Vesicular Ca(2+) participates in the catalysis of exocytosis.

Effects of vesicular monoamine transporter inhibitors on catecholamine release from bovine chromaffin cells have been examined at the level of individual exocytotic events. As expected for a depletion of vesicular stores, release evoked by depolarizing agents was decreased following 15-min incubations with reserpine and tetrabenazine, as evidenced by a decrease in exocytotic frequency and amount released per event. In contrast, two reserpine derivatives, methyl reserpate and reserpic acid, were much less effective. Surprisingly, the incubations also decreased the accompanying rise in intracellular Ca(2+) evoked by depolarizing agents. Subcellular studies revealed that reserpine and tetrabenazine at concentrations near their K(i) values not only could increase cytoplasmic catecholamines but also could displace Ca(2+) from vesicles. Furthermore, transient exposure to tetrabenazine and reserpine, but not methyl reserpate and reserpic acid, induced exocytotic release of catecholamines. Reserpine induced a rise in intracellular Ca(2+), as detected by whole-cell measurements with Fura-2. It could induce exocytosis, albeit at a lower frequency, in Ca(2+)-free solutions, supporting an internal Ca(2+) source. Depletion of endoplasmic reticulum and mitochondrial Ca(2+) pools did not eliminate the reserpine-activated release. These results indicate that vesicular Ca(2+) can play an important role in exocytosis and under some conditions may be involved in initiating this process.

Effects of vesicular monoamine transporter inhibitors on catecholamine release from bovine chromaffin cells have been examined at the level of individual exocytotic events. As expected for a depletion of vesicular stores, release evoked by depolarizing agents was decreased following 15-min incubations with reserpine and tetrabenazine, as evidenced by a decrease in exocytotic frequency and amount released per event. In contrast, two reserpine derivatives, methyl reserpate and reserpic acid, were much less effective. Surprisingly, the incubations also decreased the accompanying rise in intracellular Ca 2؉ evoked by depolarizing agents. Subcellular studies revealed that reserpine and tetrabenazine at concentrations near their K i values not only could increase cytoplasmic catecholamines but also could displace Ca 2؉ from vesicles. Furthermore, transient exposure to tetrabenazine and reserpine, but not methyl reserpate and reserpic acid, induced exocytotic release of catecholamines. Reserpine induced a rise in intracellular Ca 2؉ , as detected by whole-cell measurements with Fura-2. It could induce exocytosis, albeit at a lower frequency, in Ca 2؉ -free solutions, supporting an internal Ca 2؉ source. Depletion of endoplasmic reticulum and mitochondrial Ca 2؉ pools did not eliminate the reserpine-activated release. These results indicate that vesicular Ca 2؉ can play an important role in exocytosis and under some conditions may be involved in initiating this process.
Vesicles in neurons and secretory cells frequently contain large amounts of Ca 2ϩ (1)(2)(3)(4). For example, the vesicles of chromaffin cells isolated from the adrenal medulla contain nearly 60% of the total Ca 2ϩ contained within the cell (5). Vesicular Ca 2ϩ is incorporated during vesicle biogenesis at the Golgi-endoplasmic reticulum apparatus (6,7). Furthermore, vesicles can take up Ca 2ϩ via a Na ϩ /Ca 2ϩ exchanger. Calcium is strongly retained in chromaffin vesicles (8) because it is associated with the chromogranins, a group of acidic, soluble proteins. Chromogranins also associate with catecholamines, and this composite association allows maintenance of vesicular stores within the small volume of the vesicle.
The high concentration of vesicular Ca 2ϩ is of interest since Ca 2ϩ plays a central role in initiating the fusion between the vesicle and plasma membrane during exocytosis. Vesicular Ca 2ϩ is located at the precise location where exocytosis is occurring. If it can be mobilized, it would clearly play a major role in modulating release. Lowering of vesicular Ca 2ϩ decreases exocytotic rates in both pancreatic cells (1) and acetylcholine-containing neurons (2). These results suggest that vesicular Ca 2ϩ may frequently play a role in triggering exocytosis. In a prior work, we have shown that vesicular Ca 2ϩ in bovine chromaffin cells can be mobilized by exposure to agents that cause an alkalinization of the vesicle interior (9). Normally, the intravesicular pH is 5.5, a value that promotes the association of the intravesicular contents (7). Alkalinization of isolated vesicles with methylamine caused Ca 2ϩ efflux. Similar results were found with amphetamine exposure, a phenethylamine that can alkalize vesicles (10). In intact cells, these agents increased both Ca 2ϩ and catecholamine in the cytoplasm. With amphetamine, the amount of Ca 2ϩ displaced from the vesicles was sufficient to trigger exocytosis.
