MINIREVIEW Mitochondrial Ca 2+ sequestration and precipitation revisited Christos Chinopoulos and Vera Adam-Vizi Department of Medical Biochemistry, Semmelweis University, Neurobiochemical Group, Hungarian Academy of Sciences, Budapest, Hungary Why is it important to address Ca 2+ sequestration and precipitation? Isolated mitochondria from a variety of sources exhibit a finite capacity to accumulate and retain divalent cations, including Ca 2+ [1,2]. This capacity differs among mitochondria isolated from various tissues, and even among regions of the same tissue, for example in brain [3]. Furthermore, the mitochondrial Ca 2+ accu- mulation capacity in a tissue may change with age without measurable bioenergetic alterations [4–6]. The ability of mitochondria to act as ‘firewalls’ of intracel- lular Ca 2+ waves enables cells to survive an elevated [Ca 2+ ] emergency [7], and the factors that determine the capacity of mitochondrial Ca 2+ buffering thereby affect the cell’s fate. In addition, Ca 2+ uptake capacity may significantly decrease in pathological settings [8]. The factors that determine this capacity can be classi- fied as (a) those that define the quantity of Ca 2+ that can be retained in the mitochondrial matrix, and (b) those that define the threshold for induction of Ca 2+ release. There is no precedent for a lack of interaction between factors of the two classes. For both classes, these factors can be further categorized as ‘intrinsic’ and ‘extrinsic’. ‘Extrinsic’ factors include those that can be manipulated by experimental conditions, such as the presence and amount of phosphate and adenine nucleotides, pH and ionic strength of the medium. ‘Intrinsic’ factors refer to those that result in the vari- ous Ca 2+ accumulation capacities among the various types of mitochondria studied under identical experi- mental conditions. The intrinsic factors also include Keywords adenine nucleotides; Ca 2+ uniporter; complexation; electron microscopy; Na + /Ca 2+ exchanger; phosphocitrate; polyphosphate; precipitation; thermodynamics; uncoupler Correspondence C. Chinopoulos, V. Adam-Vizi, Department of Medical Biochemistry, Semmelweis University, Budapest, Hungary Fax: +361 2670031 Tel: +361 4591500 ext. 60024; +361 2662773 E-mail: chinopoulos.christos@eok.sote.hu; veronika.adam@eok.sote.hu (Received 9 February 2010, revised 19 May 2010, accepted 22 June 2010) doi:10.1111/j.1742-4658.2010.07755.x The ability of mitochondria to sequester and retain divalent cations in the form of precipitates consisting of organic and inorganic moieties has been known for decades. Of these cations, Ca 2+ has emerged as a major player in both signal transduction and cell death mechanisms, and, as a conse- quence, the importance of mitochondria in these processes was soon recog- nized. Early studies showed considerable effort in identifying the mechanisms of Ca 2+ sequestration, precipitation and release by uncouplers of oxidative phosphorylation; however, relatively little information was obtained, and these processes were eventually taken for granted. Here, we re-examine: (a) the thermodynamic aspects of mitochondrial Ca 2+ uptake and release, (b) the insufficiently explained effect of uncouplers in inducing mitochondrial Ca 2+ release, (c) the thermodynamic effects of exogenously added adenine nucleotides on mitochondrial Ca 2+ uptake capacity and precipitate formation, and (d) the elusive nature of the Ca 2+ -phosphate precipitates formed in the mitochondrial matrix. Abbreviations ANT, adenine nucleotide translocase; PTP, permeability transition pore. FEBS Journal 277 (2010) 3637–3651 ª 2010 The Authors Journal compilation ª 2010 FEBS 3637 those that are responsible for emerging differences in Ca 2+ accumulation capacities upon alterations in the extrinsic factors. For example, under otherwise identi- cal experimental conditions, inclusion of adenine nucle- otides increases Ca 2+ accumulation capacity in rat brain mitochondria by a factor of $ 20, while that in rat liver mitochondria is increased by a factor of < 2. Given the intimate relationship between mitochondrial Ca 2+ handling and signal transduction and cell death pathways [9–13], it is imperative to identify the mecha- nism(s) that define mitochondrial Ca 2+ accumulation capacity. This will lead to a better understanding of the contribution of mitochondria to Ca 2+ homeostasis in physiological and pathological settings, and provide the opportunity to identify potential targets for phar- macological ⁄ genetic manipulation. The purpose of this review is to address the factors that define mitochon- drial Ca 2+ retention: those affecting the induction of Ca 2+ release with respect to opening of the permeabil- ity transition pore (PTP) have been extensively reviewed elsewhere [9,14] and in the accompanying reviews [15,16]. Thermodynamic aspects of mitochondrial Ca 2+ uptake and release mechanisms As mentioned above, mitochondria contribute to signal transduction by sequestering Ca 2+ and releasing it to the cytosol in a controlled fashion. This is achieved by Ca 2+ influx and efflux pathways, namely the uniporter, PTP, the Na + ⁄ Ca 2+ exchanger, diacylglycerol-sensitive cationic channel(s), and other less-well characterized entities [17–19]. As soon as the extra-mitochondrial [Ca 2+ ] exceeds a so-called set-point value [20], Ca 2+ uptake by the uniporter, driven by the large electrical potential difference (DW m = )150 to )220 mV), results in accumulation of Ca 2+ in the matrix. The molecular identity of the mitochondrial Ca 2+ uniporter is still unknown. A recent study suggested uncoupling proteins 2 and 3 as possible components of the uni- porter [21], but this has been challenged [22] (for the response of Graier’s laboratory, see [23]). Even though the uniporter has been patch-clamped [24], current state-of-the-art methodologies stop short in the mole- cular identification of a single protein that can be captured at the tip of a glass electrode. The mitochondrial matrix can accumulate Ca 2+ up to 1 m (free and bound); however, the free [Ca 2+ ]in this compartment ([Ca 2+ ] m ) is within the low micro- molar range [25]. This results in maintenance of only a minor chemical gradient for Ca 2+ across the inner mitochondrial membrane [26]. Substantially higher transient [Ca 2+ ] m elevations have been reported using matrix-targeted aequorins [27]; however, the calibra- tion techniques used were subsequently criticized [28]. Mitochondrial Ca 2+ uptake has recently been described as relying exclusively on electrochemical diffu- sion [29]; however, this is inaccurate. According to the model by Gunter and Sheu [29], mitochondrial Ca 2+ uptake should be essentially null if DW m is more positive than )120 mV, a prediction that has repeatedly been experimentally rejected. The accumulation of Ca 2+ by mitochondria takes precedence over oxidative phos- phorylation [30], and persists even when mitochondria are severely depolarized [31] (see below). A kinetic model of the uniporter that fits the majority of experi- mental results on mitochondrial Ca 2+ uptake has been developed recently [32]. This model assumes a six-state catalytic binding and Eyring’s free-energy barrier the- ory-based transformation mechanisms, associated with carrier-mediated facilitated transport and electro-diffu- sion. The results of this modeling and how the model fits experimental data on Ca 2+ uptake in isolated respiring mitochondria from rat liver are shown in Fig. 1A. In sufficiently energized mitochondria, the release of sequestered Ca 2+ is thermodynamically unfavorable, but it is possible via the Na + ⁄ Ca 2+ exchanger, or due to the concerted action of the H + ⁄ Ca 2+ and H + ⁄ Na + exchangers [33]. The activity of the Na + ⁄ Ca 2+ exchanger in mitochondria was documented 36 years ago [34], but its molecular identity remained unknown until recently [35]. A mitochondrial H + ⁄ Ca 2+ exchan- ger termed Letm1 has also been identified recently, but its role in Ca 2+ extrusion is not clear [36]. The identity of the H + ⁄ Na + exchanger is still unknown. The thermodynamic equilibrium of the mitochondrial Na + ⁄ Ca 2+ exchanger for cytosolic [Na + ] as a function of DW m has been derived in [37] and is shown in Fig. 1B (black line). At a given DW m , the mitochondrial Na + ⁄ Ca 2+ exchanger operates in forward mode, i.e. it brings Ca 2+ into the matrix in exchange for Na + , i.e. when the concentration of extra-mitochondrial Na + falls below the black line within the dark-grey area. The mitochondrial Na + ⁄ Ca 2+ exchanger operates in reverse mode at a given DW m , i.e. brings Ca 2+ out of the matrix in exchange for Na + , when the concentration of extra-mitochondrial Na + falls above the black line within the light-grey area. Based on this model, it is immediately apparent that > 36 mm Na + need to be added to well-polarized mitochondria (i.e. DW m = )170 mV) in order to induce release of matrix Ca 2+ by reversal of the mitochondrial Na + ⁄ Ca 2+ exchanger. Without such an increase in cytosolic [Na + ], transient losses of DW m are essential for the release of sequestered Ca 2+ . Such transient losses of DW m have Mitochondrial Ca 2+ precipitation C. Chinopoulos and V. Adam-Vizi 3638 FEBS Journal 277 (2010) 3637–3651 ª 2010 The Authors Journal compilation ª 2010 FEBS been described in cells, and are attributed to low-con- ductance pore openings [38–40]. The effect of uncouplers on Ca 2+ release from mitochondria As discussed below, sequestered Ca 2+ complexes with phosphate as well as other molecules including pro- teins and ribose [41], forming an insoluble precipi- tate that is in equilibrium with a soluble pool of Ca 2+ -phosphate complex [25]. In the absence of Na + , dissipation of DW m is a prerequisite for Ca 2+ release; however, this is not sufficient in itself [31]. Addition of an uncoupler of oxidative phosphorylation, or any compound that forms a pore in the inner mitochon- drial membrane, is also required to release sequestered Ca 2+ . The ability of a pore-opening substance to affect Ca 2+ release is evident; however, the effect of an uncoupler requires further clarification. The ability of the uncoupler 2,4-dinitrophenol to cause simulta- neous discharge of Ca 2+ and phosphate in the medium from Ca 2+ -loaded mitochondria was reported 46 years ago [42], but the first attempt to explain this phenome- non was made in 2003 [25]. These authors offered the explanation that acidification results in dissolution of the Ca 2+ -phosphate complex, allowing free Ca 2+ to leave the matrix, provided that DW m is suffi- ciently diminished. A crucial aspect of the sequence of events is that the concentration of the PO 3À 4 species, which is required for complexation of Ca 2+ ,is dependent on the third power of DpH at constant external phosphate [26]. However, the DpH across the inner mitochondrial membrane is inversely related to the amount of P i in the medium [43–47], and DpH is in the range 0.11–0.15 in the presence of abundant P i [31] (see Fig. 2A). DpH also remains relatively constant within a range of pH o [47] (Fig. 2B). If matrix acidification does indeed underlie dissolution of the Ca 2+ -phosphate complex and Ca 2+ release by uncouplers, then complete depo- larization by combined inhibition of the respiratory chain plus reversal of the F 0 F 1 -ATPase would have a similar effect in acidic pH. By the same token, com- plete depolarization by uncoupling in alkaline pH should impede dissolution of this complex, thus impairing the release of sequestered Ca 2+ . The experi- mental findings [31] do not support the above expecta- tions, implying that matrix acidification by uncouplers cannot be the sole explanation for the release of sequestered Ca 2+ . Nonetheless, the rationale of Chal- mers and Nicholls is viable, but to what extent? The calculations used to estimate the intra-mitochondrial free [Ca 2+ ] increase upon pH change caused by pro- tonophores rely on a hypothetical DpH of 1, prior to addition of an uncoupler [31]. However, at DpH <0.15, the uncoupler still releases all the sequestered Ca 2+ from mitochondria. Using the standard analyti- cal calculations described previously [31], the amount of Ca 2+ that