Mitochondria and calcium signaling

3 downloads 0 Views 152KB Size Report
Keywords: Mitochondria; Calcium uniporter; Calcium phosphate; Membrane potential; Phosphate .... Ca–phosphate complex has a defined solubility product, it.
Cell Calcium 38 (2005) 311–317

Mitochondria and calcium signaling David G. Nicholls ∗ Buck Institute for Age Research, Redwood Boulevard, Novato, CA 94945, USA Received 20 June 2005; accepted 28 June 2005 Available online 8 August 2005

Abstract The kinetic properties for the uptake, storage and release of Ca2+ from isolated mitochondria accurately predict the behaviour of the organelles within the intact cell. While the steady-state cycling of Ca2+ across the inner membrane between independent uptake and efflux pathways seems at first sight to be symmetrical, the distinctive kinetics of the uniporter, which is highly dependent on external free Ca2+ concentration and the efflux pathway, whose activity is clamped over a wide range of total matrix Ca2+ by the solubility of the calcium phosphate complex provide a mechanism whereby mitochondria reversibly sequester transient elevations in cytoplasmic Ca2+ . Under non-stimulated conditions, the same transport processes can regulate matrix Ca2+ concentrations and hence citric acid cycle activity. © 2005 Elsevier Ltd. All rights reserved. Keywords: Mitochondria; Calcium uniporter; Calcium phosphate; Membrane potential; Phosphate

1. Ca2+ cycling by isolated mitochondria and the concept of the set-point Almost as soon as isolated mitochondria began to be investigated, their extraordinary capacity to accumulate and retain calcium became apparent [1,2]. In the pre-chemiosmotic era, molecular mechanisms were lacking, but many of the fundamental principles were elucidated. Thus, Ca2+ uptake could be driven either by respiration or ATP hydrolysis; the former but not the latter was sensitive to the ATP synthase inhibitor oligomycin, and uncouplers abolished both means of accumulation, showing that Ca2+ transport involved some fundamental bioenergetic property of the mitochondrion similar to ATP generation, rather than utilizing a dedicated Ca2+ -ATPase (for review see [3]). The capacity of mitochondria to accumulate Ca2+ was enormous and could exceed 1000 nmol/mg mitochondrial protein. It soon became clear that an essential role was played by parallel but independent transport of phosphate [1]. Intriguingly, the accumulation of Ca2+ in the presence of phosphate led to the appearance of protons in the extra-mitochondrial medium [1], but it was not ∗

Tel.: +1 415 209 2095; fax: +1 415 209 2232. E-mail address: [email protected].

0143-4160/$ – see front matter © 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.ceca.2005.06.011

initially realized that these were primary ‘protons’ pumped out of the mitochondrion by the respiratory chain. Early mitochondrial studies focused on isolated mitochondrial Ca2+ transport per se rather than considering the role that this process might perform in the physiological milieu of the intact cell. The seminal discovery by Crompton and Carafoli [4,5] that heart mitochondria possess a Na+ /Ca2+ exchanger in addition to the Ca2+ uniporter led to the concept that Ca2+ cycles continuously across the inner membrane between these two transporters [4,6,7], and removed a problem for the chemiosmotic theory that the Ca2+ uniporter was simply too powerful and unidirectional, since a 150 mV membrane potential (ψ) should lead to a predicted equilibrium free Ca2+ gradient of 105 between cytoplasm and matrix. However, even with this development, there was a considerable debate as to whether the ‘affinity’ of mitochondrial Ca2+ transport was sufficient for it to play a significant physiological role in the intact cell. In 1978, we started a series of studies to investigate the ability of isolated mitochondria to regulate the free Ca2+ concentration in their incubation medium [6]. We were aided in this by the development of sensitive Ca2+ -selective electrodes capable of monitoring the extra-mitochondrial free Ca2+ concentration ([Ca2+ ]e ) down to 100 nM. As long as phosphate was present as a

312

D.G. Nicholls / Cell Calcium 38 (2005) 311–317

Fig. 1. Ca2+ transport pathways in brain mitochondria. (a) Steady-state Ca2+ cycling; (b) net Ca2+ accumulation at pH 7 in the presence of excess phosphate.

permeant anion and physiological concentrations of adenine nucleotides were present in the incubation, mitochondria from liver and brain were found to buffer [Ca2+ ]e at values in the region of 0.5–1 ␮M when the total matrix Ca2+ load was varied over a wide range [6,8,9]. This buffering was dynamic, i.e. a bolus addition of Ca2+ was taken up into the mitochondrion until [Ca2+ ]e fell to precisely the previous value; alternatively, addition of a Ca2+ chelator resulted in a Ca2+ efflux from the mitochondrion to restore the initial [Ca2+ ]e . This led to the concept of a ‘set-point’ for [Ca2+ ]e at which Ca2+ uptake and efflux were equal and opposite [6]. Thus, while Ca2+ cycling across the inner membrane looks at first sight to be symmetrical across the membrane, the distinctive kinetics of the uniporter and efflux pathway and the differing extra- and intra-mitochondrial environments impose asymmetry (Fig. 1). It is apparent that the activity of the uniporter increases rapidly with [Ca2+ ]e . To quantify the kinetics of net Ca2+ accumulation by respiring mitochondria, we slowly infused Ca2+ into a mitochondrial suspension while monitoring [Ca2+ ]e which rose until a steady-state was reached at which the rate of net Ca2+ accumulation by the mitochondria exactly balanced the rate of Ca2+ infusion. By repeating the experiment at a range of infusion rates, it was possible to construct a graph of net Ca2+ uptake rate versus [Ca2+ ]e (Fig. 2a). In order to obtain values for the net uniporter activity, it was necessary to determine the activity of the Ca2+ efflux pathway. Since low concentrations of ruthenium red inhibit uniporter activity without affecting the efflux pathway [7], addition of the inhibitor to Ca2+ -loaded mitochondria allows the activity of the efflux pathway to be determined directly. A remark-

