Chromaffin-cell stimulation triggers fast millimolar mitochondrial Ca2+ transients that modulate secretion

© 1999 Macmillan Magazines Ltd articles NATURE CELL BIOLOGY | VOL 2 | FEBRUARY 2000 | cellbio.nature.com 57 Chromaffin-cell stimulation triggers fast millimolar mitochondrial Ca2+ transients that modulate secretion Mayte Montero*, Maria Teresa Alonso*, Estela Carnicero*, Inmaculada Cuchillo-Ibáñez†, Almudena Albillos†, Antonio G. García†, Javier García-Sancho* and Javier Alvarez*‡ *Instituto de Biología y Genética Molecular (IBGM), Universidad de Valladolid y Consejo Superior de Investigaciones Científicas (CSIC), Departamento de Bioquímica y Biología Molecular y Fisiología, Facultad de Medicina, Ramón y Cajal 7, E-47005 Valladolid, Spain †Instituto de Farmacología Teófilo Hernando, Departamento de Farmacología y Terapéutica, Facultad de Medicina, Universidad Autónoma de Madrid, Arzobispo Morcillo 4, E-28029 Madrid, Spain ‡e-mail: [email protected] Activation of calcium-ion (Ca2+) channels on the plasma membrane and on intracellular Ca2+ stores, such as the endoplasmic reticulum, generates local transient increases in the cytosolic Ca2+ concentration that induce Ca2+ uptake by neighbouring mitochondria. Here, by using mitochondrially targeted aequorin proteins with different Ca2+ affinities, we show that half of the chromaffin-cell mitochondria exhibit surprisingly rapid millimolar Ca2+ transients upon stimulation of cells with acetylcholine, caffeine or high concentrations of potassium ions. Our results show a tight functional coupling of voltage-dependent Ca2+ channels on the plasma membrane, ryanodine receptors on the endoplasmic reticulum, and mitochondria. Cell stimulation generates localized Ca2+ transients, with Ca2+ concentrations above 20–40 µM, at these functional units. Protonophores abolish mitochondrial Ca2+ uptake and increase stimulated secretion of catecholamines by threeto fivefold. These results indicate that mitochondria modulate secretion by controlling the availability of Ca2+ for exocytosis. uring cell activation, some mitochondria take up Ca2+ from cytosolic microdomains of high Ca2+ concentration that are generated by activation of nearby Ca2+ channels1–3. Changes in the mitochondrial Ca2+ concentration ([Ca2+]M) are therefore heterogeneous at the subcellular level, and the [Ca2+]M transient in a particular mitochondrion depends on its proximity to a Ca2+ source1–3. Increases in [Ca2+]M activate dehydrogenases2,4, coupling ATP production to energy demands. In addition, increasing amounts of evidence indicate that mitochondria may also modulate homeostasis of the cytosolic Ca2+ concentration ([Ca2+]c). For example, in neurons and chromaffin cells, mitochondria act as rapid and reversible Ca2+ buffers during cell stimulation5–9 and in the clearance of large Ca2+ loads10. Moreover, overloading of mitochondria with Ca2+ may be involved in the development of several pathological conditions, including ischaemia-reperfusion lesions, neurotoxicity and neurodegenerative diseases, where ATP depletion, overproduction of oxidative radicals and release of apoptotic factors lead to cell damage11–14. However, [Ca2+]M measurements during cell stimulation, obtained by a variety of different techniques, have provided values only in the low micromolar range1–4,9. Although these changes are enough to activate mitochondrial metabolism, their physiological significance for cellular Ca2+ homeostasis remains unclear. Similarly, the actual relevance of mitochondrial Ca2+ overload is obscure. We show here that [Ca2+]M can in fact reach the millimolar range during stimulation of chromaffin cells, and that this large mitochondrial Ca2+ uptake regulates the availability of Ca2+ for the secretory machinery. Results We used three types of mitochondrially targeted aequorin with different Ca2+ sensitivities. This allowed us to measure [Ca2+]M at a wide range of concentrations, from the submicromolar to the millimolar. Wildtype aequorin reconstituted with native coelenterazine (AEQ1) shows affinity for Ca2+ in the concentration range 0.3–8 µM; wild-type aequorin reconstituted with coelenterazine n (AEQ2) covers the range 1–40 µM Ca2+; and mutated low-Ca2+-affinity aequorin15 reconstituted with coelenterazine n (AEQ3) can measure Ca2+ in the concentration range 20 µM to 1 mM16. A further decrease in the Ca2+ sensitivity of aequorin