We have previously shown that store-associated microdomains of high Ca2+ are not essential for exocytosis in RBL-2H3 mucosal mast cells. We have now examined whether Ca2+ microdomains near the plasma membrane are required, by comparing the secretory responses seen when Ca2+ influx was elicited by two very different mechanisms. In the first, antigen was used to activate the Ca2+ release–activated Ca2+ (CRAC) current (ICRAC) through CRAC channels. In the second, a Ca2+ ionophore was used to transport Ca2+ randomly across the plasma membrane. Since store depletion by Ca2+ ionophore will also activate ICRAC, different means of inhibiting ICRAC before ionophore addition were used. Ca2+ responses and secretion in individual cells were compared using simultaneous indo-1 microfluorometry and constant potential amperometry. Secretion still takes place when the increase in intracellular Ca2+ occurs diffusely via the Ca2+ ionophore, and at an average intracellular Ca2+ concentration that is no greater than that observed when Ca2+ entry via CRAC channels triggers secretion. Our results suggest that microdomains of high Ca2+ near the plasma membrane, or associated with mitochondria or Ca2+ stores, are not required for secretion. Therefore, we conclude that modest global increases in intracellular Ca2+ are sufficient for exocytosis in these nonexcitable cells.
The intracellular free ionized calcium concentration ([Ca2+]i) is a key physiological signal in stimulus–secretion coupling in both excitable and nonexcitable cells. Theoretical models of Ca2+ diffusion within the cytoplasm of excitable cells predict that [Ca2+]i at release sites can reach hundreds of micromolar due to Ca2+ influx through voltage-gated Ca2+ channels (Chad and Eckert 1984; Neher 1986; Yamada and Zucker 1992). [Ca2+]i reported by diffusible Ca2+ indicators represents the mean average [Ca2+]i and does not reflect [Ca2+]i levels beneath the plasma membrane or near Ca2+ channels. Indeed, these microdomains of high Ca2+ are usually beyond the resolution of currently available methods of Ca2+ measurement (Neher 1998).
However, diverse indirect approaches have yielded insight into the behavior of [Ca2+]i near Ca2+ channels and its effect on the secretory machinery of excitable cells. Some of these approaches have supported theoretical predictions that [Ca2+]i at exocytotic sites can substantially exceed the spatially averaged [Ca2+]i signal measured during depolarization (Augustine and Neher 1992; Llinás et al. 1992; Zucker 1993). Furthermore, localized increases in [Ca2+]i and smaller increases in average [Ca2+]i can be used as two different modes of intracellular signalling, allowing differential activation of Ca2+-dependent physiological processes even in the same cellular compartment (Swandulla et al. 1991; Verhage et al. 1991). In nerve terminals, different modes of Ca2+ signalling allow coupling to secretion of different transmitter classes. For example, a local rise in Ca2+ may trigger secretion of amino acid neurotransmitters, whereas an increase in average [Ca2+]i stimulates secretion of neuropeptides (Verhage et al. 1991). Similarly, in human neutrophils different Ca2+ concentrations are required for the release of different types of granule (Lew et al. 1986; Nüsse et al. 1998).
We have investigated whether microdomains of high Ca2+ play a role in exocytosis in RBL-2H3 mucosal mast cells. In these nonexcitable cells, aggregation of immunoglobulin E receptors on the cell surface by multivalent antigen results in the generation of inositol 1,4,5-trisphosphate, which causes release of Ca2+ from intracellular stores and Ca2+ influx through calcium release–activated Ca2+ (CRAC) channels in the plasma membrane. Therefore, microdomains of high Ca2+ may be generated either by the local release of Ca2+ from intracellular stores or by Ca2+ entry through CRAC channels in the plasma membrane. Although the conductances of inositol trisphosphate receptor channels are very high (∼50 pS) (Bezprozvanny and Ehrlich 1994), which may allow Ca2+ microdomains to be generated near Ca2+ stores, our previous studies demonstrated that store-associated Ca2+ microdomains are not required for exocytosis in RBL-2H3 cells (Kim et al. 1997). The extremely low conductance of the CRAC channel estimated by fluctuation analysis (Zweifach and Lewis 1993) (on the order of fS as opposed to pS conductances for the voltage-gated Ca2+ channels of excitable cells) suggested that very large gradients of Ca2+ were unlikely to be generated close to the plasma membrane. More recently, extrapolation from single-channel recordings of Na+ currents through CRAC channels yielded an estimate of 1.6 pS for the Ca2+ conductance of the CRAC channel in 20 mM Ca2+ (Kerschbaum and Cahalan 1999). If this higher estimate is correct, it may be possible to generate significant microdomains of high Ca2+ near CRAC channels in mast cells.