Fig. 2. Kinetics of the Ca2+ uniporter and Ca2+ efflux pathway in rat liver mitochondria. (a) Ca2+ uniport activity was quantified as a function of external free Ca2+ ([Ca2+ ]e ) by infusing Ca2+ into the incubation at varying rates and determining the value at which [Ca2+ ]e stabilized, indicating that net uptake into the matrix exactly balanced the rate of Ca2+ infusion. This rate was then corrected for the (constant) rate of the efflux activity to obtain the absolute uniporter activity. The best fit regression line shown is proportionate to the 3.5 power of [Ca2+ ]e . (b) In the presence of excess phosphate the activity of the Ca2+ efflux pathway is constant over a wide range of total matrix Ca2+ loads. (c) Matrix free Ca2+ concentration, [Ca2+ ]m , determined with matrix-located Ca2+ indicator as a function of total matrix Ca2+ load. (d) Activity of the efflux pathway as a function of external phosphate concentration. The liver mitochondrial Ca2+ efflux pathway has a low-affinity for matrix free Ca2+ ([Ca2+ ]m ) which in turn is defined by the solubility product of the matrix Ca3 (PO4 )2 complex. As external phosphate is increased, matrix free phosphate will increase in parallel with the result that [Ca2+ ]m will decrease as the 1.5 power of the free phosphate to maintain a constant solubility product. The regression line is plotted according to this relationship. Original data from [9,16].

D.G. Nicholls / Cell Calcium 38 (2005) 311–317

able feature, to which we shall return below, is that the activity of the efflux pathway is largely independent of matrix Ca2+ load as long as the external phosphate concentration and pH are kept constant (Fig. 2b) [9]. Using this efflux data to correct the infusion experiment gave the absolute activity of the liver mitochondrial Ca2+ uniporter to vary as a 3.5 power function of [Ca2+ ]e , ranging from 1 nmol/min/mg at 1 ␮M [Ca2+ ]e to 34 nmol/min/mg at 4 ␮M [Ca2+ ]e [9]. The often repeated assertion that mitochondrial Ca2+ transport is a ‘lowaffinity’ process is highly misleading. At least in the case of brain mitochondria, Ca2+ uptake can utilize the full respiratory capacity of the mitochondrion when [Ca2+ ]e rises above about 5 ␮M [3]. The apparent constancy of the efflux pathway appeared to be related to the free matrix Ca2+ concentration rather than a property of the transporter itself. Thus, when acetate was substituted for phosphate the efflux activity increased linearly with matrix Ca2+ load [9]. This suggested that the formation of a matrix Ca–phosphate complex was controlling the free matrix Ca2+ concentration, [Ca2+ ]m . If the matrix Ca–phosphate complex has a defined solubility product, it would be predicted that a decrease in matrix free phosphate would result in an increase in [Ca2+ ]m and hence an increased activity of the efflux pathway. Because of the highly active phosphate transporter, matrix free phosphate is directly proportional to external phosphate concentration. Increasing the phosphate concentration in the medium to 3.3 mM decreased the rate of ruthenium red-induced Ca2+ efflux from Ca2+ loaded liver mitochondria by 16-fold relative to the rate with phosphate-depleted mitochondria (Fig. 2c) [9], with no change in ψ because of the very sharp dependency of uniporter activity on [Ca2+ ]e . This dramatic change in the rate of Ca2+ cycling across the membrane only caused a modest change in the set-point: from 0.78 to 0.55 ␮M.

2. The matrix Ca–phosphate complex These early findings strongly argued for the formation of a Ca2+ –phosphate complex within the matrix of Ca2+ -loaded mitochondria. However, this complex has properties that appear different from conventional Ca2+ –phosphate complexes in the test-tube. Firstly, collapse of ψ by addition of a protonophore leads to a rapid efflux of Ca2+ from the matrix; Ca2+ and phosphate exit the mitochondrion separately by, respectively, uniporter reversal and the phosphate transporter. Thus, any Ca2+ –phosphate complex within the matrix is instantly dissociable, and thus differs from test-tube precipitates that are notoriously difficult to redissolve. Secondly, a number of studies have attempted to determine the free matrix Ca2+ concentration, either from the activity of Ca2+ -activated enzymes of the citric acid cycle [10–13], by mitochondriallytargeted aequorins [14] or with matrix-loaded fluorescent Ca2+ indicators [15]. The general consensus is that the free matrix Ca2+ ([Ca2+ ]m ) is in the range 0.5–2 ␮M under a wide