can be released is predicted, but not pro- ven, to be a very minor fraction of the total. In order to measure experimentally the amount of Ca 2+ that can be released due to a collapse of DpH, the effect of nigericin, a H + ⁄ K + antiporter eliminating the DpH in high-K + ionic strength medium must be compared to that of an uncoupler in Ca 2+ -loaded mitochondria with no DW m . Such an experiment is shown in Fig. 2C. In curve ’a’, addition of nigericin to mitochondria completely depolarized by stigmatellin and oligomycin, A B Mito NCX forward 3Na + 3Na + Ca 2+ Ca 2+ Matrix Matrix Mito NCX reverse Fig. 1. (A) Fitting of the Ca 2+ uniporter model (lines) [32] to experi- mental data on Ca 2+ uptake in isolated respiring mitochondria from rat liver. Reprinted from [32] with permission from Elsevier. (B) Thermodynamic equilibrium of the mitochondrial Na + ⁄ Ca 2+ exchan- ger. The curve was calculated using an exchange ratio of 3, and the following concentrations were assumed: Ca 2þ c =54nM [37], Ca 2þ m =1lM [25] and Na þ m = 11.4 mM [37] using the resting cyto- solic concentration from [37]. This figure was used with permission from Dr Akos A. Gerencser, Buck Institute, Novato, CA, USA. C. Chinopoulos and V. Adam-Vizi Mitochondrial Ca 2+ precipitation FEBS Journal 277 (2010) 3637–3651 ª 2010 The Authors Journal compilation ª 2010 FEBS 3639 caused the release of $ 20% of the total Ca 2+ previ- ously taken up. Under these conditions (both DW m and DpH are essentially zero), subsequent addition of uncoupler still causes complete loss of sequestered Ca 2+ from mitochondria (see also curve ‘b’). There is no plausible explanation for this phenomenon based on current understanding of the chemiosmotic theory. Interpretation of the results described above may ben- efit from the observations by Kristian et al. [48] showing that a large fraction of sequestered Ca 2+ is retained even after complete depolarization by uncouplers or induction of the PTP in isolated [48–52] and in situ [53,54] mitochondria. Other studies have supported the view that sequestered Ca 2+ is released from various matrical Ca 2+ pools, implying matrical micro-compart- mentation that could promote selective Ca 2+ release [55–58]. An interaction of Ca 2+ with pyridine nucleo- tides in non-polar environments has also been proposed [59–63]. However, these reports appeared before the rec- ognition of PTP and its regulation by the mitochondrial redox state [64,65], and the release of sequestered Ca 2+ by oxidation of the NADH pool could be attributed pri- marily to opening of the PTP. Finally, it is worth men- tioning that Ca 2+ exhibits the ability to form a complex with carboxylic acids, many of which are abundant in the mitochondrial matrix [66]. The concentrations of carboxylic acids (several of which are substrates ⁄ prod- ucts of the citric acid cycle) fluctuate, and they exhibit unequal affinities for Ca 2+ , so it is difficult to predict the matrix [Ca 2+ ] free at any given time. The finding that stigmatellin plus oligomycin induced robust Ca 2+ release at pH o = 7.8, in the absence of measurable changes in light scattering, deserves further attention [31]. As there were no changes in the light scatter recordings, we were poised to accept that Ca 2+ is released through the uniporter. To this end, it is to be noted that the uniporter is sub- ject to activation ⁄ inactivation [67–70], a phenomenon that is poorly characterized. Experimental conditions that promote matrix alkalinization significantly reduce uniporter inactivation, whereas acidification allows A B C Fig. 2. (A) The influence of DpH (given as Dlog [acetate]) and DW m (given as Dlog[Rb + ]) on the internal and external ATP ⁄ ADP ratio as a function of added P i . The DpH decreases between 0.8 and 0.1 on increasing P i to 10 mM, but DW m remains largely constant. Rep- rinted from [43] with permission from Wiley-Blackwell. (B) Correla- tion of matrix pH to the pH of the experimental medium, before and after collapse of DW m by the uncoupler SF 6847. Reprinted from [47] with permission from Elsevier. (C) Reconstructed time courses of extra-mitochondrial [Ca 2+ ], calculated from calcium green 5N fluorescence. Mitochondria were added at 50 s, followed by addition of 10 l M oligomycin at 285 s and 20 lM CaCl 2 at 300 s, and additions of 1.25 n M stigmatellin (stigm) as indicated by the arrows, for both traces. After the 8th addition of stigmatellin, mito- chondria were completely depolarized (not shown). Nigericin (5 l M, trace ‘a’) or SF 6847 (100 n M, trace ‘b’) were added at 750 s, and 150 n M SF 6847 was subsequently added to both traces at 1100 s. Mitochondrial Ca 2+ precipitation C. Chinopoulos and V. Adam-Vizi 3640 FEBS Journal 277 (2010) 3637–3651 ª 2010 The Authors Journal compilation ª 2010 FEBS the uniporter to conduct Ca 2+ less readily [71]. As a Ca 2+ -selective channel, it is not surprising that the uniporter is gated by protons, something which is widely recognized for many Ca 2+ -selective channels [72–74]. It is also known that the F 1 F 0 -ATPase com- plex is an important source of protons for inactivation of the uniporter [71]. Proton coupling between the F 1 F 0 -ATPase and the uniporter channel could account for the fact that matrix Ca 2+ is released the presence of stigmatellin and oligomycin at pH o = 7.8. The collapse of DW m , combined with inhibition of the F 1 F 0 -ATP synthase complex by oligomycin in the pres- ence of a strongly alkaline matrix, could underlie the de-inhibition of the uniporter, allowing matrix Ca 2+ to be released. Thermodynamic aspects of the effect of exogenously added adenine nucleotides on mitochondrial Ca 2+ uptake capacity As discussed above, the presence of adenine nucleotides increases the Ca 2+ accumulation capacity of mitochon- dria. This has been explained previously as the result of two effects: (a) adenine nucleotides decrease the thresh- old for induction of PTP and therefore allow greater amounts of Ca 2+ to be sequestered [75], and (b) adenine nucleotides participate in formation of the matrix Ca 2+ -phosphate precipitates, thus