In this study, we have examined whether Ca2+ microdomains near the plasma membrane are required for exocytosis in RBL-2H3 mucosal mast cells. Our approach has been to compare exocytosis coupled to Ca2+ influx through CRAC channels with that evoked by a uniform increase in intracellular free ionized calcium concentration ([Ca2+]i) using a Ca2+ ionophore (Verhage et al. 1991; Fagan et al. 1996). Since store depletion by Ca2+ ionophore will also activate Ca2+ release–activated Ca2+ current (ICRAC) (Hoth and Penner 1992; Huang and Putney 1998), different means of inhibiting ICRAC before the addition of the ionophore were used. The ability of these two different pathways of Ca2+ influx to elicit Ca2+ responses and secretion was compared in individual RBL-2H3 cells.
Materials and Methods
[Ca2+]i Measurements in RBL-2H3 Cell Suspensions
RBL-2H3 mucosal mast cells (Barsumian et al. 1981) were suspended at 106 cells/ml in a modified Tyrode's solution containing 135 mM NaCl, 5 mM KCl, 1.8 mM CaCl2, 1 mM MgCl2, 5.6 mM glucose, and 10 mM Na-Hepes adjusted to pH 7.4 and supplemented with 250 μM sulphinpyrazone and 0.1% BSA. The suspended cells were sensitized with 1 μg/ml dinitrophenyl-specific immunoglobulin E (provided by Drs. Juan Rivera, National Institutes of Health, Bethesda, MD, and David Holowka, Cornell University, Ithaca, NY) and loaded with 0.5 μM indo-1 acetoxymethyl ester for 45 min at 37°C. The cells were then washed and resuspended in modified Tyrode's solution containing 250 μM sulphinpyrazone and 0.05% gelatin instead of BSA. 3 ml of cell suspension (106 cells/ml) were added to acrylic cuvettes, maintained at 37°C, and constantly stirred. Fluorescence was monitored at 405 nm with a Perkin-Elmer LS-5 fluorescence spectrophotometer (excitation at 365 nm). The [Ca2+]i was calculated as described previously, after correcting for dye leakage during the experiment (Mohr and Fewtrell 1987). The calcium-insensitive indo-1 fluorescence was determined in each experiment, and a Kd of 250 nM for indo-1 was used (Grynkiewicz et al. 1985).
β-Hexosaminidase secretion was carried out in 24-well plates in which sensitized cells had been plated overnight in culture medium. After washing the cells with modified Tyrode's solution containing 0.05% gelatin, secretion was initiated by adding 1 μg/ml antigen (bovine γ-globulin to which 10–20 dinitrophenyl groups per molecule had been coupled; Eisen et al. 1959). ICRAC was inhibited by exposing the cells to various concentrations of SK&F 96365 ∼3 min before antigen addition. Secretion was terminated after 30 min by quenching the wells with ice-cold modified Tyrode's solution. An aliquot from each supernatant was assayed fluorometrically for β-hexosaminidase using 4-methylumbelliferyl-N-acetyl β-d-glucosaminide as the substrate.
Simultaneous Measurements of [Ca2+]i and Serotonin Secretion in Individual RBL-2H3 Mucosal Mast Cells
RBL-2H3 mucosal mast cells plated at low density on glass coverslips were sensitized with antigen-specific immunoglobulin E and incubated overnight at 37°C in culture medium containing 100 μM l-5-hydroxytryptophan, a serotonin precursor (Smith et al. 1995). On the day of the experiment, cells were loaded with 0.5 μM indo-1 acetoxymethyl ester for 30–45 min at 37°C in modified Tyrode's solution containing 0.1% bovine serum albumin and 250 μM sulphinpyrazone (Millard et al. 1989). The coverslip was then mounted in a Teflon chamber on the microscope stage. The cell monolayer was washed and bathed in modified Tyrode's solution containing 0.05% gelatin and 250 μM sulphinpyrazone. The temperature was maintained at 35–37°C throughout an experiment using an open perfusion microincubator and a bipolar temperature controller (Medical Systems Corp.). Cells were stimulated by perfusing the sample chamber with warm solutions of antigen followed by inhibitor and then the Ca2+ ionophore, 4-Br A23187. Indo-1 fluorescence was excited using light from a Xenon lamp filtered at 365 nm and monitored using a Nikon CF 40×/0.85 NA Fluor DL phase objective on a Nikon Diaphot inverted microscope (MVI, Inc.). Emitted light from a single cell was split using a 455-nm dichroic mirror and directed through bandpass filters at 405 and 485 nm to a pair of photomultiplier tubes. Data were acquired and stored using Felix software (PTI, Inc.). The indo-1 ratio (R) was calculated after subtraction of background fluorescence. The [Ca2+]i was determined as described (Kao 1994) according to the following equation, assuming a Kd of 250 nM for indo-1 (Grynkiewicz et al. 1985):
R = F405/F485, Rmax = F405/F485 (with saturating Ca2+), Rmin = F405/F485 (no Ca2+), β = F485 (no Ca2+)/F485 (with saturating Ca2+). F405 indicates fluorescence intensity at 405 nm, and F485 indicates fluorescence intensity at 485 nm. Rmax was determined after adding 5 μM ionomycin to the cell, and Rmin was measured after adding ≥5 mM EGTA.
