An inhibitory effect of extracellular Ca2+ on Ca2+-dependent exocytosis.

AIM: Neurotransmitter release is elicited by an elevation of intracellular Ca(2+) concentration ([Ca(2+)](i)). The action potential triggers Ca(2+) influx through Ca(2+) channels which causes local changes of [Ca(2+)](i) for vesicle release. However, any direct role of extracellular Ca(2+) (besides Ca(2+) influx) on Ca(2+)-dependent exocytosis remains elusive. Here we set out to investigate this possibility on rat dorsal root ganglion (DRG) neurons and chromaffin cells, widely used models for studying vesicle exocytosis.
RESULTS: Using photolysis of caged Ca(2+) and caffeine-induced release of stored Ca(2+), we found that extracellular Ca(2+) inhibited exocytosis following moderate [Ca(2+)](i) rises (2-3 µM). The IC(50) for extracellular Ca(2+) inhibition of exocytosis (ECIE) was 1.38 mM and a physiological reduction (∼30%) of extracellular Ca(2+) concentration ([Ca(2+)](o)) significantly increased the evoked exocytosis. At the single vesicle level, quantal size and release frequency were also altered by physiological [Ca(2+)](o). The calcimimetics Mg(2+), Cd(2+), G418, and neomycin all inhibited exocytosis. The extracellular Ca(2+)-sensing receptor (CaSR) was not involved because specific drugs and knockdown of CaSR in DRG neurons did not affect ECIE. CONCLUSION/SIGNIFICANCE: As an extension of the classic Ca(2+) hypothesis of synaptic release, physiological levels of extracellular Ca(2+) play dual roles in evoked exocytosis by providing a source of Ca(2+) influx, and by directly regulating quantal size and release probability in neuronal cells.

Introduction

Neurotransmitter and hormone secretion are precisely controlled in neurons and endocrine cells where Ca2+ plays a pivotal role [1]–[3]. In response to action potentials (APs), the entry of extracellular Ca2+ through presynaptic Ca2+ channels results in microdomains of spatial-temporal [Ca2+]i rise which triggers synaptic transmission [2]. At the same time, the rapid influx of Ca2+ through presynaptic Ca2+ channels and then postsynaptic Ca2+-permeable channels also leads to a reduction of extracellular Ca2+ concentration ([Ca2+]o) given the limited Ca2+ storage capacity within the synaptic cleft [4], [5].

Early studies using ion-sensitive electrodes showed that a local electrical stimulus decreases [Ca2+]o by around 1/3 in cerebellar cortex [4]. Direct depolarization of postsynaptic membrane induces Ca2+ influx into the membrane and depletion of Ca2+ in the synaptic cleft. Presynaptic Ca2+ current is thus decreased by 30%, which corresponds to the more than 33% [Ca2+]o drop in a glutamatergic synapse [5]. Under pathological and injury conditions, [Ca2+]o decreases even down to 0.1–0.3 mM from 1.2 mM [6]. It is known that extracellular Ca2+ modulates ion channels in neurons; for example, lowering [Ca2+]o shifts the voltage dependence of sodium channels to more negative potentials, via a biophysical surface charge effect [7]. In addition to this biophysical effect, [Ca2+]o may have a biochemical effect on cell function by regulating non-selective cation channels and TRPM7 channel [8]–[11].

For neurotransmitter release, it is well established that an AP-induced [Ca2+]i rise directly triggers vesicle release [12]. However, it remains elusive whether extracellular Ca2+ also plays a direct role in vesicle exocytosis from outside cells. To investigate this possibility, we triggered vesicle release using photolysis of caged Ca2+ or caffeine-induced release of stored Ca2+ in the absence of membrane depolarization (without Ca2+ influx through voltage-dependent Ca2+ channels) at various levels of [Ca2+]o. We found that extracellular Ca2+ directly regulates [Ca2+]i-dependent vesicle exocytosis in the somata of sensory dorsal root ganglion (DRG) neurons and neuroendocrine chromaffin cells. Interestingly, physiological levels of [Ca2+]o inhibited [Ca2+]i-dependent exocytosis, partially by altering the quantal size and frequency of single vesicle release.

