Ca 2+ induced Ca 2+ Release in Chromaffin Cells Seen from inside the ER with Targeted Aequorin

 The Rockefeller University Press, 0021-9525/99/01/241/14 $2.00 The Journal of Cell Biology, Volume 144, Number 2, January 25, 1999 241–254 http://www.jcb.org 241 Ca 2 1 -induced Ca 2 1 Release in Chromaffin Cells Seen from inside the ER with Targeted Aequorin Maria Teresa Alonso,* Maria José Barrero,* Pedro Michelena, ‡ Estela Carnicero,* Inmaculada Cuchillo, ‡ Antonio G. García, ‡ Javier García-Sancho,* Mayte Montero,* and Javier Alvarez* *Instituto de Biología y Genética Molecular, Departamento de Bioquímica y Biología Molecular y Fisiología, Facultad de Medicina, Universidad de Valladolid y Consejo Superior de Investigaciones Científicas, E-47005 Valladolid, Spain; and ‡ Instituto de Farmacología Teófilo Hernando, Departamento de Farmacología y Terapéutica, Facultad de Medicina, Universidad Autónoma de Madrid, E-28029 Madrid, Spain Abstract. The presence and physiological role of Ca 2 1 - induced Ca 2 1 release (CICR) in nonmuscle excitable cells has been investigated only indirectly through measurements of cytosolic [Ca 2 1 ] ([Ca 2 1 ] c ). Using targeted aequorin, we have directly monitored [Ca 2 1 ] changes inside the ER ([Ca 2 1 ] ER ) in bovine adrenal chromaffin cells. Ca 2 1 entry induced by cell depolarization triggered a transient Ca 2 1 release from the ER that was highly dependent on [Ca 2 1 ] ER and sensitized by low concentrations of caffeine. Caffeine-induced Ca 2 1 release was quantal in nature due to modulation by [Ca 2 1 ] ER . Whereas caffeine released essentially all the Ca 2 1 from the ER, inositol 1,4,5-trisphosphate (InsP 3 )- producing agonists released only 60–80%. Both InsP 3 and caffeine emptied completely the ER in digitoninpermeabilized cells whereas cyclic ADP-ribose had no effect. Ryanodine induced permanent emptying of the Ca 2 1 stores in a use-dependent manner after activation by caffeine. Fast confocal [Ca 2 1 ] c measurements showed that the wave of [Ca 2 1 ] c induced by 100-ms depolarizing pulses in voltage-clamped cells was delayed and reduced in intensity in ryanodine-treated cells. Our results indicate that the ER of chromaffin cells behaves mostly as a single homogeneous thapsigargin-sensitive Ca 2 1 pool that can release Ca 2 1 both via InsP 3 receptors or CICR. Key words: endoplasmic reticulum • aequorin • chromaffin cells • calcium • ryanodine A t present, the role played by ryanodine receptors (RyR) 1 in the homeostasis of intracellular Ca 2 1 in nonmuscle cells is unclear. Mammalian tissues express three isoforms, RyR1, RyR2, and RyR3, encoded by different genes. RyR1 and RyR2 are expressed predominantly in sarcoplasmic reticulum of skeletal muscle and heart, respectively, where they have an essential role to trigger muscle contraction (Sutko and Airey, 1996; Zucchi and Ronca-Testoni, 1997). RyR3 was originally identified in brain (Hakamata et al., 1992), but in fact all three isoforms are actually expressed in brain, and the major brain isoform appears to be RyR2 (McPherson and Campbell, 1993; Sorrentino and Volpe, 1993; Giannini et al., 1995). RyRs are widely distributed in many other different tissues (Giannini et al., 1995; Mackrill et al., 1997), including the adrenal gland, and they are in many cases coexpressed with one or more isoforms of inositol 1,4,5-trisphosphate receptors (InsP 3 R) (Walton et al., 1991; Poulsen et al., 1995). The reasons for the presence of multiple Ca 2 1 release mechanisms in the same cell are not clear (for discussion see Newton et al., 1994; Sutko and Airey, 1996). A possible reason for the coexistence of both InsP 3 R and RyR in the same cells could be that they might release Ca 2 1 from different compartments, and with a different physiological significance. They may be modulated by different second messengers, such as InsP 3 in the case of InsP 3 R or cyclic adenosine diphosphate ribose (cADPR) for the RyR (Lee, 1998). However, although InsP 3 is a well-established physiological activator of the InsP 3 R, the role of cADPR as activator or modulator of RyR in the presence of physiologiM.T. Alonso and M.J. Barrero contributed equally to this work. Address correspondence to J. Alvarez, Departamento de Bioquímica y Biología Molecular y Fisiología, Facultad de Medicina, Ramón y Cajal 7, E-47005 Valladolid, Spain. Tel: (34) 983-423085. Fax: (34) 983-423588. E-mail: [email protected] 1. Abbreviations used in this paper: [Ca 2 1 ] c , cytosolic [Ca 2 1 ]; [Ca 2 1 ] ER , ER [Ca 2 1 ]; BAPTA, 1,2-bis(2-aminophenoxy)ethane-N,N,N 9 ,N 9 -tetraacetic acid; cADPR, cyclic adenosine diphosphate ribose; CICR, Ca 2 1 -induced Ca 2 1 release; CPA, ciclopiazonic acid; DMPP, 1,1-dimethyl-4-phenylpiperazinium iodide; HSV-1, herpes simplex virus type 1; InsP 3 R, InsP 3 receptor; InsP 3 , inositol 1,4,5-trisphosphate; ivu, infectious virus units; RyR, ryanodine receptor. on March 11, 2005 www.jcb.orgDownloaded from The Journal of Cell Biology, Volume 144, 1999 242 cal concentrations of ATP remains controversial (Sutko and Airey, 1996; Zucchi and Ronca-Testoni, 1997). In neuronal cells, RyR could be activated via the classical Ca 2 1 - induced Ca 2 1 release (CICR) mechanism after Ca 2 1 entry through voltage-dependent Ca 2 1 channels. Studies in some neuronal preparations have shown that depolarizing stimuli produce an increase in cytosolic [Ca 2 1 ] ([Ca 2 1 ] c ) that may be due in part to Ca 2 1 release from intracellular stores (Verkhratsky and Shmigol, 1996). However, little direct evidence has been presented for CICR in nonmuscle cells; this is due to the difficulties in separating the contribution of Ca 2 1 entry and Ca 2 1 release to the [Ca 2 1 ] c signal and to the nonspecificity of caffeine and other pharmacological agents used to activate RyR. To study CICR in neuronal cells, it would be greatly advantageous to measure [Ca 