Oxidant, Mitochondria and Calcium

Cell. Signal. Vol. 11, No. 2, pp. 77–85, 1999 Copyright  1998 Elsevier Science Inc.

ISSN 0898-6568/99 $–see front matter PII S0898-6568(98)00025-4

TOPICAL REVIEW

Oxidant, Mitochondria and Calcium: An Overview Tapati Chakraborti, Sudip Das, Malay Mondal, Sujata Roychoudhury and Sajal Chakraborti* Department of Biochemistry and Biophysics, University of Kalyani, Kalyani 742135, West Bengal, India

ABSTRACT. Mitochondria are active in the continuous generation of reactive oxygen species (ROS), (e.g., superoxide), thereby favouring a situation of mitochondrial oxidative stress. Under oxidative stress—for example, ischaemia–reoxygenation injury to cells—mitochondria form superoxide, which in turn is converted to hydrogen peroxide and the potent reactive species, hydroxyl radical. Alternatively, mitochondrial superoxide may react with nitric oxide to form potent oxidant peroxynitrite and as a consequence, mitochondrial function is altered. An increase in the release of calcium from mitochondria by oxidants stimulates calcium-dependent enzymes such as calcium-dependent proteases, nucleases, and phospholipases, which subsequently trigger apoptosis of the cells. In principle, calcium can leave mitochondria by different ways: by non-specific leakage through the inner membrane by “pore formation,” by changes in the membrane lipid phase, by reversal of the uniport influx carrier, by the specific calcium/hydrogen (or sodium) antiport system, by channel-mediated release pathways, or by a combination of two or more of these pathways. Additionally, the release of calcium from mitochondria can also occur either by oxidation of internal nicotinamide adenine nucleotides to ADP ribose and nicotinamide or by oxidation of thiols in membrane proteins. Once calcium efflux has been triggered, a series of common pathways of apoptosis are initiated, each of which may be sufficient to destroy the cell. Apoptosis requires the active participation of cellular components, and several genes have been suggested to control apoptosis. The proto-oncogene bcl-2 suppresses apoptosis through mitochondrial effects. Overexpression of bcl-2 in the mitochondrial membrane inhibits calcium efflux, but the underlying mechanisms are not clearly known. Further studies are needed to explore the nature of the apoptosis-inducing pathways, the precise mechanisms of calcium efflux, the molecular partners of bcl-2 oncoproteins at the level of the outer–inner membrane contact sites, the molecular biology of the apoptosis-inducing factor formation and release, and the essential molecular targets of apoptosis-inducing proteases. Clarification of these issues might facilitate the understanding of mitochondrial response on cellular calcium dynamics under oxidant stress. cell signal 11;2:77–85, 1999.  1998 Elsevier Science Inc. KEY WORDS. Oxidant, Mitochondria, Antioxidant, Calcium, Apoptosis, Oncogenesis, ADP-ribosylation, Pyridine nucleotides, Permeability transition, Pore formation, Na1/H1 exchanger, Na1/Ca21 exchanger

INTRODUCTION Mitochondria are active in the continuous generation of reactive oxygen species (ROS) [e.g., superoxide (O2·2)], thereby favouring a situation of mitochondrial oxidative stress (Fig. 1). Under oxidative stress—for example, ischaemia–reoxygenation injury to cells—mitochondria form O2·2, which in turn is converted into H2O2 and the potent reactive species hydroxyl radical (OH·). Alternatively, mitochondrial O2·2 may react with nitric oxide to form the potent oxidant peroxynitrite (ONOO2), and, as a consequence, mitochondrial function is altered [1]. The relationship between mitochondrial Ca21, oxidative stress and dissipation of mitochondrial membrane potential *Author to whom all correspondence should be addressed. E-mail: sajal@ klyuniv.ernet.in Received 29 December 1997; and accepted 9 February 1998.

(DC) was investigated in a variety of cells. In proximal tubular kidney cells, for example, the nephrotoxic agent, 1,2-dichlorovinyl-l-cysteine (DCVC) stimulates the formation of hydroperoxides, which has been found to be prevented by the antioxidant diphenylphenelene diamine and the iron chelator deferoxamine. DCVC also induces an increase in the intracellular free Ca21 accompanied by an alteration in the mitochondrial Ca21 dynamics. The alteration in the mitochondrial Ca21 concentration has been suggested to be an important event in the induction of oxidative stress [2]. Maintenance of low cytosolic Ca21 is necessary for proper functioning of cells. Mitochondria transport Ca21 (1) to regulate cytosolic Ca21, (2) to serve as a store of Ca21 when its concentration in the cytosol is excessive, (3) to serve as a releasable source of activator Ca21, and (4) to regulate mitochondrial matrix Ca21 and thereby control the level of activation of Ca21-sensitive metabolic enzymes. Thus, mito-

78

T. Chakraborti et al.

FIGURE 1. Mitochondrial free radical generation. The major sources of mitochondrial free radicals is the electron-transport chain lo-

cated on the inner mitochondrial membrane. Mitochondrial sources of O2·2 have been studied by using various electron-transport inhibitors (rotenone, antimycin A, potassium cyanide [KCN] and azide, dashed arrows) and substrates (NADH-linked substrates and succinate, solid arrows). NADH dehydrogenase and ubiquinone–cytochrome b region have been shown to reduce oxygen to O2·2, which, in turn, serves as a precursor for H2O2 and OH· (from Freeman & Crapo [77]).

chondria play an important role in controlling cellular Ca21 dynamics [3]. Relatively prominent features of irreversible cell damage caused by oxidative stress are morphological and functional changes in the mitochondria [4–6]. In this context, several studies have investigated the oxidative inactivation of isolated mitochondria after exposure to various toxins and enzymically produced free radicals [4–6]. These studies demonstrated that both mitochondrial matrix enzymes and membrane-bound enzymes are affected by oxidative stress. An increase in cytosolic Ca21 by oxidants may lead to apoptosis due to stimulation of intracellular protease(s), nucleases, phospholipases and other hydrolytic enzyme activities. Oxidant-induced Ca21 release from mitochondria followed by excessive Ca21 cycling and ATP depletion are suggested to be the basic events in apoptosis [7, 8]. This article briefly summarises the different mechanisms of the oxidant-mediated Ca21-release phenomenon and its consequences in apoptosis and oncogenesis.