Constant potential amperometry (Travis and Wightman 1998) was used to detect secretory events from a single cell using a carbon fiber microelectrode. The microelectrode was prepared as described previously (Kim et al. 1997) by inserting an 8 μm diameter carbon fiber into a glass capillary and pulling the glass in two stages using a patch pipette puller. The carbon fiber was sealed to the glass with cyanoacrylate glue and allowed to dry overnight. It was then trimmed as close as possible to the pipette tip, and the capillary was back filled with 3 M KCl. An Ag/AgCl wire was inserted in the capillary to allow electrical connection to the headstage of a GeneClamp 500 amplifier (Axon Instruments, Inc.). The electrode potential was held at 650 mV, and the fiber tip was placed as close as possible (∼5 μm) to the cell surface. Amperometric responses were filtered at 500 Hz, digitized, and stored on videotape for later analysis. Amperometric data were then transferred to Felix at a digitization rate of 25 Hz. Both amperometric and fluorescence data were plotted using StatMost software (DataMost Corp.). Addition artifacts were often observed in the amperometric recordings; these were deleted and replaced with straight lines in the traces.
The following protocol has been used to compare secretion that is elicited in response to Ca2+ influx through CRAC channels (ICRAC) with that seen when Ca2+ is carried into the cell by a Ca2+ ionophore. Since the Ca2+ and exocytotic responses of individual cells are quite heterogeneous, we used real-time measurements of both [Ca2+]i and secretion at the single-cell level using indo-1 fluorescence microscopy and constant potential amperometry (Kim et al. 1997). Each cell was first stimulated with antigen, allowing the generation of inositol 1,4,5-trisphosphate, which depletes stores of Ca2+ and activates ICRAC, and diacylglycerol, which activates protein kinase C. The exocytotic response resulting from the increase in intracellular Ca2+, including any microdomains of high Ca2+ close to the CRAC channel pores, and the concomitant activation of protein kinase C, was recorded for several minutes. A CRAC channel inhibitor was then added to block channels that had been activated in response to antigen and to prevent Ca2+ influx via additional CRAC channels that might be activated when intracellular stores were further depleted in response to the Ca2+ ionophore. Since highly specific CRAC channel inhibitors have yet to be identified, we used three different methods to prevent Ca2+ influx via CRAC channels. The nonfluorescent Ca2+ ionophore 4-Br A23187 (Debono et al. 1981; Deber et al. 1985) was then added to transport Ca2+ randomly across the plasma membrane into the cytosol and generate a diffuse increase in [Ca2+]i (Verhage et al. 1991; Chiono et al. 1995; Fagan et al. 1996, Fagan et al. 1998). The secretory response elicited by this uniform increase in intracellular Ca2+ and the Ca2+ concentration at which it occurred were compared with those seen when CRAC channels were activated by antigen in the same cell.
Effect of SK&F 96365 on Antigen-stimulated [Ca2+]i Responses in Indo-1–loaded RBL-2H3 Cells
The imidazole compound SK&F 96365 is known to inhibit capaciative Ca2+ entry (Merritt et al. 1990). Since SK&F 96365 has not been used before in RBL-2H3 cells, we first investigated its effects on Ca2+ responses and secretion in these cells. Fig. 1 shows typical Ca2+ responses of suspensions of indo-1–loaded RBL-2H3 cells stimulated with an optimum concentration of antigen in the presence or absence of SK&F 96365. In the absence of SK&F 96365, antigen induced a rapid increase in [Ca2+]i, which declined to a stable elevated level due to Ca2+ influx through CRAC channels (Fig. 1 A). When SK&F 96365 was added to RBL-2H3 cells before antigen stimulation, a slow increase in basal [Ca2+]i was observed (Fig. 1 B). This is probably because of some release of Ca2+ from stores by SK&F 96365, as is commonly seen with CRAC channel inhibitors (Cleveland et al. 1993). Antigen stimulation after the addition of SK&F 96365 caused a large but transient Ca2+ response that declined to the somewhat elevated basal level seen after SK&F 96365 addition. This transient increase in [Ca2+]i is due to release from stores. The absence of the elevated sustained component is attributed to inhibition of Ca2+ influx via CRAC channels. To asses the ability of SK&F 96365 to inhibit ICRAC, the drug was added 3–4 min after antigen stimulation to ensure that Ca2+ stores had released their contents and ICRAC was fully activated. Thus, the drop in the [Ca2+]i signal after the addition of SK&F 96365 is due to inhibition of ICRAC. Fig. 1 C shows that SK&F 96365 inhibited ICRAC in a dose-dependent manner with a maximum inhibition at 50 μM.