2 1 ] specifically inside the Ca 2 1 stores. We have recently reported a method to measure [Ca 2 1 ] in the lumen of the endoplasmic reticulum ([Ca 2 1 ] ER ) of intact cells, the main intracellular Ca 2 1 store, by using an ER-targeted aequorin (Montero et al., 1995, 1997a,b; Barrero et al., 1997), which can be expressed in different types of cells using a viral vector (Alonso et al., 1998). This technique is ideal to study directly CICR, because Ca 2 1 release can be measured independently of the variations in [Ca 2 1 ] c . Here we have used this technique to monitor [Ca 2 1 ] ER in chromaffin cells. These neuroendocrine cells have a potent caffeine-sensitive Ca 2 1 release mechanism (Cheek et al., 1990), and Ca 2 1 entry through several types of voltagedependent Ca 2 1 channels can be induced by K 1 depolarization or more physiologically, using nicotinic agonists (Núñez et al., 1995; Lara et al., 1998). Direct evidence for a CICR mechanism working under physiological conditions in these cells has not been provided. However, we have reported recently that the caffeine-sensitive Ca 2 1 stores may modulate catecholamine secretion induced by depolarization with high K 1 in these cells. Catecholamine secretion was reduced after store emptying with caffeine, and recovered as Ca 2 1 stores refilled during consecutive K 1 /Ca 2 1 pulses. The main conclusion from that work was that Ca 2 1 stores could have a double role, acting either as a sink or as a source of Ca 2 1 , depending of their state of filling (Lara et al., 1997). Additionally, several questions regarding the function of RyR in chromaffin cells remain unanswered. For instance, chromaffin cells have also InsP 3 R, which are at least in part colocalized with RyR2 in the ER (Poulsen et al., 1995). The presence of separate or overlapping Ca 2 1 pools responsive to either InsP 3 , caffeine, or cADPR, their differential sensitivity to inhibitors of the ER Ca 2 1 -pump such as thapsigargin, and the physiological significance of the different Ca 2 1 release mechanisms, has been a subject of debate for many years (Cheek et al., 1991; Liu et al., 1991; Robinson and Burgoyne, 1991; Stauderman et al., 1991; Morita et al., 1997). On the other hand, the mechanism of Ca 2 1 release induced by caffeine is quite particular because increasing concentrations of caffeine release Ca 2 1 in a quantal manner, a phenomenon that has been suggested to indicate that the caffeine-sensitive Ca 2 1 pool is composed of functionally discrete stores with heterogeneous sensitivities to caffeine (Cheek et al., 1993, 1994a). Here we have monitored [Ca 2 1 ] ER in chromaffin cells to investigate the mechanism of quantal Ca 2 1 release by caffeine, its relationship in terms of Ca 2 1 pools with InsP 3 - mediated Ca 2 1 release, and the presence of CICR triggered by Ca 2 1 entry. In brief, our results indicate that the ER Ca 2 1 pools responding to caffeine and InsP 3 mostly overlap. The response to caffeine was also quantal when studied from inside the ER, but this quantal response could be explained by the control of caffeine-induced Ca 2 1 release by [Ca 2 1 ] ER , with no need for separate ER compartments with heterogeneous sensitivities to caffeine, as proposed previously (Cheek et al., 1993, 1994a). We show that CICR can be induced by Ca 2 1 entry elicited either by high K 1 depolarization or by stimulation with nicotinic agonists. This is consistent with our previously proposed model for the Ca 2 1 store as a modulator of secretion (Lara et al., 1997). Additionally, using fast confocal [Ca21] measurements, we show that CICR participates in the generation and propagation of the Ca21 wave induced by cell depolarization. Materials and Methods Preparation and Culture of Bovine Chromaffin Cells Bovine adrenal medulla chromaffin cells were isolated following standard methods (Livett, 1984) with some modifications (Moro et al., 1990). Cells were suspended in Dulbecco’s modified Eagle’s medium (DME) supplemented with 5% fetal calf serum, 10 mM cytosine arabinoside, 10 mM fluorodeoxyuridine, 50 IU ml21 penicillin and 50 IU ml21 streptomycin. For secretion experiments, cells were plated in 5-cm-diam Petri dishes (5 3 106 cells per 5 ml of DME). For aequorin experiments, cells were plated on 12or 13-mm glass poly-d-lysine–coated coverslips (0.5 3 106 cells per 1 ml of DME). For measurements of [Ca21]c transients by confocal microscopy and ionic currents, cells were plated on 2.5-cm-diam glass coverslips at a density of 5 3 104 cells per ml. Cultures were maintained at 378C in a humidified atmosphere of 5% CO2. Preparation of Viral Stock and Infection of Cultures Construction, packaging, and titering of the pHSVerAEQ amplicon vector and expression in chromaffin cells has been previously described (Alonso et al., 1998). The EcoRI fragment of the erAEQmut cDNA was subcloned into the pHSVpuc vector to generate the pHSVerAEQ. As a helper virus, the herpes simplex virus type 1 (HSV-1) IE2 deletion mutant 5dl1.2 was used with a titer of 2 3 107 infectious virus units (ivu)/ml (Lim et al., 1996). Titers of viral stocks were determined by immunocytochemistry on PC12 cells. Infected cells were visualized by using a rabbit anti– HSV-1 particle antibody (1:10,000 dilution; Dako) or a mouse anti-HA1 primary antibody (1:200 dilution; Boehringer Mannheim) followed by an alkaline phosphatase-conjugated anti-mouse IgG antibody (1:200 dilution; Sigma). The titers of the vector stock were 1.1 3 106 ivu/ml pHSVerAEQ and 4.1 3 106 ivu/ml 5dl1.2. Chromaffin cell cultures (5 3 105 cells/0.5 ml) were routinely infected with 1.2 3 104 ivu 1 d before measurements. The percentage of cells expressing ER-targeted aequorin was usually z20%. Immunofluorescence revealed a typical nonnuclear reticular pattern (data not shown), similar to that previously seen in HeLa cells (Montero et al., 1995). This pattern was