313

variety of loading conditions, as long as mM phosphate concentrations are present in the incubation medium. The observation that the set-point in the presence of excess phosphate is largely independent of matrix Ca2+ loading over a range from 10 to 500 nmol Ca2+ /mg protein [16], argues that [Ca2+ ]m is similarly invariant with Ca2+ load, since otherwise the activity of the efflux pathway would increase with Ca2+ load. This constancy was confirmed in a recent study [16], where [Ca2+ ]m was found to change by less than 50% when total matrix Ca2+ was increased 50-fold from 10 to 500 nmol/mg (Fig. 2d). The rapid reversibility of matrix Ca2+ –phosphate formation led Lehninger and others [17] to search for an anti-nucleation factor; however, despite possible candidates this has not been continued. Our recent studies [16] have taken a different direction, based on the dramatic difference in the free Ca2+ concentrations in equilibrium with the Ca2+ –phosphate complex in the matrix (∼1 ␮M) and in the test-tube, for example in physiological cell incubation buffers (∼1 mM). There has been considerable uncertainty as to the structure of the matrix Ca2+ –phosphate complex, with the two most discussed forms being hydroxyapatite (Ca5 (PO4 )3 OH) and tricalcium phosphate (Ca3 (PO4 )2 ). Our proposal has been that matrix pH plays the key role in controlling the stability of these complexes. Thus, in order for Ca3 (PO4 )2 to form in the matrix, phosphate has to enter through the phosphate transporter as the electroneutral H3 PO4 (formally equivalent to a H2 PO4 − /OH− antiport) and then undergo three successive deprotonations to form the PO4 3− anion before complexing with Ca2+ . Thus, with constant extramitochondrial phosphate the concentration of the trivalent anion in the matrix will vary inversely with the cube of the proton concentration. Accurate values for the solubility products for these complexes are difficult to obtain, but a consensus value for Ca3 (PO4 )2 of 3 × 10−30 means that the [Ca2+ ]m in equilibrium with the complex in the matrix of mitochondria incubated in the presence of 2 mM phosphate would be about 2 ␮M if the matrix pH were 7.8 and ∼100 ␮M for a matrix pH of 7.0 [16]. Thus, the maintenance of the Ca3 (PO4 )2 complex in the matrix is primarily governed by matrix pH, which is in the region of 7.8 in mitochondria respiring in the presence of phosphate [18]. The question as to why protonophore addition leads to a dramatic release of Ca2+ from the matrix when the concentration gradient of free Ca2+ across the inner membrane is small, or even reversed [13] is thus answered, since protonophore addition leads to a drastic acidification of the matrix [18] destabilizing the complex and increasing the free matrix Ca2+ concentration a 100-fold.

3. Matrix Ca2+ : metabolic regulator and reversible Ca2+ store While the basic feature of mitochondrial Ca2+ transport were being elucidated, a lively debate ensued as to whether

314

D.G. Nicholls / Cell Calcium 38 (2005) 311–317

the prime physiological role of the process was to store cytoplasmic Ca2+ reversibly [19] or to regulate the activity of certain matrix-located dehydrogenases, including pyruvate dehydrogenase, NAD+ -isocitrate dehydrogenase and ␣ketoglutarate dehydrogenase [20]. These enzymes showed a substantial activation in vitro when [Ca2+ ]e was increased over the range 0.1–2 ␮M [12,20]. At a cellular level, an increase in cytoplasmic free Ca2+ , [Ca2+ ]c , would increase [Ca2+ ]m with a resulting activation of energy-producing dehydrogenases. This has been demonstrated in a number of cell types; thus, hepatocytes stimulated with high vasopressin concentrations to generate a sustained [Ca2+ ]c elevation showed an increase in [Ca2+ ]m synchronous with an activation of pyruvate dehydrogenase, enhanced NADH reduction and a raised protonmotive force [21]. This mechanism can serve to minimize the drop in protonmotive force and ATP/ADP ratio that would otherwise accompany an enhanced cellular ATP demand during hormonal activation. This regulatory mechanism can only operate over a range where [Ca2+ ]m varies, i.e. when the range of [Ca2+ ]c is below the set-point so that there is insufficient matrix Ca2+ to form a Ca2+ –phosphate complex. An elegant feature of mitochondrial Ca2+ transport is that there is a smooth transition from this ‘regulatory’ range of mitochondrial Ca2+ transport to the ‘storage’ phase when about 10 nmol Ca2+ /mg has been accumulated within the matrix and the Ca2+ –phosphate complex starts to form [16]. The physiological consequences of this reversible storage will be discussed below.