was obtained by using AEQ3 at 22 °C (ref. 17). Another property of aequorin, its irreversible consumption as it emits light, enables it to be used in detecting subcellular heterogeneities in [Ca2+] (ref. 18). Aequorin with high Ca2+ affinity is rapidly consumed in areas of high [Ca2+]. After that, measurements reflect only the behaviour of low-[Ca2+] areas. Instead, when low-Ca2+-affinity aequorin is used, only regions with high [Ca2+] contribute significantly to the luminescence. In each case, the relative fraction of aequorin consumed provides an estimation of the size of the different [Ca2+] pools. Heterogeneity of the [Ca2+]M response. With regard to their response to [Ca2+], we found two main populations of mitochondria in chromaffin cells. Figure 1a, b show similar experiments performed with either AEQ1 or AEQ2. Cell depolarization using a medium with a high potassium-ion (K+) concentration triggered a rapid consumption of 40–50% aequorin in both cases (dotted line). Calibrated in [Ca2+] (continuous line), we estimated peaks of 5 µM and 30 µM Ca2+ when using AEQ1 and AEQ2, respectively. Subsequent depolarizations produced smaller [Ca2+]M peaks of similar magnitude (about 2 µM) when calculated using both types of aequorin. Figure 1c, d show similar experiments but using caffeine instead of high [K+]. Caffeine activates ryanodine receptors (Ca2+ channels found on the endoplasmic reticulum), inducing Ca2+ release from the endoplasmic reticulum19. The first addition of caffeine resulted in the consumption of ~30% of AEQ1 or AEQ2, a value that converts to estimated [Ca2+]M peaks of 5 µM and 40 µM, respectively. Subsequent additions of caffeine produced much smaller [Ca2+]M peaks that gave similar calibrated results with both aequorins ([Ca2+]M 2 µM). The [Ca2+]c response to the same pattern of stimuli was very different, with consecutive additions of high K+ concentrations or caffeine producing similar [Ca2+]c peaks19. The discrepancy can be explained only if ~50% of chromaffin-cell mitochondria respond to stimulation with 2 µM [Ca2+]M peaks while the rest undergo a much larger increase in [Ca2+]M. Aequorin in the latter pool of mitochondria is consumed during the first stimulus, and subsequent stimuli evoke only the smaller response from the other mitochondria. In contrast to HeLa D © 1999 Macmillan Magazines Ltd articles 58 NATURE CELL BIOLOGY | VOL 2 | FEBRUARY 2000 | cellbio.nature.com cells, for which extensive mitochondrial lumenal diffusion has been reported20, large [Ca2+]M peaks were not restored in chromaffin cells after 20 min without stimulation (Fig. 1). Co-localization of Ca2+-entry and -release sites. We used the pattern of AEQ2 consumption during the consecutive addition of high K+ levels and caffeine to study the degree of overlapping among the mitochondrial pools that respond to each stimulus with large [Ca2+]M peaks. Caffeine was unable to induce any large [Ca2+]M response after treatment of cells with high [K+]; however, caffeine applied before K+ resulted in the consumption of ~60% of the aequorin that was able to respond to high [K+] with large [Ca2+]M peaks (Fig. 1e, f). If we consider the amount of AEQ2 consumption after two consecutive pulses of high [K+] and caffeine to be 100%, then the initial pulse of high [K+] resulted in the consumption of 96.5 ± 0.4% (mean ± s.e.m.; n = 10) of this aequorin. However, if caffeine were added before K+, the initial caffeine pulse resulted in the consumption of only 59 ± 2% of AEQ2 (n = 16). We conclude that about 30% of mitochondria co-localize with both plasma-membrane Ca2+ channels (which are activated by high [K+]) and ryanodine receptors (which are activated by caffeine). An additional 20% of mitochondria co-localize with plasma-membrane Ca2+ channels, but not with ryanodine receptors. The remaining 50% of the mitochondria are apparently located far away from these Ca2+ hotspots. Millimolar [Ca2+]M transients. Peak [Ca2+]M values calculated using AEQ2 (30–40 µM) are underestimated because the aequorin is consumed. Quantification of the real values of the large [Ca2+]M peaks required the use of the lower-Ca2+-affinity AEQ3. Figure 2a shows the [Ca2+]M changes induced by a series of high-[K+] pulses. The initial [Ca2+]M peak reached 200 µM, and subsequent peaks decreased in magnitude as soon as aequorin consumption approached 50–60%. These values are still