not modified by the 1-h period of ER Ca21 depletion required for ER aequorin measurements. Measurements of [Ca2 1 ]ER with Aequorin For [Ca21]ER studies, cells were infected after 1 d in culture with HSV-1 carrying the ER-targeted aequorin construct as described previously (Alonso et al., 1998). Measurements of [Ca21]ER were started z16 h after infection that were required to allow adequate expression of the targeted photoprotein. Aequorin photoluminescence measurements were performed essentially as previously described (Barrero et al., 1997). In brief, cells were depleted of Ca21 by incubation for 5–10 min at 378C with the sarco-endoplasmic reticulum Ca21 ATPase (SERCA) inhibitor 2,5-di-tertbutyl-benzohydroquinone (BHQ) 10 mM in standard medium containing 145 mM NaCl, 5 mM KCl, 1 mM MgCl2, 10 mM glucose, 10 mM Hepes, on March 11, 2005 www.jcb.orgDownloaded from Alonso et al. Ca2 1 -induced Ca2 1 Release Seen from the ER 243 pH 7.4, supplemented with 3 mM EGTA. Cells were then incubated for 1 h at room temperature in standard medium containing 0.5 mM EGTA, 10 mM BHQ, and 1 mM coelenterazine n. The coverslip was then placed in the perfusion chamber of a purpose-built thermostatized luminometer and standard medium containing 1 mM Ca21 was perfused to refill the ER with Ca21. Measurements were performed at 228C and [Ca21]ER values were calculated from the luminescence records using a computer algorithm (Brini et al., 1995) which follows the calibration curve reported before (Barrero et al., 1997). In the experiments carried out with permeabilized cells, cells were placed in the luminometer as described above and perfused for 1 min with intracellular-like medium (10 mM NaCl, 140 mM KCl, 1 mM MgCl2, 1 mM KH2PO4, 2 mM ATP, 20 mM Hepes, pH 7) containing 2 mM EGTA and 20 mM digitonin. Then, intracellular medium without digitonin and containing 100 nM EGTA-buffered Ca21 was perfused for 3–5 min to refill the ER with Ca21. The total number of counts obtained ranged between 0.3 and 2 million. Measurements of Single-cell [Ca2 1 ]c Single-cell measurements of [Ca21]c were performed at room temperature in fura-2–loaded cells as described previously (Núñez et al., 1995). Cells were epi-illuminated alternatively at 340 and 380 nm and light emitted above 520 nm was recorded by an extended ISIS-M camera (Photonic Science) and analyzed using an Applied Imaging Magical image processor (Sunderland). 16 frames excited at every wavelength were averaged by hardware, with a time resolution of z7 s for each pair of images, and [Ca21]c was estimated from the ratio F340/F380 by comparison with fura-2 standards. Field electric stimulation (McIlwain and Rodnight, 1962) was performed through a pair of silver electrodes placed 7 mm apart and 1.5 mm above the cells. Alternating positive-negative square pulses of 50-ms duration and 60-V intensity were applied at 10 Hz. The peak current was 100 mA. Confocal [Ca2 1 ]c Measurements and Electrophysiological Recordings Electrical measurements and [Ca21]c were recorded by using the wholecell patch-clamp technique (Hamill et al., 1981) in combination with fluo-3 based microfluorometry. Cells were placed in an experimental chamber that was mounted on the stage of an inverted microscope (Diaphot 200; Nikon). Cells were loaded via the patch pipette with the pentaammonium salt form of the fluorescent dye fluo-3 (100 mM). The dye was excited with a Kr-Ar laser light at 488 nm and emission was detected at 522 nm (32-nm band width). Cells were dialyzed with an intracellular solution containing 135 mM CsCl2, 8 mM NaCl; 1 mM MgCl2, 20 mM Hepes, 2 mM ATP, and 0.3 mM GTP, pH 7.3. The chamber was continuously perfused with Krebs-Hepes medium. Line-scan images (0.33-mm width) of the intracellular Ca21 distribution were acquired every 2 ms with a confocal microscope (MRC 1024; Bio-Rad), using an oil immersion, planapochromatic 603 objective (NA 5 1.4 [Nikon]). Changes in [Ca21]c were inferred from the intensity of fluo-3 fluorescence normalized to that in resting conditions (F/F0). Whole-cell currents were monitored with a DAGAN PCONE patch-clamp amplifier. Data were recorded and analyzed with Igor Pro 3.02 (Wave Metrics). Chemicals Coelenterazine n, fura-2AM, and fluo-3 were obtained from Molecular Probes. InsP3 was from Research Biochemicals International. cADPR was obtained from Sigma and from Calbiochem-Novabiochem. Other reagents were of the highest quality available from Sigma or Merck. Results Overlap between Caffeine-, Ryanodine-, Histamine-, and Thapsigargin-sensitive Components of [Ca2 1 ]ER After aequorin reconstitution with coelenterazine, with the ER completely depleted of Ca21, the experiments were started by perfusing the cells with medium containing 1 mM Ca21 to refill the ER (Fig. 1 a). As in other cells studied previously (Montero et al., 1995, 1997a; Barrero et al., 1997; Alonso et al., 1998), full refilling of the ER required 3–5 min and the steady-state [Ca21]ER reached was 500–800 mM (Fig. 1 a). Addition of histamine produced a rapid but partial (60–80%) Ca21 emptying of the ER, a new [Ca21]ER steady-state being reached at z200–300 mM. Subsequent addition of caffeine (50 mM) induced a further emptying to near background aequorin luminescence. The effects were reversible by washing, this allowing refilling of the ER that was completed within 3–5 min. Addition of caffeine at that point, when the stores were completely refilled, triggered a rapid and complete emptying of the ER. Histamine was then unable to produce any further effect. These results suggest that essentially the whole ER Ca21 pool is sensitive to caffeine and a large part of it is also sensitive to InsP3 producing agonists such as histaFigure 1. Effects of histamine and caffeine and pretreatment with ryanodine or thapsigargin on [Ca21]ER. HSV-1–infected chromaffin cells were depleted of Ca21 and reconstituted with