In principle, Ca21 can leave mitochondria by different ways: by non-specific leakage through the inner membrane by “pore formation” [10], by changes in the membrane lipid phase [11], by reversal of the uniport influx carrier [12, 13], by the specific Ca21/H1 (or Na1) antiport system [11, 14], by channel-mediated release pathways [15, 16] or by a combination of two or more of these pathways. The mitochondrial Ca21 cycle, as proposed by Gunter and Gunter [17], is schematically represented in Figure 2. Additionally, the release of Ca21 from mitochondria can also occur either by oxidation of internal nicotinamide adenine nucleotides to ADP ribose and nicotinamide [7, 8] or by oxidation of thiols in membrane proteins that are associated with the regulation of Ca21 dynamics in the mitochondria [18] (Fig. 3). The disturbance of Ca21 homeostasis resulting in Ca21 overload that occurs during ischaemia and reperfusion was

OXIDANT AND CALCIUM ION TRANSPORT PROCESSES It is now well established that Ca21 uptake and release take place by different mechanisms. The addition of Ca21 to a suspension of energised mitochondria results in the stimulation of respiration, the extrusion of protons and the uptake of Ca21 into the mitochondria. The uptake is purely an electrogenic process driven by the electrical component (membrane potential) of the total proton motive force [9]. Ca21 is transported with two positive charges through a uniport. Mitochondria, therefore, cannot take up Ca21 in unlimited amounts, and there must be Ca21 release at a rate comparable to that of the uptake. The electrophoretic Ca21 uptake operates essentially as a one-way system, and Ca21 efflux must be mediated by a separate system(s) that operates independently of the membrane potential. Furthermore, different uptake and release pathways should result in “cycling” of Ca21 across the inner mitochondrial membrane.

FIGURE 2. Ca21 transport systems in mitochondria. The mito-

chondrial Ca21 uniporter (U) facilitates the transport of Ca21 in an inward direction down the electrochemical gradient of this ion. The Na1 independent efflux mechanism (I) is depicted here as an active Ca21–2H1 exchanger, receiving energy from the electron-transport chain. The Na1-dependent efflux mechanism (D) is depicted here as a Ca21–2 Na1 exchanger. A Ca21 activated permeability transition pore (PTP) also is shown (from Gunter, K.K. & Gunter, T.E. [17]).

Oxidant, Mitochondria and Calcium

79

FIGURE 3. Schematic represen-

tation of oxidant-mediated Ca21 release from mitochondria with the involvement of oxidation of pyridine nucleotides, hydrolysis of oxidized pyridine nucleotides and subsequent ADP riboylation of intramitochondrial Ca21 regulatory proteins. Release of Ca21 enhances Ca21 cycling, which causes collapse of membrane potential (DC) and concomitantly aggravates further Ca21 release to the cytosol leading to cell death (apoptosis).

suggested to be a major pathological event [19]. The structural and functional integrity of isolated mitochondria is important because of their role in providing energy for contraction–relaxation cycles in cardiac cells. In cardiac injury by ischaemic insult, mitochondrial damage is observed early in the sequence of pathological events, as evidenced by an increase in mitochondrial swelling and a decrease in respiratory rates [20]. Several classes of ion-channel activities are associated with the outer and inner membranes of the mitochondrion. In guinea pig working hearts subjected to global ischaemia, pre-treatment with Ca21-channel antagonists of different subclasses markedly improves the recovery of myocardial function during reperfusion [19]. Nifedepine reduces the left ventricular stiffness, improves the left ventricular compliance and decreases the Ca21 content of left ventricular myocardial mitochondria [21]. In agreement with these observations, other investigators also demonstrated that the administration of Ca21-channel blockers such as nifedepine, verapamil and diltiazem prior to a period of ischaemia decreases the deleterious effects evoked by myocardial ischaemia and reperfusion in isolated hearts of rat and rabbit [19, 22].

OXIDANT AND NON-SPECIFIC CHANGE IN MITOCHONDRIAL MEMBRANE PERMEABILITY A megachannel in the inner mitochondrial membrane, known as the “permeability transition pore,” may be opened by high concentrations of inorganic phosphate due to the hydrolysis of creatine phosphate under oxidative stress—for example, reoxygenation injury to vascular cells [15, 23].

The spontaneous discharge of Ca21 from mitochondria is associated with the following sequence of events: (1) increased non-specific permeability of the mitochondrial inner membrane [24]; (2) swelling of the mitochondria [25]; (3) loss of K1 from the matrix [26, 27]; (4) loss of matrix adenine nucleotides [26, 27]; (5) oxidation [28], hydrolysis [28] or leakage of matrix nicotinamide adenine nucleotides [26] or both; (6) stimulation of the inner membrane phospholipase A2 (PLA2) activity and accumulation of unsaturated fatty acids [16]; and (7) collapse of DC [17, 23]. Haworth and Hunter [29] suggested that solutes cross through a proteinaceous pore and that the “pore” may be opened or closed by activators or inhibitors of the “permeability transition pore” in an allosteric fashion. Analogies between the permeability transition pore and channels in mitochondrial membranes are apparent from the available reports. The NMDA (N-methyl-d-aspartate) receptor channel may be taken as an example. This channel is also voltage dependent and may be modulated by the redox state of sulphhydryls that are thought to affect the voltage sensor [30]. The channel transports several cations non-selectively. The NMDA receptor channels belong to the super family of ligand-gated ion channels, and the permeability transition pore also may belong to this family [30]. The analogies that have been identified suggest the possibility of a voltagesensing mechanism modulated by the oxidation–reduction state of the dithiol(s) [31]. Interestingly, the effect of oxidizing agents was found to be dependent on DC, indicating that voltage affects the chemical reactivity of the receptor at this site [31]. Crompton’s group and others [32, 33] proposed that oxidants (e.g., t-buOOH) induce a reversible increase in the