In this work, we have taken a different approach to disrupting association vesicular contents. We have examined the effects of inhibitors of the vesicular monoamine transporter on Ca 2ϩ mobilization from vesicles and its effects upon release. Reserpine, tetrabenazine, and two reserpine derivatives were employed, all of which are specific inhibitors of catecholamine transport (11)(12)(13)(14). The results indicate that VMAT 1 inhibitors are capable of mobilizing vesicular Ca 2ϩ in addition to catecholamine, that the effects of this mobilization are reflected in altered characteristics of stimulated secretion, and that inhibitors with high affinity for VMAT can trigger exocytosis in the absence of any other secretagogue.

EXPERIMENTAL PROCEDURES
Cultured Chromaffin Cells-Primary cultures of bovine adrenal chromaffin cells were prepared as described previously (15). Briefly, chromaffin cells were isolated from several bovine adrenal glands by digestion with collagenase and density gradient centrifugation through Renografin to obtain the epinephrine-enriched fraction. Cells were plated at a density of 3 ϫ 10 5 cells/35-mm diameter plate and maintained in a controlled atmosphere with 5% CO 2 in air at 37°C. Experiments were performed during days 3 through 7 of culture.
Electrodes and Electrochemical Procedures-Microelectrodes were fabricated by sealing 6-m diameter carbon fibers (Thornel T-650, Amoco Corp., Greenville, SC) in glass as described previously (16). The electrodes were polished at a 45°angle on a diamond dust-embedded micropipette-beveling wheel (Sutter Instrument Co., Novato, CA) and soaked in isopropyl alcohol before use. Electrodes were calibrated using a flow-injection apparatus. Amperometry employed an Axopatch 200B (Axon Instruments, Foster City, CA) in the voltage clamp mode with the whole cell configuration (␤ ϭ 1). The low pass filter was set to 10 kHz, and the gain was 10 mV/pA for Ca 2ϩ -containing buffers and 20 mV/pA for all others. The applied potential was ϩ650 mV with respect to a sodium-saturated calomel electrode. The amperometric signal was digitized (PCM-2, Medical Systems Corp., Greenvale, NY) and stored on video cassette recorder tape.
Amperometric data recorded at cells were played back into a PC with commercially available hardware and software (Cyberamp 320 and Axotape, Axon Instruments). Normally, current records were digitally * This research was supported by National Institutes of Health Grant 1R01 NS38879. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ To whom correspondence should be addressed. Tel.: 919-962-1472; Fax: 919-962-2388; E-mail : rmw@unc.edu. filtered at 400 Hz before analysis. Individual spikes (S/N ϭ 5) were located by locally written software. However, after 15-min incubations with VMAT inhibitors, spikes were sufficiently smaller so that 10-Hz filtering was required to detect the spikes. This over-filtering does not affect the area of each spike (Q), a measure of the quantal size. Data sets were analyzed by single factor ANOVA for statistical significance.
Single Chromaffin Cell Experiments-Before the experiments, the cells were washed three times and then maintained in a pH 7.4 buffer on the culture plate. The primary buffer contained 145 mM NaCl, 5 mM KCl, 11.2 mM glucose, 10 mM HEPES, 2 mM CaCl 2 , and 1 mM MgCl 2 ⅐6 H 2 0. Chemical stimulations and incubations of the cells were performed with the desired agents dissolved in this solution for the Ca 2ϩ -containing experiments. Ca 2ϩ -free experiments were performed in the primary buffer that was made without the addition of CaCl 2 . In every case, the application of buffers alone via pressure ejecting did not elicit release of catecholamine from the cells nor did it cause changes in [Ca 2ϩ ] i .
The culture plate was placed on the stage of an inverted microscope (Axiovert 35, Zeiss, Eastern Microscope, Raleigh, NC). A pressure ejection device (Picospritzer, General Value Corp., Fairfield, NJ) connected to a micropipette with a 10-m diameter tip was used to apply secretagogues onto individual cells. The carbon-fiber microelectrode and micropipette were mounted on micromanipulators (Patch Clamp Driver PCS-250, Burleigh, Instruments Inc., Fishers, NY). The micropipette was positioned 25-35 m from the cell, and the microelectrode was positioned firmly against the cell surface. In experiments that involved incubation of the cells, cells were first tested by pressure ejection of nicotine (typically 4 times for 4 -5 s with 2 min between each ejection) followed by the addition of drug and incubation for 15 min. For experiments involving the transient exposure of cells to drugs, the drug of interest was applied via pressure ejection for 4 to 5 s. Cytoplasmic catecholamine levels were determined by incubating the cells with the indicated drug for 15 min and then permeabilizing it via the pressure ejection of 10 M digitonin (17).