coelenterazine n. (a) The ER was refilled by incubation with medium containing 1 mM Ca21, then either 10 mM histamine or 50 mM caffeine were perfused as indicated. (b) Where indicated, ryanodine-pretreated cells were treated before aequorin reconstitution with medium containing 50 mM caffeine and 10 mM ryanodine for 2 min. Cells were then washed and the same treatment was repeated four times at 2-min intervals. Where indicated, thapsigargin-pretreated cells were incubated with 1 mM thapsigargin for 10 min before starting the record. During the experiments, medium containing 1 mM Ca21 and either 50 mM caffeine, 1 mM bradykinin, or 10 mM histamine was perfused as indicated. on March 11, 2005 www.jcb.orgDownloaded from The Journal of Cell Biology, Volume 144, 1999 244 mine. Similar results were obtained using 1 mM bradykinin instead of histamine (data not shown). To obtain more information about the nature of the caffeine-sensitive Ca21 pool, we studied the ability of the ER to refill in the presence of caffeine, ryanodine, or the Ca21ATPase inhibitor thapsigargin. Fig. 1 b, left panel shows that when Ca21-depleted cells were incubated with Ca21containing medium but in the presence of 50 mM caffeine, refilling was almost abolished until caffeine was washed away. The small increase observed in [Ca21]ER in the presence of caffeine was insensitive to histamine. Ryanodine has been reported to lock open irreversibly the RyR in a use-dependent manner, that is, when ryanodine is present while the channels have been opened by caffeine (Ehrlich et al., 1994). If all the ER had functional caffeine and ryanodine-sensitive RyR, we would then expect that pretreatment of the cells with caffeine and ryanodine would inhibit also refilling of the ER with Ca21. Fig. 1 b, middle panel, shows that this is the case. Cells were treated with five pulses of 50 mM caffeine and 10 mM ryanodine, and then the drugs were removed before aequorin reconstitution. Addition of 1 mM Ca21 to these cells produced only an small increase in [Ca21]ER, that was little sensitive to caffeine or bradykinin. The same results were obtained (Fig. 1 b, right panel) if the cells were pretreated with the SERCA inhibitor thapsigargin (1 mM). Similar effects were obtained using lower (20 nM) thapsigargin concentrations (data not shown). Therefore, the whole ER has caffeineand ryanodine-sensitive RyRs, and refills with Ca21 via thapsigargin-sensitive Ca21 pumps. Fig. 2 illustrates the time course of the use-dependent effect of ryanodine on RyR. Fig. 2 a shows that consecutive additions of 50 mM caffeine produced comparable decreases in [Ca21]ER if an interval of 3–5 min was left between two consecutive additions to allow refilling with Ca21 of the ER. If 10 mM ryanodine was present during the caffeine pulses (Fig. 2 b), the first pulse was identical to the control but then the ER became progressively unable to refill. After four pulses, the ER remained at near background [Ca21]ER levels, and only the increase in [Ca21]c elicited by depolarization with 70 mM K1 was able to activate the Ca21 pump and produce a small and transient increase in [Ca21]ER. The effect of caffeine was not inhibited by incubation with 20 mM dantrolene (data not shown), an inhibitor of RyR that is particularly effective on the skeletal muscle RyR1 (Van Winkle, 1976). Quantal Effect of Caffeine The effect of caffeine has been reported to be quantal (Cheek et al., 1993, 1994a), meaning that low caffeine concentrations release only part of the caffeine-sensitive pool. The same phenomenon was observed here. Fig. 3 a shows that addition of submaximal caffeine concentrations induced a rapid but partial emptying of the ER, leading within 30 s to a new lower steady-state of [Ca21]ER. At that point, only the addition of a higher caffeine concentration was able to produce further emptying of the ER. Similar quantal effects were also observed during refilling of the ER when it was carried out in the presence of caffeine. Fig. 3 b shows that the ER did not refill in the presence of 50 mM caffeine, but refilled about halfway once the caffeine concentration was dropped to 5 mM, and completely when caffeine was washed away. Subsequent addition of 5 and 50 mM caffeine released Ca21 and reached the same [Ca21]ER levels obtained during refilling in the presence of these caffeine concentrations. The degree of emptying induced by a particular caffeine concentration was quite reproducible in consecutive additions. Fig 3 c shows that consecutive additions of 5 mM caffeine produced always z50% emptying of the ER, and only the addition of a higher caffeine concentration was able to produce further emptying. Fig. 3 d shows the effect of ryanodine added in the presence of a submaximal dose of caffeine. We can see that the first pulses were identical to the control, but again here the ER refilled progressively more slowly after each new caffeine addition. In this case, in contrast to the experiment shown in Fig. 2, finally a [Ca21]ER steady-state corresponding to about half-filling was reached, in fact, the same [Ca21]ER obtained initially after addition of 5 mM caffeine. Once at this point, addition of a maximal dose of caffeine was required to induce emptying of the remaining portion of the ER. The experiments of Fig. 3 are consistent with previous results obtained looking at [Ca21]c in the same cell prepaFigure 2. Use-dependent inhibition of ER refilling by ryanodine. The ER was refilled by perfusing with medium containing 1 mM Ca21. Then, several consecutive stimulations with 50 mM caffeine were performed as indicated, either in the absence (a) or in the presence (b) of 10 mM ryanodine. In b, standard medium containing 70 mM KCl (replacing an equimolar amount of NaCl) was perfused when indicated (K1). Other details are as in Fig. 1. on March 11, 2005 www.jcb.orgDownloaded from Alonso et al. Ca2 1 -induced Ca2 1 Release Seen from the ER 245 ration by Cheek et al., 1994a. Those experiments led the authors to suggest that there should be different compartments within the ER having different sensitivities to caffeine. This hypothesis would explain why ryanodine only empties the pool sensitive to 5 mM caffeine but leaves the rest of the pool untouched, which could be only released with a higher caffeine dose. However, there is an alternative explanation for these results, based on the regulation of caffeine-induced Ca21 release by the lumenal [Ca21]. In this hypothesis, the results can be explained with only one ER compartment if we assume that submaximal caffeine concentrations can only release Ca21 until [Ca21]ER is reduced to a certain level. The higher the caffeine concentration, the lower the [Ca21]ER level attained. Both alternative hypotheses lead to different predictions under some experimental conditions. In particular, if we obtain halffilled stores by different procedures, e.g., by emptying them with an agonist acting via InsP3 production or by refilling the stores only halfway, the hypothesis of several compartments predicts that we should have all of them half-filled. Therefore, a submaximal dose of caffeine should still release Ca21 from half of them. On the contrary, if the release was directly controlled by [Ca21]ER, we would expect the effect of 5 mM caffeine to be independent of the procedure used to reach that half-filling. The experiments shown in Fig. 4 indicate that the last hypothesis is the correct one. In Fig. 4 a, half-filling was obtained by emptying the ER with histamine. After that, 5 mM caffeine had no effect. Instead, if 5 mM caffeine was added with the ER full of Ca21, it was able to empty it exactly down to the same point. Fig. 4 b shows the effect of halffilling the ER by reducing the time of refilling. Again, the effect of 5 mM caffeine was strictly dependent on the level of [Ca21]ER reached at the point it was added. It had no effect at half-filling, but released 50% of the pool when the Ca21 stores were completely filled. Fig. 4 c shows a similar approach but made in cells preloaded with the Ca21 chelator 1,2-bis(2-aminophenoxy)ethane-N,N,N9,N9-tetraacetic acid (BAPTA) to slow the refilling. We can observe again that the effect of 5 mM caffeine was strictly dependent on the [Ca21]ER at the moment of addition. In addition, this experiment also shows that Ca21 release induced by caffeine requires only resting [Ca21]c. Fura-2 measurements performed in parallel showed that in cells loaded with BAPTA, the [Ca21]c changes induced by caffeine were almost abolished (data not shown). This result points out also that quantal Ca21 release by caffeine is due to the regulation of Ca21 release by the lumenal [Ca21], and suggests that changes in [Ca21]c do not play a major role in the development of the quantal effect. Ca2 1 Entry Activates CICR The next step in this study was to investigate the presence of CICR activated by Ca21 entry through the plasma membrane Ca21 channels. Depolarization with high K1 medium or by stimulation with nicotinic acetylcholine agonists such as 1,1-dimethyl-4-phenyl-piperazinium iodide (DMPP), produced large [Ca21]c peaks (Núñez et al., 1995), that were not significantly modified by previous Ca21 depletion of the ER with caffeine or with the ER Ca21 pump inhibitor thapsigargin (see below). Therefore, the possible contribution of CICR to these [Ca21]c peaks cannot be estimated from conventional [Ca21]c studies, and direct measurement of [Ca21]ER becomes essential. Fig. 5 a shows that 10-s pulses of depolarization with high K1 medium induced a transient Ca21 release from the ER, which could be triggered repetitively by consecutive pulses. The [Ca21]ER decrease was of 60–100 mM (10–15% of the steady-state [Ca21]ER). Therefore, in spite of the large increase in [Ca21]c, the activation of CICR produced a much smaller [Ca21]ER decrease than treatment with cafFigure 3. Quantal response to submaximal concentrations of caffeine. The ER was refilled by perfusing with medium containing 1 mM Ca21 either in the presence (b) or in the absence of caffeine. Then, different concentrations of caffeine were added as indicated, either in the presence or in the absence of 10 mM ryanodine. Other details are as in Fig. 1. on March 11, 2005 www.jcb.orgDownloaded from The Journal of Cell Biology, Volume 144, 1999 246 feine. Increasing the duration of the high K1 pulse increased the magnitude of Ca21 release little. Fig. 5 b illustrates the effect of depolarization with longer (1 min) high K1 pulses. In this case, the first high K1 pulse produced the same effect as in Fig. 5 a, and then [Ca21]ER increased more rapidly, probably as a result of the prolonged stimulation of the ER Ca21 pump by the sustained high [Ca21]c levels. This led to a new [Ca21]ER steady-state at z800–900 mM. After that, subsequent K1 pulses induced somewhat larger [Ca21]ER decreases of 150–200 mM, but corresponding still to only 20% of the steady-state [Ca21]ER. To investigate if increased Ca21 pumping could be responsible for the incomplete Ca21 release, the effect of K1 depolarization was tested in the presence of the ER Ca21 pump inhibitor ciclopiazonic acid (CPA). Fig. 5 c shows that CPA itself induces a slow Ca21 release from the ER, and that simultaneous addition of high K1 medium induced a fast initial Ca21 release of z20% of the [Ca21]ER, followed by a slower release at a rate comparable to that induced by CPA alone. This suggests that CICR induced by a maximal K1 depolarization is able to produce only a decrease of z20% of the [Ca21]ER, even in the absence of Ca21 pumping. CICR, however, was potentiated by simultaneous addition of a low caffeine concentration. Fig. 5 d shows that addition of 1 mM caffeine produced little effect by itself, but strongly potentiated the effect of K1 