditions, caspase 3 (Li et al., 1997; Liu et al., 1996). It is very TNF-induced cytotoxicity was reduced (Matthews et likely that cytochrome c release from mitochondria can al., 1987). Furthermore, manganese superoxide dis- induce apoptosis, however several other reports imply mutase (Mn SOD), located in mitochondria, dismu- di€erent roles of cytochrome c in apoptosis signaling tates mitochondrial superoxide anion, which has leaked (Adachi et al., 1997; Tang et al., 1998). In from the electron transport chain, and abrogates TNF- dexamethasone-induced apoptosis, MPT is induced mediated apoptosis (Wong et al., 1989; Wong and (Macho et al., 1995; Zamzami et al., 1995) and the Goeddel, 1988). In contrast, Zn/Cu SOD, which is located in the cytosol, does not inhibit TNF-induced apoptosis. Furthermore, TNF increased hydrogen peroxide generation in the mitochondria in certain *Correspondence: M Higuchi, Institute of Molecular Medicine for systems (Hennet et al., 1993) and inhibited MRC the Prevention of Human Diseases, 2121 W. Holcombe, Houston, through antioxidant-sensitive pathway (Higuchi et al., TX 77030, USA 1992). These results indicate that superoxide anion {This is publication number 135-IMM from the institute of from MRC is involved in TNF-induced cell death. Molecular Medicine for the Prevention of Human Diseases, It is likely that respiratory function may a€ect University of Texas-Houston Health Science Center Received 28 July 1998; revised 2 October 1998; accepted 9 October oxidative stress. In a previous study to investigate the 1998 requirements of mitochondrial respiratory function in Mitochondrial respiration chains and apoptosis MHiguchiet al 2516 TNF-induced apoptosis signaling, we generated a panel Results of respiration de®cient clones and their reconstituted cybrids with normal mitochondria (Higuchi et al., Inhibition of MRC complex I by TNF and serum 1997). We showed that mitochondrial respiratory starvation activity correlated with the ability of TNF, anti-Fas antibody, and serum starvation, but not staurosporine, We have shown previously that mitochondrial respira- to induce apoptosis. Our study placed mitochondrial tory function is involved in the induction of apoptosis respiratory function at a step prior to the activation of by TNF, anti-Fas, and serum starvation (Higuchi et a caspase 3-like protease in TNF-induced apoptosis al., 1997). Inhibition of certain components of MRC signaling. Therefore, it is very likely that TNF a€ects can a€ect electron ¯ow and regulate apoptotic mitochrondrial respiratory function and then activates response. Figure 1 shows the MRC complexes caspase 3 through a respiratory function dependent involved in the mitochondrial electron transfer fashion. However, changes in mitochondrial respira- system. Initially, we investigated the e€ect of TNF on tory function by TNF-induced apoptosis have not been MRC complexes in ML-1a cells. A 90 min treatment of elucidated. ML-1a cells with TNF plus protein synthesis inhibitor Mitochondrial respiratory chain is composed of ®ve cycloheximide induced apoptosis (Higuchi et al., 1995). enzyme complexes: NADH:ubiquinone oxidoreductase Cells were treated with TNF and cycloheximide for the (complex I), succinate:ubiquinone oxidoreductase times indicated in Figure 2 and then tested for their (complex II), ubiquinone:ferricytochrome c oxidore- e€ect on MRC activity. A 10 min treatment of ML-1a ductase (complex III), ferrocytochrome c: oxygen cells with TNF abrogated 25% of the malate/ oxidoreductase (complex IV), and F0F1-ATPase glutamate-dependent oxygen consumption, which (complex V). As shown in Figure 1, complexes I, II, measures electron ¯ow through complexes I, III and III and IV plus coenzyme Q and cytochrome c make IV (see Figure 1), whereas the same treatment showed up MRC and complex V generates ATP by coupling no e€ect on succinate-dependent oxygen consumption, with the proton energy generated at complex I, III and which measures electron ¯ow through complexes II, III IV. Several reports imply that the damage to complex- and IV (Figure 2). A 20 min treatment inhibited 84% es I and III can induce apoptosis (Guidarelli et al., of malate/glutamate-dependent oxygen consumption 1996; Hartley et al., 1994) and that complex V might and 41% of succinate-dependent consumption (Figure be involved in apoptosis (Matsuyama et al., 1998). 2). These results indicate that MRC complex I was In this paper, we investigated the role of MRC inhibited prior to the inhibition of complexes II, III during an early step of apoptotic signaling. We and IV. demonstrated that the inhibition of MRC complex I We also investigated whether the speci®c inhibition occurs during the early stages of mitochondrial changes of MRC complex I could be observed in apoptosis in apoptosis. We also showed that inhibition of MRC complex I released cytochrome c from mitochondria and induced MPT. Furthermore, we investigated the involvement of an antioxidant-sensitive pathway and caspases in cytochrome c release and MPT induction by complex I inhibition leading to apoptosis.