In Ca 2ϩ -free buffer, pressure ejection of 10 mM caffeine was used to test the exocytotic viability of the cells (18). To chelate intracellular Ca 2ϩ , the cells were incubated in a Ca 2ϩ -containing buffer containing 50 M BAPTA-AM for 10 min (procedure from Ref. 19). Cells were then rinsed with a Ca 2ϩ -free buffer and maintained in that solution while agents were pressure-ejected onto the cells. To investigate thapsigargin-sensitive stores, cells were permeabilized by incubation with 1 M digitonin in Ca 2ϩ -free buffer for 2 to 5 min followed by a 5-min incubation with 100 nM thapsigargin in the Ca 2ϩ -free buffer. The cells were then rinsed, and release was tested in Ca 2ϩ -free buffer (the procedure was modified from Ref. 20). To investigate Ca 2ϩ stores in mitochondria, cells were permeabilized by incubation with 1 M digitonin for 4 min to allow entry of 50 M ruthenium red. After 5 min, the cells were stimulated via pressure ejection (procedure modified from Ref. 3).
Fura-2 Ca 2ϩ Measurements-Fura-2 AM was dissolved in a 10 l of 10% Pluronic® F-127 and 40 l of dry Me 2 SO mixture. Chromaffin cells were incubated in a buffer solution containing 1 M Fura-2 AM and 0.1% bovine albumin for 20 min at 22°C. The cells were then rinsed with fresh buffer, and the Fura-2 AM was allowed to de-esterify for 20 min followed by two additional washes with buffer.
Whole-cell fluorescence measurements were made with a pinhole that allowed light from a 27-m diameter spot on the plate to be detected by a photomultiplier tube (EMPIX Photometer System, Mississaugua, Canada). Cells were alternately excited at 340 Ϯ 7.5 and 380 Ϯ 7.5 nm, whereas emission at 510 Ϯ 10 nm was measured. Light was collected through a 40ϫ oil immersion objective (Carl Zeiss Fluar 40ϫ, Thornwood, NY). To reduce photobleaching, excitation light from the Xenon arc lamp passed through a 0.5 neutral density filter (30% transmission) positioned between the arc lamp and the 40ϫ objective. Ratios of F 340 /F 380 were determined at 750-ms intervals.
Ca 2ϩ in Isolated Vesicles-Vesicles from epinephrine-containing cells were isolated and purified as described previously (9,21). The vesicle pellet was resuspended in 70 ml of Ca 2ϩ -free isotonic buffer without a chelating agent. The vesicles were incubated with the desired compound for 10 min and centrifuged for 5 min, and pelleted vesicles were resuspended in the Ca 2ϩ -free isotonic buffer. Calcium levels were determined by atomic absorption spectroscopy (Instrumental Laboratory S-12 AA/AE Spectrophotometer, ϭ 422.7 nM; the gas mixture was approximately 5:1 air to acetylene). LaCl 3 was added to obtain a final concentration of 1% (w/v) to minimize interference by anions such as phosphates. Standard additions of Ca 2ϩ were utilized to determine the mol of Ca 2ϩ /mg of protein.
Materials-Nicotine and BAPTA-AM were obtained from RBI (Natick, MA). Fura-2 AM and Pluronic® F-127 were purchased from Molecular Probes (Eugene, OR). Tetrabenazine was from Fluka (Buchs, Switzerland). Methyl reserpate and reserpic acid were prepared as described (22). Penicillin/streptomycin was obtained from the Lineberger Cancer Research Center (Chapel Hill, NC). The chromaffin cell culture medium was Dulbecco's modified Eagle's medium/Ham's F-12 medium from Life Technologies, Inc. Collagenase was acquired from Worthington Chemicals (Freehold, NJ). Renografin-60 was from Squibb Diagnostics (New Brunswick, NJ). Dimethyl sulfoxide was purchased from Mallinckrodt Chemical Works. All other chemicals were purchased from Sigma. All solutions were prepared in distilled, deionized water (Mega Pure System MP-3A, Corning Glass).