4. The permeability transition From the earliest days of mitochondrial Ca2+ transport, it was apparent that their capacity to accumulate Ca2+ was finite; when this was exceeded, Ca2+ was released together with low molecular weight matrix contents and solute entered resulting in swelling of the matrix and rupture of the outer membrane. This phenomenon, the permeability transition, has led to an enormous literature (for reviews see [22]), particularly in the last 10 years when induction of the permeability transition pore (PTP) has been studied as one mechanism to initiate the release of cytochrome c and other pro-apoptotic proteins. In reviewing this vast literature, it is important to distinguish studies that make an attempt to reproduce physiologically (or pathophysologically) relevant conditions, since there are many papers that incubate mitochondria in artificial media (often sucrose in the absence of adenine nucleotides), expose them to bolus Ca2+ additions far higher than would be encountered in the cell and simply measure the decrease in scattered light as solute enters the matrix equalizing the refractive indices of the medium and matrix. Under more relevant conditions (including the presence of adenine nucleotides), the PTP can either be induced by massive loading of the matrix with Ca2+ (in the case of brain mitochondria by exceeding 1500 nmol Ca2+ accumulated/mg protein [16])

or by thiol oxidation in which case the Ca2+ requirement is greatly reduced [23]. The best attested context for a pathophysiological induction of the permeability transition is in ischemia-reperfusion of the heart [24] and perhaps the brain [25]. 5. Mitochondrial Ca2+ buffering in the intact cell As discussed above, Ca2+ transport by isolated mitochondria can either function to regulate matrix dehydrogenase regulation or to buffer extra-mitochondrial Ca2+ . There is no contradiction between these functions, since there is a smooth transition between modes at a matrix Ca2+ load of about 10 nmol/mg protein when the Ca2+ –phosphate complex starts to form [16]. Thus, matrix dehydrogenase regulation is the dominant mode in non-excitable cells where [Ca2+ ]c remains below the mitochondrial set-point, whereas in many excitable secretory cells matrix Ca2+ sequestration is an important mechanism to prolong, blunt, or remove feed-back inhibition upon physiological Ca2+ transients that play key regulatory roles. Because other chapters in this issue focus on endoplasmic reticulum Ca2+ transport and its interaction with mitochondrial Ca2+ regulation, this discussion will be limited to studies with neurons where mitochondrial Ca2+ buffering appears to play key physiological and patho-physiological roles in the life and death of the cell. With the advent of digital Ca2+ imaging, it became apparent that in intact cells, particularly from excitable tissues such as muscle and brain, [Ca2+ ]c could undergo large transient increases in response to stimulation [26–28], taking the parameter well above the ‘set-point’ values of 0.5–1 ␮M. Our studies with isolated mitochondria led us to predict in 1983 [19] that in situ they would become net accumulators of the cation when [Ca2+ ]c was above the set-point, and that they would slowly release Ca2+ to the cytoplasm when [Ca2+ ]c was lowered below this value and uniporter activity fell below that of the efflux pathway. Thus, mitochondria would tend to limit the peak [Ca2+ ]c elevation occurring as a consequence of plasma membrane entry and would broaden the time profile of the elevation [19]. Importantly, the mitochondrial buffer system would only be activated when [Ca2+ ]c rose above 0.5 ␮M, whereas the plasma membrane Ca2+ transport processes were capable of maintaining basal [Ca2+ ]c at about 0.1 ␮M; thus, mitochondria should be largely depleted of Ca2+ under basal conditions—as has been repeatedly found [29]. This is important, since a set-point at or below the basal [Ca2+ ]c in a cell would result in an inexorable loading of the matrix with Ca2+ and the rapid death of the cell. In neurons, Thayer and Miller [30] induced varying Ca2+ loads in dorsal root ganglion cells by plasma membrane depolarization with a whole-cell patch electrode. They found that small [Ca2+ ]c transients decayed monophasically as the cation is pumped out of the cell; however, larger Ca2+ elevations were buffered by a ruthenium red-sensitive process, while recovery to basal [Ca2+ ]c was delayed by a ‘shoulder’

