Calcium dynamics in catecholamine-containing secretory vesicles

Cell Calcium 37 (2005) 555–564 Calcium dynamics in catecholamine-containing secretory vesicles Alfredo Morenoa, Carmen D. Lobat´ ona, Jaime SantoDomingoa, Laura Vaya, Esther Hern´ andez-SanMiguela, Rosario Rizzutob, Mayte Monteroa, Javier Alvareza,∗ aInstituto de Biolog´ıa y Gen´etica Molecular (IBGM), Departamento de Bioqu´ımica y Biolog´ıa Molecular y Fisiolog´ıa, Facultad de Medicina, Universidad de Valladolid and Consejo Superior de Investigaciones Cient´ıficas (CSIC), Ram´on y Cajal 7, E-47005 Valladolid, Spain bExperimental and Diagnostic Medicine Section of General Pathology, University of Ferrara, 44100 Ferrara, Italy Received 19 October 2004; received in revised form 17 February 2005; accepted 21 February 2005 Available online 8 April 2005 Abstract Wehaveusedanaequorinchimeratargetedtothemembraneofthesecretorygranulestomonitorthefree[Ca2+]insidetheminneurosecretory PC12 cells. More than 95% of the probe was located in a compartment with an homogeneous [Ca2+] around 40␮M. Cell stimulation with either ATP, caffeine or high-K+depolarization increased cytosolic [Ca2+] and decreased secretory granule [Ca2+] ([Ca2+]SG). Inositol-(1,4,5)- trisphosphate, cyclic ADP ribose and nicotinic acid adenine dinucleotide phosphate were all ineffective to release Ca2+ from the granules. Changes in cytosolic [Na+] (0–140mM) or [Ca2+] (0–10␮M) did not modify either ([Ca2+]SG). Instead, [Ca2+]SG was highly sensitive to changes in the pH gradient between the cytosol and the granules. Both carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone (FCCP) and nigericin, as well as cytosolic acidification, reversibly decreased [Ca2+]SG, while cytosolic alcalinization reversibly increased [Ca2+]SG. These results are consistent with the operation of a H+/Ca2+ antiporter in the vesicular membrane. This antiporter could also mediate the effects of ATP, caffeine and high-K+on [Ca2+]SG, because all of them induced a transient cytosolic acidification. The FCCP-induced decrease in [Ca2+]SG was reversible in 10–15min even in the absence of cytosolic Ca2+ or ATP, suggesting that most of the calcium content of the vesicles isbound toaslowlyexchangingCa2+ buffer. Thislargestorebuffers[Ca2+]SG changesinthelong-term butallowshighly dynamicfree[Ca2+]SG changes to occur in seconds or minutes. © 2005 Elsevier Ltd. All rights reserved. Keywords: Calcium; Secretory vesicles; PC12; Catecholamine; Aequorin 1. Introduction Secretory vesicles constitute one of the less known intracellular Ca2+ pools. It has been known for more than 20 years that they contain large amounts of total calcium (20–60mM), in particular chromaffin granules of the adrenal medulla and zymogen granules of the exocrine pancreas [1–3]. In addition, chromaffin granules contain also large amounts of proteins (100mg/ml, of which 40% is chromoAbbreviations: [Ca2+]SG, secretory granule [Ca2+]; [Ca2+]c, cytosolic [Ca2+]; cADPR, cyclic ADP ribose; EGFP, enhanced green fluorescent protein; FCCP, carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone; InsP3, inositol-(1,4,5)-trisphosphate; NAADP, nicotinic acid adenine dinucleotide phosphate; VAMP, vesicle-associated membrane protein; VAMP-mutaeq, VAMP-mutated aequorin chimera ∗Corresponding author. Tel.: +34 983 423085; fax: +34 983 423588. E-mail address: [email protected]a.es (J. Alvarez). granin A), catecholamines (0.5–0.6M) and ATP (130mM) [2,4]. The physiological significance of intragranular Ca2+ ([Ca2+]SG) within these densely packed vesicles, either to play a role in secretion, to be destined for secretion or just structural to keep granule contents together, is still obscure. Similarly, the mechanisms for Ca2+ uptake and release from the granules are poorly known or controversial and may be different depending on the cell type [5–9]. Recently, using a vesicle-associated membrane protein (VAMP)–aequorin chimera targeted to the membrane of the secretory vesicles, it was possible for the first time to monitortheintragranular[Ca2+]inMIN6␤-cells[8].Intragranular free [Ca2+] was around 50␮M, Ca2+ was accumulated into the vesicles via a P-type Ca2+-ATPase and released through ryanodine receptors and nicotinic acid adenine dinucleotide phosphate (NAADP) receptors [8,9]. We have used here the same VAMP–aequorin chimera to monitor the dynamics of 0143-4160/$ – see front matter © 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.ceca.2005.02.002 556 A. Moreno et al. / Cell Calcium 37 (2005) 555–564 [Ca2+] inthecatecholamine-containingsecretory granules of neurosecretory PC12 cells. Our results show that, although the free [Ca2+] is similar to that found in MIN6 ␤-cells, the mechanisms of Ca2+ uptake and release and the response to extracellular stimuli are largely different. 