increasing the amount of Ca 2+ that can be retained [41,48,76]. The effect of adenine nucleotides is thought to be mediated by binding to either an atractylate-sensitive site, i.e. the adenine nucleotide translocase, or an atractylate-insensitive site, the identity of which is still unknown. Research on the atractylate-insensitive site has yielded scarce, moderately conflicting, but important information. The presence of an additional ADP-binding component other than adenine nucleotide translocase (ANT) has been suggested [77,78]. ADP has also been shown to exert an effect on the permeability transition by interaction at two binding sites, one that is carboxyatractyloside-sensitive, most likely ANT, and another that shows low affinity for adenine nucleotides but is insensitive to atractylates [79]. In another study, addition of 2 mm ADP after 5 lm carboxyatractyloside did not change respiration rates in mouse liver mito- chondria, but increased mitochondrial Ca 2+ uptake capacity more than fivefold [80]. In this study, the K i for PTP inhibition in Mg 2+ -free medium was estimated to be 0.9 mm ADP. Using de-energized rat liver mito- chondria, Halestrap calculated a K i for PTP inhibition as low as 0.025 mm ADP, but this was in the presence of Mg 2+ in the medium [85]. In another study, Mg 2+ and ADP were found to close the pore in rat liver mito- chondria in a carboxyatractyloside-insensitive manner [81]. Furthermore, the presence of a low-affinity ADP- binding (K m 77 lm) non-carboxyatractyloside binding site appeared to confer increased sensitivity to cyclo- sporin A [82–84]. This low-affinity ADP-binding site was suggested to reside on the matrix side of the inner mitochondrial membrane. However, Mg 2+ was impli- cated in modulation of PTP opening by binding to a site located on the outer part of the inner mitochondrial membrane [46]. However, in other studies on heart mito- chondria, ADP was found to be ineffective [85–87]. In rat liver mitochondria, an ADP binding site other than ANT has been reported to confer resistance to PTP opening, with a K i of 0.07 mm [83]. ATP was shown to delay PTP opening, but only at 0.3 mm concentration and in the presence of cyclosporin A. In hamster brown adipose tissue mitochondria, an atractylate-insensitive site located on the outer face of the inner mitochondrial membrane with an affinity for purine nucleotides (ADP and GDP) binding and a capacity of 0.7 nmolÆmg )1 pro- tein has been reported [88]. GDP was found to compete with ADP. The affinity constant was dependent on pH, that for GDP being 4.2 lm at pH 6.7 and 34 lm at pH 7.9. However, no such site was found in rat liver [88]. In beef heart and pig heart mitochondria, ADP was found to inhibit Ca 2+ -induced PTP opening in the presence of carboxyatractyloside [84], and ADP was more potent than ATP. The EC 50 for ADP was estimated as 0.09 mm and that for ATP was estimated as 0.18 mm. If ATP uptake were a requirement for an increase in maximum Ca 2+ uptake capacity, ATP hydrolysis by F 0 F 1 -ATPase (provided that mitochondrial membrane potential is sufficiently decreased, see below) should provide inorganic phosphate for formation of precipi- tates; however, in the presence of ATP, inorganic phosphate was found not to be absolutely critical for Ca 2+ uptake [89,90], and the presence of oligomycin did not alter these outcomes. DeLuca and Engstrom [90] reported that inorganic phosphate is not necessary in the presence of ATP, while Vasington and Murphy [89], using similar conditions, showed that omission of inorganic phosphate decreased the amount of Ca 2+ taken up by 40–70%. However, experiments investigat- ing the effect of exogenously added adenine nucleo- tides on Ca 2+ -phosphate precipitation in the matrix must be evaluated by electron microscopy, not by following changes in light scattering. This is because the inner mitochondrial membrane is known to con- tract in response to addition of adenine nucleotides, an effect that is mediated by ANT, creating changes in mitochondrial optical density, but this is unrelated to matrix precipitate formation [77,78]. C. Chinopoulos and V. Adam-Vizi Mitochondrial Ca 2+ precipitation FEBS Journal 277 (2010) 3637–3651 ª 2010 The Authors Journal compilation ª 2010 FEBS 3641 The above reports indicate that there is a site on mitochondria other than ANT that increases maximum Ca 2+ uptake capacity, and that it binds adenine nucle- otides with an affinity lower than that of ANT. How- ever, there is no consensus on whether this site is located on the outer or the inner leaflet of the inner mitochondrial membrane in mitochondria of the vari- ous tissue types, whether it binds both ADP (or GDP) and ATP and with what affinities, and what the role of Mg 2+ is in this binding, if any. Additionally, there is still no information regarding the identity of this atractylate-insensitive adenine nucleotide binding site. With regard to the uptake of adenine nucleotides through ANT as a prerequisite for operation of the atractylate-sensitive site, many reports seem to disre- gard the influence of DW m on the directionality of ANT [91–93]. ANT operates in forward mode, i.e. it brings ADP into the matrix in exchange for ATP, if DW m is more negative than $ )100 mV, and works in reverse, i.e. brings ATP into the matrix in exchange for ADP, if DW m is more positive than $ )100 mV [47,94–96]. A typical ADP–ATP exchange rate–DW m profile for isolated rat liver mitochondria is shown in Fig. 3A. Mitochondria expel ATP in exchange for ADP if their membrane potential is in the range from )145 to )100 mV, but consume extra-mitochondrial ATP at more positive DW m values. The membrane potential value at which there is no net transfer of ADP–ATP across the inner mitochondrial membrane is the reversal potential of ANT (E rev_ANT ). By ther- modynamic deduction, E rev_ANT is given by: where ‘out’ indicates outside the matrix, ‘in’ indicates inside the matrix, R is the universal gas constant (8.31 JÆmol )1 ÆK )1 ), F is the Faraday constant (9.64 · 10 4 CÆmol )1 ) and T is the temperature (in Kelvin) [94]. There is no