D.G. Nicholls / Cell Calcium 38 (2005) 311–317

that was abolished in the presence of a protonophore, arguing that the endogenous mitochondria were reversibly sequestering the ‘excess’ cytoplasmic Ca2+ . Interestingly, the [Ca2+ ]c at which this shoulder appeared, approximately 0.5 ␮M [30] was close to the set-point found in experiments with isolated brain mitochondria [8]. Mitochondria in resting neurons contain no Ca2+ that can be detected by a transient increase in [Ca2+ ]c on addition of protonophore [31] or by direct X-ray probe analysis [29]. In such studies, it is important to distinguish between a transient spike of Ca2+ as it is released from the matrix into the cytoplasm and then extruded across the plasma membrane, and a sustained [Ca2+ ]c elevation after protonophore that usually denotes cytoplasmic ATP deficiency. Prior inclusion of oligomycin so that ATP is supplied by glycolysis avoids a confounding change in ATP levels when the protonophore is added. Alternatively, ψm can be selectively depolarized by the combination of a respiratory chain inhibitor and oligomycin [31]. Studies with lizard motor nerve terminals [32,33] provide more evidence for the physiological importance of reversible mitochondrial Ca2+ sequestration. In response to high-frequency stimulation [Ca2+ ]c initially increased rapidly; when [Ca2+ ]c had increased by 200 nM, mitochondrial [Ca2+ ]m , detected by the mitochondrially-localized rhod-2, started to rise. On termination of stimulation Ca2+ fell more rapidly than [Ca2+ ]m . In excellent agreement with the studies with isolated brain mitochondria [16], [Ca2+ ]m seems to plateau at 1–2 ␮M free Ca2+ regardless of the total matrix content due to the formation of the Ca2+ –phosphate complex [34,35]. Interestingly, a mutation in superoxide dismutase 1 that underlies some familial forms of amyotrophic lateral sclerosis was associated with defective matrix buffering of free Ca2+ concentration [35]. Cultured neurons can be loaded with Ca2+ by activation of voltage-dependent Ca2+ channels (VDCCs) or Ca2+ permeant ionotropic receptors, notably the NMDA receptor or AMPA receptors lacking a GluR2 subunit. A variety of techniques can be employed to monitor matrix accumulation of Ca2+ . Energy dispersive X-ray microanalysis of cryo-sectioned slice cultures [36–38] provides a means for the direct quantitation of total Ca2+ deposits in sub-cellular organelles. Interestingly, Ca2+ sequestration in the dendrites of hippocampal CA3 neurons following trains of action potentials was initially reported to occur in a sub-population of endoplasmic reticulum rather than mitochondria [36], but this was subsequently revised when improved time-resolution showed an initial mitochondrial accumulation followed by a slower translocation from mitochondria to the dendritic e.r. [38]. Discrete calcium deposits within the mitochondria were associated with phosphate, again emphasizing the importance of Ca2+ –phosphate complex formation. When X-ray microanalysis was applied to frog sympathetic neurons depolarized by elevated KCl so that [Ca2+ ]c rose to 600 nM a large increase in mitochondrial Ca2+ was seen, that was abolished by protonophore [37]. X-ray micro-

315

analysis allows spatial resolution within the cell, and it was seen that mitochondria close to the plasma membrane preferentially loaded with the cation, suggesting that mitochondria can trap Ca2+ entering through VDCCs before it diffuses uniformly throughout the cell [37]. It should be noted once again that the in situ set-point for net mitochondrial Ca2+ accumulation fits with that determined earlier for isolated mitochondria [6,8]. The dependency of mitochondrial Ca2+ accumulation on the sub-cellular localization of the organelles is also seen in pancreatic acinar cells, where mitochondria located close to the plasma membrane, secretory granules or nucleus each respond preferentially to [Ca2+ ]c transients in their immediate vicinity [39].

6. Physiological roles of neuronal mitochondria Ca2+ transport In the neuronal context, mitochondrial Ca2+ uptake could increase the inward Ca2+ current by providing a sink preventing feed-back inhibition of the channel, or blunting and extending the duration of [Ca2+ ]c elevations by reversible uptake and release. There is convincing evidence for both mechanisms. Prior mitochondrial depolarization in the absence of ATP depletion decreases the total 45 Ca uptake into neurons induced by NMDA receptor activation [40] implying the existence of a feed-back mechanism that is inactivated in the presence of functional mitochondrial Ca2+ transport. In mouse motor nerve terminals, David and Barrett [41,42] found that mitochondrial depolarization without ATP depletion caused a large increase in asynchronous transmitter release during low frequency stimulation in parallel with a greater increase in bulk cytoplasmic Ca2+ concentration, suggesting that a major role of mitochondrial Ca2+ transport in the nerve terminal was to sequester Ca2+ subsequent to action potentials to limit asynchronous transmitter release.

7. Mitochondrial calcium transport and excitotoxic neuronal death Pathological activation of NMDA-selective glutamate receptors results in a massive entry of Ca2+ into neurons and its accumulation by the in situ mitochondria [29,40,43–45]. Unless the rate of Ca2+ entry totally overwhelms the maximal activity of the plasma membrane Ca2+ -ATPase (PMCA) mitochondrial Ca2+ accumulation seems to play a key role in the subsequent fate of the neuron [46]. Since no cell-permeant selective inhibitor of the mitochondrial Ca2+ uniporter has been described, an indirect strategy has to be adopted to prevent mitochondria from accumulating Ca2+ . One approach that has been successful with cell preparations with sufficiently high glycolytic ATP generation is to compare the fates of neurons in the presence of oligomycin (where ATP generation is glycolytic but mitochondrial Ca2+

316

D.G. Nicholls / Cell Calcium 38 (2005) 311–317

accumulation is unaffected) with that in the presence of both oligomycin and a respiratory chain inhibitor, resulting in a decay of ψm and inhibition of mitochondrial Ca2+ uptake. In such preparations, mitochondrial depolarization is protective [29,40,47,48]. The mechanism by which mitochondrial Ca2+ loading induces neuronal necrosis as a consequence of NMDA receptor activation remains unclear. The simplest explanation for the effects of continuous NMDA receptor activation, sufficient to maintain [Ca2+ ]c above the mitochondrial set-point, would be ultimate Ca2+ overload of the mitochondria and induction of the permeability transition [49]. Certainly the in situ mitochondria show a reversible change in shape from thread-like to swollen during neuronal glutamate exposure [29,50,51] consistent with their matrix Ca2+ loading. Mitochondrial depolarization is synchronous with the loss of cytoplasmic Ca2+ homeostasis (delayed Ca2+ deregulation, DCD) that is the first commitment point of excitotoxic cell death [40,47]. Under carefully controlled conditions of NMDA receptor activation, a significant proportion of neurons undergo apoptotic rather than necrotic cell death [52]. Even transient exposure of cultured neurons to glutamate can cause a delayed loss of mitochondrial cytochrome c consistent with activation of the permeability transition [53,54] and a question is how the resulting collapse of ψm can be reconciled with the maintenance of sufficient ATP levels for the processing of programmed cell death. A recent analysis by Pivavorova et al. [29] showed that under such conditions, only a sub-population of mitochondria within the glutamateexposed neurons swelled and were damaged. Thus, sufficient mitochondrial release of cytochrome c can occur to trigger apoptosis while the residual functional mitochondria continue to maintain the cell.