2. Experimental procedures 2.1. Cell culture and transfection The construction strategy of the 2-synaptobrevin-mutated aequorin (VAMP-mutaeq) chimera has been described previously [8]. PC12 rat pheochromocytoma cells were grown in Dulbecco’s Modified Eagle’s Medium supplemented with 7.5% fetal calf serum, 7.5% horse serum and 2mM glutamine.Theconstructforcytosolicaequorinhasbeenalsodescribed previously [10]. The VAMP-enhanced green fluorescentprotein(EGFP)constructwasmadebyinsertinginframe the VAMP sequence in the HindIII site of the pEGFP-N1 plasmid. Transfections were carried out using Metafectene (Biontex, Germany) or ExGen500 (Fermentas, Germany). 2.2. [Ca2+]Mand [Ca2+]cmeasurements with aequorin PC12 cells plated onto 13mm plastic round coverslips were transiently transfected with the corresponding plasmid. For aequorin reconstitution, PC12 cells expressing cytosolic aequorin were incubated for 1–2h at room temperature with 1␮M of wild-type coelenterazine, and cells expressing VAMP-mutaeq were incubated for 1–2h at room temperature with 1␮M of coelenterazine n, in standard medium (145mM NaCl, 5mM KCl, 1mM MgCl2, 1mM CaCl2, 10mM glucose and 10mM Hepes, pH 7.4). Cells were then placed in the perfusion chamber of a purpose-built luminometer thermostatized at 37◦C. For experiments with permeabilized cells, cells expressing VAMP-mutaeq were reconstituted with 1␮M coelenterazine n for 1–2h and then placed in the luminometer as above. Then, standard medium containing 0.5mM EGTA instead of Ca2+ was perfused for 1min, followed by 1min of intracellular medium (130mM KCl, 10mM NaCl, 1mM MgCl2, 1mM K3PO4, 0.5mM EGTA, 1mM ATP, 20␮M ADP, 2mM succinate and 20mM Hepes,pH7)containing20␮Mdigitonin.Then,intracellular medium without digitonin and containing 100nM [Ca2+] (prepared with an EGTA/Ca2+ mixture) was perfused. Control experiments performed in PC12 cells expressing mitochondrially targeted mutated aequorin showed that this permeabilization procedure rendered all the mitochondrial space accessible to the perfusion medium (data not shown). Calibration of the luminescence data into [Ca2+] was made using an algorithm as previously described [11]. 2.3. Confocal measurements PC12 cells expressing VAMP-EGFP were imaged on a BioRad Radiance 2000 confocal spectrophotometer using a 60×oil immersion objective mounted in a Nikon Eclipse TE2000microscope.Cellswerealsoloadedwith50nMlysotrackerredfor1minatroomtemperatureandimagedsequentially for both EGFP and lysotracker red. EGFP was excited with the 488nm line of the Argon laser, and the fluorescence emitted between 500 and 530nm was collected. In the case of lysotracker red, the fluorescence excited with the 543nm line of the green He–Ne laser and emitted between 570 and 700nm was measured. 2.4. Measurement of cytosolic pH pH measurements were carried out by microfluorimetry of bcecf-loaded HeLa cells. Cells were loaded with bcecf by incubationwith0.8␮M bcecf-AMfor30min at30◦C.Then, the dye was washed out and cells were incubated for 30min at 30◦C before mounting into the perfusion chamber of an Axiovert 200 microscope. Single cell fluorescence was excitedat490and440nmusingaCairnmonochromator.Apair of images was taken every 2s by a Hamamatsu ORCA-ER camera and ratioed after background substraction using the Metafluor program (Universal Imaging). Experiments were performed at 37◦C. For calibration of pH measurements, solutions containing 150mM KCl, 10mM Hepes, 10mM glucose,1mMCaCl2,1mMMgCl2,5␮Mnigericin andpH6.6, 7.0 or 7.4 were perfused at the end of every experiment. 2.5. Materials Wild-type coelenterazine, coelenterazine n, lysotracker red and bcecf-AM were obtained from Molecular Probes, OR, USA. Bafilomycin A was from LC laboratories, Woburn, MA, USA or from Sigma, Madrid. Inositol-(1,4,5)- trisphosphate (InsP3), cyclic ADP ribose (cADPR), nicotinic acid adenine dinucleotide phosphate (NAADP), carbonyl cyanide 4-(trifluoromethoxy)phenylhydrazone (FCCP) and nigericin were from Sigma, Madrid. Other reagents were from Sigma, Madrid or Merck, Darmstadt. 