Effects of Incubation with VMAT Inhibitors on Nicotineevoked Catecholamine
Release-To evaluate the effects of incubation with VMAT inhibitors on subsequent release, nicotine was employed as the secretagogue. Individual cells were stimulated 4 times with 50 M nicotine (with 2-min intervals between each 5-s stimulation), and catecholamine release was monitored. The drug of interest was then added to the buffer. After 15 min, another round of nicotine stimulations was performed as described above. In the absence of drug, no spikes were observed in the 15-min interval between nicotine stimulation (Fig. 1a), and the second series of nicotine stimulations evoked a similar spike frequency and average quantal size to that obtained before the 15-min interval (Fig. 2). In contrast, incubation with 90 nM reserpine evoked spikes during the first 8 to 10 min of incubation, and the frequency of spikes was lower (Fig. 1B). Analysis of the spikes revealed that this incubation significantly lowered the average quantal size as well as the spike frequency evoked by the second nicotine exposure (Fig.  2).
Similar experiments were conducted with a lower dose of reserpine, with tetrabenazine, another VMAT2 inhibitor at two different doses, and with structural analogues of reserpine, methyl reserpate and reserpic acid. The latter two compounds did not evoke spikes or alter subsequent depolarization-induced release at doses up to 500 nM. Even at 1 M they only had slight effects. However, at both concentrations tested, reserpine and tetrabenazine elicited release during the first few minutes of the incubation period (see below). Additionally, the frequency of nicotine-evoked spikes and mean quantal size ( Fig. 2) were significantly reduced following incubation with all VMAT inhibitors. Incubation effects with amphetamine were also examined. As with the VMAT inhibitors, it induced release during the incubation interval, diminished quantal size, and decreased the frequency of subsequent nicotine-evoked release (Fig. 2).
Effect of Incubation with VMAT Inhibitors on Intracellular Ca 2ϩ Response to K ϩ Stimulation-The effect of incubation with VMAT inhibitors and amphetamine on [Ca 2ϩ ] i was examined by monitoring the fluorescence ratio of intracellular Fura-2, alternately excited at two wavelengths (340 and 380 nm). In this case, 60 mM K ϩ was used as a secretagogue, an agent that causes robust changes in the fluorescence ratio, indicating elevated intracellular Ca 2ϩ (23). For cells that were not exposed to VMAT inhibitors, the rise in [Ca 2ϩ ] i evoked by 60 mM K ϩ was unchanged compared with that seen before the 15-min incubation (Fig. 2C). In contrast, the rise in [Ca 2ϩ ] i was diminished on the second K ϩ exposure following incubation with reserpine and tetrabenazine at the doses shown (Fig. 2C). The decreased frequency of evoked spikes after incubation correlates very well with the decreased [Ca 2ϩ ] i evoked by a depolarizing stimulus. Identical results were obtained with amphetamine (Fig. 2C).
Effect of VMAT Inhibitors on [Ca 2ϩ ] in Isolated Vesicles-To explore the effects of reserpine, tetrabenazine, reserpic acid, and methyl reserpate on [Ca 2ϩ ] vesicle , the Ca 2ϩ content of isolated vesicles was examined by atomic absorption spectroscopy after exposure to these inhibitors. Isolated vesicles were maintained in Ca 2ϩ -free buffer with or without VMAT inhibitors for a total of 15 min. Ca 2ϩ -free buffer was utilized to replicate the extremely low amounts of Ca 2ϩ in the cytoplasm of the cell. The vesicular content found, 124.9 nmol of Ca 2ϩ /mg of protein, is quite similar to those found previously for untreated vesicles: 125 Ϯ 15 nmol of Ca 2ϩ /mg of protein (24) and 112 Ϯ 6.3 nmol of Ca 2ϩ /mg of protein (25). Reserpine (9 and 90 nM) as well as tetrabenazine (50 nM and 10 M) caused a decrease in [Ca 2ϩ ] vesicle (Fig. 3A). These results indicate that reserpine and tetrabenazine, like amphetamine (9), are capable of causing Ca 2ϩ displacement during extended exposure. Methyl reserpate and reserpic acid, both at 1 M, did not cause significant changes in the Ca 2ϩ content of isolated vesicles.