80

permeability of the inner mitochondrial membrane (“pore” formation) that is non-specific and characterised by leakiness to small ions and proteins and swelling of the mitochondria, resulting in a loss of DC. Oxidative stress is believed to cause “pore” opening, leading to the release of Ca21 from mitochondria of different origin. However, the free radical inducing neurotoxin 6-hydroxydopamine (6HD) induces Ca21 efflux from mitochondria, and such efflux was found to be prevented by cyclosporin A. The 6HD-mediated Ca21 efflux was not found to be accompanied by mitochondrial swelling, depolarisation of the mitochondrial inner membrane or movement of radiolabelled sucrose into the mitochondrial matrix [34]. These findings provided evidence indicating that the opening of mitochondrial permeability transition pores may not be the only pro-oxidant mechanism—for example, 6HD-induced Ca21 efflux [34]. Recently, Vercesi and colleagues [35] demonstrated that, at low Pi concentrations (about 1 mM), mitochondrial swelling is reversible and prevented by cyclosporin A but not by butyl hydroxytoluene (BHT), an inhibitor of the mitochondrial permeability transition pore. At high Pi concentrations, however, mitochondrial swelling is only partly prevented by BHT. In both cases (i.e., at low or high phosphate concentrations), exogenous catalase prevented mitochondrial swelling, which suggests that ROS participate in these mechanisms. At low Pi, therefore, membrane permeabilisation is reversible and mediated by opening of the mitochondrial permeability transition pore, whereas, at high Pi concentrations, membrane permeabilisation is irreversible because of the occurrence of lipid peroxidation [35]. In rat liver mitochondria, Ca21-mediated induction of permeability transition can be affected by the lipid peroxidation by-products—for example, 4-hydroxyhexenal [36]. Pfeiffer and coworkers [3] suggested that the defects in the membrane lipid phase could arise owing to the accumulation of PLA2 reaction products—for example, free fatty acids. In this view, a substrate cycle of phospholipid de-acylation–re-acylation appears to establish the steady state of free fatty acids and lysophospholipids, which, in turn, alters the permeability of the mitochondrial inner membrane. This provides a possible mechanistic explanation for the effects of lipid peroxidation on the induction of mitochondrial membrane permeability transition [36]. The mitochondrial membrane permeability transition is proposed as a mechanism of cell necrosis. It has also been suggested that an enhanced activity of mitochondrial protease(s) causes cell necrosis through the induction of the mitochondrial membrane permeability transition [37]. Chakraborti et al. [38] demonstrated that oxidants such as H2O2 and t-buOOH stimulate PLA2 activity in pulmonary cells through the involvement of a protease. Therefore, protease(s) may play a pivotal role in the induction of the mitochondrial membrane permeability transition pathway(s) in mitochondria under oxidant-triggered conditions. The membrane-phase-pathway-associated theory is supported by the appearance of PLA2 reaction products during the transition and by the ability of the commonly used

T. Chakraborti et al.

PLA2 inhibitors (e.g., mepacrine) to inhibit the transition [3]. The potency of cyclosporin A as an inhibitor of the transition and the lack of its effect on PLA2 activity apparently favour the “pore” concept [10]. However, the inhibition produced by cyclosporin A is transient and persists when used with an inhibitor of PLA2. Conversely, the protection afforded by PLA2 inhibition also is transient unless cyclosporin A is present [3]. These results support the findings that oxidants can damage mitochondria by peroxidation-dependent and -independent mechanisms. When peroxidation is inhibited by the appropriate agents, cyclosporin A protects hepatocytes against oxidants—for example, menadione toxicity [39]. Furthermore, it appears that mitochondrial de-energisation per se correlates closely with cell death induced by several agents that affect the membrane permeability transition [39]. The uncoupling effects of free radicals observed at the three coupling sites of oxidative phosphorylation do not cause collapse of the proton electrochemical gradient and do not destroy the membrane permeability barrier. This view is supported by several lines of evidence. First, the addition of NAD1 completely prevents the increase in DNP-uncoupled respiration of pyruvate and b-OH butyrate. Second, the activities of NADH oxidase and NADH-cytochrome c reductase are unaffected in sonicated mitochondria exposed to free radicals. Thirdly, reverse electron flow and ATPase activity assessed by the energydependent reduction of NAD1 by succinate in sub-mitochondrial particles are unaffected. Fourthly, the accumulation of NADH that could have been seen after complex I inhibition (i.e., rotenone or amytal treatment) is not observed [3, 39]. The development of mitochondrial swelling, Ca21 efflux, sucrose entrapment and an alteration of the mitochondrial inner membrane phase transition by an increase in the PLA2 activity under oxidant stress may cause mitochondrial membrane permeability alteration leading to the efflux of NAD1 and NADH from the mitochondria. In addition to such leakage, the hydrolysis of NAD1 to nicotinamide and ADP ribose by oxidative stress may be involved in the loss of intramitochondrial nicotinamide nucleotides. The loss of nicotinamide nucleotides leads to selective alterations of mitochondrial respiration of site I substrates [3, 28, 39]. Together, these studies suggest that both transmembrane pathways (i.e., the permeability transition pore and the mitochondrial membrane permeability transition) are interactive and interdependent. PYRIDINE NUCLEOTIDES AND OXIDANT-INDUCED CALCIUM ION RELEASE The oxidation–reduction state of the co-enzyme NAD(H) also is clearly a major factor that regulates the transition. With this parameter, multiple sites of action are likely to be involved. Using colloid osmotic shrinkage to assay solute flux through pores, Haworth and Hunter [29] suggested that NADH acts synergistically with ADP to inhibit the pores whereas either compound alone gives a mixed type of inhibition, which indicates that ADP and NADH may act at