Figure 2 Kinetics of TNF-induced inhibition of the respiration 7 Figure 1 Electron ¯ow through mitochondrial respiratory chain chains. 3610 ML-1a cells were incubated with 1 nM TNF for the complexes. Malate/glutamate, succinate a-glycerophosphate, and indicated times and then the rates of O2 uptake was measured as ascorbate/TMPD are substrates. CoQ: coenzyme Q, cyt. C: described under Materials and methods. The concentrations of cytochrome C substrates were 5 mM malate and 5 mM succinate respectively Mitochondrial respiration chains and apoptosis MHiguchiet al 2517 induced by serum starvation. ML-1a cells were none, an inhibitor of MRC complex I, induced apoptotic 48 h after serum starvation (Higuchi et al., apoptosis. However we were unable to observe 1997). Twenty-four hours after serum starvation, 95% induction of apoptosis by TTFA (inhibitor of MRC of electron ¯ow through MRC complexes I, III and IV complex II), antimycin A (inhibitor of MRC complex was inhibited (Table 1). In comparison, 49% of III), or sodium azide (inhibitor of MRC complex IV) electron ¯ow through MRC complexes II, III and IV, (Figure 3a). The majority of speci®c oxygen consump- and 19% of electron ¯ow through MRC complexes III tion was blocked by 2 mM rotenone, 100 mM TTFA, and IV was inhibited (Table 1). These results indicate 2 mg/ml antimycin A, or 100 mM sodium azide that complex I is the most signi®cantly inhibited MRC respectively. Similar to TNF, apoptosis induced by component during apoptotic conditions induced by rotenone was greatly increased in the presence of serum starvation. protein synthesis inhibitor cycloheximide (Figure 3b). Similar morphological changes associated with TNF- induced apoptosis including nuclear condensation and Induction of apoptosis by the inhibition of MRC blebbing (Figure 4) and formation of membrane complex I blebbing and apoptotic bodies (data not shown) were To investigate whether apoptosis could be induced by also observed by rotenone. To determine if rotenone the inhibition of MRC, the e€ect of MRC inhibitors acted speci®cally on mitochondrial respiratory func- in the presence of cycloheximide was tested. Rote- tion, its e€ect on ML-1a cells and its respiration de®cient and re-constituted clones was examined. Rotenone induced apoptosis in ML-1a cells but not in respiration-de®cient clone 19, furthermore, rote- Table 1 The e€ect of depletion of serum on mitochondrial respiration chain in ML-1a none-induced apoptosis reappeared in the cybrid 7 clones (Figure 5 left panel). In addition, we Respiration (nM/min/10 cells) Substrate Control Serum depletion investigated the e€ect of caspase 3-like protease activation by rotenone in these cells. Rotenone Malate 3.9 0.2 Succinate 15.0 7.6 activated caspase 3-like protease in ML-1a cells and Ascorbate/TMPD 24.0 19.5 cybrid clones, but not in respiration de®cient clone 19 ML-1a cells were incubated for 24 h in the presence or absence of (Figure 5 right panel). These results indicate that FCS, and state 3 oxygen consumption was detected in the presence of speci®c inhibition of MRC complex I could activate a malate, succinate, and ascorbate/TMPD as substrate as described caspase 3-like protease and induce apoptosis through under Materials and methods respiratory function.

Figure 3 Induction of apoptosis by rotenone. (a) 3H-TdR prelabeled ML-1a cells (46104) were incubated with the indicated amounts of inhibitors of MRC complexes and 1 mg/ml cycloheximide for 90 min, and fragmented DNA was determined as 3 4 described under Materials and methods. (b) H-TdR prelabeled ML-1a Cells (4610 ) were incubated with 10 mM rotenone in the presence or absence of 1 mg/ml cycloheximide for 90 min, and fragmented DNA was determined as described under Materials and methods Mitochondrial respiration chains and apoptosis MHiguchiet al 2518 Several reports have shown the involvement of Involvement of electron transfer regulated by complex I mitochondrial ROS in TNF-induced apoptosis. An inhibition in apoptosis increase in electron ¯ow upregulates the generation of Inhibition of MRC complex I may increase electron ROS from MRC (Ksenzenko et al., 1983; Turrens and ¯ow through complexes II, III and IV to compensate Boveris, 1980). To investigate whether oxidative stress for the decrease in ATP synthesis (Figure 1). Such an is involved at a step subsequent to complex I increase might be inhibited by TTFA, an inhibitor of inhibition, the e€ect of antioxidant PDTC on TNF- complex II. TTFA nearly completely inhibited both and rotenone-induced apoptosis was tested. PDTC TNF-induced apoptosis and rotenone-induced apopto- nearly completely inhibited the apoptotic e€ect of TNF sis (Table 2), supporting the idea that a change in and that of rotenone (Table 2), suggesting the in- electron ¯ow through complexes II, III and IV could volvement of an antioxidant-sensitive pathway follow- cause apoptosis. ing complex I inhibition in apoptotic signaling. Recently, several reports suggest the involvement of mitochondrial membrane potential (Shimizu et al., 1998) and F0F1-ATPase (complex V) (Matsuyama et al., 1998) on apoptosis. We tested whether the uncoupler FCCP and oligomycin A, an inhibitor of F0F1-ATPase, could inhibit TNF- and rotenone- induced apoptosis. As shown in Table 2, both of the inhibitors inhibited TNF and rotenone-induced apop- tosis. These results indicate that mitochondrial membrane potential and complex V might be involved in apoptosis signaling subsequent to complex I inhibition.