3. Results 3.1. Targeting of VAMP-mutaeq in PC12 cells The specificity of targeting of the VAMP-mutaeq chimera to secretory vesicles has been previously shown by immunofluorescence and immunoelectron microscopy in insulin containing secretory vesicles of MIN6 ␤-cells [8].We have used here a VAMP-EGFP chimera to investigate the targeting of the probe in PC12 cells. Fig. 1 (left panels) shows confocalimagesofthefluorescenceofaPC12cellexpressing VAMP-EGFP at three different planes. The pattern is typically granular and consistent with the targeting to the secretory granules. In addition, VAMP-EGFP colocalizes with the marker for acidic compartments lysotracker red. The middle panels show the fluorescence of lysotracker red obtained in A. Moreno et al. / Cell Calcium 37 (2005) 555–564 557 Fig. 1. Confocal fluorescence co-localization of VAMP-EGFP and lysotracker red in PC12 cells. PC12 cells expressing VAMP-EGFP were loaded with lysotracker red as described in Section 2. The left panels show three different confocal planes of the fluorescence of VAMP-EGFP. The middle panels show the fluorescence of lysotracker red in the same confocal planes. The right panels show the overlap of both fluorescence images. the same confocal planes and the right panels show that the fluorescences of both VAMP-EGFP and lysotracker red are mostly overlapping. Although the VAMP-EGFP fluorescence images suggest that the probe is correctly targeted to secretory granules also in PC12 cells, it is difficult to estimate from these images if there may be a small percentage of the chimera in a different location. To investigate this point, we have used the properties of aequorin to determine the degree of homogeneity of the [Ca2+] in the compartment expressing VAMP-mutaeq. Emission of light by aequorin is irreversible from an experimental point of view. Thus, aequorin is being continuously consumed along the experiments at a rate which is highly dependent on the local [Ca2+]. When aequorin is distributed in compartments with different [Ca2+], it will be consumed faster in those compartments with higher [Ca2+]. This can be easily detected and quantified in the aequorin luminescence records, and this property has been used previously to estimate the presence and size of compartments with different [Ca2+] in mitochondria and endoplasmic reticulum [12,13]. Fig.2 (upperpanel) showsacruderecordofluminescence in PC12 cells expressing VAMP-mutaeq. As in the previous 558 A. Moreno et al. / Cell Calcium 37 (2005) 555–564 Fig. 2. Steady-state [Ca2+] in the secretory granules. PC12 cells expressing VAMP–aequorin were reconstituted with coelenterazine n as described in Section 2and then the aequorin luminescence was monitored under continuous perfusion with external medium. Upper panel: crude luminescence record. When indicated “cell lysis”, standard medium containing 100␮M digitonin and 10mM Ca2+ was perfused. Middle panel: calibrated [Ca2+] values from the same experiment. Lower panel: calibrated [Ca2+] values from the same experiment obtained not taking into account the final 3.5% of the emitted luminescence. This experiment is representative of seven similar ones. work in MIN6 ␤-cells, we have reconstituted VAMP-mutaeq expressingcellswith coelenterazinen inorder tomeasure the high-[Ca2+] presumablycontainedwithinthe secretory granules. However, we have preferred to avoid Ca2+-depletion of the secretory granules prior to the measurements, by two reasons: (i) the rate of aequorin consumption at the [Ca2+] measured allowed monitoring the [Ca2+]SG during reasonable times (20–30min) and (ii) it has been described that refilling ofthesecretory granule Ca2+ storeinPC12 cellsisextremely slow [3]. Because of the lack of previous Ca2+-depletion, the Ca2+ contentofthesecretorygranulesishighatthebeginning of the experiment, and the record of luminescence shows a continuously decreasing trend due to aequorin consumption in a high-[Ca2+] environment. Once the luminescence had reached nearly zero values, cells were permeabilized with digitonin to reveal all the aequorin luminescence still present in the sample. In case a significant proportion of aequorin was present (due to mistargeting) in compartments having low resting [Ca2+], its luminescence should be released here. However, only a small percentage of the total luminescence, below 5% (4.3±0.3%, mean±S.E.M., n=7), was released at this