Methods

Cell preparation and patch clamp recordings

The use and care of animals in this study was approved and overseen by the Institutional Animal Care and Use Committee of Peking University. The permit number for this project is IMM-zhouz-11. Freshly isolated DRG neurons from 120–150 g Wistar rats were prepared as described previously [13], [14]. Cells were used 1–8 h after preparation. Only small neurons (15–25 µm, C-fiber) without apparent processes were selected for experiments.

Rat adrenal medulla slices were prepared as described previously [15]. Briefly, the adrenal glands were removed from adult Wistar rats of 250–300 g. The glands were immediately immersed in ice-cold, low-calcium Ringer's saline (in mM: 125 NaCl, 2.5 KCl, 0.1 CaCl2, 5 MgCl2, 1.25 NaH2PO4, 26 NaHCO3, 12 glucose, pH 7.4) gassed with 95%O2/5%CO2. Single glands were glued with cyanoacrylate to the stage of a vibratome chamber and covered with the same Ringer's saline. Slices of 200–300 µm were cut parallel to the larger base of the gland (Vibrotome 1000, St. Louis, MO). The slices were incubated for 30 min at room temperature in normal Ringer's saline (in mM: 125 NaCl, 2.5 KCl, 2.5 CaCl2, 1 MgCl2, 1.25 NaH2PO4, 26 NaHCO3, 12 glucose) bubbled with 95%O2/5%CO2. The slices were used for up to 8 h after cutting. For amperometric recordings, slices were transferred to a recording chamber attached to the stage of an upright microscope and continuously superfused with Ringer's saline at room temperature (∼22°C) [15].

Standard external medium contained (in mM) 150 NaCl, 5 KCl, 2.5 CaCl2, 1 MgCl2, 10 HEPES, 10 glucose, pH 7.4. Ca2+-free medium was the same except that CaCl2 was removed and 1 mM EGTA was added. Ca2+ and Mg2+ concentrations were changed with osmotic compensation. For solutions with Ca2+ and Mg2+ concentrations between 1 and 100 µM, EGTA was added for Ca2+ and EDTA for Mg2+ to obtain accurate concentrations. Cells were initially bathed in normal external solution and low Ca2+, Ca2+-free solutions, and drugs were locally puffed onto cells using a multichannel microperfusion system (MPS–2; INBIO, Wuhan, China) [16], [17]. Standard intracellular pipette solution contained (in mM) 153 CsCl, 1 MgCl2, 10 HEPES, 4 Mg-ATP, pH 7.2. For flash photolysis experiments, the intracellular pipette solution contained (in mM) 110 Cs-glutamate, 5 NP-EGTA, 8 NaCl, 2.5 CaCl2, 2 Mg-ATP, 0.3 GTP, 0.2 fura-6F, and 35 HEPES, pH 7.2. All chemicals were from Sigma (St. Louis, Missouri), except for NP-EGTA (5 mM), fura-2 AM (final concentration 15 µM), fura-2 salt (20 µM), fura-6F (0.2 mM), and fluo-4 AM (15 µM) (Molecular Probes, Eugene, Oregon).

Membrane capacitance (Cm) measurements

Cm was measured using a software lock-in module of Pulse 8.76 together with an EPC10/2 amplifier, as described previously [14]. Briefly, a 1 kHz, 20 mV peak-to-peak sinusoid was applied around a DC holding potential of −70 mV. The resulting current was analyzed using the Lindau-Neher technique to estimate Cm, membrane conductance and series resistance [18]. Flash-induced maximum capacitance changes were measured as ΔCm.

Amperometry

Highly sensitive, low-noise, 5-µm diameter carbon fiber electrodes (ProCFE, Dagan, Minneapolis, MN) were used for electrochemical monitoring of quantal release of catecholamine from chromaffin cells in slices [19]. Long-tip CFEs with 200 µm sensor tips were used to detect local catecholamine release from many cells [15]. Amperometric currents were recorded at +780 mV, a potential sufficient to oxidize catecholamines [20], [21], low-pass filtered at 300 Hz, and sampled by the EPC-9/2 at 5 kHz.