Oxidant, Mitochondria and Calcium

the same site. NADPH further aggravates the inhibition caused by NADH [40]. This scheme does not explain why induction of the transition correlates better with the oxidation of NADPH rather than with that of NADH [40]. Oxidation of NADPH may favour the transition in the lipid phase pathway secondary to GSSG accumulation and the attendant effects on sulphhydryl groups required for lysophospholipid reacylation [3]. It has therefore been suggested that, unlike ADP and Pi, the oxidation state of the mitochondrial pyridine nucleotides may be the controlling factor for both pore- and lipid-phase-dependent mechanisms [3]. Recently, monobromobimane, a thiol reagent and a selective blocker of dithiol(s), was used in an attempt to examine whether its oxidation–reduction status modifies voltage sensing by the mitochondrial permeability transition pore, a cyclosporin A-sensitive channel. Monobromobimane did not inhibit the phosphate carrier, Ca21 transport, energy coupling or ATP production and transport. Monobromobimane selectively prevents the shift in the pore-gating potential caused by some dithiol oxidants or cross-linkers but not by increasing Ca21, allowing a clear distinction of the pore agonists that act at this site [41, 42]. The molecules harbouring the sulphhydryl groups is currently unknown. Attractive candidates are cyclophilin or the NAD1-hydrolysing enzyme itself [43]. At low concentrations, the thiol oxidant peroxynitrite [ONOO2; formed from the reaction between nitric oxide (NO) and superoxide (O·2)] stimulates Ca21 release. The pyridine nucleotide hydrolysis and the formation of peroxynitrite and Ca21 release were inhibited by cyclosporin A. Peroxynitrite appears to be ineffective when pyridine nucleotides are kept at their reduced states. Additionally, Ca21 release induced by peroxynitrite occurs with the maintenance of the mitochondrial membrane potential and is not accompanied by the entry of sucrose into mitochondria. These studies suggest that peroxynitrite stimulates Ca21 release from intact mitochondria by modifying critical mitochondrial thiols other than glutathione (conceivably protein thiols) in such a way that hydrolysis of oxidised pyridine nucleotides is achieved [44]. In mitochondria not loaded with Ca21, the oxidant t-buOOH-induced pyridine nucleotide oxidation is reversible owing to the enzymic rereduction with succinate after hydroperoxide consumption [45]. In Ca21-loaded mitochondria, the oxidant causes pyridine nucleotide oxidation and enzymic hydrolysis with the formation of ADP ribose and nicotinamide along with Ca21 release [45, 46] (Fig. 3). Oxidation alone of pyridine nucleotides is not sufficient to induce Ca21 release, which was shown with ATP [47]. In its presence, a decrease in pyridine nucleotide hydrolysis and Ca21 release was observed. Similar observations were seen during menadione- and alloxan-induced Ca21 release [8]. Importantly, neither a decrease nor an increase in oxidized mitochondrial glutathione favour Ca21 release from mitochondria under the oxidant (e.g., tert-butyl hydroperoxide) triggered conditions [28, 48]. The functional link between pyridine nucleotide hydrolysis and Ca21 release in

81

mitochondria under oxidant-triggered conditions was suggested to be due to the protein ADP ribosylation in the inner mitochondrial membrane [28] and is schematically represented in Figure 3. Pyridine nucleotides are important co-factors in mitochondria and are presumably synthesised in the sub-organelle, because the intact inner mitochondrial membrane is not permeable to these polar molecules. However, no report has been available about the origin of mitochondrial pyridine nucleotides, and little is known about their degradation. Thus, pyridine nucleotide metabolism in mitochondria under oxidant-triggered conditions deserves further attention. OXIDANT, PROTEIN ADP RIBOSYLATION AND CALCIUM ION RELEASE The intramitochondrial pyridine nucleotide hydrolysis and the release of nicotinamide from mitochondria exposed to oxidants suggest the existence of NAD glycohydrolase activity in mitochondria. The enzyme is localised in the inner side of the inner mitochondrial membrane [49] and is inhibited by ATP [50]. Hydrogen peroxide is enzymically reduced in mitochondria to water at the expense of NADH and NADPH owing to the combined action of glutathione peroxidase, glutathione reductase and the energy-linked transhydrogenase. When Ca21 is present in mitochondria, the NAD1 formed by this or other reactions is hydrolysed to ADPribose and nicotinamide. NAD1 hydrolysis allows protein mono-ADP ribosylation and the release of Ca21 from mitochondria [45, 51]. This Ca21 release is specific and does not require the socalled mitochondrial permeability transition, as judged from the following evidence. When re-uptake of Ca21 by mitochondria is prevented, release occurs with maintenance of the mitochondrial membrane potential (DC). The release of Ca21 was neither associated with that of K1 efflux nor paralleled with sucrose entry into mitochondria, and mitochondria did not swell [3, 45]. Incubation of rat sub-mitochondrial particles with NAD1 labelled at the adenine part leads to a time-dependent incorporation of radioactivity into proteins. With nicotinamide-labelled NAD1, no incorporation is observed. Analysis of labelled SMP on SDS-PAGE revealed that the radioactivity is almost exclusively in the region of protein with Mr of about 32,000. Treatment of the labelled protein with snake venom phosphodiesterase liberates mainly 59-AMP, indicating mono-ADP ribosylation of the protein [28, 45, 51]. When Ca21-loaded mitochondria were challenged with the oxidant t-BuOOH, a moderate increase in protein-bound mono-ADP ribose was observed during Ca21 release. Its level remained enhanced after the oxidant addition until the release of Ca21 was complete. This finding is in keeping with the participation of ADP ribosylation in oxidant-mediated Ca21 release from mitochondria [28, 45]. Cyclosporin A inhibits oxidant-induced Ca21 release from rat liver mitochondria by preventing the initial reaction during protein-ADP ribosylation (i.e., pyridine nucleotide