8. The roles of reactive oxygen species in excitotoxic neuronal death A working hypothesis has been that pathological NMDA receptor activation and the resulting matrix Ca2+ accumulation leads to mitochondrial superoxide (O2 •− ) generation, damage to the organelle and the cell and subsequent cell death [47,55–57]. One problem is that there is no clear mechanism by which Ca2+ loaded but otherwise functional mitochondria should produce excess O2 •− , indeed insofar that mitochondria actively accumulating Ca2+ have a lower ψm than controls, the prediction, which has been confirmed [58], is that O2 •− levels are unchanged or lower in these mitochondria. Furthermore, potent cell-permeant superoxide dismutase mimetics effectively trap superoxide and yet have no effect on the timing or extent of DCD [59], while O2 •− levels in individual glutamate-exposed neurons only increase when DCD occurs [59], suggesting that the increase is a consequence, rather than a cause, of mitochondrial depolarization.

Acknowledgment Part of this study was funded by NIH grant R01 NS4 1908. References [1] C.S. Rossi, A.L. Lehninger, Stoichiometry of respiratory stimulation, accumulation of Ca2+ and phosphate and oxidative phosphorylation in rat liver mitochondria, J Biol. Chem. 239 (1964) 3971–3980. [2] B. Chance, The energy-linked reaction of calcium with mitochondria, J. Biol. Chem. 240 (1965) 2729–2748. ˚ [3] D.G. Nicholls, KE.O. Akerman, Mitochondrial calcium transport, Biochim. Biophys. Acta 683 (1982) 57–88. [4] M. Crompton, I. Heid, The cycling of calcium, sodium, and protons across the inner membrane of cardiac mitochondria, Eur. J. Biochem. 91 (1978) 599–608. [5] M. Crompton, R. Moser, H. Ludi, E. Carafoli, The interrelations between the transport of sodium and calcium in mitochondria of various mammalian tissues, Eur. J. Biochem. 82 (1978) 25–31. [6] D.G. Nicholls, The regulation of extra-mitochondrial free calcium by rat liver mitochondria, Biochem. J. 176 (1978) 463–474. [7] J.S. Puskin, T.E. Gunter, K.K. Gunter, P.R. Russell, Evidence for more than one Ca2+ transport mechanism in mitochondria, Biochemistry 15 (1976) 3834–3842. [8] D.G. Nicholls, I.D. Scott, The regulation of brain mitochondrial calcium-ion transport: the role of ATP in the discrimination between kinetic and membrane-potential-dependent Ca efflux mechanisms, Biochem. J. 186 (1980) 833–839. [9] F. Zoccarato, D.G. Nicholls, The role of phosphate in the regulation of the Ca efflux pathway of liver mitochondria, Eur. J. Biochem. 127 (1982) 333–338. [10] J.G. McCormack, H.M. Browne, N.J. Dawes, Studies on mitochondrial Ca2+ -transport and matrix Ca2+ using fura-2-loaded rat heart mitochondria, Biochim. Biophys. Acta 973 (1989) 420–427. [11] S.P. Allen, U. Darley, J.G. McCormack, D. Stone, Changes in mitochondrial matrix free calcium in perfused rat hearts subjected to hypoxia-reoxygenation, J. Mol. Cell. Cardiol. 25 (1993) 949–958. [12] R.G. Hansford, F. Castro, Intramitochondrial and extramitochondrial free calcium ion concentrations of suspensions of heart mitochondria with very low, plausibly physiological, contents of total calcium, J. Bioenerg. Biomembr. 14 (1982) 361–376. [13] H. Miyata, H.S. Silverman, S.J. Sollott, E.G. Lakatta, M.D. Stern, R.G. Hansford, Measurement of mitochondrial free Ca2+ concentration in living single rat cardiac myocytes, Am. J. Physiol. 261 (1991) H1123–H1134. [14] R. Rizzuto, AWM Simpson, M. Brini, T. Pozzan, Rapid changes of mitochondrial Ca2+ revealed by specifically targeted recombinant aequorin, Nature 358 (1992) 325–327. [15] S.P. Allen, D. Stone, J.G. McCormack, The loading of fura-2 into mitochondria in the intact perfused rat heart and its use to estimate matrix Ca2+ under various conditions, J. Mol. Cell. Cardiol. 24 (1992) 765–773. [16] S. Chalmers, D.G. Nicholls, The relationship between free and total calcium concentrations in the matrix of liver and brain mitochondria, J. Biol. Chem. 279 (2003) 19062–19070. [17] W.P. Tew, C. Mahle, J. Benavides, J.E. Howard, A.L. Lehninger, Synthesis and characterization of phosphocitric acid, a potent inhibitor of hydroxylapatite crystal growth, Biochemistry 19 (1980) 1983–1988. [18] D.G. Nicholls, The influence of respiration and ATP hydrolysis on the proton electrochemical potential gradient across the inner membrane of rat liver mitochondria as determined by ion distribution, Eur. J. Biochem. 50 (1974) 305–315. ˚ [19] KEO Akerman, D.G. Nicholls, Physiological and bioenergetic aspects of mitochondrial Ca transport, Rev. Physiol. Biochem. Pharmacol. 95 (1983) 149–201.