Antigen-induced Secretion Is Inhibited by SK&F 96365
It is well established that inhibition of Ca2+ influx prevents secretion in RBL-2H3 cells (Mohr and Fewtrell 1987; Fewtrell et al. 1989). Having shown that SK&F 96365 can inhibit ICRAC in RBL-2H3 cells, we examined its effect on antigen-induced secretion of β-hexosaminidase in adherent RBL-2H3 cells. The acid hydrolase β-hexosaminidase is known to be present in mast cell granules (Schwartz et al. 1979), together with histamine and serotonin, and its release is widely used as an assay for exocytosis (Mohr and Fewtrell 1987; Beaven and Ozawa 1996). As expected, SK&F 96365 inhibited antigen-stimulated secretion in a dose-dependent manner (Fig. 1 D). The inhibition of β-hexosaminidase secretion correlated well with the inhibition of ICRAC, with complete inhibition of secretion at 50 μM SK&F 96365. Therefore, it seems likely that inhibition of secretion by SK&F 96365 is due to blockade of Ca2+ entry through CRAC channels.
The Ca2+ Ionophore 4-Br A23187 Is Effective in Increasing [Ca2+]i and Eliciting Secretion when ICRAC Is Inhibited by SK&F 96365
Antigen stimulation of individual RBL-2H3 cells elicited an increase in intracellular Ca2+ as well as a burst of secretory events (Fig. 2). Four typical cell responses are presented. As we have previously shown (Kim et al. 1997), secretion occurred when [Ca2+]i was elevated and, when [Ca2+]i oscillated, secretory events were clustered around the peaks of oscillations. When Ca2+ influx was inhibited by adding 50 μM SK&F 96365, intracellular [Ca2+]i dropped and secretion was immediately halted. When the ionophore 4-Br A23187 was subsequently added, an increase in intracellular Ca2+ and resumption of secretory events were observed. A concentration of 1–2 μM 4-Br A23187 was needed in order to achieve a significant increase in the mean [Ca2+]i after inhibition of ICRAC. These concentrations were not cytotoxic to the cells, as assessed by the retention of intracellular indo-1. In different responsive cells, the average [Ca2+]i that was effective in triggering secretion in response to 4-Br A23187 was in the range of 300–600 nM. This is equal to or less than the effective [Ca2+]i observed in response to antigen at the beginning of the experiment, which served as an internal control. However, when [Ca2+]i reached very high levels (>1,200 nM) during exposure to 4-Br A23187, secretion was never observed. Therefore, secretion is inhibited at extremely high concentrations of Ca2+, which do not usually occur in response to physiological stimuli. One notable feature, however, is that when [Ca2+]i rises rapidly to maximum levels, a burst of secretory events can often be seen on the rising phase (Fig. 2B and Fig. D). This supports the idea that, although an increase in [Ca2+]i is needed for secretion, once [Ca2+]i reaches unphysiologically high levels, the cells stop secreting. Under these conditions, the cells remain viable since there is no loss of intracellular indo-1. At the end of each experiment, saponin lysis released the remaining intracellular serotonin as well as the fluorescent dye indo-1.
4-Br A23187 also Triggers Exocytosis When ICRAC is Inhibited by Depolarizing the Cells in High K+
A residual Ca2+ signal was usually observed after ICRAC inhibition by SK&F 96365 (Fig. 2). Although this might simply be due to an elevation in the resting level of Ca2+, as described earlier (Fig. 1 B), it could also result from incomplete inhibition of ICRAC. Other means of inhibiting Ca2+ influx via CRAC channels were therefore explored. Depolarizing the cells in high potassium is known to inhibit Ca2+ influx (Mohr and Fewtrell 1987), and we confirmed that both the increase in [Ca2+]i and secretory events induced by antigen are inhibited in single RBL-2H3 cells upon depolarization in a high potassium (140 mM) saline solution (Fig. 3). Subsequent addition of 4-Br A23187 increased [Ca2+]i and was effective in triggering additional secretory events. Individual RBL-2H3 cell responses show that depolarization in high K+ does not completely inhibit the Ca2+ response at the physiological Ca2+ concentration used (1.8 mM), whereas exocytotic events were completely abolished (Fig. 3).
Is the Ca2+ Response to 4-Br A23187 Entirely Due to Ca2+ Influx Via the Ionophore?