Induction of MPT and cytochrome c release by the inhibition of complex I Release of cytochrome c from mitochondria and induction of MPT were considered to be the causes of apoptosis. We tested whether TNF and rotenone in the presence of cycloheximide could induce these changes. Figure 6 shows that cytochrome c was released when ML-1a cells were treated with TNF or rotenone in the presence of cycloheximide, whereas TNF, rotenone, and cycloheximide alone had no e€ect. Induction of MPT was estimated by detecting mitochrondrial membrane potential (Zamzami et al., 1995). Induction of MPT was only observed when cells were stimulated by TNF or rotenone in the presence of cycloheximide (Figure 7). TNF, rotenone, or cyclohex- imide alone were ine€ective at least within 90 min treatment. These results indicate that inhibition of MRC complex I could cause the release of cytochrome c and induction of MPT. If cytochrome c release and MPT induction in TNF- induced apoptosis were the result of MRC complex I inhibition, TNF may induce complex I inhibition which would then lead to cytochrome c release and MPT induction. We investigated these time-dependent mitochondrial changes. Figure 8 shows that complex I inhibition was ®rst observed, then cytochrome c was released and MPT was induced. Next, we examined whether a caspase 3-like protease and antioxidant-sensitive pathway were involved in the signal sequences between complex I inhibition and the induction of MPT or cytochrome c release. To investigate the role of caspase 3-like protease and

ROS, a caspase 3 inhibitor, z-DEVD-CH2F, and antioxidant pyrrolidine dithiocarbamate (PDTC) were

used respectively. z-DEVD-CH2F and PDTC nearly Figure 4 Induction of apoptosis by TNF and rotenone. ML-1a completely inhibited TNF- and rotenone-induced cells were incubated with or without 1 nM TNF or 10 mM apoptosis in ML-1a cells (Table 2). ML-1a cells were rotenone in the presence or absence of 1 mg/ml cycloheximide, treated with TNF and rotenone in the presence of stained by propidium iodide and Hoechst 33432 and investigated visually by ¯uorescent microscopy as described under Materials cycloheximide, and the e€ect of z-DEVD-CH2Fand and methods PDTC on the induction of cytochrome c release and Mitochondrial respiration chains and apoptosis MHiguchiet al 2519

Figure 5 Induction of apoptosis and activation of capase 3-like protease by rotenone in ML-1a, respiration de®cient clone 19, and reconstituted clones. 3H-TdR prelabeled (DNA fragmentation) and nonlabeled (capase 3-like protease) ML-1a, clone 19, and reconstituted clones p1, p2, and p3 were incubated with 10 mM rotenone for 90 min and tested for DNA fragmentation and capase 3-like protease activity as described under Materials and methods

Table 2 Inhibition of TNF- and rotenone-induced apoptosis by several inhibitors DNA Fragmentation (%) Inhibitors Control TNF Rotenone

Control 5.8+0.2 85.3+0.9 70.7+0.8 +TTFA 12.6+0.9 18.8+0.2 15.1+0.7 Control 7.6+0.2 84.9+0.8 38.7+0.5 +FCCP 16.0+0.8 34.8+0.5 23.6+0.6 +Oligomycin A 5.2+0.1 37.6+2.0 17.4+0.2 Control 6.3+0.4 78.2+2.0 27.4+0.5 +PDTC 4.7+0.2 6.1+0.2 8.4+0.9 +z-DEVD-CH2F 4.2+0.6 7.8+0.1 5.2+0.3 3 4 H-TdR prelabeled ML-1a cells (4610 ) were incubated with 1 nM TNF or 10 mM rotenone plus 1 mg/ml cyclohexidimide in the absence or presence of TTFA (1 mM), FCCP (2.5 mM), oligomycin A (5 mM), PDTC (50 mM), and z-DEVD-CH2F (50 mM) for 90 min, and fragmented DNA was determined as described under Materials and methods