point. Fig.2(middlepanel)showstherecordof[Ca2+]calculated from the same experiment. It shows a steady-state [Ca2+] level around 35␮M, which keeps stable for 15–20min and then starts to decrease slowly. This final decrease is due to the presence of a small percentage of aequorin in a compartment with low-[Ca2+]. As most of the aequorin present in the high-Ca2+ compartment is being consumed, the relative size of the low-Ca2+ compartment increases progressively and this is reflected as a decrease in the calibrated [Ca2+] values. The percentage of aequorin present in that low-[Ca2+] compartment can be estimated from the relative amount of luminescence emitted after cell permeabilization, that is, less than 5%. In fact, if that percentage of luminescence was not takenintoaccountduringthecalibrationofluminescenceinto [Ca2+] values, the record of calibrated [Ca2+] remained stable for nearly 1h around 40␮M(Fig. 2, lower panel). In this case, calculation of [Ca2+] values was made not taking into accountthefinal3.5%of theluminescence. Inseveralsimilar experiments, this percentage was always below 5%. Therefore, these results constitute functional evidence that more than95%ofour VAMP-mutaeqprobeismeasuring [Ca2+]in a compartment with an homogeneous [Ca2+] around 40␮M (43±3␮M, mean±S.E.M., n=7). 3.2. Dynamic measurements of [Ca2+] in the secretory vesicles Wehavenow investigatedthe effecton thesecretorygranule [Ca2+] of a series of agents known to induce cytosolic Ca2+ transients by different mechanisms: ATP, acting on purinergic receptors, triggersInsP3productionandrelease of Ca2+ from the ER; caffeine, acting on ryanodine receptors, induces also Ca2+ release from the ER; and high-K+,bydepolarizingthecells,inducesCa2+ entryviaplasmamembrane voltage-dependent Ca2+ channels. Fig. 3 shows the effects of these stimuli on both cytosolic and secretory granule [Ca2+]. As expected, all of the stimuli induced transient [Ca2+] increases in the cytosol. However, it came as a surprise to see thatalloftheminducedalsoatransientdecreaseinthe[Ca2+] of secretory granules. These results suggested that secretory granules could release Ca2+ during the process of stimulated secretion. To investigate the mechanism of stimulated Ca2+ release from the granules, we used permeabilized cells to test the effects of several second messengers: InsP3, cADPR and NAADP.Fig.4showsthatnoneofthemproducedanychange inthe[Ca2+]SG.Similarly,perfusionofa10␮M[Ca2+]buffer A. Moreno et al. / Cell Calcium 37 (2005) 555–564 559 Fig. 3. Effects of ATP, caffeine and high-K+on both [Ca2+]cand [Ca2+]SG. PC12 cells were transfected with either cytosolic aequorin (left panels) or VAMP–aequorin (right panels), and reconstituted with coelenterazine as described in Section 2. The left panels show the cytosolic [Ca2+] peaks induced by 1min stimulation with 100␮M ATP or 10mM caffeine, or 30s stimulation with a solution containing 70mM K+. The right panels show the effect of the same stimuli on [Ca2+]SG. These experiments are representative of 5–10 similar ones of each kind. tomimic localincreasesinthe[Ca2+] aroundthe vesicleshad no effect on [Ca2+]SG. Therefore, these second messengers are apparently not involved in the mechanism of the decrease in [Ca2+]SG induced by the agonists. Regarding alternative Ca2+ transport systems in the granules, the presence of either Ca2+-ATPases, Na+/Ca2+ or H+/Ca2+ exchange systems have been proposed in different preparations [14–18]. Inhibition of sarcoendoplasmic reticulumCa2+-ATPasewith 10␮M benzohydroquinoneproduced little effects on [Ca2+]SG (data not shown), suggesting that this type of Ca2+-ATPase was not involved. On the other hand, in experiments performed with permeabilized cells, changes in the concentration of Na+ion from 10 to 0mM or from 10 to 140mM produced also no change in the [Ca2+]SG (data not shown), suggesting that Na+/Ca2+ exchange is not operating in these granules. Instead, changes in the pH in the cytosolic side of the vesicles induced fast and reversible changes in [Ca2+]SG consistent with the operation of a H+/Ca2+ exchange system. Fig. 5 shows experiments performed in permeabilized PC12 cells perfused with 560 A. Moreno et al. / Cell Calcium 37 (2005) 555–564 Fig. 4. Effect of InsP3, cADPR, NAADP and Ca2+ on [Ca2+]SG. PC12 cells expressing VAMP–aequorin were reconstituted with coelenterazine n and permeabilized with digitonin as described in Section 2immediately