The residual Ca2+ signal seen in both depolarized and SK&F 96365–treated RBL-2H3 cells could result from incomplete inhibition of ICRAC. This is of concern since subsequent addition of the ionophore 4-Br A23187 might further deplete stores and activate additional CRAC channels. This would then argue against our assumption that the increase in [Ca2+]i is due solely to ionophore-mediated Ca2+ entry across the plasma membrane. In order to test this hypothesis, [Ca2+]i measurements in RBL-2H3 cell populations were performed in the absence of extracellular Ca2+. Fig. 4 A confirms that, after antigen addition, chelation of extracellular Ca2+ with EGTA completely abolishes the sustained increase in Ca2+, which is due to Ca2+ influx. Subsequent addition of 4-Br A23187 elicited a transient increase in [Ca2+]i, which could be due to incomplete Ca2+ release from stores in response to antigen. Indeed, when the Ca2+–ATPase inhibitor thapsigargin was added to antigen-stimulated cells, addition of 4-Br A23187 in the presence of EGTA elicited a much smaller Ca2+ response (Fig. 4 B). Prior exposure of the cells to the mitochondrial inhibitors, antimycin and oligomycin, caused a slight further reduction in the Ca2+ response to 4-Br A23187 after antigen and thapsigargin (Fig. 4 C). This suggests that only a very small part of the Ca2+ response to 4-Br A23187 could be due to release of Ca2+ from mitochondria. Since mitochondria are known to accumulate Ca2+ when the cells are stimulated with antigen (Mohr and Fewtrell 1990) or thapsigargin (Ali et al. 1994), this contribution might be greater if the Ca2+ ionophore was added while intracellular [Ca2+]i was still high, and the mitochondria were loaded with calcium. However, under the conditions used in our experiments (Fig. 2,Fig. 3,Fig. 4,Fig. 5), intracellular [Ca2+]i had returned close to resting levels before the ionophore was added. The mitochondria should therefore have released most of their calcium. The tiny residual response to 4-Br A23187 seen in Fig. 4 C presumably originates from a different population of stores (Pizzo et al. 1997; Huang and Putney 1998).
The transient Ca2+ signal due to release from intracellular stores is not normally sufficient to support secretion in the absence of Ca2+ influx (Mohr and Fewtrell 1987; Kim et al. 1997). Furthermore, we confirmed that in the absence of extracellular Ca2+, 2 μM 4-Br A23187 was not able to evoke exocytosis in cells that had been activated by antigen (data not shown). Although the additional release from stores (Fig. 4) will contribute to the global increase in [Ca2+]i, store-associated microdomains of high Ca2+ should not be generated in response to the Ca2+ ionophore. Furthermore, intracellular calcium release by the Ca2+ ionophore is complete within 1 min (Fig. 4), so it cannot contribute to the secretory events seen several minutes after the addition of 4-Br A23187 in Fig. 2, Fig. 3, and Fig. 5.
Lanthanum Inhibition of ICRAC Does Not Prevent 4-Br A23187 from Inducing Exocytosis
Since we were unable to demonstrate unequivocally that Ca2+ ionophore-induced secretion in the presence of SK&F 96365 or in K+-depolarized cells was not due to Ca2+ influx through a few remaining CRAC channels, a third method of inhibiting ICRAC that completely inhibited Ca2+ influx was needed. It is well established that the lanthanide metal ion La3+ blocks Ca2+ influx (Millard et al. 1989) through CRAC channels in mast cells (Hoth and Penner 1993). Fig. 5 shows that addition of 10 μM La3+ completely abolished the Ca2+ response induced by antigen and interrupted secretion. Subsequent addition of the Ca2+ ionophore 4-Br A23187 generated an increase in [Ca2+]i comparable to or less than that induced by antigen. As a result, another train of secretory events occurred. Secretion was then halted at very high [Ca2+]i levels that exceeded 1,200 nM. These results are consistent with those observed using SK&F 96365 and depolarization in high K+, which showed that the mean [Ca2+]i levels at which secretion occurs in response to the diffuse Ca2+ signal generated by the Ca2+ ionophore are no greater than those measured during exocytosis in response to CRAC channel activation.
Inhibiting Mitochondria Has Little Effect on Secretion
The possibility that mitochondrial Ca2+ sequestration and/or release could affect secretion (Montero et al. 2000) in response to 4-Br A23187 was also explored. Fig. 6 shows the secretory responses of control cells and of cells that had been exposed to antimycin and oligomycin in the continued presence of glucose. Under the latter conditions, ATP levels are maintained by glycolysis, but mitochondrial function is inhibited and mitochondria are no longer able to sequester Ca2+ or synthesize ATP (Mohr and Fewtrell 1990). As expected, inhibition of antigen-induced Ca2+ influx with lanthanum completely abolished β-hexosaminidase secretion from control cells, unless the Ca2+ ionophore 4-Br A23187 was used to carry Ca2+ into the cells (Fig. 6, white bars). Inhibiting mitochondrial function (Fig. 6, black bars) only slightly reduced secretion in response to antigen, which is consistent with the somewhat lower Ca2+ response seen under these conditions (Fig. 4 C). Again, blocking CRAC channels with lanthanum completely inhibited antigen-induced secretion. Furthermore, the Ca2+ ionophore was still able to restore the secretory response in cells whose mitochondria were inhibited and therefore unable to sequester Ca2+ (Fig. 6, black bars). This clearly confirms that secretion in response to 4-Br A23187 is not due to release of Ca2+ from mitochondria.