FIG. 1. Exocytotic rate of single chromaffin cells exposed to nicotine (NIC) at 15-min intervals measured by amperometry. Initially, each cell was exposed 4 times to a 5-s pressure ejection of 50 M nicotine. In A, the cell was left untreated for 15 min and stimulated again with nicotine. The inset is a representative amperometric trace induced by nicotine. In B, 90 nM reserpine (RES) was added to the culture plate after the initial nicotine exposure, and then it was exposed to nicotine again after 15 min. Individual spikes were summed in 30-s interval bins.

FIG. 2. Effects of VMAT inhibitors on quantal size, spike frequency, and ⌬[Ca
2؉ ] i . Cells were stimulated with nicotine before and after a 15-min incubation period as described in Fig. 1. During the incubation period, the drug indicated was present in the buffer. Amperometric spikes were characterized with respect to quantal size (Q) and spike frequency in the minute immediately after nicotine application (spikes/minute). The data shown in A are the average quantal size after the 15-min interval normalized by the values obtained initially. In this way, each cell serves as its own control. In B, the data is the average spike frequency normalized in the same way. A minimum of 450 spikes after drug application were used in this analysis. Measurements of internal Ca 2ϩ were made with Fura-2 during exposures to K ϩ before and after a 15-min interval in which the drug indicated was placed in the buffer are shown in C. All results are the average ratios from at least four cells, and the error bars represent the S.E. All ratios with added drug are statistically significant from the control as analyzed by ANOVA (p Ͻ 0.04, degrees of freedom Ͼ10). The mean spike characteristics evoked by 50 M nicotine were typical of previous reports (20 ms, 1.42 pC (corresponding to 4.4 ϫ 10 6 molecules), and 80 pA, for t1 ⁄2 , quantal size, and amplitude, respectively, when filtered at 400 Hz). RES, reserpine; TBZ, tetrabenazine; AMPH, amphetamine.
Cells Incubated with VMAT Inhibitors Increased Cytoplasmic Catecholamine Content-Cells were incubated with the drug of interest in Ca 2ϩ -containing buffer for 15 min. To evaluate the effects of VMAT inhibitors on cytoplasmic catecholamine levels, the cells were then transferred to Ca 2ϩ -free buffer and permeabilized with 10 M digitonin delivered from a pressure ejection pipette for 5 s. In some cases, multiple exposures to digitonin were required to induce membrane permeabilization. Under these conditions, the cytoplasmic contents of the cell can leak out, and the electroactive species can be detected with an adjacent carbon-fiber electrode (26). The digitonin-induced efflux of cytoplasmic catecholamine is prolonged over several seconds, quite different from the sharp concentration spikes observed during exocytosis. When untreated cells were permeabilized with digitonin, no cytoplasmic catecholamines were detected (9). However, measurable cytoplasmic catecholamines were present in cells incubated with both doses of tetrabenazine and reserpine (Fig. 3B). For reserpine and tetrabenazine, the detected amount of cytoplasmic catecholamine was concentration-dependent. For reserpine, a 9 nM concentration (n ϭ 5) resulted in (5.52 Ϯ 1.65) ϫ 10 7 molecules being detected, and a 90 nM concentration (n ϭ 4) released (8.03 Ϯ 0.21) ϫ 10 7 molecules. For tetrabenazine, a 50 nM concentration (n ϭ 4) resulted in (5.46 Ϯ 1.09) ϫ 10 6 molecules being detected, and a 10 M concentration (n ϭ 4) released (2.47 Ϯ 0.43) ϫ 10 7 molecules. Note the strong correlation between amounts of Ca 2ϩ and catecholamine displaced.
Exocytosis Induced by Transient Exposure to VMAT Inhibitors-Reserpine pressure ejected onto intact cells for 5 s evoked exocytosis with a frequency that was concentration-dependent (Fig. 4). Concentrations of tetrabenazine lower than 10 M did not elicit release (Fig. 4). To determine if the lipophilicity of reserpine was a factor in inducing the release of vesicles from chromaffin cells, two derivatives were employed: 1 M methyl reserpate and 1 M reserpic acid. These drugs are less effective than reserpine in inhibiting transport of catecholamines by VMAT. Neither is a substrate for transport by VMAT, and both are relatively impermeant when compared with reserpine (12,27). The reserpine derivatives (1 M) generated only one or two spikes when pressure-ejected onto the cells. Most cells did not respond to this transient exposure to the reserpine derivatives.