Figure 6 Cytochrome c release from mitochondria by TNF and Figure 7 Induction of MPT by TNF and rotenone. ML-1a cells rotenone. ML-1a cells were incubated with or without TNF or were incubated for 90 min in the presence or absence of TNF or rotenone in the presence or absence of cycloheximide for 90 min. rotenone with or without cycloheximide and induction of MPT After the cytosolic fraction was obtained, cytochrome c release was detected by FACS analysis by using 3,3'-dehexyloxacarbo- from mitochondria was detected by Western blotting using anti- carnine iodine (DiOC6) as a probe as described under Materials cytochrome c antibody as described under Materials and methods and methods Mitochondrial respiration chains and apoptosis MHiguchiet al 2520 the induction of MPT was investigated. As shown in mitochondrial respiratory chain complex I can be Figure 9, rotenone- and TNF-induced cytochrome c inhibited prior to the other changes tested, such as release was inhibited by PDTC but not by z-DEVD- the inhibition of MRC complexes II, III and IV, MPT

CH2F, suggesting that an antioxidant-sensitive pathway induction, and cytochrome c release. Rotenone, an but not caspase 3-like proteases is involved in the inhibitor of MRC complex I but not the other signal sequences prior to cytochrome c release. Figure complexes, induced apoptosis in ML-1a cells but not 10 shows that MPT induction could be inhibited by in respiration de®cient cells. These results clearly

both z-DEVD-CH2F and PDTC, suggesting that indicate that TNF induces MRC complex I inhibition caspase 3-like protease and an antioxidant-sensitive and that it is the cause of the induction of apoptosis in pathway are involved prior to the induction of MPT. TNF-induced apoptosis signaling. Furthermore, we In summary, TNF-induced MRC complex I inhibi- also showed that complex I inhibition can cause tion induces cytochrome c release through an antiox- cytochrome c release through an antioxidant-sensitive idant-sensitive pathway and MPT through an rather than a caspase-dependent fashion, whereas MPT antioxidant-sensitive and caspase dependent way. induction by complex I inhibition requires both of the antioxidant-sensitive and caspase-dependent signals. We and others reported that MRC is inhibited in the Discussion early stage of necrotic cell death induced by TNF in L.P3 cells, a clone of L929, and in L-929 cells (Higuchi In this report, we investigated the e€ect of TNF on et al., 1992; Schulze-Ostho€ et al., 1992). We also MRC complexes and their role in apoptosis signaling. showed that this step of MRC inhibition is blocked in To study the e€ect of TNF on MRC, we showed that TNF resistant cells, and is needed to induce the late

event of cell death, which includes activation of PLA2. The role of mitochondria in TNF-induced necrosis may be explained by one of the following two mechanisms: (1) Damage to MRC will decrease cellular ATP level, and in turn induce necrosis; (2) Damage to MRC releases Ca2+. The latter mechanism can also induce apoptosis (Richter, 1993). Induction of apoptosis by complex I inhibitors such as rotenone and 1-methyl-4-phenylpyridinium (Hartley et al., 1994) has been reported. In such apoptotic reactions, apoptosis does not always result from a decrease in cellular ATP level (Pastorino et al., 1995), nor from an increase in Ca2+ (Gabai et al., 1992), which are generally considered to be induced by MRC inhibitors. Oxidative energy is generated by F0F1- ATPase via coupled vectorial electron translocation at three stages along the respiratory chain, represented by complexes I, III and IV (Figure 1). When MRC complex I is inhibited, the succinate concentration might increase in the TCA cycle and electron ¯ow through complexes II, III and IV might increase to compensate for the decrease in ATP synthesis through complexes I, III and IV. TTFA, an inhibitor of complex II, strongly inhibited TNF- and rotenone- induced apoptosis, supporting the idea that the change Figure 8 Time dependent induction of complex I inhibition, in the metabolism through complexes II, III and IV is cytochrome c release, and MPT. ML-1a cells were incubated with the cause of apoptosis induction. Anti-oxidant PDTC TNF or rotenone in the presence of 1 mg/ml cycloheximide for indicated times, and complex I inhibition, cytochrome c release, inhibited TNF- and rotenone-induced apoptosis (Table and MPT induction was detected as described under Materials 2), suggesting that an antioxidant-sensitive pathway is and methods involved in apoptosis. The change in electron transfer

Figure 9 The e€ect of z-DEVD-CH2F and PDTC on TNF- and rotenone-induced cytochrome c release. ML-1a cells were incubated for 90 min in the presence or absence of TNF or rotenone plus cycloheximide with and without 25 mM z-DEVD-CH2F or 25 mM PDTC and the release of cytochrome c from mitochondria was detected as described under Materials and methods Mitochondrial respiration chains and apoptosis MHiguchiet al 2521