prior to the experiment. Then, the effects of either 5␮M InsP3,5␮M cADPR, 10␮MNAADPor10␮MCa2+ weretested, as indicated. Theseexperiments are representative of five similar ones of each kind. buffers containing 100nM [Ca2+] and different pH. It can be observed that a decrease in pH from 7.0 to 6.6 induced a decrease in [Ca2+]SG from 40␮M to below 30␮M, reversible after returning the cytosolic pH to 7.0. Vice versa, an increase in pH from 7.0 to 7.4 induced an increase in [Ca2+]SG from40␮M tomore than50␮M,reversibleafter returningto pH 7.0. The effects of nigericin, an ionophore that catalizes H+/K+exchange, were also consistent with the presence of a H+/Ca2+ exchange. Fig. 5 (lower panel) shows that nigericin induced a decrease in [Ca2+]SG, which was reversible after washing the ionophore with a 0.1% albumin containing solution. The decrease in [Ca2+]SG induced by ATP, caffeine and high-K+could be also mediated by a vesicular H+/Ca2+ antiporter in case these stimuli induce a cytosolic acidification. This was the case. Fig. 6 shows the mean response of all the cells present under the microscope field in experiments performed with each agonist. Most of the cells responded to the stimuli with a transient or prolonged acidification and the mean response was in all the cases an acidification of around 0.1 pH unit. Fig. 7 (upper panel) shows that the protonophore FCCP produced also a decrease in [Ca2+]SG, which was reversible after washing the ionophore. The effect of FCCP is again consistent with the presence of a H+/Ca2+ exchange operating in the membrane of the secretory granules. In case this system was responsible for the accumulation of Ca2+ into the vesicles, we reasoned that the reversibility of the effect of FCCP on [Ca2+]SG should depend on the restoration of the vesicular H+gradient via the bafilomycin-sensitive vacuolar H+-ATPase. However, as shown in the lower panel of Fig. 7, bafilomycin had no effect on this process. Similarly, incubation with 200nM bafilomycin for 15min produced no effect on the resting [Ca2+]SG (data not shown). Therefore, restoration of the vesicular H+gradient was not required to revert the changes in [Ca2+]SG induced by Fig. 5. Effect of pH changes and nigericin on [Ca2+]SG. PC12 cells expressing VAMP–aequorin were reconstituted with coelenterazine n. In the upper andmediumpanels,theywerealsopermeabilizedwithdigitoninasdescribed in Section 2immediately prior to the experiment. Then, solutions containing 100nM Ca2+ and the different pH indicated in the figure were perfused. In the lower panel, intact PC12 cells were treated with 2␮M nigericin as indicated. The ionophore was then washed by perfusing medium containing 0.1% bovine serum albumin. These experiments are representative of four to eight similar ones of each kind. FCCP. Alternatively, the presence of P-type Ca2+-ATPases, as suggested for the insulin containing vesicles of MIN6 ␤- cells [8], cannot be excluded. To obtain further evidence on thispoint,we perfusedintracellularmedium containingADP (and 5␮M oligomycin to abolish internal ATP production) instead of ATP in permeabilized cells, but we did not observe any changes in [Ca2+]SG. In addition, Fig. 8 (upper and A. Moreno et al. / Cell Calcium 37 (2005) 555–564 561 Fig. 6. Effects of ATP, caffeine and high-K+on cytosolic pH. PC12 cells were loaded with bcecf as described in Section 2and then stimulated with either 100␮M ATP for 1min, 10mM caffeine for 1min or with a solution containing 70mM K+for 30s, as indicated in the figure. Traces shown correspond to the mean of all the cells present under the microscope field (high-K+, 20 cells; caffeine, 19 cells; ATP, 40 cells). These experiments are representative of at least three separate experiments of each kind. Fig.7. EffectofFCCPon[Ca2+]SG.PC12cellsexpressingVAMP–aequorin were reconstituted with coelenterazine n as described in Section 2. Then, 2␮M FCCP was perfused as indicated. In the lower panel, 200nM bafilomycin was also present before, during and after FCCP addition, as indicated. These experiments are representative of six similar ones of each kind. Fig. 8. Effect of FCCP on [Ca2+]SG in permeabilized cells. PC12 cells expressing VAMP–aequorin were reconstituted with coelenterazine n and permeabilized with digitonin as described in Section 2immediately prior to the experiment. Then, in the upper and medium panels, solutions containing 100nM Ca2+ and either