doubt that both ADP and ATP accumulate in mitochondria during Ca 2+ loading [97], but the question is how to reconcile the entry of both adenine nucleotide species with the mutually exclu- sive forward and reverse modes of ANT operation. To explain this, we show the results of superimposed time courses of DW m and extra-mitochondrial Ca 2+ during multiple additions of CaCl 2 to rat brain mito- chondria in the presence of 3 mm ATP and 0.8 mm ADP. It is a widely acknowledged, but, to the best of our knowledge, unreferenced concept, that measuring maximum Ca 2+ uptake capacity by monitoring DW m using a potential-sensitive fluorescent probe yields lower values than if a Ca 2+ -sensitive probe is used that is distributed in the extra-mitochondrial space. There is only a single report showing that safranine O, a fluorescent probe that responds to changes in DW m , A C B D Fig. 3. (A) Plot of the ATP–ADP exchange rate mediated by ANT versus DW m in isolated rat liver mitochondria depolarized to various voltages by various amounts of the uncoupler SF 6847. (B) Combined traces of DW m (gray line) and extra-mitochondrial Ca 2+ (black line) during stepwise additions of 20 l M CaCl 2 to isolated rat brain mitochon- dria in high-K + ionic strength medium sup- plemented with 3 m M ATP, 0.8 mM ADP and 2 m M MgCl 2 . (C) Lead-contrasted image of a mitochondrion visualized by electron microscopy, sampled after time point ‘c’ as shown in (B) (magnification: · 50 000). (D) Lead-contrasted image of a mitochondrion visualized by electron microscopy, sampled from the time interval ‘a’–‘b’ as shown in (B) (magnification · 50 000). E rev ANT ¼ 2:3RT F  log  ½ADP 3À  free out ½ATP 4À  free in 0 ½ADP 3À  free in ½ATP 4À  free out ! Mitochondrial Ca 2+ precipitation C. Chinopoulos and V. Adam-Vizi 3642 FEBS Journal 277 (2010) 3637–3651 ª 2010 The Authors Journal compilation ª 2010 FEBS decreases maximum Ca 2+ uptake capacity if used at a concentration above 5 lm [98]. If safranine O is used properly, i.e. at 2.5 lm (as for the results shown in Fig. 3B), it does not affect mitochondrial bioenergetics. As shown in Fig. 3B, mitochondria accumulate nine pulses of 20 lm CaCl 2 without any alterations in the baseline of DW m (up to point ‘a’). From point ‘a’ to point ‘b’, mitochondria accumulate ten additional equimolar CaCl 2 pulses with an apparently undimin- ished avidity, but DW m starts to become more positive. The reasons for the gradual decrease in DW m upon excessive Ca 2+ accumulation is beyond the scope of this review, but may include inhibition of a-ketogluta- rate dehydrogenase [99] and complex I [100] [101] by high amounts of Ca 2+ . In the interval between points ‘a’ and ‘b’, DW m remains in the range within which ANT operates in the forward mode, and thus mito- chondria are able to take up ADP in exchange for ATP. However, this does not apply for F 0 F 1 -ATPase, as the reversal potential of this complex (E rev_ATPase )is more negative than that of ANT [94]. By thermody- namic deduction, E rev_ATPase is given by: and ½P À  in ¼½P total  in  1 þ 10 pH i ÀpK a2 ÀÁ where ‘o’ or ‘out’ indicate outside the matrix, ‘i’ or ‘in’ indicateinside the matrix, n is the H + ⁄ ATP coupling ratio, R is the universal gas constant (8.31 JÆmol )1 ÆK )1 ), F is the Faraday constant (9.64 · 10 4 C mol )1 ), T is the temperature (in Kelvin), [P ) ] is the free phosphate concentration in Molar, and pK a2 = 7.2 for phosphoric acid [94]. The reversal of F 0 F 1 -ATPase due to a high rate of Ca 2+ uptake has been known for 35 years [102]. However, under the conditions described above, a paradox emerges in which the activ- ity of F 0 F 1 -ATPase is reversed, thereby consuming mitochondrial ATP, but ANT is still operating in the forward mode [94], and therefore extra-mitochondrial ATP cannot be provided for the F 0 F 1 -ATPase. Six additional equimolar Ca 2+ pulses are sequestered by the mitochondria after point ‘b’ (although the uptake rate starts to decrease), and after point ‘b’, DW m attains values at which ANT also reverses, thereby allowing extra-mitochondrial ATP to enter the matrix. At point ‘c’, mitochondria are completely depolarized, although they are still able to remove extra-mitochon- drial Ca 2+ , albeit at a rapidly deteriorating uptake rate. These results indicate that when mitochondria are challenged with sufficiently high amounts of Ca 2+ , DW m will eventually decrease to a degree that allows reversal of ANT and import of extra-mitochondrial ATP into the matrix. The fate of adenine nucleotides that have been take up with respect to formation of the precipitates is discussed below. It is worth empha- sizing the ability of mitochondria with no measurable membrane potential to remove extra-mitochondrial Ca 2+ . Mitochondria sampled for electron microscopic inspection from such an experiment from point ‘c’ onwards exhibit classic signs of PTP. Many of them appear as shown in Fig. 3C, broken open but with pre- cipitates in them. Mitochondria fixed before point ‘b’ and visualized by electron microscopy appear as shown in Fig. 3D. The property of mitochondria to undergo permeability transition but retain Ca 2+ -phosphate pre- cipitates has been reported previously [48]. However, we propose that broken mitochondria might chelate exogenously added Ca 2+ , as exogenously added CaCl 2 has unobstructed access to the precipitation machinery in broken mitochondria. That could account for the paradox that extra-mitochondrial Ca 2+ is sequestered by mitochondria with no membrane potential that exhibit obvious signs of PTP opening (from point ‘c’ until the beginning of Ca 2+ release). What is the nature of the mitochondrial precipitates formed upon Ca 2+ uptake? Precipitation of Ca 2+ during various pathological conditions has been observed in several intracellular locations, including mitochondria [10,11,103,104]. Within isolated or in situ mitochondria, the precipi- tates take the form of granules that almost always show electron-transparent cores [42], often in associa- tion with the inner membranes [42]. This pattern is similar even if the divalent