D.G. Nicholls / Cell Calcium 38 (2005) 311–317 [20] R.M. Denton, J.G. McCormack, N.J. Edgell, Role of calcium ions in the regulation of intramitochondrial metabolism. Effects of Na+ , Mg2+ and ruthenium red on the Ca2+ -stimulated oxidation of oxoglutarate and on pyruvate dehydrogenase activity in intact rat heart mitochondria, Biochem. J. 190 (1980) 107–117. [21] L.D. Robb-Gaspers, P. Burnett, G.A. Rutter, R.M. Denton, R. Rizzuto, A.P. Thomas, Integrating cytosolic calcium signals into mitochondrial metabolic responses, EMBO J. 17 (1998) 4987–5000. [22] P. Bernardi, E. Basso, R. Colonna, et al., Perspectives on the mitochondrial permeability transition, Biochim. Biophys. Acta BioEnergetics 1365 (1998) 200–206. [23] A.P. Halestrap, K.Y. Woodfield, C.P. Connern, Oxidative stress, thiol reagents, and membrane potential modulate the mitochondrial permeability transition by affecting nucleotide binding to the adenine nucleotide translocase, J. Biol. Chem. 272 (1997) 3346–3354. [24] M. Crompton, The mitochondrial permeability transition pore and its role in cell death, Biochem. J. 341 (1999) 233–249. [25] M.J. Hansson, T. Persson, H. Friberg, et al., Powerful cyclosporin inhibition of calcium-induced permeability transition in brain mitochondria., Brain Res. 960 (2003) 99–111. [26] J.A. Connor, Digital imaging of free Ca changes and of spatial gradients in growing processes in single mammalian CNS cells, Proc. Natl. Acad. Sci. U.S.A. 83 (1986) 6179–6183. [27] J.A. Connor, H.Y. Tseng, P.E. Hockberger, Depolarization- and transmitter-induced changes in intracellular Ca2+ of rat cerebellar granule cells in explant culture, J. Neurosci. 7 (1987) 1384–1400. [28] J.A. Connor, W.J. Wadman, P.E. Hockberger, R.K.S. Wong, Sustained dendritic gradients of Ca2+ induced by excitatory amino acids in CA1 hippocampal neurons, Science 240 (1988) 649–653. [29] N.B. Pivovarova, H.V. Nguyen, C.A. Winters, C.A. Brantner, C.L. Smith, S.B. Andrews, Excitotoxic calcium overload in a subpopulation of mitochondria triggers delayed death in hippocampal neurons, J. Neurosci. 24 (2004) 5611–5622. [30] S.A. Thayer, R.J. Miller, Regulation of the intracellular free calcium concentration in single rat dorsal root ganglion neurones in vitro, J. Physiol. (Lond) 425 (1990) 85–115. [31] S.L. Budd, D.G. Nicholls, A re-evaluation of the role of mitochondria in neuronal calcium homeostasis, J. Neurochem. 66 (1996) 403–411. [32] G. David, J.N. Barrett, E.F. Barrett, Evidence that mitochondria buffer physiological Ca2+ loads in lizard motor nerve terminals, J. Physiol. (Lond) 509 (1998) 59–65. [33] G. David, E.F. Barrett, Stimulation-evoked increases in cytosolic [Ca2+ ] in mouse motor nerve terminals are limited by mitochondrial uptake and are temperature-dependent, J. Neurosci. 20 (2000) 7290–7296. [34] G. David, J. Talbot, E.F. Barrett, Quantitative estimate of mitochondrial [Ca2+ ] in stimulated motor nerve terminals, Cell Calc. 33 (2003) 197–206. [35] L. Vila, E.F. Barrett, J.N. Barrett, Stimulation-induced mitochondrial [Ca2+ ] elevations in mouse motor terminals: comparison of wild-type with SOD1-G93A, J. Physiol. (Lond) 549 (2003) 719–728. [36] L.D. Pozzo-Miller, N.B. Pivovarova, R.D. Leapman, R.A. Buchanan, T.S. Reese, S.B. Andrews, Activity-dependent calcium sequestration in dendrites of hippocampal neurons in brain slices, J. Neurosci. 17 (1997) 8729–8738. [37] N.B. Pivovarova, J. Hongpaisan, S.B. Andrews, D.D. Friel, Depolarization-induced mitochondrial Ca accumulation in sympathetic neurons: spatial and temporal characteristics, J. Neurosci. 19 (1999) 6372–6384. [38] N.B. Pivovarova, L.D. Pozzo-Miller, J. Hongpaisan, S.B. Andrews, Correlated calcium uptake and release by mitochondria and endoplasmic reticulum of CA3 hippocampal dendrites after afferent synaptic stimulation, J. Neurosci. 22 (2002) 10653–10661. [39] M.K. Park, M.C. Ashby, G. Erdemli, O.H. Petersen, A.V. Tepikin, Perinuclear, perigranular and sub-plasmalemmal mitochondria have