underestimates, because large [Ca2+]M changes take place in only half of the mitochondria. Taking this fact into account (Fig. 2b), we estimate [Ca2+]M peaks of similar magnitude for equal repetitive stimuli, with peaks being ~300 µM (340 ± 20 µM, n = 15) for 10-s pulses and ~700 µM (670 ± 30 µM, n = 11) for 30-s pulses. Figure 2c shows the effect of caffeine, which produced a first [Ca2+]M peak of 200 µM and then decreasing peaks as consumption approached 40%. Again, if data are recalibrated considering only the active mitochondrial space (Fig. 2d), we obtain a series of similar fast [Ca2+]M transients reaching peak values of near 500 µM (430 ± 30 µM, n = 8). The effect of caffeine resulted from Ca2+ release from intracellular stores, because it produced the same [Ca2+]M transient in Ca2+-free medium (data not shown).The physiological agonist acetylcholine also produced a large [Ca2+]M peak (280 ± 50 µM, n = 4; Fig. 2e). Finally, the high-[K+]-induced [Ca2+]M transient was reversibly abolished by the protonophore carbonyl cyanide p-trifluoromethoxy phenyl hydrazone (FCCP) (Fig. 2f). Return of [Ca2+]M to resting levels after stimulation was fast, reaching background values within 1–2 min. CGP37157, an inhibitor of the mitochondrial Na+/Ca2+ exchanger, reduced the rate of [Ca2+]M decrease by 60% (half-time 62 ± 3 s, n = 6; control, half-time 25 ± 1 s, n = 9; Fig. 3a). The [Ca2+]M and [Ca2+]c transients induced by high [K+] follow very similar time courses (Fig. 3b). This means that [Ca2+]M is able to follow rapid oscillations in [Ca2+]c, but at concentrations more than two orders of magnitude higher. Ca2+ exit from mitochondria and Ca2+ uptake by mitochondria differed in sensitivity to temperature. The rate of [Ca2+]M increase was very similar at 22 °C (67 ± 2 µM s–1, n = 17) and 37 °C (58 ± 3 µM s–1, n = 15). The Figure 1 Effects of high K+ concentrations or caffeine on [Ca2+]M, measured with AEQ1 or AEQ2. The effects of 10-s pulses with medium containing 70 mM K+ (K+) and/or 30-s pulses with 50 mM caffeine (C) on [Ca2+]M (continuous line) and the aequorin consumption (values shown are the percentage of aequorin consumed; dotted line), measured either with AEQ1 (a, c) or with AEQ2 (b, d–f). The total amount of aequorin consumed was measured at the end of every experiment by perfusion of 100 µM digitonin plus 10 mM Ca2+. The breaks in a–d indicate 15 min without stimulation. The temperature was 37 °C. aAEQ1 5 min 0 1 2 3 4 5 0 20 40 60 80 100 K+ K+ K+K+K+K+K+ 0 10 20 30 5 min K+ K+K+K+K+K+K+ 0 20 40 60 80 100 bAEQ2 2 4 65 min 0 20 40 60 80 100 [Ca ] (µM) 2+ M Percentage consumed C CCCCCC 5 min 00 20 40 60 80 100 [Ca ] (µM) 2+ M Percentage consumed 10 20 30 40 C CCCC CC 00 20 40 60 80 100 5 min CCC 10 20 30 K+K+ C C 0 10 20 30 0 20 40 60 80 100 5 min K+CK+ K+ cd ef AEQ1 AEQ2 AEQ2 AEQ2 K+ Figure 2 Effects of high K+ concentrations or caffeine on [Ca2+]M measured with AEQ3. a, c, The effects of 10-s or 30-s pulses with a high K+ concentration (K+) or 30-s pulses with 50 mM caffeine (C) on [Ca2+]M, measured with AEQ3 (continuous line), and on the percentage of aequorin consumption (dotted line). b, d, Results of experiments that were the same as those in a, c but recalibrated by assuming that [Ca2+]M changes occurred in only 58% and 48%, respectively, of the total space. e, The effect of a 10-s pulse with 10 mM acetylcholine (ACh) followed by 30-s K+ pulses. f, 2 µM FCCP and 10-s K+ pulses were perfused as indicated. Temperature was 37 °C. CCC CCCC C CCCC a c e b d f 5 min 5 min 5 min 5 min 5 min 5 min 0 50 100 150 200 0 20 40 60 80 100 K+ K+ K+ K+ K+ K+ 00 50 100 100 80 60 40 20 150 200 250 [Ca ] (µM) 2+ M 0 100 200 300 400 500 [Ca ] (µM) 2+ M 0 200 400 600 800 10 s K+ 10 s 0 200 400 600 ACh 0 200 100 300 400 K+ 30 s K+ 30 s K+ 30 s K+ 30 s K+ 10 s K+ 10 s FCCP K+ 10 s K+ 10 s K+ 10 s K+ 10 s 10 s 10 s 30 s 30 s K+ 30 s K+ 30 sK+ 30 s 30 s Percentage consumed © 1999 Macmillan Magazines Ltd articles NATURE CELL BIOLOGY | VOL 2 | FEBRUARY 2000 | cellbio.nature.com 59 kinetics of the [Ca2+]c transient were also similar at both temperatures (Fig. 3c). However, the rate of Ca2+ exit from mitochondria was nearly fourfold slower at 22 °C than at 37 °C (half-times 92 ± 4 s, n = 11, and 25 ± 1 s, n = 9, at 22 °C and 37 °C, respectively). Therefore, at 22 °C, the [Ca2+]M transient lasted for much longer than the cytosolic one. The height of