82

FIGURE 4. Summary of how intramitochondrial Ca21 may affect

the overall process of oxidative phosphorylation in a coordinated manner in mammalian tissues. Abbreviations: CaBI–Ca21-binding inhibitory subunit; PDH–pyruvate dehydrogenase; PDHPPase–PDH phosphate phosphatase; NAD-ICDH–NAD1-linked isocitrate dehydrogenase; OGDH–2-oxoglutarate dehydrogenase complex (from McCormack, J.G. & Denton, R.M. [55]).

hydrolysis). The action of cyclosporin A may therefore be comparable to that of ATP, which also inhibits pyridine nucleotide hydrolysis, but different from that of other known Ca21-release inhibitors—for example, meta-iodobenzylguanidine, which inhibits presumably by competing with the natural acceptor proteins for ADP ribose [52]. OXIDANT-INDUCED CALCIUM ION RELEASE FROM MITOCHONDRIA: CONSEQUENCES IN THE CELL Physiologically, Ca21 is important for mitochondrial functioning because it affects energy metabolism at the level of ATP synthesis and substrate oxidation and it regulates nucleic acid synthesis [45, 53]. Hansford [54] and McCormack and Denton [55] found that the three Ca21-stimulated mitochondrial dehydrogenases are coupled to the electron-transport system. These dehydrogenases are pyruvate dehydrogenase, a-ketoglutarate dehydrogenase and isocitrate dehydrogenase. Activation of these dehydrogenases results in stimulation of both electron transport and ADP phosphorylation. Figure 4 summarises the role that intramitochondrial Ca21 plays to influence the overall process of oxidative phosphorylation in mammalian tissues. Reaction of cytochrome oxidase with H2O2 produces a peroxy intermediate, and its excessive generation inhibits the enzyme activity, which in turn lowers the ATP level under oxidant stress [56, 57]. Intramitochondrial free Ca21 has thus been considered a mechanism that controls the rate of ATP output, independently of substrate stimulation or product inhibition. These dehydrogenases have been shown to be activated by Ca21. In the heart, mitochondria are able to control matrix Ca21 so that it is roughly the same as an average of cytosolic Ca21 under physiological conditions. A rapid and frequent Ca21 pulse would engender a higher average level of mitochondrial matrix Ca21, a greater activation of dehydrogenases and faster ATP production. In contrast, a decrease in matrix Ca21 due to its release from mitochondria causes inhibition of the dehydrogenases and reduces ATP production [55]. Recent research demonstrated that an a2-macroglobulin-sensitive protease plays a major role in regulating mitochondrial free

T. Chakraborti et al.

Ca21 by acting on the mitochondrial Ca21-binding proteins (e.g., calmitine) and subsequently controls various enzyme activities in the mitochondrial matrix that take part in ATP synthesis [58]. Thus, it is obvious that the dehydrogenases should be affected by oxidant-induced release of Ca21 from mitochondria. In Ca21-cycling mitochondria, the matrix dehydrogenases, the citric acid cycle enzymes and the respiratory-chain components are seriously compromised. Another deleterious consequence of excessive Ca21 cycling is the lowering of membrane potential, the inability of mitochondria to synthesise ATP and uncoupling of respiration. Additionally, the oxidant-induced hydrolytic product of NAD (i.e., the cyclic ADP ribose), produced by the mitochondrial enzyme NAD1 glycohydrolase, stimulates Ca21 release from the ryonidine-sensitive Ca21 release pool(s) of the endoplasmic reticulum [59], which further contributes to the Ca21 overload in the cell. Orrenius and colleagues [60] demonstrated that the release of mitochondrial Ca21 in hepatocytes by the oxidant t-buOOH results in blebbing of the plasma membrane and eventually causes cell death. Blebbing is probably indicative of the inability of hepatocyte cytoskeleton to maintain normal surface morphology owing to an increase in extramitochondrial Ca21. Oxidants impair the ability of mitochondria to retain Ca21. The observed killing of the cell may therefore be due to the loss of the ability of mitochondria to act as a safety device under these conditions. The increase in cytosolic Ca21 may play a role in initiating the final common pathway(s) for the toxic cell death. The effects of oxidants, therefore, appear to be specifically due to activation of the mitochondrial Ca21 release pathway(s). OXIDANTS AND CALCIUM IONS: APOPTOSIS AND ONCOGENESIS Many agents that induce apoptosis are either oxidants or stimulators of cellular oxidative metabolism. Conversely, many inhibitors of apoptosis have anti-oxidant activities or enhance cellular anti-oxidant defences. Mammalian cells exist in a state of oxidative siege in which survival requires an appropriate balance of oxidants and anti-oxidants. Eukaryotic cells may benefit from this perilous existence by involving oxidative stress as a common mediator of apoptosis [7]. Mitochondria take up and buffer cytosolic Ca21 when its concentration increases to levels that allow the operation of the mitochondrial low-affinity uptake system. Because the influx of Ca21 across a damaged plasma membrane is a frequent and early final common pathway in which cells are killed, mitochondria act as a safety device against a toxic increase in cytosolic Ca21. When this ability to retain Ca21 is compromised or lost, the injured cells will die [61]. Apoptosis (programmed cell death, or physiological cell death) is a general property of most cells [28]. Apoptosis was in the past mainly seen as a means to remove cells. However, it is now realised that inhibition of apoptosis may contribute to oncogenesis, because tissue homeostasis is a fragile balance between cell proliferation and death [2]. Cells go into apoptosis in response to a variety of primary triggers