Figure 10 Inhibition of TNF- and rotenone-induced MPT by z-DEVD-CH2F and PDTC. ML-1a cells were incubated for 90 min in the presence or absence of TNF or rotenone plus cycloheximide with and without 25 mM z-DEVD-CH2F or 25 mM PDTC and induction of MPT was detected by FACS analysis by using 3,3'-dehexyloxacarbocarnine iodine (DiOC6) as a probe as described under Materials and methods in MRC might regulate this antioxidant-sensitive pathway by one of the following ways: (1) Reduced MRC might generate more ROS; (2) Inhibition of oxygen consumption in MRC increases oxygen concentration in mitochondria and in turn increases ROS generation; (3) A change in electron transfer may a€ect oxidation/reduction control in mitochondria and cytosols which may regulate apoptosis. In the last case the generation of ROS is not always necessary. An increase in ROS was also observed in septic shock (Taylor et al., 1995). Recently, the role of F0F1- ATPase in apoptosis has been suggested (Matsuyama et al., 1998). F0F1-ATPase generates ATP by coupling with the electron transport chain, and this coupling is essential to generate ROS from MRC (Skulachev, 1996). Taken together, MRC complex I inhibition may induce apoptosis through an antioxidant-sensitive pathway by coupling with F0F1-ATPase. However, induction of apoptosis by ROS alone required a longer incubation time than that with TNF (data not shown); ROS generation alone may not completely explain the rapid induction of apoptosis by TNF. Complex III inhibitor antimycin A has been known to induce apoptosis. In our system, antimycin A alone could not induce apoptosis up to 90 min but could induce apoptosis 180 min after treatment (Higuchi Figure 11 Mechanistic scheme of TNF-induced apoptosis unpublished results). We also observed that cyclohex- signaling Mitochondrial respiration chains and apoptosis MHiguchiet al 2522 imide had no e€ect on antimycin A-induced apoptosis from mitochondria. Association of cytochrome c to (Higuchi unpublished results). Possibly, di€erent mitochondrial inner membrane is dependent on its mechanisms are involved in each apoptosis induction interaction with acidic lipid cardiolipin and their and further investigation is now underway. association is downregulated by ROS (Soussi et al., In addition to TNF, rotenone induced cytochrome c 1990). It is possible that an antioxidant-sensitive release from mitochondria and MPT, indicating that mechanism induced by complex I inhibition can complex I inhibition could be the cause of these dissociate cytochrome c from mitochondrial inner mitochondrial changes. MPT induction is estimated membrane. Possibly, both of these mechanisms, either