cation is Ba 2+ or Sr 2+ (see Fig. 4A). The electron opacity of the rim of these granules does not depend upon heavy-metal staining but is intrinsic [76]. However, in isolated mitochon- dria from certain tissues (such as rabbit heart) that sequester large amounts of Ca 2+ in the absence of Mg 2+ , the precipitates appear as needles instead (Fig. 4B,C) [105]. The granules formed in the mitochondrial matrix upon Ca 2+ sequestration contain an inorganic and an organic moiety [41,76]. Depending on the method of E rev ATPase ¼Àð316=nÞÀ  2:3RT F=n   log ½ATP 4À  free in 0 ½ADP 3À  free in ½P À  in ! À 2:3RT F ÂðpH o À pH i Þ C. Chinopoulos and V. Adam-Vizi Mitochondrial Ca 2+ precipitation FEBS Journal 277 (2010) 3637–3651 ª 2010 The Authors Journal compilation ª 2010 FEBS 3643 granule isolation, the organic moiety accounts for 16–60% of the total (Fig. 4D). The organic moiety appears to occupy the electron-transparent core, while the inorganic moiety is located in the electron-dense rim. The physico-chemical properties of the granules imply an intimate association of the organic with the inorganic constituents [76]. The inorganic moiety is rich in Ca 2+ ,P i ,Mg 2+ and CO 2À 3 ,corresponding to hydroxy- apatite Ca 10 (PO 4 ) 6 (OH) 2 , whitelockite Ca 3 (PO 4 ) 2 ,or a mixture of both, as well as traces of MgO, presumably derived from MgCO 3 [41]. Subsequent studies have concluded that the hydroxyapatite present in the inor- ganic moiety of the granules is Ca 2+ -deficient [76]. The composition of the constituents of the inorganic moiety can be manipulated by the rate of Ca 2+ infusion to isolated mitochondria [48], and also by endogenous factors (see below). This may be due to alleviation against bursts of metabolic compensations, as moni- tored by alterations in state 4 respiration [25,106]. The organic moiety contains nitrogen and tests positive in a biuret test, indicating the presence of protein(s), and also for ribose, implying the presence of RNA [41]. Most of the P i in the inorganic moiety found in the granules originates from the medium, and only 16% of the initial specific activity of labeled ATP is found in the granules [41]. However, during loading of mitochondria with Ca 2+ , there is an unexplained anion deficit that cannot be fully accounted for by complex- ation to phosphate [107]. Complexation of Ca 2+ and its salts to the organic moiety may account for this anion deficit. However, as noted previously [25], there is an addi- tional puzzle regarding Ca 2+ -phosphate complexation, namely reconciliation of the apparent properties of the matrix Ca 2+ -phosphate complex with those of known complexes in solution. ‘Physiological’ incuba- tion media for cells contain millimolar [Ca 2+ ] in the presence of millimolar [P i ]; furthermore, in experi- ments with isolated mitochondria, CaCl 2 is added in the submillimolar concentration range in media that also contain millimolar amounts of P i , and yet an osmotically inactive complex forms in the matrix when [Ca 2+ ] m rises above 1–5 lm [25,26]. Why is there no formation of precipitates outside the matrix? At least one study has shown precipitate formation outside the matrix, but it used the pyroantimonate technique [108], the validity of which is disputed [48]. An obvious conclusion is that ‘the mitochondrial matrix is perhaps as far from an ideal solution as it is possible to imagine’ [25]. Perhaps the organic moi- ety serves the purpose of a ‘scaffold’ upon which Ca 2+ precipitates with phosphate, given the low con- centration of the former compared to the latter. If a protein exists that plays this scaffolding role, it would be of great value to know its identity. However, the mystery of Ca 2+ -phosphate precipita- tion in the mitochondrial matrix is only one side of the coin. The other is the reason(s) behind the lack of transition of calcium phosphate deposits to hydroxyap- atite, best exemplified by the late Albert Lehninger, as ‘why we do not all turn into stone’ [109]. Theories of bone [110] and teeth [111] calcification that implicated A C B D Fig. 4. (A) Smooth muscle cell from a toad urinary bladder incubated for 6 h in calcium- free Ringer’s solution containing 2 m M bar- ium acetate, showing dense intramitochond- rial granules, most of which appear hollow (magnification: · 210 000). Reprinted from [2] with permission from Rockefeller Univer- sity Press. (B,C) Rabbit cardiac mitochondria fixed after active Ca 2+ uptake in the pres- ence (B) and absence (C) of Mg 2+ . Rep- rinted from [135]. (D) Effect of incineration temperature on the fine structure of dense granule residues isolated from formalde- hyde-fixed, calcium phosphate-loaded mito- chondria from rat liver (magnification: · 63 000). The granule residues are bubble- like; the granule mass appears to fuse at high incineration temperature and bubbles are formed as the organic component vapor- izes. Reprinted from [76] with permission from Rockefeller University Press. Mitochondrial Ca 2+ precipitation C. Chinopoulos and V. Adam-Vizi 3644 FEBS Journal 277 (2010) 3637–3651 ª 2010 The Authors Journal compilation ª 2010 FEBS mitochondria enjoyed wide attention until the begin- ning of the 1980s [109,112–115], when it was realized that ossification and enamel-forming mechanisms are separate from the calcification processes that occur within the mitochondrial matrix [116]. However, because of these studies, the concept of ‘calcific dis- eases’ emerged; these included major maladies of our times, such as arthritis, atherosclerosis, urolithiasis, calcific valvular sclerosis and tumor calcification [117– 119]. Research on these ‘calcific diseases’ yielded dis- covery of a factor (originally termed ‘Howard factor’) with the ability to prevent calcification in buffered solutions containing Ca 2+ and phosphate, preventing the formation of the hydroxyapatite lattice [109,120]. This factor, which is normally present in urine, blood, milk and saliva, is absent from individuals who suffer from repeated calcium oxalate stones in their kidneys [109]. Unexpectedly, Becker et al. found that rat liver mitochondria