Flash photolysis of caged Ca2+ and [Ca2+]i measurements

For measurements of [Ca2+]i and UV flashes in DRG cells, we used an IX-71 inverted microscope (Olympus, Tokyo, Japan) equipped with a monochromator-based system (TILL Photonics, Planegg, Germany). NP-EGTA is very selective for Ca2+ and intracellular dialysis of NP-EGTA generated a small loading transient, which did not affect membrane capacitance (data not shown) [3], [22], [23]. The Ca2+ indicator fura-6F was excited at 1 Hz with 5-ms light pulses at 340 and 380 nm for the measurement of [Ca2+]i in the presence of NP-EGTA [24]. The excitation intensity was far below the level for evident photolysis of NP-EGTA. The emitted fluorescence was detected by a photomultiplier tube (TILL), sampled by EPC10/2, and acquired with a Fura extension of the Pulse software (HEKA). UV light from a xenon arc flash lamp (Rapp, Hamburg, Germany) was combined with the excitation light provided by a monochromator before entering an epifluorescence port of the IX-71 microscope. For all flash experiments we used an Olympus Uapo/340 40×, 1.35 NA oil-immersion objective. The calibration method used for caged Ca2+ experiments was similar to that reported [24]. The first flash was triggered 2–3 min after the whole-cell configuration was established. This allowed recovery of [Ca2+]i to ∼300 nM after the loading transient.

For measurements of [Ca2+]i in cells in slices, we used the Uniblitz shutter-based fluorescence system (Vincent, Rochester, NY), a DVC1412 CCD camera (DVC, Austin, TX), and a BX-51 upright microscope (Olympus). The images were acquired with TILLvisION software (TILL Photonics, Planegg, Germany). The adrenal slice was loaded with fluo-4 AM or fura-2 AM by local puffer for ∼120 s via a patch pipette tip of ∼5 µm diameter. Individual cells were imaged by DIC optics using a 40× water-immersion lens (NA = 0.8). For fluo-4, Ca2+ transients were recorded by measuring the emission at 525 nm, resulting from excitation at 488 nm [25]. For fura-2, Ca2+ transients were recorded by measuring the ratio of emission at 510 nm, resulting from excitation at 340 and 380 nm [16].

CaSR knockdown by shRNAs

shRNAs for rat CaSR were designed using Invitrogen BLOCK-iT™ RNAi Designer (https://rnaidesigner.invitrogen.com/rnaiexpress/). Knockdown efficiency was first tested in HEK 293 cells using over-expressed rat CaSR [26]. Five shRNAs were designed and the most efficient two were chosen for knockdown experiments in DRG neurons. The two shRNAs were: sh 1, 5′ -GATCCGCAGGCTCCTCAGCAATAACGAATTATTGCTGAGGAGCCTGCTTTTTTGAATTCA- 3′, and sh 2, 5′ -GATCCGCGCATGCCCTACAAGATATACGAATATATCTTGTAGGGCATGCGCTTTTTTGAATTCA- 3′. Electrophysiology was done 4–5 days after shRNA transfection.

For Western blot, HEK293 cells were lysed in 20 mM HEPES, 1 mM EDTA, 0.1 g/L PMSF, 100 mM NaCl, 1% NP40, pH 7.4, with protease inhibitor cocktail (Calbiochem, San Diego, USA) [27], [28]. Proteins were solubilized in SDS-PAGE buffer. After running SDS-PAGE, they were analyzed by immunoblotting with antibodies to CaSR and β-actin. CaSR antibody was from Affinity BioReagents (Colorado, USA). β-actin antibody was from Sigma (St. Louis, Missouri).

Immunoprecipitations

DRGs were dissected and lysed in 20 mM HEPES, 1 mM EDTA, 0.1 g/L PMSF, 100 mM NaCl, 1% NP40, pH 7.4, with protease inhibitor cocktail (Calbiochem, San Diego, USA) [27], [28]. The complexin antibody was incubated with DRG extract overnight before Protein A beads (GE Healthcare, Uppsala, Sweden) were added. Bound proteins were solubilized in SDS-PAGE buffer. After running SDS-PAGE, they were analyzed by immunoblotting with monoclonal antibodies to syntaxin 1, SNAP 25, synaptotagmin 1, GAPDH and CaSR, and a polyclonal antibody to complexin 2.