depolarization, that was now able to release rapidly z50% of the stored Ca21. Finally, CICR could also be triggered in a more physiological way using an agonist for the nicotinic acetylcholine receptor. Addition of DMPP induced a rapid and partial Ca21 release from the ER, very similar to that shown above for K1 depolarization, and which was also potentiated by low concentrations of caffeine (data not shown). The effect of a low caffeine concentration as positive modulator of CICR might reproduce the action of a physiological modulator of this phenomenon. Phosphorylation by cAMP-dependent protein kinase and production of the b-NAD1 metabolite cyclic ADP ribose (cADPR) have been reported to act as physiological modulators for RyR in bovine chromaffin cells (Morita et al., 1997). However, incubation for 3–5 min with the adenylate cyclase activator forskolin (20 mM) had neither any significant effect on the sensitivity to caffeine of Ca21 release nor on the magnitude of high K1 depolarization-induced CICR (data not shown). Regarding cADPR, it has been reported that acetylcholine, high K1 depolarization, and forskolin all stimulate its synthesis by ADP ribosyl cyclase in bovine chromaffin cells (Morita et al., 1997). Therefore, production of this mediator should already be stimulated in our CICR experiments. Nevertheless, to study directly the effect of cADPR on Ca21 release from the ER, we performed experiments in permeabilized cells. Cells were depleted of Ca21 and reconstituted with coelenterazine as usual. Recording of luminescence was started and the cells were permeabilized by perfusion with intracellular-like medium containing 20 mM digitonin and 2 mM EGTA for 1 min. Then, intracellular-like medium containing 100 nM Ca21 (buffered with EGTA) and 2 mM ATP-Mg was perfused. Fig. 6 a shows that [Ca21]ER increased in digitonin-permeabilized cells with very similar kinetics to that found in intact cells after addition of 1 mM extracellular Ca21. The steady-state [Ca21]ER reached was also similar (compare with Fig. 1 a). Fig. 6 a also shows that 2 mM InsP3 and 50 mM caffeine produced a rapid and near complete release of Ca21 from the ER whereas 5 mM cADPR had no effect. Two different commercial sources of cADPR were tested with the same results. In some experiments, cADPR was added in the presence of 1 mM calmodulin and no effect was found either. A possible explanation for the discrepancy among our results and those of Morita et al. (1997) would be that cADPR releases Ca21 from a different (non-ER) Ca21 pool. In fact, these authors report that InsP3 releases Ca21 from a pool sensitive to 20 nM thapsigargin, whereas cADPR and caffeine release Ca21 from a pool only sensitive to .200 nM thapsigargin. In our hands, refilling of the ER was completely inhibited by either 20 nM or 1 mM thapsigargin in both intact (see above) and permeabilized cells (data not shown). Therefore, in order to make compatible our own results and those of Morita et al. (1997), cADPR and caffeine should be able to release Ca21 from an additional non-ER Ca21 pool in the presence of 20 nM thapsigargin. We then performed single-cell Figure 4. Regulation of caffeine-induced Ca21 release by [Ca21]ER. The ER was refilled by perfusing with medium containing 1 mM Ca21 and then 1 mM histamine or different concentrations of caffeine were added, as indicated. In the experiment shown in c, cells were incubated with 10 mM BAPTA-AM during aequorin reconstitution in order to load the cytosol with this Ca21 chelator. Other details are as in Fig. 1. on March 11, 2005 www.jcb.orgDownloaded from Alonso et al. Ca2 1 -induced Ca2 1 Release Seen from the ER 247 fura-2 imaging experiments in intact cells looking at the effect of caffeine on [Ca21]c in cells pretreated with 20 nM thapsigargin. Fig. 6 b shows that addition of caffeine or histamine in Ca21-free medium produced no increase in [Ca21]c under these conditions (the observed decrease is due to the perfusion of Ca21-free medium), whereas K1 depolarization still produced the usual [Ca21]c peak due to Ca21 entry. Effect of Ryanodine on CICR On the basis of the use-sensitive action of ryanodine, illustrated here for caffeine stimulation (Figs. 2 and 3), we should expect also to find a use-sensitive inhibition of ER refilling after repeated stimulation with high K1 pulses inducing CICR. Under this rationale, experiments similar to those shown in Fig. 5, but in the presence of 10 mM ryanodine, were performed. In these experiments we were not able to detect any significant effect of ryanodine on the [Ca21]ER decrease induced by four or five consecutive K1 pulses (data not shown). However, as aequorin consumption limits the sensitivity of the measurement at the end of these experiments, we also decided to perform single-cell imaging experiments looking at the magnitude of the [Ca21]c peak induced by caffeine after several K1 pulses in the presence of ryanodine. Fig 7 a shows that the [Ca21]c increase induced by caffeine was little modified after five consecutive 30-s K1 pulses given in the presence of ryanodine (compare with the initial three caffeine additions). It is interesting to note that the second caffeine addition after the K1 1 ryanodine pulses produced no [Ca21]c increase, even though caffeine was always added in the absence of ryanodine. In fact, perfusion with ryanodine for a short period (30–60 s) before any caffeine addition was effective in promoting the typical use-dependent inhibition of ER refilling on application of caffeine (without ryanodine) 30–60 min later. This means that, even though ryanodine by itself does not produce any apparent effect on ER filling, it remains within the cells after washing and acts later, on stimulation of Ca21 release (see below). Depolarization can also be produced using the nicotinic acetylcholine agonist DMPP instead of K1. Again here, consecutive pulses