Effect of Transient Exposure to VMAT Inhibitors on [Ca 2ϩ ] i -
The effect of pressure ejection of the VMAT inhibitors on [Ca 2ϩ ] i was examined by monitoring the fluorescence of intracellular Fura-2 alternately at two excitation wavelengths (340 and 380 nm). Transient exposure to 90 nM reserpine caused a brief increase in [Ca 2ϩ ] i (Fig. 5A), although the change was lower than that evoked by the pressure ejection of 60 mM K ϩ in Ca 2ϩ -containing medium. Ca 2ϩ effects induced by tetrabenazine (Fig. 5B), methyl reserpate, and reserpic acid were not detectable. Additionally, reserpine-evoked increases in [Ca 2ϩ ] i were not observed in Ca 2ϩ -free medium.
External Ca 2ϩ Requirement for Reserpine-induced Exocytosis-To test whether external calcium was required for the release response to transient reserpine exposure, this Ca 2ϩ source was eliminated by chelation with 0.5 mM EGTA for 5 min, and either reserpine or caffeine was applied onto the cell via pressure ejection for 5 s. Caffeine evokes release of Ca 2ϩ from internal thapsigargin-sensitive stores, serving as a positive control (Table I). Reserpine also induced exocytosis under these conditions, implying that external Ca 2ϩ facilitates more frequent release, but is not required for reserpine-induced release. To chelate internal Ca 2ϩ , cells were incubated for 10 min with 50 M BAPTA-AM and rinsed with a Ca 2ϩ -free buffer solution. Caffeine-and reserpine-induced exocytosis decreased (Table I). This reduction in spike frequency indicates that an internal Ca 2ϩ source is involved in exocytosis. No VMAT-induced changes in internal Ca 2ϩ concentration were detectable by Fura-2 measurements in the absence of external Ca 2ϩ (data not shown).
Internal Ca 2ϩ Source for Reserpine-induced Exocytosis-To determine which internal store of Ca 2ϩ could be involved in the ⌬[Ca 2ϩ ] i and subsequent release of catecholamine induced by reserpine, pharmacological intervention was employed. Thapsigargin is an inhibitor of the sarco(endo)plasmic reticulum calcium ATPase (SERCA) family of intracellular Ca 2ϩ pumps, which account for Ca 2ϩ uptake into inositol 1,4,5-triphospate, as well as caffeine-sensitive Ca 2ϩ stores in bovine chromaffin cells (20). Cells were permeabilized with 1 M digitonin, treated with 100 nM thapsigargin for 5 min, and then maintained in Ca 2ϩ -free buffer. Pressure ejection of reserpine (9 nM) induced several vesicles to be released, whereas caffeine (10 mM) caused a few smaller amplitude spikes to be detected (Fig. 6B). This result indicates that the source of Ca 2ϩ for reserpine-induced release is different from the thapsigargin-sensitive store. Next, the mitochondrial store of Ca 2ϩ was investigated with ruthenium red, an inhibitor of Ca 2ϩ transport in mitochondria (3). Cells were permeabilized with digitonin and then incubated with 50 M ruthenium red for 4 min. Pressure ejection of 9 nM reserpine evoked several concentration spikes (Fig. 6C), revealing that mitochondrial Ca 2ϩ stores were not necessary for the evoked release. These results provided pharmacological evidence that the vesicles were one of the organelles that could contain the reserpine-sensitive Ca 2ϩ store. DISCUSSION A growing body of evidence supports the notion that storage vesicles serve as important depots for intracellular calcium and that storage and release of vesicular calcium may play a significant role in exocytosis from neurons and other secretory cells. Chromaffin storage vesicles contain nearly 60% of the total Ca 2ϩ contained within the cell (5). Several channels and pumps have been observed in storage vesicle membranes (28), including Na ϩ /Ca 2ϩ exchange activity (3) and a Ca 2ϩ -H ϩ antiporter (29). In PC12 cells, Ca 2ϩ influx due to plasma membrane depolarization appears to trigger an intravesicular increase in pH, possibly via exchange of vesicular protons for cytoplasmic Ca 2ϩ (30). Manipulation of intracellular Ca 2ϩ stores can have dramatic effects on the release response, possibly affecting both b ANOVA reveals value is statistically different from value with 9 nM reserpine in Ca 2ϩ -free buffer (p Ն 0.019, degrees of freedom Ն 9). fusion and docking steps of the secretory process (31). A number of reports suggest a physiologically relevant role for vesicular calcium in exocytosis (1,4,32). Gerasimenko et al. (33) observe that cyclic ADP-ribose and inositol 1,4,5-triphospate treatment of isolated zymogen granules led to a rapid efflux of stored Ca 2ϩ , implicating these storage granules as the Ca 2ϩ source for the agonist-induced rise in cytoplasmic Ca 2ϩ observed in pancreatic acinar cells. In several preparations, an inhibitor of vesicular calcium storage enhanced acetylcholine release, suggesting a role for synaptic vesicles in buffering of local calcium concentrations at sites of exocytosis (2).