also contain a substance that inhibits precipitation of calcium phosphate and its conversion to hydroxyapatite [109]. Subsequent chromatographic analysis, mass spectrometry and proton NMR identi- fied this calcification inhibitor, which present in body fluids and in the mitochondrial matrix, as phosphoci- tric acid [120–122]. Phosphocitrate was quickly realized to be the most potent inhibitor of hydroxyapatite crys- tal growth [120,123] (Fig. 5A,B). Other naturally occurring substances found in mitochondria are also known to inhibit hydroxyapatite formation, such as inorganic pyrophosphate [124], ATP and Mg 2+ [125], although these are far less potent than phosphocitrate [120]. Another endogenous substance that disrupts hydroxyapatite formation is inorganic polyphosphate. Polyphosphate is a polymer comprising as few as ten to several hundred phosphate molecules linked by ATP-like high-energy bonds, and has been found in all eukaryotic organisms tested, localized in various com- partments, including the mitochondria [126]. Polyphos- phate has strong links to the mitochondrial Ca 2+ sequestration system: (a) it is implicated in composi- tion of the ion-conducting module of the PTP [127,128], (b) reduction of polyphosphate levels increases mitochondrial Ca 2+ uptake capacity and decreases the probability of pore opening [129], (c) it is a chelator of Ca 2+ , among other divalent ions [129], and (d) it inhibits calcium hydroxypatite crystal growth [130] (Fig. 5C,D). Furthermore, polyphosphate levels and mitochondrial bioenergetic parameters are recipro- cally regulated [131,132], and parameters such as the mitochondrial membrane potential and ATP produc- tion by the F 0 F 1 -ATPase are important determinants of mitochondrial Ca 2+ uptake capacity [31]. Phospho- citrate [133] and polyphosphate occur naturally within mitochondria, however, it is not known how they are produced [126]; phosphocitrate can easily be chemically synthesized [134]. cm mm mm A B C D Fig. 5. (A,B) Electron microscopic study of the effects of phosphocitrate on parathyroid hormone-induced nephrocalcinosis. Thin sections from mice treated with phosphocitrate and parathyroid hormone or saline and parathyroid hormone for 4 days were fixed and stained with uranyl acetate and lead citrate (magnification: · 12 400). The control (saline ⁄ parathyroid hormone) sections (A) contained numerous heavily mineralized mitochondria (mm) as well as extensive areas of calcification within the cytoplasm (cm). Mineral deposits were not observed in sections from those animals given phosphocitrate prior to parathyroid hormone (B). Adapted from [122]. (C,D) The appearance of freshly pre- cipitated calcium orthophosphate in solution (magnification: · 190 000) (C) and freshly precipitated calcium orthophosphate inhibited by poly- phosphate in solution (magnification: · 190 000) (D). Adapted from [130]. C. Chinopoulos and V. Adam-Vizi Mitochondrial Ca 2+ precipitation FEBS Journal 277 (2010) 3637–3651 ª 2010 The Authors Journal compilation ª 2010 FEBS 3645 Conclusions In this review, we have re-examined the thermo- dynamic aspects of mitochondrial Ca 2+ uptake and release mechanisms, processes that have been investi- gated for decades and still generate a vast amount of literature. The major ‘players’ in Ca 2+ uptake and release mechanisms are still unknown: the molecular identities of the uniporter and PTP are unknown, and the identities of the Na + ⁄ Ca 2+ exchanger and possi- bly the Ca 2+ ⁄ H + antiporter have only been revealed very recently. The action of uncouplers in induction of mitochondrial Ca 2+ release also remains inadequately explained, although it is now accepted that the effect on matrix acidification accounts for only a fifth of the total amount of Ca 2+ that can be released. Further- more, the thermodynamic aspects of the role of adenine nucleotides in mitochondrial Ca 2+ uptake capacity and precipitate formation have been exam- ined, and placed under the perspective of the direc- tionality of ANT operation. The possible existence of an atractylate-insensitive, adenine nucleotide binding site that modulates Ca 2+ uptake capacity is re- appraised. Finally, the nature of the Ca 2+ -phosphate precipitates formed in the mitochondrial matrix has been re-addressed, and it will be insightful to deter- mine the composition of the organic moiety, and to resurrect the concepts on the origin and regulation of the endogenous Ca 2+ -P i hydroxyapatite lattice break- ers, such as phosphocitrate and polyphosphate. The present review provides more questions than answers, but the key to fruitful research is to ask the right question! Acknowledgements We thank Dr Akos A. Gerencser, Buck Institute, Novato, CA, USA for generating the model of the mitochondrial Na + ⁄ Ca 2+ exchanger. The work by our group referred to in the text was supported by grants from Orsza ´ gos Tudoma ´ nyos Kutata ´ si Alapprogram (OTKA), Magyar Tudoma ´ nyos Akade ´ mia (MTA), Nemzeti Kutata ´ si e ´ s Technolo ´ giai Hivatal (NKTH) and Egeszsegu ¨ gyi Tudoma ´ nyos Tana ´ cs (ETT) to V.A V., and by OTKA-NKTH grant number NF68294 and OTKA grant number NNF78905 to C.C. References 1 Rossi CS & Lehninger AL (1963) Stoichiometric rela- tionships between accumulation of ions by mitochon- dria and the energy-coupling sites in the respiratory chain. Biochem Z 338, 698–713. 2 Peachey LD (1964) Electron microscopic observations on the accumulation of divalent cations in intramitoc- hondrial granules. J Cell Biol 20, 95–111. 3 Brustovetsky N, Brustovetsky T, Purl KJ, Capano M, Crompton M & Dubinsky JM (2003) Increased susceptibility of striatal mitochondria to calcium- induced permeability transition. J Neurosci 23, 4858–4867. 4 Damiano M, Starkov AA, Petri S, Kipiani K, Kiaei M, Mattiazzi M, Flint BM & Manfredi G (2006) Neural mitochondrial Ca 2+ capacity impairment precedes the onset of motor symptoms in G93A Cu ⁄ Zn-superoxide dismutase mutant mice. J Neurochem 96, 1349–1361. 5 LaFrance R, Brustovetsky N, Sherburne C, Delong D & Dubinsky JM (2005) Age-related changes in regional brain mitochondria from Fischer 344 rats. Aging Cell 4, 139–145. 