of 10 mM DMPP in the presence of ryanodine did not have any effect on a later [Ca21]c peak induced by caffeine (data not shown). An alternative and perhaps more physiological depolarizing maneuver is field electric stimulation. Fig. 7 b shows the effect of several consecutive 10-s pulses at 10 Hz, before and after the addition of ryanodine. Again, we can observe that 5 mM caffeine produced the same [Ca21]c peak after and before field electric stimulation. As in Fig. 7 a, consecutive additions of caffeine produced inhibition long after washing of ryanodine. The lack of effect of ryanodine in these experiments may be attributed to the much smaller activation of Ca21 release by high K1-induced Ca21 entry compared with that induced by caffeine. An alternative possibility to activate RyR by Ca21 would be to release Ca21 directly from the ER using an InsP3-producing agonist in order to produce a big increase in [Ca21]c just besides the RyR. In fact, it has been reported that histamine-induced Ca21 release could be inhibited partially with ryanodine after five consecutive stimulation pulses (Stauderman and Murawsky, 1991). In our hands, stimulation by five consecutive pulses of histamine and ryanodine had little effect on the first subsequent [Ca21]c peak obtained by caffeine stimulation (Fig. 7 c). Histamine was always added in the absence of extracellular Ca21, as it also activates Ca21 entry (Cheek et al., 1994b). Figure 5. Activation of CICR by the Ca21 entry elicited by high K1-induced cell depolarization. The ER was refilled by perfusing medium containing 1 mM Ca21. Then, standard medium containing 70 mM KCl was perfused as indicated. In c, 10 mM CPA was added as indicated with or without high K1 medium. In this panel, the [Ca21]ER scale has been normalized as percentage of the maximum [Ca21]ER level in order to facilitate comparison. In the presence of CPA and high K1 medium the initial rate of [Ca21]ER decrease is much higher that with CPA alone, but once [Ca21]ER is below 80% of the initial level, the rates of release in both cases turn similar and can be nearly superimposed. In d, 1 mM caffeine was also added as indicated. Other details are as in Fig. 1. on March 11, 2005 www.jcb.orgDownloaded from The Journal of Cell Biology, Volume 144, 1999 248 Similarly, caffeine was added also in Ca21-free medium to avoid the Ca21 entry induced by caffeine, similar to that reported in GH3 pituitary cells (Villalobos and GarcíaSancho, 1996). As before, the second caffeine stimulation after histamine and ryanodine pulses was completely abolished. Effects of [Ca2 1 ]ER on CICR In spite of the lack of inhibition by ryanodine, the potentiation by caffeine of Ca21 release induced by K1 depolarization (as shown in Fig. 5 d) suggests that CICR may take place through the same RyR activated by caffeine. An important property of the RyR activated by caffeine is the regulation by lumenal [Ca21] shown in Figs. 3 and 4. In the case of CICR, data of Fig. 5 b also suggest that Ca21 release may be stronger at higher lumenal [Ca21] since the first stimulation with high K1, carried out when the ER was only half-filled, was less efficient than the subsequent ones. To obtain more evidence on this point, we investigated the effects of 10-s pulses of high K1/Ca21-containing medium at different [Ca21]ER levels (Fig. 8). At the beginning of the experiment the ER was completely depleted of Ca21 and the first two K1 pulses induced an increase in [Ca21]ER. Since the cells were kept in EGTA-containing medium during the intervals between the K1/Ca21 pulses, refilling of the ER took place only from Ca21 entering into the cells during the pulses. Subsequent K1 pulses produced also [Ca21]ER increase, although progressively smaller, reaching finally a steady-state [Ca21]ER at z300 mM, where the K1/Ca21 pulses had almost no effect. We then completely refilled the ER with Ca21 by perfusing the cells with Ca21-containing medium, thus reaching the usual Figure 6. (a) Effects of cADPR, InsP3, and caffeine on [Ca21]ER in permeabilized cells. Cells were permeabilized by perfusion with 20 mM digitonin for 1 min as indicated. Then the ER was refilled by perfusion with medium containing 100 nM Ca21 (buffered with EGTA). Finally, either 5 mM cADPR, 2 mM InsP3, or 50 mM caffeine were perfused as indicated. (b) Effect of preincubation with 20 nM thapsigargin on the [Ca21]c responses to histamine, caffeine, and high K1 medium. Cells were loaded with fura-2, preincubated for 20 min with 20 nM thapsigargin, and then suspended in standard medium containing 1 mM CaCl2. Then, either 10 mM histamine, 50 mM caffeine, or standard medium containing 1 mM CaCl2 and 70 mM KCl were perfused as indicated. Both histamine and caffeine were perfused in Ca21free medium (containing 100 mM EGTA) to avoid Ca21 entry. Perfusion with Ca21-free medium was started 15 s before and continued for 15 s after stimulation with histamine or caffeine. The trace shown corresponds to the average of 42 cells present in the microscope field. Other details are as in Fig. 1. Figure 7. Effects of ryanodine and either depolarization with high K1 medium (a), field electric stimulation (b), or histamine (c) on the [Ca21]c responses induced by caffeine. Cells were loaded with fura-2 and placed under the microscope in standard medium containing 1 mM CaCl2. Then, different stimuli were given as indicated: different concentrations of caffeine (Caf, in mM), medium containing 70 mM KCl (K1), 10 mM histamine (His), 10 mM ryanodine (Ry), or field electric stimulation (E.S. 10 Hz for 10 s, arrows). Caffeine and histamine were added in Ca21-free medium (containing 100 mM EGTA). Perfusion with Ca21 free medium was started 15 s before and continued for 15 s after the stimuli. Transition from Ca21-containing to Ca21-free medium sometimes produced a small [Ca21]c peak. The traces shown are the average of 