the [Ca2+]M peaks induced by the stimuli was also larger at 22 °C than at 37 °C (10-s high-[K+] pulse, [Ca2+]M peak = 550 ± 30 µM, n = 10; 30-s high-[K+] pulse, [Ca2+]M peak = 870 ± 30 µM, n = 8; caffeine, [Ca2+]M peak = 520 ± 50 µM, n = 4; acetylcholine, [Ca2+]M peak = 670 ± 40 µM, n = 6). These findings are consistent with the reported ability of mitochondria to prolong the [Ca2+]c peaks by sequestering Ca2+ during the stimulation and releasing it afterwards, in experiments performed at room temperature5–9. Our results indicate that the ability of mitochondria to prolong [Ca2+]c transients should be much less at 37 °C. Figure 3c also shows the inhibition by CGP37157 of mitochondrial Ca2+ release at 22 °C (half-time 380 ± 60s, n = 3; 76% inhibition). Mitochondria as biosensors of local [Ca2+]c. The large [Ca2+]M peaks that we have observed can be generated only if mitochondria are exposed to very high cytosolic Ca2+ concentrations. As the rate of Ca2+ uptake by mitochondria is proportional to [Ca2+]c (refs 1, 10, 21, 22), we can estimate the size of the local [Ca2+]c transients by comparing the rates of [Ca2+]M increase observed in intact cells with those obtained in permeabilized cells perfused with known Ca2+ concentrations (Fig. 4a). A similar approach has been used in permeabilized RBL mucosal mast cells to show that inositol-1,4,5-trisphosphate produces local [Ca2+]c transients above 16 µM (ref. 21). We found that the consumption of aequorin by permeabilized cells exposed to Ca2+ buffers followed a monotonic rate, as would be expected for a single mitochondrial pool with a homogeneous rate of Ca2+ uptake. This result indicates that the heterogeneous responses shown in Fig. 1 are due to the subcellular localization of each mitochondrial pool and not to differences in Ca2+-uptake properties. The relationship between the rate of [Ca2+]M increase and [Ca2+]c is plotted in Fig. 4b. The rate of [Ca2+]M increase was negligible below 4 µM [Ca2+]c, but increased steeply above these values and was fully blocked by ruthenium red, a blocker of the mitochondrial Ca2+ uniporter. Fitted values for the maximal rate of uptake and for the [Ca2+]c that gives a halfmaximal uptake rate were 225 µM s–1 and 43 µM, respectively. This KM value is consistent with previous measurements made in chromaffin cells10,22. The rate of increase of [Ca2+]M obtained in cells stimulated with high [K+] corresponded to ~20 µM [Ca2+]c. The rates of [Ca2+]M increase observed in intact cells stimulated with caffeine or acetylcholine were, respectively, 110 ± 9 µM s–1 (n = 11) and 94 ± 7 µM s–1 (n = 6), which correspond to cytosolic Ca2+ concentrations of 40 µM and 30 µM, respectively. In those mitochondria undergoing small (1–2 µM) [Ca2+]M transients on stimulation (Fig. 1), the rate of high-[K+]-induced [Ca2+]M increase was only 0.29 ± 0.02 µM s–1 (n = 35). This rate corresponds to a [Ca2+]c of 2–3 µM, close to the mean [Ca2+]c values measured using Fura-2 (Fig. 3b). Figure 3 Effects of CGP37157, an inhibitor of the mitochondrial Na+/Ca2+ exchanger, and temperature on the rate of [Ca2+]M decrease after stimulation. a, Control cells and cells incubated at 37 °C for 2 min with 20 µM CGP37157 (CGP) were stimulated with a 10-s pulse of high [K+] (K+). The inhibitor was also present after the stimulus. b, Comparison of the kinetics of the [Ca2+]M (continuous line) and [Ca2+]c (dotted line) transients at 37 °C after stimulation with a 10-s pulse of high [K+]. c, 10-s pulses of high [K+] were given at 22 °C in the presence or absence of 20 µM CGP37157, as indicated. The kinetics of the [Ca2+]c transient at 22 °C is also shown (dotted line). Experiments were performed using AEQ3 and calibrated using the assumption that [Ca2+]M changes occurred in only 55% of the total space. 2 min [Ca ] (µM) 2+ M[Ca ] (µM) 2+ M [Ca ] (µM) 2+ M [Ca ] (µM) 2+ M 800 400 300 200 100 0Control 22 C 600 400 200 0K+ K+ K+ K+ 400 300 200 100 0 0 0.5 1 1.5 2 CGP CGP [Ca ] (µM) 2+ c 0 0.5 1 1.5 2 o 37 C o ab c Figure 4 Dependence of the rate of mitochondrial Ca2+ uptake on [Ca2+]c. a, An increase in [Ca2+]M was induced by perfusion of permeabilized cells with buffers of known Ca2+ concentration, as indicated. In the curve labelled ‘20 µM+RR’, the buffer containing 20 µM Ca2+ was added in the presence of 4 µM ruthenium red. These experiments were carried out using AEQ3 at 22 °C, to increase the sensitivity of the measurements at high uptake rates. b, Plot of the maximum rates of uptake obtained at every Ca2+ concentration. Data at 2 µM and 4 µM [Ca2+]c (rates of uptake 0.2 and 0.45 µM s–1, respectively) were obtained using AEQ2. Error bars show s.e.m. of four determinations made for each point. 