[40]

[41]

[42]

[43]

[44]

[45]

[46] [47]

[48]

[49]

[50]

[51]

[52]

[53]

[54]

[55]

[56]

[57]

[58]

[59]

317

distinct functions in the regulation of cellular calcium transport, EMBO J. 20 (2001) 1863–1874. S.L. Budd, D.G. Nicholls, Mitochondrial calcium regulation and acute glutamate excitotoxicity in cultured cerebellar granule cells, J. Neurochem. 67 (1996) 2282–2291. G. David, E.F. Barrett, Mitochondrial Ca2+ uptake prevents desynchronization of quantal release and minimizes depletion during repetitive stimulation of mouse motor nerve terminals, J. Physiol. (Lond) 548 (2003) 425–438. J.D. Talbot, G. David, E.F. Barrett, Inhibition of mitochondrial Ca2+ uptake affects phasic release from motor terminals differently depending on external [Ca2+ ], J. Neurophysiol. 90 (2003) 491–502. S.A. Thayer, G.J. Wang, Glutamate-induced calcium loads: effects on energy metabolism and neuronal viability, Clin. Exp. Pharmacol. Physiol. 22 (1995) 303–304. G.J. Wang, S.A. Thayer, Sequestration of glutamate-induced Ca2+ loads by mitochondria in cultured rat hippocampal neurons, J. Neurophysiol. 76 (1996) 1611–1621. R.J. White, I.J. Reynolds, Mitochondria accumulate Ca2+ following intense glutamate stimulation of cultured rat forebrain neurones, J. Physiol. (Lond) 498 (1997) 31–47. D.G. Nicholls, S.L. Budd, Mitochondria and neuronal survival, Physiol. Rev. 80 (2000) 315–360. R.F. Castilho, O. Hansson, M.W. Ward, S.L. Budd, D.G. Nicholls, Mitochondrial control of acute glutamate excitotoxicity in cultured cerebellar granule cells, J. Neurosci. 18 (1998) 10277–10286. A.K. Stout, H.M. Raphael, B.I. Kanterewicz, E. Klann, I.J. Reynolds, Glutamate-induced neuron death requires mitochondrial calcium uptake, Nat. Neurosci. 1 (1998) 366–373. C.C. Alano, G. Beutner, R.T. Dirksen, R.A. Gross, S.S. Sheu, Mitochondrial permeability transition and calcium dynamics in striatal neurons upon intense NMDA receptor activation, J. Neurochem. 80 (2002) 531–538. G.L. Rintoul, A.J. Filiano, J.B. Brocard, G.J. Kress, I.J. Reynolds, Glutamate decreases mitochondrial size and movement in primary forebrain neurons, J. Neurosci. 23 (2003) 7881–7888. J.M. Dubinsky, Y. Levi, Calcium-induced activation of the mitochondrial permeability transition in hippocampal neurons, J. Neurosci. Res. 53 (1998) 728–741. M. Ankarcrona, J.M. Dypbukt, E. Bonfoco, et al., Glutamate-induced neuronal death: a succession of necrosis or apoptosis depending on mitochondrial function, Neuron 15 (1995) 961–973. S.L. Budd, L. Tenneti, T. Lishnak, S.A. Lipton, Mitochondrial and extramitochondrial apoptotic signaling pathways in cerebrocortical neurons, Proc. Natl. Acad. Sci. U.S.A. 97 (2000) 6161–6166. C.M. Luetjens, N.T. Bui, B. Sengpiel, et al., Delayed mitochondrial dysfunction in excitotoxic neuron death: cytochrome c release and a secondary increase in superoxide production, J. Neurosci. 20 (2000) 5715–5723. M. Lafon-Cazal, S. Pietri, M. Culcasi, J. Bockaert, NMDAdependent superoxide production and neurotoxicity, Nature 364 (1993) 535–537. V.P. Bindokas, J. Jordan, C.C. Lee, R.J. Miller, Superoxide production in rat hippocampal neurons: selective imaging with hydroethidine, J. Neurosci. 16 (1996) 1324–1336. I.J. Reynolds, T.G. Hastings, Glutamate induces the production of reactive oxygen species in cultured forebrain neurons following NMDA receptor activation, J. Neurosci. 15 (1995) 3318–3327. A.A. Starkov, B.M. Polster, G. Fiskum, Regulation of hydrogen peroxide production by brain mitochondria by calcium and Bax, J. Neurochem. 83 (2002) 220–228. S. Vesce, L. Kirk, D.G. Nicholls, Relationships between superoxide levels and delayed calcium deregulation in cultured cerebellar granule cells exposed continuously to glutamate, J. Neurochem. 90 (2004) 683–693.