36 (a), 49 (b), and 61 (c) cells present in the microscope field. on March 11, 2005 www.jcb.orgDownloaded from Alonso et al. Ca2 1 -induced Ca2 1 Release Seen from the ER 249 [Ca21]ER steady-state levels at z700 mM. At that point, we restarted the protocol of repeated stimulation with 10-s high K1/Ca21 pulses, using EGTA-containing medium for the intervals. The pulses produced now a rapid Ca21 release that could be clearly distinguished from the slower [Ca21]ER decrease induced by the lack of extracellular Ca21. After three K1 pulses, [Ca21]ER was reduced again to z300 mM, and at that point the last pulse produced little effect. This result shows clearly that Ca21 release induced by K1 depolarization, similarly to that induced by caffeine, is strictly dependent on the [Ca21]ER. Differential Inhibition of CICR by Type-specific Ca2 1 Channel Inhibitors We have tested the effect of several inhibitors of Ca21 entry on both refilling of the ER and high K1 depolarizationinduced CICR. A combination of inhibitors of all the voltage-dependent Ca21 channels including 1 or 2 h of preincubation with v-conotoxin GVIA (1 mM), v-conotoxin MVIIC (3 mM), and v-agatoxin IVA (1 mM), together with preincubation and perfusion along the experiment with 3 mM nisoldipine, had little effect on the rate of refilling of the ER with Ca21. Instead, refilling was almost completely blocked by perfusion of 100 mM Cd21 (data not shown). Next, the effects of the inhibitors of voltage-gated Ca21 channels on CICR were tested. For these purposes, we stimulated the cells with high K1 medium in the presence of 1 mM caffeine (as in the experiments shown in Fig. 5 d) in order to potentiate the mechanism and increase the sensitivity of the measurements. The decreases in [Ca21]ER observed, normalized as percentage of those obtained in the controls, were (mean 6 SEM): control, 100 6 11 (n 5 6); 3 mM nisoldipine, 103 6 6 (n 5 5); 1 mM v-conotoxin GVIA, 93 6 6 (n 5 6); 3 mM v-conotoxin MVIIC, 42 6 6 (n 5 7); 1 mM v-agatoxin IVA, 41 6 9 (n 5 5); a combination of all three toxins and nisoldipine, 20 6 6 (n 5 5). Toxins and nisoldipine were preincubated with the cells for 1 or 2 h before the measurements. When the effect of nisoldipine was tested, this inhibitor was also perfused along the experiment. Ryanodine Treatment Modifies the [Ca2 1 ]c Wave Induced by Short Depolarizations as Visualized by Confocal Microscopy Under physiological conditions, cell stimulation is triggered by short depolarizations lasting a few milliseconds. To estimate the contribution of CICR to the Ca21 transient under these conditions, we have compared the rate of diffusion of the Ca21 wave induced by a short (100 ms) cell depolarization both in control cells or in cells in which the Ca21 stores had been blocked by previous treatment with caffeine and ryanodine. We combined the whole-cell patch-clamp technique with fluo-3–based microfluorimetry using a confocal microscope. Cells were line-scanned along 100-ms square depolarizing pulses from a holding potential of 270 to 110 mV. The recorded inward currents showed two typical components: a initial transient peak (INa) followed by a slow inactivating phase (ICa) (data not shown). The ryanodine treatment did not affect the total stimulated Ca21 entry, calculated as the integral of the last 90 ms of the recorded inward current (mean 6 SEM: control cells, 7.15 6 0.42 pC [n 5 34]; ryanodine-treated cells, 6.68 6 1.03 pC [n 5 21]). In spite of this, line scan images representing [Ca21]c showed clear differences between control and ryanodine-treated cells. Fig. 9 a shows the spatiotemporal pattern of [Ca21]c increase in control cells, codified in pseudocolor. [Ca21]c increased first near the plasma membrane and then the Ca21 wave propagated intracellularly. Fig. 9 b shows the results obtained in cells with the Ca21 stores previously emptied by treatment with caffeine and ryanodine. In this case, the [Ca21]c increase was smaller and the propagation of the Ca21 wave delayed. Fig. 9, panels c–e detail the behavior of several parameters that quantify the phenomenon described above in terms of peak [Ca21]c rise (Fig. 9 c), maximum rate of [Ca21]c increase (Fig. 9 d), and time required to increase fluorescence by 10% (Fig. 9 e) at different intracellular locations. Fig. 9 c shows that the maximum fluo-3 fluorescence (indicating the maximum [Ca21] peak) was reached near the plasma membrane. An 80% increase was found in control cells compared with only a 40% increase in ryanodine-treated cells. The fluorescence peaks were smaller as we move deep inside the cell, but the difference among control and ryanodine-treated cells was maintained. Fig. 9 d shows that the maximum rate of fluorescence increase was located near the plasma membrane and decreased steeply as we move into the cell. Again here, the rates were two to three times faster in the control cells than in the ryanodine-treated ones. Fig. 9 e shows the time required for the fluorescence to be increased by 10% at different locations. This parameter is very sensitive to the intracellular propagation of the [Ca21]c wave. We find that the [Ca21]c wave propagates about twice as fast in control cells than in cells treated with ryanodine. These results indicate that CICR significantly contributes to the Ca21 signal induced by cell Figure 8. Effect of [Ca21]ER on CICR induced by depolarization with high K1. Cells depleted of Ca21 and reconstituted with coelenterazine n were placed in the luminometer in 0.5 mM EGTA containing standard medium. Then, 10-s pulses of medium containing 70 mM KCl and 2 mM CaCl2 were given as indicated and 0.5 mM EGTA containing standard medium was perfused during the intervals. After five pulses, Ca21-containing (1 mM) medium was perfused for 3 min to refill the ER, and then the previous protocol was started again. on March 11, 2005 www.jcb.orgDownloaded from