0 20 40 60 80 100 0 50 100 150 200 a b 0 500 1,000 1,500 60 µM Uptake rate (µM s ) –1 [Ca ] (µM) 2+ M [Ca ] (µM) 2+ c 100 µM 40 µM 20 µM 10 µM 7 µM 20 µM + RR 10 s © 1999 Macmillan Magazines Ltd articles 60 NATURE CELL BIOLOGY | VOL 2 | FEBRUARY 2000 | cellbio.nature.com Modulation of secretion by mitochondria. We analysed the physiological relevance of mitochondrial Ca2+ uptake by comparing catecholamine secretion induced by stimulation with high [K+], caffeine or acetylcholine, both in the presence and in the absence of the protonophore carbonyl cyanide m-chlorophenyl-hydrazone (CCCP). We expected that CCCP, by preventing Ca2+ sequestration by mitochondria, would allow a larger local increase in subplasmalemmal [Ca2+]c and hence potentiate secretion. This was the case (Fig. 5). In response to cellular challenge with acetylcholine (100 µM for 2 s, at 2min intervals), fairly reproducible secretory peaks of around 300 nA were produced; CCCP (2 µM) reversibly increased the secretory response to ~1,000 nA (Fig. 5a). Similar results were obtained when cells were challenged with high [K+] (70 mM K+, 2 s; Fig. 5b) or caffeine (50 mM, 2 s; Fig. 5c). Figure 5d shows the averaged total secretion obtained from ten repetitions of each of these experiments. The small effect on secretion of CCCP alone disappeared when the interval between stimuli was increased to 5 min, but potentiation of secretion remained unchanged. Potentiation of secretion by CCCP was not due to increased Ca2+ uptake from extracellular sources. Experiments in which the uptake of 45Ca2+ was measured showed that CCCP did not increase the net Ca2+ uptake induced by any of the stimuli (data not shown). Neither was the effect of CCCP mediated by a possible Ca2+ release from the endoplasmic reticulum. When cells were treated with a combination of 10 mM caffeine, 10 µM ryanodine and 1 µM thapsigargin in Ca2+-free medium for 30 min, caffeine produced no secretory response, but the potentiation by CCCP of K+-induced catecholamine secretion was unmodified (data not shown). Finally, mitochondrial inhibitors produced the same effects as CCCP. We stimulated cells with 2-s high-[K+] pulses every 2 min. Oligomycin (3 µM, preincubation 90 s), which blocks ATP production but should not decrease the mitochondrial membrane potential, had little effect on K+-stimulated secretion. The mean normalized peak secretion was 100 ± 12% in control pulses, and 139 ± 21% in the presence of oligomycin (mean ± s.e.m, n = 11; P>0.1, Student’s ttest). However, a combination of 4 µM rotenone (which blocks respiration) and 3 µM oligomycin produced a large potentiation of high-K+-induced secretion. The mean normalized peak secretion in control pulses was 100 ± 6%, and it increased to 204 ± 21% (n = 11; P<0.001, t-test) in the presence of oligomycin plus rotenone. Discussion It was astonishing to find that some mitochondria can undergo, in stimulated intact cells, rapid reversible increases in Ca2+ concentration from the submicromolar to the millimolar range. Such high mitochondrial Ca2+ concentrations were previously associated only with pathological phenomena related to cell damage or the initial stages of apoptosis10–13. These [Ca2+]M values contrast with those obtained in a similar preparation using the fluorescent dye rhod2 (ref. 9). The discrepancy can be explained by differences in the specificity of targeting and range of measurement of the probes. Co-localization of those mitochondria undergoing large [Ca2+]M transients with both plasma-membrane Ca2+ channels and ryanodine receptors reveals the presence of functional units of Ca2+-induced Figure 5 CCCP enhances secretory responses evoked by acetylcholine, high K+ concentration or caffeine. a–c, Cells were superfused with Krebs–HEPES medium. After a 10-min equilibration period, to get a stable baseline secretion, cells were challenged at 2-min intervals with 2-s pulses of a, 100 µM acetylcholine (ACh), b, medium with a high K+ concentration (70 mM, with iso-osmotic reduction of Na+), or c, caffeine (50 mM). The dots at the bottom of the panels each