In previous work, we demonstrated that alkalinization of the vesicle interior can cause efflux of vesicular Ca 2ϩ to the cytoplasm in adrenal chromaffin cells and that the amounts of Ca 2ϩ released from this source are sufficient to elicit exocytosis (9). In this work, we examined the effects of reserpine and tetrabenazine, specific inhibitors of the vesicular monoamine transporter (11)(12)(13)(14), upon release and Ca 2ϩ mobilization from vesicles. Incubations with these agents for 15 min decreased both the exocytotic frequency and amount released per event evoked by exposure to depolarizing agents (Fig. 2), similar to that seen at PC12 cells (34). The decreased spike frequency is not simply a reflection of a shift in spike amplitudes toward smaller, undetectable, values after incubation, since the amperometric traces were purposely over-filtered to facilitate detection of small spikes.
However, the data also indicate that the VMAT inhibitors are releasing internal Ca 2ϩ in quantities sufficient to evoke exocytotic release. The decreased [Ca 2ϩ ] i evoked by depolaring agents following long term incubation, the exocytosis evoked by transient exposure to reserpine and tetrabenazine, and the increase in [Ca 2ϩ ] i evoked by reserpine all indicate that VMAT inhibitors are capable of altering a Ca 2ϩ storage pool that plays a significant role in stimulated secretion. The ability of reserpine and its analogs to cause exocytosis, efflux of catecholamines into the cytoplasm (Fig. 3B), and to decrease nicotine-evoked exocytosis correlates well with their potencies as inhibitors of vesicular storage via VMAT (35,36). Methyl reserpate is both less lipophilic and less potent than reserpine as an inhibitor of VMAT (12), whereas reserpic acid has a very low inhibitory potency, and its water solubility is relatively high (27). Since incubation of isolated chromaffin vesicles with VMAT inhibitors causes a decrease in [Ca 2ϩ ] vesicle (Fig. 3A), the results strongly suggest that this storage pool is the vesicles themselves. Furthermore, the direct correlation between diminished nicotine-evoked cytoplasmic Ca 2ϩ levels and frequency of exocytosis after long term incubation with VMAT inhibitors suggests that this reserpine-sensitive Ca 2ϩ store is normally involved in exocytosis.
The dynamic interaction between Ca 2ϩ entry and intracellular stores plays an important role in sustaining release (37). The most well characterized mechanism for exocytosis is elevation of [Ca 2ϩ ] i arising from influx of Ca 2ϩ though voltagegated or ligand-operated Ca 2ϩ channels (38). However, when this source of Ca 2ϩ is no longer available, Ca 2ϩ from intracellular stores can evoke exocytosis (19,39). The mobilization of intracellular Ca 2ϩ in bovine chromaffin cells also can lead to the opening of Ca 2ϩ channels in the plasma membrane to admit Ca 2ϩ from the exterior (40). Vesicles have been found to dock near these Ca 2ϩ channels (19). This proximity to the source would enable a small amount of Ca 2ϩ released from a vesicle either to trigger influx of external Ca 2ϩ or to initiate and to potentiate exocytosis. Consistent with this effect, the reserpine-induced exocytotic frequency is reduced, but not eliminated, in Ca 2ϩ -free medium, where influx of external Ca 2ϩ is greatly diminished. Chelation of internal Ca 2ϩ by BAPTA-AM further lowers reserpine-evoked exocytosis (Table  I). This decrease in frequency supports an internal Ca 2ϩ source that would cause a small rise in localized Ca 2ϩ even though changes in intracellular Ca 2ϩ levels were undetected with Fura-2, an indicator of global rises in [Ca 2ϩ ] i . A small, localized Ca 2ϩ source such as a vesicle could induce other vesicles within close proximity to fuse with the plasma membrane. The proteins that enable this fusion half maximally bind Ca 2ϩ in the micromolar range (41)(42)(43). However, this localized rise in Ca 2ϩ levels would only be high sufficient to affect the readily releasable pool (44).