standard intracellular medium containing 1mM ATP (upper panel) or intracellular medium containing 1mM ADP and 5␮M oligomycin instead of ATP (medium panel) were perfused. In the lower panel, standard intracellular medium containing 2mM EGTA and no Ca2+ was perfused. When indicated, 2␮M FCCP was perfused in the same medium used for each experiment. These experiments are representative of three to five similar ones of each kind. 562 A. Moreno et al. / Cell Calcium 37 (2005) 555–564 medium panels) shows that the recovery from the [Ca2+]SG decrease induced by FCCP occurred similarly in permeabilized cells perfused with medium with or without ATP, suggesting that ATP hydrolysis was also not necessary to restore the [Ca2+]SG levels. Finally, Fig. 8 (lower panel) shows that recovery of [Ca2+]SG after FCCP action took place unmodified even in permeabilized cells perfused with medium in the absence of Ca2+ (EGTA 2mM). 4. Discussion We show in this paper dynamic measurements of [Ca2+] in the catecholamine-containing secretory granules of PC12 cellsusingaVAMP–aequorinchimera.Regardingthequality of the targeting to the secretory granules, we provide functionalevidencethatourprobe(atleastthatabletoreconstitute with coelenterazine n in the absence of Ca2+-depletion) is more than 95% located in a compartment containing around 40␮M[Ca2+]understeady-stateconditions.Noothersubcellular compartment known has [Ca2+] levels in this range. Endoplasmic reticulum and Golgi apparatus have [Ca2+] 5–10fold higher [19,20], while cytosol, nucleus and mitochondria stay under resting conditions at [Ca2+] below 1␮M. Our data of resting [Ca2+]SG are consistent with those obtained with the same probe in MIN6 ␤-cells (around 50␮M, see ref [8]) and confirm that secretory granules are the organelle with the largest bound/free Ca2+ relationship. Estimations for the total calcium content of secretory granules in chromaffin or PC12 cells range between 20 and 40mM [3,21]. Most of this Ca2+ must be bound to proteins such as chromogranin A, the major intravesicular protein of adrenal chromaffin granules. This protein, which constitutes 40% of the soluble intravesicular proteins [22], is known to bind 152nmol of Ca2+/mg of protein with a Kdof 54␮M, apart of several additional low affinity (millimolar) binding sites [4,21,23]. Given that the free [Ca2+]SG is around 40␮M, only the high affinity sites can bind significant amounts of Ca2+.Inaddition, a significant proportion of Ca2+ must be bound to ATP, as chromaffin granules may contain up to 130mM ATP [2]. The rapid release of Ca2+ from the secretory vesicles induced by the different agonists suggests that Ca2+ release from the granules during cell stimulation could play a role in secretion, in particular if the observed global decrease in [Ca2+]SG induced by the agonists could be restricted to a small population of vesicles undergoing larger changes in their[Ca2+].However,noneoftheknownsecondmessengers tested (InsP3, cADPR, NAADP or Ca2+) produced any effect on [Ca2+]SG in permeabilized cells. Thus, unless additional cellular components (perhaps lost during permeabilization) mayberequired,ourdata suggestthatneitherInsP3receptors nor ryanodine receptors (at least those activated by cADPR or Ca2+) or NAADP-activated channels are involved in the mechanism of agonist-induced Ca2+ release from PC12 secretory granules. Our data show also that changes in the pH gradient between the cytosol and the secretory granules immediately trigger changes in [Ca2+]SG that are consistent with the operation of a H+/Ca2+ exchange. These effects cannot be attributedtotheeffectsofintragranularpHchangesinaequorin luminescence,becauseaequorinislittlesensitivetopHinthis range[8].Inaddition, it shouldberemarked thatbothlow pH andnigericinorFCCPinduceddecreasesin[Ca2+]SG (Fig.5), although their effects in intragranular pH should be opposite. pH-dependent Ca2+ fluxes thus appear to be the only mechanism able to generate fast changes in [Ca2+]SG under our conditions, and could be responsible of agonist-induced Ca2+ release from the granules. We show here that cell stimulation with ATP, caffeine or high-K+induces a transient acidification of the cytosol of about 0.1 pH unit. It is difficult toassessifthisacidificationisenoughtoinducethemeasured decreases in [Ca2+]SG, but obviously the changes go in the rightdirection.Inaddition,localpHchangesinthevicinityof the secretory granules