represent one challenge with one of these secretagogues. In addition, 2 µM CCCP was applied in two phases, each of 90 s, as indicated by the bars. Graphs represent original traces from three typical experiments using pulses of acetylcholine (a), high [K+] (b) or caffeine (c). The amplitudes of secretion peaks are expressed as oxidation currents (measured in nA). d, Averaged results obtained from ten repetitions each of the experiments described in a–c. Total secretion was calculated by integration of the areas of the secretory peaks, and is expressed as µC (µA × s). a CCCP CCCP 200 nA 4 min ACh b 120 nA High K + High K + d Secretion (µC) 0 2 4 6 8 1210 864 20 ACh Caffeine CCCP CCCP Peak number c Caffeine CCCP CCCP 4 min 120 nA 4 min CCCP CCCP Figure 6 The complex functional unit responsible for the generation of local high [Ca2+]c transients and catecholamine secretion in a bovine chromaffin cell. The voltage-dependent Ca2+ channel (VDCC), the Ca2+ uniporter (U) of the upper mitochondria, the ryanodine receptor (RyR) from the endoplasmic reticulum (ER) and the secretory vesicle are all strategically located close together beneath the plasma membrane. Ca2+ entry through VDCCs triggers Ca2+-induced Ca2+-release from the ER through RyRs and generates a local [Ca2+]c hotspot of about 50 µM. In the rest of the cell, the [Ca2+]c increase is much less, the resulting [Ca2+]c being 1–2 µM. Mitochondria take up Ca2+ through the uniporter, with uptake depending on the surrounding [Ca2+]c. The uniporter located close to the [Ca2+]c hotspot in this diagram will take up Ca2+ very quickly and the mitochondrial [Ca2+] will reach near millimolar levels. This uptake effectively reduces the amount of Ca2+ available for the secretory vesicle, allowing a tight control of local [Ca2+]c and hence of the secretory response. Ca2+ is then released through the Na+/Ca2+ exchanger. The other mitochondrion, placed far away from the region of high [Ca2+]c, increases its [Ca2+]M to only about 2 µM. [Ca2+]ER, the ER Ca2+ concentration; [Ca2+]o, the extracellular Ca2+ concentration. VDCC Secretory vesicle RyR Endoplasmic Reticulum 50 µM Mitochondrion [Ca ] 2+ 2+ M [Ca ] 2+ o [Ca ] 2+ c 1 µM [Ca ] 2+ c 500 µM [Ca ] 2+ M 2 µM [Ca ] 2+ ER 500 µM 1,000 µM U U ~ Ca Na+ 2+ Ca Na+ ~ ~ © 1999 Macmillan Magazines Ltd articles NATURE CELL BIOLOGY | VOL 2 | FEBRUARY 2000 | cellbio.nature.com 61 Ca2+ release (CICR). Our data indicate that most of the ryanodine receptors in chromaffin cells are either located within these CICR functional units or have no mitochondria placed near them. A precise spatial relationship between ryanodine receptors and mitochondria occurs only at these functional CICR units. The large size of the mitochondrial pool close to the Ca2+-entry and -release sites (50%) indicated that it might be important for rapid buffering of Ca2+ in local subplasmalemmal regions. We confirmed this by showing that catecholamine secretion induced by several stimuli depended markedly on mitochondrial Ca2+ uptake. When this uptake was blocked, either by collapsing the mitochondrial transmembrane H+ gradient with CCCP or by blocking respiration with rotenone and preventing reversal of the mitochondrial ATPase with oligomycin, the secretory response increased several-fold. These results indicate that mitochondrial Ca2+ uptake controls the increase of [Ca2+]c at exocytic sites. In addition, blockade of mitochondrial Ca2+ uptake could allow spreading of the Ca2+ wave to deeper cytosolic areas, thus favouring mobilization of the reserve pool of secretory vesicles23. In any case, our results support the idea that chromaffin cells possess complex functional units (Fig. 6) that cluster together all the elements needed to control the subplasmalemmal [Ca2+]c transients near exocytic sites. The interplay between Ca2+-channel activation, mitochondrial Ca2+ uptake and CICR will thus modulate the rate and extent of exocytosis and vesicle transport to exocytic sites. We also used mitochondria as biosensors to quantify the localized [Ca2+]c transients generated during cell stimulation. We found that about 50% of cell mitochondria become exposed to a cytosolic Ca2+ concentration of 20–40 µM. These values are consistent with the upper reported estimates of the local [Ca2+]c reached during physiological stimulation24. We cannot estimate the size of these hotspots. However, half of the