Ca2+ Dependence of Exocytosis
The Ca2+ dependence of secretion in response to activation of ICRAC by antigen and when Ca2+ influx occurs via the ionophore 4-Br A23187 is shown in Fig. 7. It is clear that the mean [Ca2+]i levels at which secretory events are detected are quite similar for the two pathways of Ca2+ influx. When CRAC channels were activated, the frequency of secretory events peaked between 0.6 and 0.8 μM Ca2+. When influx occurred via the Ca2+ ionophore 4-Br A23187, the frequency of secretion was somewhat greater at all Ca2+ concentrations, and it peaked between 0.4 and 0.6 μM Ca2+. We have previously shown that secretion is not seen in the troughs between Ca2+ oscillations or in cells where Ca2+ oscillates from baseline levels (Kim et al. 1997). It is therefore possible that the enhanced secretion seen with the Ca2+ ionophore occurs because intracellular Ca2+ levels remain elevated and no longer oscillate, as they usually do in response to antigen. Alternatively, we found that the average delay between the initial increase in intracellular Ca2+ and the onset of exocytosis in response to antigen was decreased from 34 to 17 s when protein kinase C was preactivated with phorbol myristate acetate (Kim et al. 1997). If, as these results suggest, protein kinase C is not fully activated during the early part of the Ca2+ response to antigen, this might explain why the frequency of secretion is somewhat higher when the Ca2+ ionophore is subsequently added.
Although [Ca2+]i rarely exceeds 0.8 μM under physiological conditions, it is interesting to note that the frequency of exocytosis, in response to both antigen and Ca2+ ionophore, apparently decreases at these high Ca2+ concentrations (Fig. 7). Furthermore, as noted above and shown in Fig. 2, Fig. 3, and Fig. 5, exocytosis is completely inhibited at the very high [Ca2+]i (>1.2 μM) that are sometimes seen in response to the Ca2+ ionophore.
It is well established that the transient signal due to release of Ca2+ from intracellular stores, in the absence of Ca2+ influx, is unable to trigger exocytosis in RBL-2H3 mucosal mast cells (Mohr and Fewtrell 1987; Kim et al. 1997). However, the precise role played by Ca2+ influx is not yet understood. One possibility is that it is simply required to sustain the global Ca2+ response. Alternatively, it is possible that Ca2+ influx is needed to generate domains of high Ca2+. Many lines of evidence suggest that Ca2+ concentrations of 10–100 μM can be generated transiently near the pores of open voltage-gated Ca2+ channels in excitable cells and that these high Ca2+ concentrations are essential for exocytosis (Augustine and Neher 1992; Llinás et al. 1992; Zucker 1993; Chow et al. 1994; von Gersdorff and Matthews 1994; Bollmann et al. 2000; Schneggenburger and Neher 2000). Since the conductance of the inositol trisphosphate receptor channel (53 pS) (Bezprozvanny and Ehrlich 1994) is much higher than that of the CRAC channel, we initially proposed that Ca2+ microdomains associated with stores might be more critical for secretion in mucosal mast cells. However, our results showed that even when refilling and continued release of Ca2+ from stores was prevented by the Ca2+–ATPase inhibitor thapsigargin, Ca2+ influx through CRAC channels could still trigger secretion, suggesting that store-associated microdomains of Ca2+ are not required for exocytosis (Kim et al. 1997). Furthermore, we have shown that inhibition of Ca2+ uptake and release by mitochondria in stimulated mucosal mast cells (Mohr and Fewtrell 1990; Ali et al. 1994) appears to have little effect on Ca2+ responses (Millard et al. 1989; Fig. 4) and secretion (Mohr and Fewtrell 1990; Fig. 6).
At a membrane potential of −70 mV, the unitary current amplitude associated with a 1.6 pS CRAC channel (Kerschbaum and Cahalan 1999) would correspond to a flux of ∼3 × 105 Ca2+ ions/s. In contrast, Ca2+ ionophores are lipophilic mobile ion carriers that transport Ca2+ randomly across the plasma membrane into the cytoplasm (Pressman 1976; Debono et al. 1981). The turnover number of ionophore-mediated Ca2+ transport was reported to be 45 ions/s for the ionophore A23187 and 22 ions/s for its nonfluorescent 4-bromo derivative (Pfeiffer et al. 1978; Debono et al. 1981). Thus, a Ca2+ ionophore will induce a relatively uniform increase in intracellular Ca2+ with no microdomains of high Ca2+ (Verhage et al. 1991; Fagan et al. 1996). We have therefore compared the exocytotic response elicited as a result of Ca2+ influx through CRAC channels with that evoked by a Ca2+ ionophore.