Oxidant, Mitochondria and Calcium

83

FIGURE 5. Hypothetical model of oxidant-mediated regulation of apoptosis. Apoptosis can be induced by sub-necrotic cell damage or

by receptor-mediated signals through the mediation of ROS or RNS or both that conceivably enter mitochondria through ANC [72, 73]. The particular apoptosis second-messenger system depends on the stimulus and thus constitutes apoptosis-inducing pathways involving bcl-2 and bax genes. These genes may act as a cellular autonomous rheostat regulating cell death. After ROS signal for programmed cell death (PCD), cells die if BAX is in excess but live if Bcl-2 predominates. Excess BAX product stimulates Ca21 release from mitochondria through yet unidentified pathways and eventually stimulates protease(s) activity thereby altering the protease–antiprotease balance in favour of proteases. As a consequence, a further production of ROS, cytoplasmic changes and nuclear apoptosis could occur. Abbreviations: ROS–reactive oxygen species; RNS–reactive nitrogen species, ANC–anion channel; AIF–apoptosis-inducing factor; PCD–programmed cell death.

such as the appearance or disappearance of hormonal signals, changes in cytokine and growth-factor levels or a marked increase in cytosolic Ca21 levels. Ca21 release from mitochondria by oxidants can trigger apoptosis (Fig. 3). Oxidative stress, excessive generation of nitric oxide and its co-geners, Ca21, proteases, nucleases and mitochondria are considered mediators of apoptosis [7, 28, 62–64]. At present, their importance and exact roles are elusive, but it is clear that mitochondria are both the target and the source of oxidative stress. The mitochondrial membrane potential (DC), which is the driving force of mitochondrial ATP synthesis, declines during apoptosis, and restoration of DC prevents apoptosis. Because apoptosis is highly regulated and likely to have a high energy demand, cellular ATP is an important determinant of cell death [7, 28, 63, 65, 66] (Fig. 3). Apoptosis requires the active participation of cellular components, and several cellular genes, such as bcl-2, c-myc, p53, TRPM-2, RP-2, RP-8, APO-IIFAS, ced-3, ced-4, and ced-9, have been suggested to control apoptosis (suppress or

stimulate) [67–69]. The proto-oncogene bcl-2 suppresses apoptosis through mitochondrial effects. In leukaemia cells, bcl-2 hyper-expression correlates with enhanced ATP:ADP ratios [67–69]. Over-expression of bcl-2 in the mitochondrial membrane inhibits Ca21 efflux, but the underlying mechanism(s) is not clearly known. Bcl-2 prevents Ca21 efflux from the mitochondrial matrix under oxidant stress apparently by inhibiting the apoptosis-inducing factor in the mitochondrial intermembrane space [69, 70]. Recent research demonstrated that the bcl-2 homologue BAX possesses a structure resembling pore-forming bacterial proteins, translocates from the outer to the inner membrane and disrupts the ion gradient build-up on the inner membrane. Bcl-2 has been suggested to regulate apoptosis by modulating the permeability transition pore [70–71]. However, the mechanism(s) of action of bcl-2 products on other Ca21-efflux mechanisms that are largely based on PLA2 activity, membrane fludity, NAD1-glycohydrolase activity, protein mono-ADP ribosylation and protein thiol–disul-

84

phide exchange must be examined further before a definite conclusion can be made. A proposed schema of the overall cellular autonomous rheostat of programmed cell death circuited by mitochondrial bcl-2 and BAX involving ROS and Ca21 is presented in Figure 5. Given that bcl-2 elicits an anti-oxidative response in cells, one mechanism by which bcl-2 prevents apoptosis is the prevention of ROS-induced mitochondrial Ca21 cycling, a process that results in a collapse of DC and in cellular ATP depletion. Thus, bcl-2 prevents disturbances of cellular Ca21 homeostasis and ROS production at the mitochondrial level. On the basis of these and other findings, it was suggested that oxidant-induced Ca21 release from mitochondria followed by Ca21 cycling and ATP depletion is a common cause of apoptosis. Accordingly, the maintenance of DC stabilises mitochondria and thereby prevents apoptosis. bcl-2 thus provides the link between the anti-oxidant defence system, Ca21 and DC [74, 75]. Another mechanism by which bcl-2 prevents apoptosis could relate to a shift in the NADH/NAD1 ratio in favour of the reduced form. Bcl-2 prevents Ca21 release from the endoplasmic reticulum through the inositol trisphosphate-insensitive pathway [76] known to be stimulated by cyclic ADP ribose, derived from the hydrolysis of NAD1 by the mitochonderial enzyme NAD1 glycohydrolase. Conceivably, bcl-2 prevents this release by shifting NAD1 to NADH, thereby preventing the formation of cyclic ADP ribose [76]. CONCLUSION AND FUTURE PROSPECTS Reactive oxygen species not only act as physiological modulators of some mitochondrial function, but may also damage mitochondria when generated in excessive amounts. H2O2, which originates in mitochondria predominantly from the dismutation of superoxide, causes oxidation of mitochondrial pyridine nucleotides and thereby stimulates Ca21 release from the mitochondria. Stimulation of mitochondrial ROS production followed by an enhanced Ca21 release and re-uptake (Ca21 cycling) by mitochondria causes apoptosis and necrosis and contributes to cellular injury. Oxidative stress causes the sustained elevation of cellular Ca21 levels with subsequent disruption of the cytoskeleton and activation of Ca21-dependent catabolic enzymes, including phospholipases, kinases, proteases and endonucleases. The adverse effects of free radicals are countered in part by enzymic (e.g., glutathione peroxidase and catalase) and by non-enzymic means, such as tocopherols and ascorbic acid. In general, tissue levels of these defences vary from one tissue to another between species and with age. Changes in these defences are likely to be due to an increase in the production of the superoxide radical and hydrogen peroxide and a decrease in the formation of ATP by the mitochondria. The available data suggest that mitochondrial Ca21 efflux pathway(s) is a central coordinating event of the apoptotic effector phase (Fig. 5). The hypothesis predicts that various pathways of apoptosis converge at the level of Ca21 efflux.