Syntaxin 1, SNAP 25, and complexin 2 antibodies were from Synaptic Systems (Göttingen, Germany). Synaptotagmin 1 antibody was from Stressgen Bioreagents (British Columbia, Canada).

Statistics

All experiments were carried out at room temperature (22–25°C), and data are presented as mean ± SEM. The peak values for evoked ΔCm signals were used for analysis, and cell numbers are shown as “n” in brackets. Data were analyzed using Igor software (Wavemetrics, Lake Oswego, Oregon). Significance of difference was tested using either Student's t-test or ANOVA followed by post hoc tests (*p<0.05, **p<0.01, ***p<0.001, n.s. = not significant (p>0.05)).

Results

Extracellular Ca2+ inhibited somatic exocytosis in DRG neurons

DRG neurons are primary sensory cells that release pain-related transmitters including CGRP, substance P and ATP from their terminals and somata in response to APs [13], [16], [29]. There are two kinds of depolarization-induced exocytosis in the somata of DRG neurons [16], [30]; in the present study, we focused on the Ca2+-dependent exocytosis by holding the soma at resting potential [13], [16], [30].

By raising [Ca2+]i through photolysis of caged Ca2+ [24] or caffeine-induced store release [31], we determined whether extracellular Ca2+ directly regulates the [Ca2+]i-dependent exocytosis. We did not use APs, because AP-induced Ca2+ influx causes microdomain changes in [Ca2+]i which cannot be controlled under various levels of [Ca2+]o [5], [32]. Isolated DRG neurons underwent whole-cell dialysis with 5 mM of the caged-Ca2+ compound nitrophenyl-EGTA. We elicited a [Ca2+]i rise using photolysis to a level near the threshold of evoked exocytosis. A train of low-energy UV flashes was applied to produce a [Ca2+]i plateau of ∼2 µM, resulting in a large Cm increase of ∼2 pF in a Ca2+-free bath (Figure 1). This Cm increase reflects a substantial amount of exocytosis (∼4000 vesicles, assuming 0.5 fF per 140 nm vesicle [16]). However, 2 min after adding 2.5 mM Ca2+ to the bath, similar flash-induced [Ca2+]i transients nearly failed to induce Cm increases in the same neuron (Figure 1). The evoked ΔCm was fully recovered after removing external Ca2+. In this particular cell, physiological levels of extracellular Ca2+ (2.5 mM) fully inhibited the [Ca2+]i-dependent ΔCm. In a second cell, the UV-flash evoked a similar level of [Ca2+]i rise and ΔCm in 2.5 mM [Ca2+]o was inhibited again, albeit by 40% compared to the Ca2+-free bath (Figure 1). These two cells represent the maximal and minimal inhibition we observed. To compare the rates of photolysis-induced exocytosis in normal external Ca2+ solution and Ca2+-free solution, we averaged ΔCm responses within 9 s after photolysis. Photolysis induced a much larger ΔCm in Ca2+-free solution (Figure 1, upper panel). We then used normalized peak ΔCm to quantify extracellular Ca2+ inhibition of exocytosis (ECIE). Compared to 0 mM [Ca2+]o, the evoked [Ca2+]i rise was similar (106±8%), but ΔCm was reduced by 82±8% in 2.5 mM [Ca2+]o (Figure 1, lower panel). Similar ECIE was also found when a [Ca2+]i rise was elicited by caffeine, which releases Ca2+ from the intracellular ryanodine-sensitive Ca2+ store [31] (Figure 1). The elicited [Ca2+]i was nearly identical, while ΔCm was reduced by 58±5% in 2.5 mM [Ca2+]o (Figure 1). Thus, extracellular Ca2+ inhibited exocytosis in the somata of rat DRG neurons.

Figure 1

Extracellular Ca2+ inhibited [Ca2+]i-dependent exocytosis in DRG neurons.