by the decrease in DiOC6 which detects mitochondrial dependently or independently, work to release cyto- membrane potential. Since we did not have a direct chrome c from mitochondria leading to apoptosis. method to detect MPT induction in intact cells, we How complex I is inhibited after TNF treatment has utilized this method. Therefore, although several not been elucidated. Interestingly, Bcl-2 can inhibit reports indicate the induction of MPT by TNF complex I damage by a Ca2+ overload and an increase (Pastorino et al., 1996), we cannot distinguish MPT in Ca2+ uptake capacity when the electron is induction and the decrease in mitochondrial membrane transferred through complexes I, III and IV (Murphy potential, which is recently considered to be important et al., 1996). Therefore, it is very possible that Bcl-2 for apoptosis signalings (Shimizu et al., 1998). Since can regulate the activity of MRC complex I through PDTC can inhibit rotenone-induced apoptosis, cyto- unknown mechanisms, and thus regulate apoptosis. chrome c release, and MPT induction, an antioxidant- Further analysis is necessary to elucidate the exact sensitive pathway might be involved between complex I mechanisms. inhibition and cytochrome c release or MPT induction in TNF signalings. In addition, caspase 3 inhibitor z- DEVD-CH F inhibited rotenone-induced MPT induc- 2 Materials and methods tion but not cytochrome c release. Involvement of a caspase 3-like protease in MPT induction was sup- Materials ported by other investigators (Marzo et al., 1998). These results indicate that caspase 3-like protease is Gentamicin, RPMI-1640 medium, and fetal calf serum involved between complex I and MPT induction and (FCS) were obtained from GIBCO (Grand Island, NY, subsequent to cytochrome c release. However, it is still USA).Ethidiumbromide,uridine,glucose,pyruvate, dimethyl sulfoxide, sucrose, EGTA, adenine diphosphate possible that caspases which cannot be inhibited by z- (ADP) okadaic acid, H-7, rotenone, theonyltri¯uoro- DEVD-CH2F, such as caspase 1-like protease, may acetone (TTFA), antimycin A, sodium azide, oligo- work prior to cytochrome c release. From these results, mycin A, carbonylcyanide p-trifluoromethoxyphenylhydra- we derived the mechanistic scheme in Figure 11. TNF zone (FCCP) malate, glutamate, succinate, ascorbate, treatment inhibits MRC complex I, which then induces N,N,N',N'-tetramethyl-p-phenylenediamine (TMPD) and the release of cytochrome c through a PDTC-sensitive pyrrolidine dithiocarbamate were obtained from Sigma fashion (possibly by ROS). Cytochrome c release may Chemical Co. (St. Louis, MO, USA). activate caspase 3-like protease to induce MPT and apoptosis. We demonstrated that inhibition of MRC Cell lines complex I is an mitochondrial event prior to cytochro- Myelogenous leukemia ML-1a were obtained from Dr Ken me c release and MPT induction and that MRC Takeda(ShowaUniversity,Tokyo,Japan).Allcelllines, complex I inhibition is the cause of the release of except clone 19, were grown in RPMI-1640 medium cytochrome c and MPT induction. The order of supplemented with 10% FCS and gentamicin (50 mg/ml) apoptotic signalings of cytochrome c ± caspase (essential medium). Respiration de®cient clone 19 was activation ± MPT induction was also observed in UV grown in RPMI-1640 medium with 10% FCS and irradiation- or staurosporine-induced apoptosis (Bossy- gentamicin (50 mg/ml) supplemented with 4.5 mg/ml Wetzel et al., 1998). In conclusion, complex I inhibition glucose, 5 mg/ml uridine, and 100 mg/ml pyruvate is considered to be an important turning point of (enriched medium). The cells were seeded at a density of apoptotic signaling. Although an inducer of MPT can 16105 cells/ml in T-25 ¯asks (Falcon 3013, Becton induce apoptosis (Susin et al., 1996), we are not sure Dickinson Labware, Lincoln Park, NJ, USA) containing 10 ml of essential medium and grown at 37 Cinan whether MPT induction is responsible for the signal 8 atmosphere of 95% air and 5% CO2.Cellcultureswere sequences between cytochrome c release and apoptosis, split every 3 to 5 days. since it is still possible that MPT induction down- stream of caspase 3 activation can further activate caspase and increase apoptosis signaling. In addition, Establishment of respiration de®cient and reconstituted clones the possibility of cytochrome c independent mechan- The methods to establish respiration de®cient clone 19 and isms (Tang et al., 1998) after complex I inhibition still reconstituted clones p1, p2, and p3 are shown elsewhere remains. (Higuchi et al., 1997). Complex I inhibition a€ected electron transfer through complex II, III and IV, and may release DNA fragmentation assay cytochrome c possibly through antioxidant-sensitive DNA fragmentation induced by TNF was assayed by the fashion. Cytochrome c is located on the mitochondrial modi®ed method as described (Higuchi et al., 1995). inner membrane. Cytochrome c can be released by the Brie¯y, ML-1a cells were prelabeled with 3H-TdR by disruption of outer membrane (Reed, 1997) or through incubating 26105 cells/ml in essential medium with certain transport. If the latter is the case, the 0.5 mCi/ml 3H-TdR at 378C for 16 h. The cells were dissociation of cytochrome c from mitochondrial inner washed three times and resuspended in RPMI-1640 membrane is also necessary to release cytochrome c mediumandplatedin96-wellplates(46104/well, total Mitochondrial respiration chains and apoptosis MHiguchiet al 2523 volume 200 ml) with or without the test samples. One mg/ml 460 nM using Fluoroscan II (Labsystems, Helsinki, Fin- of cycloheximide was added to increase the sensitivity of land). TNF (Kull and Cuatrecasas, 1981). After incubation for 90 min or the indicated times, they were lysed by the Detection of cytochrome c release from mitochondria by addition of 50 ml of detergent bu€er (10 m Tris-HCl M Western blotting pH 8.0 containing 5 mM EDTA and 2.5% Triton X-100) and incubated an additional 15 min at 48C. High-speed After incubation of 16107 cells under certain conditions, centrifugation was performed in an Eppendorf microcen- ML-1a cells were harvested by centrifugation at 1800 g for trifuge at 12 000 g for 1 min. The radioactivity in the 10 min. After washing with PBS, 20 ml of bu€er (sucrose supernatant represents DNA release into cells due to DNA 250 mM,Tris12.5mM,pH7.0,DTT1mM,EDTA fragmentation. For the total count, cells were lysed by the 0.125 mM,Glycerol5%,PMSF1mM, leupeptin 1 mg/ml, addition of 50 ml of 0.5% sodium dodecyl sulfate. pepstatin 1 mg/ml, and aprotinin 1 mg/ml) was added, and The percent DNA release was calculated as follows: % cells were disrupted by douncing 80 times in a 0.3 ml DNA fragmentation=(c.p.m. in test sample supernatant/total Kontes douncer with the B pestle (Kontes Glass c.p.m.)6100. Each assay was done in triplicate and results Company). Protein gel electrophoresis was carried out are shown as the mean+s.e. using 20% polyacrylamide slab gels in the presence of 0.1% SDS. After electrophoresis, the proteins were transferred electrophoretically to nitrocellulose mem- Respiration measurement brane. After inhibiting the nonspeci®c binding with dried Oxygen consumption was measured with a Clark oxygen milk, the membrane was incubated with anity-puri®ed electrode (model 5300: Yellow Spring Instrument Co., anti-cytochrome c antibody 7H8.22C12 (Pharmingen, San Yellow Spring, OH, USA) as described (Granger and Diego, CA, USA). The membrane was washed three times Lehninger, 1982). Cells were permibilized with 0.0025% in TTBS, and then incubated with rabbit anti-mouse IgG- digitonin in respiration medium (0.25 M sucrose, 20 mM horseradish peroxidase conjugate. The membrane was then