close to the plasma membrane may be larger than the global changes measured here. Consistently, it has been shown that stimulation of PC12 cells with highK+depolarizationinducesalcalinizationofsecretoryvesicles [24]. A similar finding has been described in mast cells [25]. Operation of a H+/Ca2+ exchange has been previously described in synaptic vesicles [18,26] and yeast vacuoles [27]. As mentioned above, our data in PC12 secretory granules share with those described in MIN6 ␤-cells [8,9] the similar steady-state [Ca2+] levels. However, there are many differences in the dynamics, such as the lack of effect of FCCP on MIN6 ␤-cells [Ca2+]SG, the increase in [Ca2+]SG induced by an InsP3-producing agonist in those cells and the release of Ca2+ from the MIN6 ␤-cells through both ryanodine receptors and NAADP receptors. On the contrary, none of these receptorsappears tobe presentin thePC12 secretoryvesicles and the protonophore FCCP, the H+/K+exchange ionophore nigericine, and the different extracellular stimuli induced all of them rapid decreases of [Ca2+]SG in PC12 cells. These differences probably result from the presence of a H+/Ca2+ exchange system in PC12 granules. Regarding the mechanism of accumulation of Ca2+ into PC12 granules, our data on the restoration of [Ca2+]SG levels after FCCP action are intriguing. In permeabilized cells, FCCP induced a decrease in [Ca2+]SG that was completely reversible even in the absence of Ca2+ and ATP in the perfusion medium. The only explanation to these results would be the presence in the secretory granules of a large capacity, low affinity and slow-exchanging Ca2+ buffer, that would be able to release Ca2+ during the following 10–15min after FCCP action to restore [Ca2+]SG levels in the absence of any Ca2+ influx from the cytosol. This possibility is consistent with previous data showing that calcium binds and precipitates chromogranins at affinities compatible with their physiological concentrations and that this tight binding could account for more than 90% of intravesicular calcium [28]. The picturethatemerges nowisthatof avesiclecontainingrelatively small amounts of free or rapidly exchanging bound Ca2+ and A. Moreno et al. / Cell Calcium 37 (2005) 555–564 563 a large store of tightly bound slow-exchanging Ca2+. Accordingly, disruption of the vesicular pH gradient with weak amines induces a very slow release of calcium from the vesicles, about 20% in 15min [29]. The steady-state Ca2+ accumulation into the vesicles may be due to the operation of H+/Ca2+ exchangecombinedwiththeenergyprovidedbythe H+gradient generated by the vacuolar H+-ATPase. However, thepresenceofP-typeCa2+-ATPasescannotbeexcluded,and both systems could in fact coexist, as has been described in synaptic vesicles [30] and yeast vacuoles [27]. In conclusion, the main findings of this work are: (i) catecholamine-containing secretory granules have an homogeneous steady-state [Ca2+] around 40␮M; (ii) cell stimulation by different mechanisms, either mediated by InsP3 receptors, ryanodine receptors or plasma membrane Ca2+ channels, induces a transient decrease in secretory granule [Ca2+] that occurs simultaneously to the increase in cytosolic[Ca2+];(iii)changesin cytosolicpHor abolition ofthepH gradient between the cytosol and the granules induce rapid changes of [Ca2+] in the secretory vesicles that are consistent with the operation of a H+/Ca2+ antiporter in the vesicular membrane; (iv) reversibility of FCCP-induced [Ca2+]SG decreaseoccurssimilarlyeitherinthepresenceorintheabsence of Ca2+ or ATP in the cytosol, suggesting the presence of a large and slow-exchanging Ca2+ buffer in the vesicle matrix. This large store of tightly bound calcium buffers [Ca2+]SG changes in the long-term, but allows rapid changes in free [Ca2+]SG occurring in times of seconds or minutes. Secretory granules are therefore a much more dynamic organelle in terms of free [Ca2+]SG than previously thought. The functional significance of these Ca2+ changes is still unknown, but evidence for a role of vesicular Ca2+ in exocytosis has been provided in several types of secretory vesicles [31–33]. A paper recently published by Mahapatra et al. [34] explores also the dynamics of [Ca2+] in the secretory granules of PC12 cells using a chromogranin–aequorin chimeric protein. However, their results are largely different from ours both in terms of the free [Ca2+] in the secretory