(A) Membrane capacitance responses to Ca2+ spikes induced by photolysis in the presence (2.5 mM) or absence (0 mM) of external Ca2+. The internal solution in the whole-cell recording pipette contained the photolabile Ca2+ chelator nitrophenyl-EGTA. The arrowheads mark the application of four UV flashes at 0.5 Hz. In one DRG neuron, from the top, individual traces show changes in membrane capacitance (ΔCm), series resistance (Gs), membrane conductance (Gm), and [Ca2+]i. There were 2-min breaks between recordings. The initial values of Cm, Gs, Gm, and Im are noted. [Ca2+]i was monitored by Fura-6F measurements. (B) In a second DRG neuron, for clarity, only Cm and [Ca2+]i are shown. (C) Upper panel: averaged Cm responses in 0 and 2.5 mM [Ca2+]o within the first 9 s after photolysis (n = 10). Averaged initial Cm was 21.4±1.2 pF. Lower panel: statistics of normalized peak ΔCm and [Ca2+]i following photolysis. Compared to 0 mM [Ca2+]o, the evoked [Ca2+]i rise was similar (106±8%), but exocytosis was reduced by 82±8% in 2.5 mM [Ca2+]o (n = 10, p<0.001). (D) Membrane capacitance responses to Ca2+ rise induced by local caffeine application (caff, 20 mM) in the presence (2.5 mM) or absence (0 mM) of external Ca2+ in a DRG neuron. There were 90-s breaks between recordings. (E) Normalized peak ΔCm and [Ca2+]i following caffeine stimulation. Compared to 0 mM [Ca2+]o, the [Ca2+]i rise was similar, while exocytosis was reduced by 58±5% in 2.5 mM [Ca2+]o (n = 14, p<0.001).

Exocytosis at both 0 and 2.5 mM [Ca2+]o was facilitated by cAMP elevation with forskolin, presumably via activation of protein kinase A (PKA), which is a well-known feature of Ca2+-dependent exocytosis for synaptic transmission in hippocampal neurons and somatic release in chromaffin cells (Figure S1) [33]. Thus, the ECIE-targeted vesicles use the classic type of Ca2+-dependent exocytosis.

Because membrane capacitance measures the sum of membrane-expanding exocytosis and membrane-retrieving endocytosis, and a 0 Ca2+ external solution may arrest the internalization of vesicles [34], the lack of a Cm response to UV flashes at 2.5 mM [Ca2+]o could be due to a similar-sized endocytosis. We examined this possibility by inhibiting dynamin, a GTPase essential for most endocytotic pathways [35]. Endocytosis was greatly accelerated when the external solution was shifted to 2.5 mM Ca2+ from 0 Ca2+ (Figure 2, right panel), which corresponds with a previous report [34]. Intracellular dialysis of GTPγS (200 µM) inhibited endocytosis without affecting the ECIE in DRG neurons (Figure 2). Note that GTPγS also reduced partial exocytosis in DRG neurons (Figure 2), implying that GTP affected exocytosis in these neurons [36], [37]. Since endocytosis was increased in 2.5 mM Ca2+ (Figure 2), ECIE measured by Cm could be partially contaminated by exocytosis-coupled endocytosis.

Figure 2

GTPγS had no effect on ECIE in DRG neurons.

(A) Responses of ΔCm and [Ca2+]i in a control neuron without GTPγS. The peak of flash-induced exocytosis is Cm1 while Cm2 represents a robust endocytosis in 40 s. The ratio Cm2/Cm1 (Rendo/exo) reflects the degree of endocytosis compared to the preceding exocytosis [14]. The initial value of Cm is noted. [Ca2+]i was monitored by Fura-6F measurements. Note the shift of extracellular solution from 0 Ca2+ to 2.5 mM Ca2+ accelerated endocytosis. (B) A different neuron dialyzed with 200 µM GTPγS, where exocytosis was induced only in the absence of external Ca2+. However, following the evoked exocytosis in 0 Ca2+, there was little decline in Cm (endocytosis) when the external bath was changed to 2.5 mM Ca2+. (C) Average results of ECIE in the presence of GTPγS. Compared with 0 Ca2+, ΔCm in 2.5 mM Ca2+ was reduced to 18±14% (n = 5, p<0.001), while [Ca2+]i was not reduced (121±13%). (D) Average ratios of the evoked endocytosis and exocytosis (Rendo/exo). Rendo/exo was 45±8% with GTPγS but 136±8% without GTPγS (n = 6, p<0.001).