Since both the Ca2+ responses and secretion in individual mast cells are extremely heterogeneous (Millard et al. 1989; Kim et al. 1997), it was important to carry out these experiments at the single-cell level. This allowed us to compare the secretory response seen when ICRAC is activated by antigen with that seen in the same cell when Ca2+ is carried by the ionophore 4-Br A23187. Both an increase in intracellular Ca2+ (Mohr and Fewtrell 1987; Kim et al. 1997) and activation of protein kinase C (Ozawa et al. 1993; Wolfe et al. 1996) are essential for exocytosis in mucosal mast cells, and it has been suggested that other as yet unidentified signals may also be involved (Beaven et al. 1987; Cunha-Melo et al. 1989). Addition of antigen will therefore generate the full physiological response, including activation of ICRAC, protein kinase C, and any other necessary signals. Furthermore, when Ca2+ influx via CRAC channels is inhibited and the Ca2+ ionophore is added, the other signals generated in response to antigen should remain activated. Since there are no highly selective inhibitors of CRAC channels, three different pharmacological approaches were used to inhibit ICRAC in antigen-stimulated mucosal mast cells. All three manipulations, SK&F 96365, potassium depolarization, and lanthanum, blocked the Ca2+ influx and secretion seen in response to antigen. However, in all cases diffuse Ca2+ entry via the ionophore 4-Br A23187 was effective in restoring exocytosis. These results clearly suggest that secretion is not tightly coupled to Ca2+ entry through CRAC channels and that it is the global Ca2+ concentration that is being sensed by the exocytotic machinery in mast cells.
Several studies have demonstrated the validity of the approach of using a Ca2+ ionophore to generate a uniform increase in intracellular Ca2+ with no microdomains of high Ca2+. The first of these compared the abilities of voltage-sensitive Ca2+ channel activation and the Ca2+ ionophore, ionomycin, to evoke secretion of synaptic vesicles and dense core vesicles from isolated nerve terminals (Verhage et al. 1991). For a given increase in mean [Ca2+]i, the Ca2+ ionophore was much less effective than Ca2+ channel activation in releasing amino acids from synaptic vesicles, whereas neuropeptide release occurred at similar global Ca2+ concentrations with both stimuli. It was therefore concluded that exocytosis of amino acid–containing synaptic vesicles requires the higher Ca2+ concentrations generated in the vicinity of Ca2+ channels in the active zone, whereas neuropeptide secretion from large dense core vesicles is triggered by much smaller elevations in Ca2+ in the bulk cytoplasm. In other studies (Chiono et al. 1995; Fagan et al. 1996, Fagan et al. 1998), it has been shown that Ca2+ influx via capacitative Ca2+ entry channels is required to regulate the activity of several Ca2+-regulated adenylyl cyclases in glioma cells, whereas Ca2+ signals generated by a Ca2+ ionophore or by release of stored Ca2+ were ineffective. It was therefore proposed that Ca2+-sensitive adenylyl cyclases are functionally colocalized with capacitative Ca2+ entry channels and that their activity is regulated by microdomains of high Ca2+ associated with these channels (Cooper et al. 1995).
The localized nature of Ca2+ microdomains associated with voltage-gated Ca2+ channels in neuronal cells led to the suggestion that these domains of [Ca2+]i may trigger secretion from synaptic vesicles in close proximity or tightly coupled to these channels (Matthews 1996; Bennett 1997). The fast time course of secretion at synapses was also attributed to the tight association between vesicles and Ca2+ channels (Simon and Llinas 1985; Adler et al. 1991; Heidelberger et al. 1994). This is in contrast to the slower time course of secretion in neuroendocrine cells, where vesicles and channels are not strictly colocalized (Chow et al. 1994). By analogy, the very slow time course of secretion in mucosal mast cells, and the ability of diffuse Ca2+ entry to elicit secretion in the absence of ICRAC, is most likely to reflect spatial separation between the sites of Ca2+ entry and exocytosis in mast cells. This is consistent with the fact that mast cell granules are distributed throughout the cytoplasm and are not clustered at the plasma membrane. Although we have shown that microdomains of high Ca2+ are not required for exocytosis, this does not rule out the possibility that Ca2+ microdomains associated with CRAC channels (Zweifach and Lewis 1995) are important in regulating other cell functions (Chiono et al. 1995; Fagan et al. 1998) in mast cells.