T. Chakraborti et al.

When Ca21 efflux has been triggered, a series of common pathways of apoptosis are initiated, each of which may be sufficient to destroy the cell. Further studies will explore the nature of the apoptosis-inducing pathway(s), the precise mechanism(s) of Ca21 efflux, the molecular partner of bcl-2 oncoproteins at the level of the outer–inner membrane contact sites, the molecular biology of the apoptosis-inducing factor formation and release and the essential molecular targets of apoptosis-triggering protease(s). Clarification of these issues might facilitate an understanding of the mitochondrial response under oxidant stress and associated molecular mechanisms. Financial assistance from the Indian Council of Medical Research (ICMR) and the Council of Scientific and Industrial Research (CSIR), New Delhi, are acknowledged. Thanks are due to Professor Kasturi Datta (School of Environmental Sciences, Jawaharlal Nehru University, New Delhi), Dr. Vasantha Muthuswami (Deputy Director General, ICMR, New Delhi) and Dr. S. Mullick (HRDG, CSIR, New Delhi) for their help. This article is dedicated to our teacher, Professor Jagat J. Ghosh (Department of Biochemistry, University of Calcutta, Calcutta), for his teaching of the ABCs of biochemistry to us.

References 1. Gutteridge J. M. C. (1994) Chem.-Biol. Interact. 9, 133–144. 2. Van de Walter B. and Zoeterveiz J. P. (1994) Lipids 269, 14546–14552. 3. Gunter T. E. and Pfeiffer D. R. (1990) Am. J. Physiol. 258, C755–C786. 4. Takeyama N., Matsno N. and Tanaua T. (1993) Biochem. J. 294, 719–725. 5. Carini R., Parola M., Dianzani M. U. and Albano E. (1992) Arch. Biochem. Biophys. 297, 110–118. 6. Matsumoto N., Tasaka K., Miyane A. and Tanizawa O. (1990) J. Biol. Chem. 265, 22533–22536. 7. Richter C. (1993) FEBS Lett. 325, 104–107. 8. Richter C. and Kas G. E. N. (1991) Chem.-Biol. Interact. 77, 1–23. 9. Carafoli E. (1987) Annu. Rev. Biochem. 56, 395–433. 10. Crompton M., Goldstone T. P. and Al-Nasser I (1986) In: Intracellular Calcium Regulation (Badert H., Gietzen K., Rosenthal J., Ruden R. and Wolf H. U., Eds), pp. 67–78. Manchester University Press, Manchester. 11. Gunter T. E. and Pfeiffer D. R. (1990) Am. J. Physiol. 258, C755–C785. 12. Nicholls D. G. (1978) Biochem. J. 170, 511–522. 13. Igbavgoa M. and Pfeiffer D. R. (1991) J. Biol. Chem. 266, 938–947. 14. Packer C. S. (1994) Proc. R. Soc. Exp. Biol. Med. 207, 148–174. 15. Cox D. and Matlib M. A. (1993) J. Biol. Chem. 268, 938–947. 16. Kinally K. W., Antonenko Y. N. and Zorov D. B (1992) J. Bioenerg. Biomembr. 24, 99–110. 17. Gunter K. K. and Gunter T. E. (1994) J. Bioenerg. Biomembr. 26, 471–485. 18. Fagian M. M., Silva L. P., Martins I. S. and Vercesi A. (1990) J. Biol. Chem. 265, 19955–19960. 19. Huttenberg J. G., Mario J. D., Voorst V., Marle J. V., Mathy M. J., Beckeringh J. JH., Van Deenen H. A., Hendricks M. G. C. and Zwieten P. A. V. (1990) Eur. J. Pharmacol. 178, 71–78. 20. Miller D. J. and Macfarlane N. G. (1995) J. Hum. Hypertens. 9, 465–473. 21. Xi T. and Rao M. R. (1995) Yao Hsneh Hsneh Pao (China) 30, 6–11. 22. Saris N. E. and Erricksson K. O. (1995) Acta Anaesthesiol. Scand. Suppl. 107, 171–176.