6 Brustovetsky N, LaFrance R, Purl KJ, Brustovetsky T, Keene CD, Low WC & Dubinsky JM (2005) Age- dependent changes in the calcium sensitivity of striatal mitochondria in mouse models of Huntington’s disease. J Neurochem 93, 1361–1370. 7 Walsh C, Barrow S, Voronina S, Chvanov M, Petersen OH & Tepikin A (2009) Modulation of calcium signalling by mitochondria. Biochim Biophys Acta 1787, 1374–1382. 8 Panov AV, Gutekunst CA, Leavitt BR, Hayden MR, Burke JR, Strittmatter WJ & Greenamyre JT (2002) Early mitochondrial calcium defects in Huntington’s disease are a direct effect of polyglutamines. Nat Neurosci 5, 731–736. 9 Chinopoulos C & Adam-Vizi V (2006) Calcium, mitochondria and oxidative stress in neuronal pathology. Novel aspects of an enduring theme. FEBS J 273, 433–450. 10 Buchs PA, Stoppini L, Parducz A, Siklos L & Muller D (1994) A new cytochemical method for the ultrastructural localization of calcium in the central nervous system. J Neurosci Methods 54, 83– 93. 11 Gajkowska B & Mossakowski MJ (1992) Calcium accumulation in synapses of the rat hippocampus after cerebral ischemia. Neuropatol Pol 30, 111–125. 12 Takeyama Y, Ozawa K & Katagiri T (1980) Studies on the subcellular localization of electrolytes in normal and infarcted canine myocardium. With special reference to calcium ion. Jpn Heart J 21, 859–872. 13 Reynolds ES (1963) Liver parenchymal cell injury. I. Initial alterations of the cell following poisoning with carbon tetrachloride. J Cell Biol 19, 139–157. 14 Halestrap AP (2006) Calcium, mitochondria and reperfusion injury: a pore way to die. Biochem Soc Trans 34 232–237. Mitochondrial Ca 2+ precipitation C. Chinopoulos and V. Adam-Vizi 3646 FEBS Journal 277 (2010) 3637–3651 ª 2010 The Authors Journal compilation ª 2010 FEBS [...].. .Mitochondrial Ca2+ precipitation C Chinopoulos and V Adam-Vizi 15 Pivovarova NB & Andrews SB (2010) Calciumdependent mitochondrial function and dysfunction in neurons FEBS J 277, 3622–3636 16 Starkov AA (2010) The molecular identity of the mitochondrial Ca2+ sequestration system FEBS J 277, 3652–3663 17 Chinopoulos C, Starkov AA,... Long-term modulation of mitochondrial Ca2+ signals by protein kinase C isozymes J Cell Biol 165, 223–232 Moreau B, Nelson C & Parekh AB (2006) Biphasic regulation of mitochondrial Ca2+ uptake by cytosolic Ca2+ concentration Curr Biol 16, 1672–1677 FEBS Journal 277 (2010) 3637–3651 ª 2010 The Authors Journal compilation ª 2010 FEBS Mitochondrial Ca2+ precipitation C Chinopoulos and V Adam-Vizi 71 Moreau... of some carboxylic acids derived from d-glucose and d-fructose Can J Chem 73, 1338–1347 Collins TJ, Lipp P, Berridge MJ & Bootman MD (2001) Mitochondrial Ca2+ uptake depends on the spatial and temporal profile of cytosolic Ca2+ signals J Biol Chem 276, 26411–26420 Maechler P, Kennedy ED, Wang H & Wollheim CB (1998) Desensitization of mitochondrial Ca2+ and insulin secretion responses in the beta cell... compilation ª 2010 FEBS 3649 Mitochondrial Ca2+ precipitation C Chinopoulos and V Adam-Vizi 99 Lai JC & Cooper AJ (1986) Brain alpha-ketoglutarate dehydrogenase complex: kinetic properties, regional distribution, and effects of inhibitors J Neurochem 47, 1376–1386 100 Sadek HA, Szweda PA & Szweda LI (2004) Modulation of mitochondrial complex I activity by reversible Ca2+ and NADH mediated superoxide... Transient mitochondrial depolarizations reflect focal sarcoplasmic reticular calcium release in single rat cardiomyocytes J Cell Biol 142, 975–988 O’Reilly CM, Fogarty KE, Drummond RM, Tuft RA & Walsh JV Jr (2003) Quantitative analysis of spontaneous mitochondrial depolarizations Biophys J 85, 3350–3357 Gerencser AA & Adam-Vizi V (2005) Mitochondrial Ca2+ dynamics reveals limited intramitochondrial Ca2+. .. Journal 277 (2010) 3637–3651 ª 2010 The Authors Journal compilation ª 2010 FEBS 3647 Mitochondrial Ca2+ precipitation C Chinopoulos and V Adam-Vizi 44 Chance B & Mela L (1966) Hydrogen ion concentration changes in mitochondrial membranes J Biol Chem 241, 4588–4599 45 Petronilli V, Cola C & Bernardi P (1993) Modulation of the mitochondrial cyclosporin A-sensitive permeability transition pore II The minimal... by cyclosporin A and FK506 following ischemia corresponds with differing abilities to inhibit calcineurin and the mitochondrial permeability transition Neurobiol Dis 10, 219–233 Vinogradov A, Scarpa A & Chance B (1972) Calcium and pyridine nucleotide interaction in mitochondrial membranes Arch Biochem Biophys 152, 646–654 Lehninger AL, Vercesi A & Bababunmi EA (1978) Regulation of Ca2+ release from... activate mitochondrial cationic channel(s) and release sequestered Ca2+ J Bioenerg Biomembr 37, 237–247 18 Gunter TE & Pfeiffer DR (1990) Mechanisms by which mitochondria transport calcium Am J Physiol 258, C755–C786 19 Bernardi P (1999) Mitochondrial transport of cations: channels, exchangers, and permeability transition Physiol Rev 79, 1127–1155 20 Nicholls DG (1978) The regulation of extramitochondrial... proteins 2 and 3 are fundamental for mitochondrial Ca2+ uniport Nat Cell Biol 9, 445–452 22 Brookes PS, Parker N, Buckingham JA, Vidal-Puig A, Halestrap AP, Gunter TE, Nicholls DG, Bernardi P, Lemasters JJ & Brand MD (2008) UCPs – unlikely calcium porters Nat Cell Biol 10, 1235–1237 23 Trenker M, Fertschai I, Malli R & Graier WF (2008) UCP2 ⁄ 3 – likely to be fundamental for mitochondrial Ca2+ uniport... to Ca2+ already in the submicromolar range: correlation with redox state Cell Calcium 31, 97–104 29 Gunter TE & Sheu SS (2009) Characteristics and possible functions of mitochondrial Ca2+ transport mechanisms Biochim Biophys Acta 1787, 1291– 1308 30 Rossi CS & Lehninger AL (1964) Stoichiometry of respiratory stimulation, accumulation of Ca++ and 31 32 33 34 35 36 37 38 39 40 41 42 43 phosphate, and . MINIREVIEW Mitochondrial Ca 2+ sequestration and precipitation revisited Christos Chinopoulos and Vera Adam-Vizi Department of Medical. between mitochondrial Ca 2+ handling and signal transduction and cell death pathways [9–13], it is imperative to identify the mecha- nism(s) that define mitochondrial