To confirm the ECIE with a non-capacitance assay and investigate its role in transmitter release, we used a neuroendocrine cell which is a widely used model for exocytosis [31], [38], [39], the rat adrenal chromaffin cell (RACC). Combining membrane capacitance and amperometric recording of catecholamine release, we found that ECIE occurred in chromaffin cells too. Compared to photolysis-induced exocytosis in Ca2+-free solution, the photolysis-induced ΔCm was inhibited to 69±5%, the number of amperometric spikes was reduced to 68±6%, and the integrated amperometric signal was inhibited to 59±9% (Figure S2). The photolysis-induced [Ca2+]i was not changed (107±4%). Taken together, extracellular Ca2+ inhibited Ca2+-regulated vesicle exocytosis.

Attenuation of ECIE by high [Ca2+]i rise

The above experiments showed ECIE in moderate [Ca2+]i (2–3 µM). These data are not consistent with other studies using photolysis of caged Ca2+ to much higher [Ca2+]i levels (6–25 µM) [24], [40], [41]. Thus, we investigated ECIE at higher [Ca2+]i by using high-energy UV flashes and 10 mM nitrophenyl-EGTA. Extracellular Ca2+ still reversibly inhibited exocytosis when the UV flash-induced [Ca2+]i rise was increased to ∼4 µM (Figure 3). Statistically, the evoked exocytosis at [Ca2+]o of 2.5 mM was inhibited by 40±5% of that at a [Ca2+]o of 0 (Figure 3), much less than the 82±8% inhibition at a 2–3 µM [Ca2+]i rise, indicating that ECIE was attenuated by higher [Ca2+]i rises (p<0.001, Figure 3, Figure S3). Therefore, the inhibition by [Ca2+]o could be overcome by higher [Ca2+]i.

Figure 3

High [Ca2+]i attenuated ECIE.

(A) ΔCm and [Ca2+]i signals in response to Ca2+ release by trains of UV flashes at high energy. The initial value of Cm is noted. [Ca2+]i was monitored by Fura-6F measurements. (B) Statistics of normalized ΔCm and [Ca2+]i (n = 14, p<0.001). UV flash-induced [Ca2+]i rises were similar, 6.3±0.4 µM ([Ca2+]o = 0) and 6.4±0.4 µM ([Ca2+]o = 2.5 mM). (C) Comparison of ECIE (or normalized exocytosis ΔCm) in high and low [Ca2+]i. Compared to 0 mM [Ca2+]o, evoked exocytosis was reduced by 82±8% ([Ca2+]i = 2 µM, n = 10) and 40±5% ([Ca2+]i = 6 µM, n = 14) in 2.5 mM [Ca2+]o.

Changes of [Ca2+]o within the physiological range modulated exocytosis

[Ca2+]o varies during normal physiological neural activity [4], [5], [42]. Direct depolarization of the postsynaptic membrane causes a more than 33% drop of [Ca2+]o in a glutamatergic synapse [5]. In single DRG neurons, we reduced [Ca2+]o 34% (from 2.5 mM to 1.65 mM) and tested its effect on evoked exocytosis. The caffeine-induced exocytosis was 35% greater in 1.65 mM than in 2.5 mM [Ca2+]o, while the evoked [Ca2+]i rise was similar (Figure 4). Thus, these experiments strongly suggested that ECIE occurred under physiological conditions in DRG neurons.

Figure 4

Physiological levels of extracellular Ca2+ decrease modulated exocytosis in DRG neurons.

(A) Combined ΔCm and [Ca2+]i recordings in a DRG neuron. A reduction of [Ca2+]o from 2.5 mM to 1.65 mM (34%) increased the caffeine-induced (20 mM) ΔCm signal. Upper, ΔCm signals following caffeine stimulation. Lower, corresponding [Ca2+]i rise monitored by Fura-2 measurements. There were 90-s breaks between recordings. The initial value of Cm is noted. AU, arbitrary units. (B) Statistics of normalized exocytosis and [Ca2+]i rise signals. A 34% reduction of [Ca2+]o increased caffeine-induced exocytosis signals by 35±8% (n = 9, p<0.01). The caffeine-induced [Ca2+]i rise was similar (99±2%, n = 8) when [Ca2+]o was changed between 1.65 mM and 2.5 mM.