mitochondria in the same cells detected only 2–3 µM [Ca2+]c during stimulation, indicating the presence of a steep subcellular [Ca2+]c gradient. Finally, we speculate that if either the location or the Ca2+- uptake properties of those mitochondria located in the functional units could be modulated, this would be an effective mechanism by which to regulate secretion. If similar phenomena were applicable to neurons, it would be also a very simple mechanism by which to generate synaptic plasticity. Under pathological conditions or during ageing, mitochondrial damage may reduce the ability of mitochondria to take up Ca2+. This would lead to increased secretion of neurotransmitters and increased cell activation, a vicious cycle that may trigger processes leading to necrosis or apoptosis. h Methods Cell culture and expression of aequorin. Bovine adrenal medulla chromaffin cells were obtained as described19 and used within 1–3 days. The mitochondrial aequorin1,2 was a gift from T. Pozzan. Mutated (Asp 119 → Ala) mitochondrial aequorin was obtained by replacing, in-frame, complementary DNA encoding the wild-type aequorin with that encoding the mutated aequorin15. Expression in chromaffin cells was achieved by infecting the cells with a defective herpes simplex virus type 1 containing the mitochondrial aequorin gene (pHSVmitAEQ). Virus packaging and titring have been described25. Chromaffin-cell cultures (5 × 105 cells per 0.5 ml) were routinely infected with 2 × 103 infectious virus units 12–24 h before measurements. FCCP and CCCP were obtained from Sigma. CGP37157 was from Tocris (Bristol, UK). [Ca2+]M and [Ca2+]c measurements. To reconstitute aequorins, cells expressing mitochondrial aequorin were incubated for 1–2 h at room temperature with 1 µM of either wild-type coelenterazine or coelenterazine n, in standard medium containing 145 mM NaCl, 5 mM KCl, 1 mM MgCl2, 1 mM CaCl2, 10 mM glucose, and 10 mM HEPES, pH 7.4. Cells were then placed in the perfusion chamber of a purpose-built thermostatic luminometer at 22 °C or 37 °C. For experiments with permeabilized cells, reconstitution was performed as above and cells were placed in the luminometer. Standard medium containing 0.5 mM EGTA instead of CaCl2 was then perfused for 1 min, followed by a 1-min perfusion of intracellular medium (130 mM KCl, 10 mM NaCl, 1 mM MgCl2, 1 mM K3PO4, 0.2 mM EGTA, 1 mM ATP, 20 µM ADP, 2 mM succinate, 20 mM HEPES, pH 7) containing 20 µM digitonin. Intracellular medium without digitonin was perfused for 1 min, followed by buffers of known [Ca2+], between 2 and 100 µM, prepared using HEDTA/Ca2+/Mg2+ mixtures. Calibration was done using the calibration curves corresponding to each aequorin type. Calibration curves for AEQ3 at both 22 °C and 37 °C have been described16,17. For AEQ1 and AEQ2, new calibration curves at 37 °C were generated as described16. Aequorin consumption was calculated as the integral of the luminescence measured during the experiment, normalized as a percentage. [Ca2+]c measurements were obtained using Fura-2 as described19. Statistical values are given as means ± s.e.m. On-line measurements of catecholamine release. Cells were scraped off carefully from the bottom of the Petri dish with a rubber policeman, and centrifuged at 800 r.p.m. (Heraeus centrifuge) for 10 min. The cell pellet was resuspended in 200 µl KrebsHEPES medium (144 mM NaCl, 5.9 mM KCl, 2 mM CaCl2, 1.2 mM MgCl2, 11 mM glucose, 10 mM HEPES, pH 7.4). Cells were introduced in a jacketed microchamber for superfusion at 37 °C. The superfusion rate was 2 ml min–1. The liquid flowing from the superfusion chamber reached an electrochemical detector (model Metrohn AG CH-9100 Hersau), placed just at the outlet of the microchamber, that monitors ‘on-line’, under the amperometric mode, the amount of catecholamines secreted26. Cells were stimulated to secrete with short pulses (2 s) of Krebs-HEPES solution containing the stimuli. RECEIVED 18 JUNE 1999; REVISED 1 NOVEMBER 1999; ACCEPTED 6 DECEMBER 1999; PUBLISHED 20 DECEMBER 1999. 1. Rizzuto, R., Brini, M., Murgia, M. & Pozzan, T. Microdomains with high Ca2+ close to IP3-sensitive channels that are sensed by neighboring mitochondria. Science 262, 744–747 (1993). 2. Rizzuto, R., Bastianutto, C., Brini, M., Murgia, M. & Pozzan, T. Mitochondrial Ca2+ homeostasis in intact cells. J. Cell Biol. 126, 1183–1194 (1994). 