granules (only 1.4␮M) and the response to different stimuli. The discrepancyisclearlydue tothedifferentCa2+ affinity ofthe aequorinprobesinbothcases.Mahapatraetal.usedanaequorin chimera with much higher Ca2+ affinity, which is unable to reconstitute with coelenterazine in areas where the [Ca2+]is aboveafewmicromolar.Thus,theirmeasurementsreflectthe behavior of the small compartment (less than 5% of the totalspace)havinglowresting[Ca2+]andfunctionalproperties completelydifferentfromthoseshownhereforthebulkofthe granular space. We have reported before similar discrepancies in the dynamics of [Ca2+] in the endoplasmic reticulum using aequorin chimeras with different Ca2+ affinities [12]. Acknowledgements This work was supported by grants from Direcci´ on GeneraldeEnse˜ nanzaSuperior(BFI2002-01397)andfromJunta deCastillayLe´ on(VA005/02).J.S.andE.H.-S.holdFPIand FPU fellowships, respectively, from the Spanish Ministerio de Educaci´ on y Ciencia. References [1] F. Clemente, J. Meldolesi, Calcium and pancreatic secretiondynamics of subcellular calcium pools in resting and stimulated acinar cells, Br. J. Pharmacol. 55 (1975) 369–379. [2] H. Winkler, E. Westhead, The molecular organization of adrenal chromaffin granules, Neuroscience 5 (1980) 1803–1823. [3] C. Fasolato, M. Zottini, E. Clementi, D. Zacchetti, J. Meldolesi, T. Pozzan, Intracellular Ca2+ pools in PC12 cells. Three intracellular pools are distinguished by their turnover and mechanisms of Ca2+ accumulation, storage, and release, J. Biol. Chem. 266 (1991) 20159–20167. [4] D. Bulenda, M. Gratzl, Matrix free Ca2+ in isolated chromaffin vesicles, Biochemistry 24 (1985) 7760–7765. [5] T. Pozzan, R. Rizzuto, P. Volpe, J. Meldolesi, Molecular and cellular physiology of intracellular calcium stores, Physiol. Rev. 74 (1994) 595–636. [6] G.A. Rutter, C. Fasolato, R. Rizzuto, Calcium and organelles: a twosided story, Biochem. Biophys. Res. Commun. 253 (1998) 549–557. [7] S.H. Yoo, Coupling of the IP3receptor/Ca2+ channel with Ca2+ storage proteins chromogranins A and B in secretory granules, Trends Neurosci. 23 (2000) 424–428. [8] K.J. Mitchell, P. Pinton, A. Varadi, C. Tacchetti, E.K. Ainscow, T. Pozzan, R. Rizzuto, G.A. Rutter, Dense core secretory vesicles revealed as a dynamic Ca2+ store in neuroendocrine cells with a vesicle-associated membrane protein aequorin chimaera, J. Cell Biol. 155 (2001) 41–51. [9] K.J. Mitchell, F.A. Lai, G.A. Rutter, Ryanodine receptor type I and nicotinic acid adenine dinucleotide phosphate receptors mediate Ca2+ release from insulin-containing vesicles in living pancreatic beta-cells (MIN6), J. Biol. Chem. 278 (2003) 11057–11064. [10] M. Brini, R. Marsault, C. Bastianutto, J. Alvarez, T. Pozzan, R. Rizzuto, Transfected aequorin in the measurement of cytosolic Ca2+ concentration ([Ca2+]c). A critical evaluation, J. Biol. Chem. 270 (1995) 9896–9903. [11] J. Alvarez, M. Montero, Measuring [Ca2+] in the endoplasmic reticulum with aequorin, Cell Calcium 32 (2002) 251–260. [12] M. Montero, J. Alvarez, W.J. Scheenen, R. Rizzuto, J. Meldolesi, T. Pozzan, Ca2+ homeostasis in the endoplasmic reticulum: coexistence of high and low [Ca2+] subcompartments in intact HeLa cells, J. Cell Biol. 139 (1997) 601–611. [13] M. Montero, M.T. Alonso, E. Carnicero, I. Cuchillo-Ibanez, A. Albillos, A.G. Garcia, J. Garcia-Sancho, J. Alvarez, Chromaffin-cell stimulation triggers fast millimolar mitochondrial Ca2+ transients that modulate secretion, Nat. Cell Biol. 2 (2000) 57–61. [14] H. Krieger-Brauer, M. Gratzl, Uptake of Ca2+ by isolated secretory vesicles from adrenal medulla, Biochim. Biophys. Acta 691 (1982) 61–70. [15] H. Krieger-Brauer, M. Gratzl, Effects of monovalent and divalent cations on Ca2+ fluxes across chromaffin secretory membrane vesicles, J. Neurochem. 41 (1983) 1269–1276. [16] J.D. Troadec, S. Thirion, J.P. Laugier, G. Nicaise, Calcium-induced calcium increase in secretory vesicles of permeabilized rat neurohypophysial nerve terminals, Biol. Cell. 90 (1998) 339–347. [17] S. Thirion, J.D. Troadec, N.B. Pivovarova, S. Pagnotta, S.B. Andrews, R.D. Leapman, G. Nicaise, Stimulus-secretion coupling in neurohypophysial nerve endings: a role for intravesicular sodium? Proc. Natl. Acad. Sci. U.S.A. 96 (1999) 3206–3210. [18] P.P. Goncalves, S.M. Meireles, C. Gravato, M.G. Vale, Ca2+–H+ antiport activity in synaptic vesicles isolated from sheep brain cortex, Neurosci. Lett. 247 (1998) 87–90.