Oxidant, Mitochondria and Calcium 23. McCord J. M. (1987) Fed. Proc. 46, 2402–2406. 24. Haworth R. A. and Hunter D. R. (1979) Arch. Biochem. Biophys. 195, 460–467. 25. Gunter T. E. and Pfeiffer D. R. (1990) Am. J. Physiol. 258, C755–C786. 26. Zoccarato F., Pugolo M., Siliprandi D. and Siliprandi N. (1981) Eur. J. Biochem. 114, 195–199. 27. Schlegel J., Schweizer M. and Richter C (1992) Biochem. J. 285, 65–69. 28. Richter C. and Frei B (1988) Free Radical Biol. Med. 4, 365–375. 29. Haworth R. A. and Hunter D. R. (1980) J. Membr. Biol. 54, 231–236. 30. Scatton B (1993) Fund. Clin. Pharmacol. 7, 383–400. 31. Bernardi P., Brockneir K. M. and Pfeiffer D. R. (1994) J. Bioenerg. Biomembr. 26, 509–517. 32. Reichman N., Porteons C. M. and Murphy M. P. (1994) Biochem. J. 297, 151–155. 33. Crompton M. and Costi A. (1990) Biochem. J. 226, 33–39. 34. Halistrap A. P. and Davidson A. M. (1990) Biochem. J. 266, 33–39. 35. Kowaltowski A. J., Castilho R. F. and Vercesi A. (1996) FEBS Lett. 378, 150–152. 36. Kristal B. S., Park B. K. and Yu B. P. (1996) J. Biol. Chem. 271, 6033–6038. 37. Aguilar H. I., Bolta R., Arora A. S., Bronk S. F. and Gores G. J. (1996) Gastroenterology 110, 558–566. 38. Chakraborti S., Michael J. R., Gurtner G. H., Ghosh S. S., Datta G. and Merker A. (1993) Biochem. J. 292, 585–589. 39. Broekemeir K. M., Carpenter-Deyo L., Reed D. I. and Pferiffer D. R. (1992) FEBS Lett. 304, 192–194. 40. Glin M. A., Lee C. P. and Ernster L. (1996) Biochem. Biophys. Acta 1318, 246–254. 41. Beatrice M. C., Stiers D. L. and Pfeifler D. R. (1994) J. Biol. Chem. 259, 1279–1287. 42. Costantini P., Chernyak B. V., Petronilli V. and Bernardi P. (1995) FEBS Lett. 362, 239–242. 43. Tarveer A., Virij S., Andreera L., Totty N. F., Hsnan J. J., Ward J. M. and Crompton M. (1996) Eur. J. Biochem. 238, 166–172. 44. Schweizer M. and Richter C. (1996) Biochemistry 35, 4524– 4528. 45. Richter C., Gogradge V., Laffranchi R., Schlapbach R., Schweizer M., Sutter M. W., Alter P. and Yaffee M. (1995) Biochim. Biophys. Acta 1271, 67–74 46. Lotscher H. R., Winterhalter K. H., Carafoli E. and Richter C. (1979) Proc. Natl. Acad. Sci. USA 76, 4340–4344. 47. Ceccavelli D., Gallesi D., Giovarni F., Ferrali M. and Masini A. (1995) Pharmacol. Res. 12, 518–522. 48. Roychoudhury S., Chakraborti T., Ghosh S. K., Ghosh J. J. and Chakraborti S. (1996) Indian J. Biochem. Biophys. 33, 218–222. 49. Zharg Z., Ziegler M., Schneider R., Klocker H., Aner B. and Schweizer M. (1995) FEBS Lett. 377, 530–534.

85 50. Hofsletter N., Muchlebach T., Lotscher H. R., Winterhalter K. A. and Richter C. (1981) Eur. J. Biochem. 117, 361–367. 51. Richter, C. (1992) In: New Comprehensive Biochemistry (Neuberger A. and van Deenen L. L. M., Eds). pp. 349–358. Elsevier, Amsterdam. 52. Richter C., Theus M. and Schlegel I. J (1990) Biochem. Pharmacol. 40, 779–782. 53. Slater A. F., Nobel C. S. and Orrenius S. (1995) Biochim. Biophys. Acta 1271, 59–62. 54. Hansford R. G. (1991) Bioenerg. Biomembr. 23, 823–854. 55. MacCormack J. G. and Denton R. M. (1993) Biochem. Soc. Trans. 221, 793–798. 56. Brown G. C. (1995) FEBS Lett. 369, 136–139. 57. Naqui A., Chance B. and Cadenas E. (1990) Annu. Rev. Biochem. 55, 137–166. 58. Lucas Heron B., Ray B. L. and Schmitt N. (1995) FEBS Lett. 374, 309–311. 59. Zhang J., Ziegler M., Schneider R., Klocker H., Auer B. and Schereizer M. (1995) FEBS Lett. 377, 530–534. 60. Jewell S. A., Bellomo G., Thor H., Orrenius S. and Smith M. (1982) Science 217, 1257–1259. 61. Carbonera D., Angrilli A. and Azzone G. F. (1988) Biochim. Biophys. Acta 936, 139–147. 62. Richter C., Schweizer M., Cossarizza A. and Franchesi C. (1996) FEBS Lett. 378, 107–110. 63. Chacon E., Ohata H., Harper I. S., Trollinger D. R., Herman B. and Lamsters J. J. (1996) FEBS Lett. 382, 31–36. 64. Zhivetovsky B., Burgess D. H., Vanags D. M. and Orrenius (1997) Biochem. Biophys. Res. Commun. 230, 481–488. 65. Richter C., Theus M. and Schlegel I. J. (1990) Biochem. Pharmacol. 40, 779–782. 66. Kroemer G., Petit P., Zamzami N., Vayssiere J. and Mignotte B. (1995) FASEB J. 9, 1277–1287. 67. Levine B., Huang Q., Isaacs J. T., Reed J. C., Griffin D. E. and Hardwick J. M. (1993) Nature 36, 739–742. 68. Schulze-Osthoff K., Bakker A. C., Vanhaesbrocen B., Bayaert R. and Jacob W. A. (1992) J. Biol. Chem. 267, 5317–5323. 69. Korsmeyer S. J., Yin X. M., Oltvai Z. N., Veis-Novack D. J. and Linette G. P. (1995) Biochim. Biophys. Acta 1271, 63–66. 70. Kroemer G., Zamzami N. and Susin S. R. (1997) Immunol. Today 18, 44–51. 71. Hermis-Lima M., Castilho R. F., Valle V. G. R., Bechara E. G. H. and Vercesi A. E. (1992) Biochim. Biophys. Acta 1180, 201–206. 72. Roychoudhury S, Ghosh S. K., Chakraborti T. and Chakraborti S. (1996) Indian J. Exp. Biol. 34, 1220–1223. 73. Chakraborti S., Batabyal S. K. and Chakraborti T. (1995) Mol. Cell. Biochem. 146, 91–98. 74. Richter C., Scheveizer M., Cossarizza A. and Franchesi C. (1996) FEBS Lett. 378, 107–110. 75. Chen L. B. (1988) Annu. Rev. Cell Biol. 4, 155–181. 76. Lam M., Dubyak G., Chen L., Nunez G., Miesfeld R. L. and Diestelhorst C. W. (1994) Proc. Natl. Acad. Sci. USA 91, 6569–6573. 77. Freeman B. A. and Crapo J. D. (1982) Lab. Invest. 47, 412–426.