3. Brini, M. et al. Subcellular analysis of Ca2+ homeostasis in primary cultures of skeletal muscle myotubes. Mol. Biol. Cell 8, 129–143 (1997). 4. Robb-Gaspers, L. D. et al. Integrating cytosolic calcium signals into mitochondrial metabolic responses. EMBO J. 17, 4987–5000 (1998). 5. Werth, J. L. & Thayer, S. A. Mitochondria buffer physiological calcium loads in cultured rat dorsal root ganglion neurons. J. Neurosci. 14, 346–356 (1994). 6. White, R. J. & Reynolds, I. J. Mitochondria accumulate Ca2+ following intense glutamate stimulation of cultured rat forebrain neurons. J. Physiol. (Lond.) 498, 31–47 (1997). 7. Park, Y. B., Herrington, J., Babcock, D. F. & Hille, B. Ca2+ clearance mechanisms in isolated rat adrenal chromaffin cells. J. Physiol. (Lond.) 492, 329–346 (1996). 8. Herrington, J., Park, Y. B., Babcock, D. F. & Hille, B. Dominant role of mitochondria in clearance of large Ca2+ loads from rat adrenal chromaffin cells. Neuron 16, 219–228 (1996). 9. Babcock, D. F., Herrington, J., Park, Y.-B. & Hille, B. Mitochondrial participation in the intracellular Ca2+ network. J. Cell. Biol. 136, 833–843 (1997). 10. Xu, T., Naraghi, M., Kang, H. & Neher, E. Kinetic studies of Ca2+ binding and Ca2+ clearance in the cytosol of adrenal chromaffin cells. Biophys. J. 73, 532–545 (1997). 11. Duchen, M. R. Contributions of mitochondria to animal physiology: from homeostatic sensor to calcium signalling and cell death. J. Physiol. (Lond.) 516, 1–17 (1999). 12. Schinder, A. F., Olson, E. C., Spitzer, N. C. & Montal, M. Mitochondrial dysfunction is a primary event in glutamate neurotoxicity. J. Neurosci. 16, 6125–6133 (1996). 13. Di Lisa, F. & Bernardi, P. Mitochondrial function as a determinant of recovery or death in cell response to injury. Mol. Cell Biochem. 184, 379–391 (1998). 14. Green, D. R. & Reed, J. C. Mitochondria and apoptosis. Science 281, 1309–1312 (1998). 15. Montero, M. et al. Monitoring dynamic changes in free Ca2+ concentration in the endoplasmic reticulum of intact cells. EMBO J. 14, 5467–5475 (1995). 16. Montero, M., Barrero, M. J. & Alvarez, J. [Ca2+] microdomains control agonist-induced Ca2+ release in intact HeLa cells. FASEB J. 11, 881–885 (1997). 17. Barrero, M. J., Montero, M. & Alvarez, J. Dynamics of [Ca2+] in the endoplasmic reticulum and cytoplasm of intact HeLa cells. J. Biol. Chem. 272, 27694–27699 (1997). 18. Montero, M. et al. Ca2+ homeostasis in the endoplasmic reticulum: coexistence of high and low [Ca2+] subcompartments in intact HeLa cells. J. Cell Biol. 139, 601–611 (1997). 19. Alonso, M. T. et al. Ca2+-induced Ca2+ release in chromaffin cells seen from inside the ER with targeted aequorin. J. Cell Biol. 144, 241–254 (1999). 20. Rizzuto, R. et al. Close contacts with the endoplasmic reticulum as determinants of mitochondrial Ca2+ responses. Science 280, 1763–1766 (1998). 21. Csordás, G., Thomas, A. P. & Hajnóczky, G. Quasi-synaptic calcium signal transmission between endoplasmic reticulum and mitochondria. EMBO J. 18, 96–108 (1999). 22. Uceda, G., García, A. G., Guantes, J. M., Michelena, P. & Montiel, C. Effects of Ca2+ channel antagonist subtypes on mitochondrial transport. Eur. J. Pharmacol. 289, 73–80 (1995). 23. von Rüden, L. & Neher, E. A Ca-dependent early step in the release of catecholamines from adrenal chromaffin cells. Science 262, 1061–1065 (1993). 24. Neher, E. Vesicle pools and Ca2+ microdomains: new tools for understanding their roles in neurotransmitter release. Neuron 20, 389–399 (1998). 25. Alonso, M. T. et al. Functional measurements of [Ca2+] in the endoplasmic reticulum using a herpes virus to deliver targeted aequorin. Cell Calcium 24, 87–96 (1998). 26. Borges, R., Sala, F. & García, A. G. Continuous monitoring of catecholamine release from perfused cat adrenals. J. Neurosci. Meth. 16, 389–400 (1986). ACKNOWLEDGEMENTS We acknowledge financial support from the Dirección General de Enseñanza Superior (grant PM98/ 0142 to J.A. and grant PB97/0474 to J.G.-S.), from the Dirección General de Investigación Científica y Técnica (grant PB94/0150) and from Janssen-Cilag to A.G.G., and from Junta de Castilla y León (grant VA19/99 to J.A. and grant VA62/99 to M.T.A.). I.C.-I. and A.A. hold fellowships from the Ministerio de Educación y Ciencia. We thank C. González and T. Pozzan for helpful comments, and J. Fernández for technical assistance. Correspondence and requests for materials should be addressed to J.A.