An important question arising from results presented here is the mechanism for Ca 2ϩ mobilization by VMAT inhibitors from chromaffin vesicles. Reserpine is a highly lipophilic and very potent inhibitor of vesicular catecholamine transport (45,46). At high concentrations (Ͼ200 pmol/mg of membrane protein) it may act as a detergent, causing release of freely diffusable substance within the vesicle (47). However, all of the concentrations utilized in pressure ejection were below this value (a 5-s, 90 nM pressure ejection corresponds to 163 pmol/mg of membrane protein), making a detergent action unlikely. Another possible mechanism is direct interaction of inhibitors with an ion channel or pump involved in Ca 2ϩ mobilization or storage. Mahata et al. (48) describe a dose dependent reserpineinduced release of norepinephrine from PC12 and bovine chromaffin cells preloaded with the radiolabeled catecholamine and reserpine inhibition of secretion stimulated by nicotine or membrane depolarization (48). Although attributed to an inhibition of L-type Ca 2ϩ channels, no demonstration of a direct inhibitory action on channel currents by reserpine or tetrabenazine was provided.
FIG. 6. Release in Ca 2؉ -free medium. A, amperometric response following pressure ejection of 10 mM caffeine (CAF) and 9 nM reserpine (RES) onto the same cell in Ca 2ϩ -free medium. B, a cell was permeabilized with 1 M digitonin for 5 min and then incubated with 100 nM thapsigargin for 5 min in the absence of Ca 2ϩ . Amperometric recordings were made during pressure ejection of 10 mM caffeine and 9 nM reserpine in the absence of Ca 2ϩ . C, prior to the traces the cell was permeabilized by incubation with 1 M digitonin for 5 min in Ca 2ϩ -free buffer and then incubated with 50 M ruthenium red for another 5 min. During the amperometric recordings, 10 mM caffeine and 9 nM reserpine were applied onto the cell via pressure ejection. At each cell, 2 min elapsed between caffeine and reserpine drug exposure.
A third possibility and one we consider more likely is that efflux of both Ca 2ϩ and catecholamine results directly or indirectly from inhibition of VMAT. Michalke et al. (35) proposed a "pump and leak system" involving reserpine-sensitive accumulation but reserpine-insensitive catecholamine efflux as a result of studies on chromaffin granule ghosts. Two efflux mechanisms were observed in highly purified synaptic vesicles from rat brain (36); one was inhibited by reserpine and tetrabenazine, whereas the other was not. These leak pathways may also directly mobilize calcium. Alternatively, the inhibitor-induced efflux of catecholamines from the vesicle could cause Ca 2ϩ mobilization by perturbation of the vesicle matrix. The matrix of the vesicle is composed of chromogranins, catecholamine, Ca 2ϩ , Mg 2ϩ , and ATP that make up a noncrystalline, partially dehydrated gel-like polymer that confers osmotic stability despite the high concentrations of solutes within the vesicle (49,50). Since the decrease in the free catecholamine concentration within the vesicle would alter catecholamine matrix association, this may alter the binding of Ca 2ϩ to the matrix as well. Indeed, reserpine and tetrabenazine treatment of bovine chromaffin cells leads to a decrease in ATP, another component of the vesicular matrix (13). Similarly, removal of vesicular Ca 2ϩ with the selective ionophore, A23187, causes isolated vesicles to dissociate (51). Changes in vesicle pH have profound effects on the association of the vesicle components (7,52) and account for effects of amine weak bases on secretion and Ca 2ϩ mobilization observed previously (9).
In summary, the results here and our prior work with weak bases show that Ca 2ϩ from vesicles can participate in the induction of exocytosis. Indeed, as shown here, incubations with amphetamine lead to a decreased frequency of release that is accompanied by decreased quantal size and decreased nicotine-evoked [Ca 2ϩ ] i , behaviors identical to that of the VMAT inhibitors. Although the VMAT inhibitors and weak bases have similar effects, the underlying mechanisms for Ca 2ϩ mobilization in each case probably differs. However, the results with both classes of drugs clearly show that alteration of vesicular catecholamine content is accompanied by liberation of sufficient Ca 2ϩ to cause exocytosis. More importantly, it appears that this source of Ca 2ϩ normally participates in the catalysis of exocytosis since the exocytotic rate decreases when the store is diminished. Thus, chromaffin cells as well as other secretory cells share the attribute that the trigger for vesicleplasma membrane fusion is contained within the vesicle.