HEPES, pH 7.2, 2 mM KH2PO4 and 1 mM EGTA), washed stainedwithanenhancedchemiluminescencesystem andthensuspendedat36107 cells/ml in respiration (Amersham Inc., UK) according to the supplier's instruc- medium. 0.8 ml of cell suspension was injected into the tions and exposed to X-ray ®lm. For Figure 8 and Table 3, respiration chamber (378C) together with 1 mM ADP and integrated density value of cytochrome c band was detected then a mitochondrial substrate was added; 5 mM for by using a densitometer. For Figure 8, per cent of MPT malate/glutamate, 5 mM succinate or 1.0 mM ascorbate/ induction was calculated as 0% at time 0 min and 100% at 0.2 mM TMPD to initiate State 3 respiration. Respiration time 90 min. was calculated as the rate of change in the O2 concentration following the addition of a substrate, Detection of the induction of MPT assuming an initial O2 concentration as 217 mM/ml. Induction of MPT was evaluated by the decrease in Assay for caspase 3-like protease mitochondrial membrane potential as described (Zamzami et al., 1995). Brie¯y, after stimulation with reagents, ML- Caspase 3-like protease activity was measured with a 1a cells were incubated with DiOC6 for 15 min at 378C, modi®ed method by Enari et al. (1995). Brie¯y, after cells followed by analysis on a cyto¯uorometer (Facscalibar, (16106) were incubated with test samples, cytosolic excitation; 488 nm; emmission: 552 nm). For Figure 8, per extracts were prepared by repeated freezing and thawing cent of MPT induction was calculated as 0% at time 0 min in 300 ml extraction bu€er (Tris 12.5 mM,pH7.0,DTT and 100% at time 90 min. 1mM, EDTA 0.125 mM,Glycerol5%,PMSF1mM,1mg/ ml leupeptin, 1 mg/ml pepstatin, and 1 mg/ml aprotinin). Cell lysate was diluted with the bu€er (in mM) Tris 50, pH 7.0, DTT 1, EDTA 0.5, 20% glycerol, and incubated at 378C in the presence of 20 mM acetyl-Asp-Glu-Val-Asp- Acknowledgements aminomethylcoumarin, a caspase 3 substrate, and acetyl- We are very thankful to Ms Laura S Caskey for her Tyr-Val-Ala-Asp-aminomethylcoumarin, a caspase 1 sub- thoughtful reviews of the manuscript. This research was in strate. Fluorescent aminomethylcoumarin, a product part supported by a grant from American Heart Associa- formed, was measured at excitation 355 nM, emission tion (98BG287) to MH.