The [Ca2+]i at which secretion is greatest in mucosal mast cells is well below 1 μM both when Ca2+ enters the cell through CRAC channels and when it enters via the Ca2+ ionophore 4-Br A23187 (Fig. 7). Consistent with this finding, it has been shown that similarly low concentrations of Ca2+ support secretion in washed permeabilized RBL-2H3 cells, provided that protein kinase C is replenished and activated (Ozawa et al. 1993; Beaven and Ozawa 1996). These concentrations of Ca2+ are remarkably similar to those required for the exocytosis of large dense core vesicles from nerve terminals (Verhage et al. 1991). In neutrophils, the Ca2+ dependence of exocytosis depended on the type of granule being secreted and the way in which the Ca2+ dependence was measured (Lew et al. 1986; Nüsse et al. 1998). In all cases, half-maximal Ca2+ concentrations were significantly higher, ranging from 1 to 100 μM Ca2+. However, neutrophils, like mast cells, require other synergistic signals, in addition to Ca2+, for optimal secretion (Lew et al. 1986; Nüsse et al. 1997). Since Ca2+ was the only signal provided in these studies, the lack of additional signals (Lew et al. 1986) or their loss during microperfusion of the cell (Nüsse et al. 1998) may explain why higher [Ca2+]i were required for exocytosis in neutrophils.
Current models for exocytosis suggest that synaptic vesicles have developed low-affinity Ca2+ sensors that respond to locally high increases in [Ca2+]i generated near the mouths of Ca2+ channels at the active zone of a nerve terminal. Our results, obtained using a Ca2+ ionophore with single mucosal mast cells, suggest that granule exocytosis occurs at Ca2+ concentrations between 0.3 and 0.8 μM. Furthermore, these Ca2+ concentrations are very similar to those at which secretion in response to antigen was seen in the same cells, before ionophore addition, and also in our earlier studies (Kim et al. 1997). This suggests that the secretory granules of mast cells are spatially well separated from CRAC channels and respond to the global [Ca2+]i rather than microdomains of high [Ca2+]i associated with CRAC channels. They would therefore be expected to have a much higher affinity Ca2+ sensor than synaptic vesicles.
It has been shown that mast cells express three different isoforms of the Ca2+-binding protein synaptotagmin (Baram et al. 1999), which has been proposed as the Ca2+ sensor for exocytosis (Südhof and Rizo 1996). Synaptotagmin II, the most abundant isoform in mast cells, is associated with lysosomes rather than with secretory granules (Baram et al. 1999). The low sensitivity (200 μM) of synaptotagmin II for Ca2+ (Südhof and Rizo 1996) would be consistent with the high Ca2+ requirement for lysosomal exocytosis (Rodríguez et al. 1997; Baram et al. 1999). Synaptotagmin III, which has a high sensitivity (<1 μM) for Ca2+ (Südhof and Rizo 1996), is another of the isoforms identified in mast cells. It is not known whether it is associated with the secretory granules, but the relatively low cytoplasmic Ca2+ concentrations (0.3–1 μM) at which most granule exocytosis occurs (Kim et al. 1997; this paper) make synaptotagmin III an attractive candidate for the Ca2+ sensor for exocytosis in mast cells.
In summary, we have shown that microdomains of high Ca2+ associated with CRAC channels are not essential for exocytosis in RBL-2H3 mucosal mast cells. Therefore, both release from stores and influx across the plasma membrane can contribute to the global increase in [Ca2+]i that leads to exocytosis. Furthermore, it is clear that the optimal [Ca2+]i for secretion in these cells is <1 μM. This is in contrast to neuronal cells, where microdomains of very high [Ca2+]i (10–100 μM) associated with voltage-dependent Ca2+ channels are required for the rapid release of neurotransmitters at the neuromuscular junction and other synapses. Our findings are consistent with the fact that exocytosis in mast cells occurs relatively slowly. Furthermore, mast cell granules are distributed throughout the cytoplasm, and there is no evidence for active zones of secretion where granules might be in very close proximity to CRAC channels. The relatively low [Ca2+]i at which secretion occurs in mast cells may also explain why concomitant activation of protein kinase C is also necessary (Ozawa et al. 1993; Wolfe et al. 1996), whereas a large increase in [Ca2+]i is sufficient for secretion in excitable cells.
We thank Drs. Juan Rivera and David Holowka for gifts of immunoglobulin E.
This work was supported by grants from the National Science Foundation.
Sahar Mahmoud's current address is Department of Neurobiology and Behavior, Cornell University, W 149 Mudd Hall, Ithaca, NY 14853.
Abbreviations used in this paper: [Ca2+]i, intracellular free ionized calcium concentration; CRAC, calcium release–activated Ca2+; ICRAC, Ca2+ release–activated Ca2+ current.