References

Adachi S, Cross AR, Babior BM and Gottlieb RA. (1997). J. Higuchi M, Aggarwal BB and Yeh ET. (1997). J. Clin. Biol. Chem., 272, 21878 ± 21882. Invest., 99, 1751 ± 1758. Bossy-Wetzel E, Newmeyer DD and Green DR. (1998). Higuchi M, Shirotani K, Higashi N, Toyoshima S and EMBO J., 17, 37 ± 49. Osawa T. (1992). J. Immunother., 12, 41 ± 49. Enari M, Hug H and Nagata S. (1995). Nature, 375, 78 ± 81. Higuchi M, Singh S and Aggarwal BB. (1995). J. Immunol. Gabai VL, Kabakov AE and Mosin AF. (1992). Tissue & Meth., 178, 173 ± 181. Cell, 24, 171 ± 177. Kluck RM, Bossy-Wetzel E, Green DR and Newmeyer DD. Granger DL and Lehninger AL. (1982). J. Cell. Biol., 95, (1997). Science, 275, 1132 ± 1136. 527 ± 535. Ksenzenko M, Konstantinov AA, Khomutov GB, Tikhonov Guidarelli A, Brambilla L, Rota C, Tomasi A, Cattabeni F AN and Ruuge EK. (1983). FEBS Letts., 155, 19 ± 24. and Cantoni O. (1996). J. Biochem., 317, 371 ± 375. Kull Jr FC and Cuatrecasas P. (1981). Cancer Res., 41, HartleyA,StoneJM,HeronC,CooperJMandSchapira 4885 ± 4890. AH. (1994). J. Neurochem., 63, 1987 ± 1990. Lennon SV, Martin SJ and Cotter TG. (1991). Cell Prolif., Hennet T, Richter C and Peterhans E. (1993). J. Biochem., 24, 203 ± 214. 289, 587 ± 592. Mitochondrial respiration chains and apoptosis MHiguchiet al 2524 Li P, Nijhawan D, Budihardjo I, Srinivasula SM, Ahmad M, Shimizu S, Eguchi Y, Kamiike W, Funahashi Y, Mignon A, Alnemri ES and Wang X. (1997). Cell, 91, 479 ± 489. Lacronique V, Matsuda H and Tsujimoto Y. (1998). Proc. Liu X, Kim CN, Yang J, Jemmerson R and Wang X. (1996). Nat. Acad. Sci. USA, 95, 1455 ± 1459. Cell, 86, 147 ± 157. Skulachev VP. (1996). Quarterly Reviews of Biophysics, 29, Macho A, Castedo M, Marchetti P, Aguilar JJ, Decaudin D, 169 ± 202. Zamzami N, Girard PM, Uriel J and Kroemer G. (1995). Soussi B, Bylund-Fellenius AC, Schersten T and Angstrom J. Blood, 86, 2481 ± 2487. (1990). J. Biochem., 265, 227 ± 232. MarchettiP,HirschT,ZamzamiN,CastedoM,Decaudin Susin SA, Zamzami N, Castedo M, Daugas E, Wang HG, D, Susin SA, Masse B and Kroemer G. (1996). J. Geley S, Fassy F, Reed JC and Kroemer G. (1997). J. Exp. Immunol., 157, 4830 ± 4836. Med., 186, 25 ± 37. Marzo I, Brenner C, Zamzami N, Susin SA, Beutner G, Susin SA, Zamzami N, Castedo M, Hirsch T, Marchetti P, Brdiczka D, Remy R, Xie Z-H, Reed JC and Kromer G. Macho A, Daugas E, Geuskens M and Kroemer G. (1996). (1998). J. Exp. Med., 187, 1261 ± 1271. J. Exp. Med., 184, 1331 ± 1341. Matsuyama S, Xu Q, Velours J and Reed JC. (1998). Mol. Tang D, Li L, Zhu Z and Joshi B. (1998). Biochem. Biophys. Cell., 3, 327 ± 336. Res. Com., 242, 380 ± 384. Matthews N, Neale ML, Jackson SK and Stark JM. (1987). Taylor DE, Ghio AJ and Piantadosi CA. (1995). Arch. Immunol., 62, 153 ± 155. Biochem. Biophys., 316, 70 ± 76. Murphy AN, Bredesen DE, Cortopassi G, Wang E and Turrens JF and Boveris A. (1980). J. Biochem., 191, 421 ± Fiskum G. (1996). Proc.Nat.Acad.Sci.USA,93, 9893 ± 427. 9898. WongGH,ElwellJH,OberleyLWandGoeddelDV.(1989). Pastorino JG, Simbula G, Yamamoto K, Glascott Jr PA, Cell, 58, 923 ± 931. Rothman RJ and Farber JL. (1996). J. Biol. Chem., 271, Wong GH and Goeddel DV. (1988). Science, 242, 941 ± 944. 29792 ± 29798. Yang J, Liu X, Bhalla K, Kim CN, Ibrado AM, Cai J, Peng Pastorino JG, Wilhelm TJ, Glascott Jr PA, Kocsis JJ and TI, Jones DP and Wang X. (1997). Science, 275, 1129 ± Farber JL. (1995). Arch. Biochem. Biophys., 318, 175 ± 181. 1132. Reed JC. (1997). Cell, 91, 559 ± 562. Yang JC and Cortopassi GA. (1998). Free Rad. Biol. Med., Richter C. (1993). FEBS Letts., 325, 104 ± 107. 24, 624 ± 631. Rutka JT, Giblin JR, Berens ME, Bar-Shiva E, Tokuda K, ZamzamiN,MarchettiP,CastedoM,DecaudinD,Macho McCulloch JR, Rosenblum ML, Eessalu TE, Aggarwal A, Hirsch T, Susin SA, Petit PX, Mignotte B and Kroemer BB and Bodell WJ. (1988). Int. J. Cancer, 41, 573 ± 582. G. (1995). J. Exp. Med., 182, 367 ± 377. Schulze-Ostho€ K, Bakker AC, Vanhaesebroeck B, Beyaert R, Jacob WA and Fiers W. (1992). J. Biol. Chem., 267, 5317 ± 5323.