Calcimimetics inhibited [Ca2+]i-dependent exocytosis in DRG neurons

In search of the molecular mechanism underlying ECIE, we investigated the extracellular calcium-sensing receptor (CaSR), which inhibits secretion in parathyroid cells [43]. The CaSR is expressed in nerve terminals [26] and in DRG neurons [44], [45]. First, we used CaSR agonists (or “calcimimetics”), the divalent ions, Ca2+, Mg2+, and Cd2+, G418, and neomycin [43], [46]. Similar to Ca2+, G418 or other calcimimetics also inhibited caffeine-induced exocytosis in DRG neurons, while the caffeine-induced [Ca2+]i rise was unaffected (Figure 5). In these experiments, G418 was the most potent inhibitor with an EC50 of 28 µM, followed by Cd2+ (47 µM), Mg2+ (1.26 mM), and Ca2+ (1.38 mM) (Figure 5). Second, since these calcimimetics may not be specific to the CaSR [9], [47], we assessed the effects of the more specific CaSR antagonist calhex 231 (calhex) and the agonist calindol [48], [49]. In HEK 293 cells expressing rat CaSR, calhex inhibited the [Ca2+]i rise induced by high [Ca2+]o perfusion while calindol increased [Ca2+]i (Figure S4). However, calhex failed to affect caffeine-induced exocytosis in 2.5 mM [Ca2+]o and calindol did not affect exocytosis in 0 mM [Ca2+]o (Figure 5), or in 0.5 mM [Ca2+]o (data not shown). Neither calhex (1 µM) nor calindol (1 µM) affected the caffeine-induced [Ca2+]i rise in DRG neurons (Figure S5). Third, in cortical synaptosomes, a non-selective cation channel (NSCC) coupled with the CaSR is modulated by [Ca2+]o [9]. NSCC may modulate synaptic transmissions in cortical neurons [50]. We determined whether NSCC existed in DRG neurons and found that it was not detectable (Figure S6). Fourth, we determined whether the CaSR was physically linked to the SNARE complex, which is responsible for all vesicle exocytosis [51]. Co-immunoprecipitation with complexin failed to detect the CaSR in the presence or absence of 3.5 mM Ca2+, while syntaxin1 and SNAP25 were found (Figure S7) as reported in brain [52].

Figure 5

Characterization of ECIE.

(A) Combined ΔCm and [Ca2+]i measurements in a DRG neuron. A representative neuron showing that G418 inhibited the exocytosis induced by caffeine (20 mM). Upper traces: ΔCm responses to application solutions containing caffeine and G418 (0.8 mM). The initial value of Cm is noted. Lower traces: corresponding [Ca2+]i monitored by Fura-2 measurements. There were 90-s breaks between recordings. (B) Dose-dependence of inhibition of exocytosis by [Ca2+]o (EC50 = 1.38 mM, n = 6), [Mg2+]o (EC50 = 1.26 mM, n = 6), [Cd2+]o (EC50 = 47 µM, n = 6), and [G418]o (EC50 = 28 µM, n = 6). (C) Cm responses to application solutions containing caffeine and calhex (1 µM) or calindol (1 µM). (D) Statistics of the caffeine-induced ΔCm signals. Compared to exocytosis in 0 mM Ca2+ and 2.5 mM Ca2+, caffeine-induced ΔCm was 98±1% in calindol (n = 7) and 107±3% in calhex (n = 6).

Finally, we undertook a RNAi knockdown approach to test whether the CaSR was responsible for ECIE. Two short hairpin RNAs (shRNAs) for rat CaSR were designed and tested in HEK 293 cells. Both shRNAs efficiently knocked down over-expressed rat CaSR (Figure S8A). However, CaSR knockdown in DRG neurons had no effect on ECIE (Figure S8B, C).

Taken together, we concluded that ECIE was not mediated through the CaSR, although the ECIE sensor and CaSR shared agonists of Ca2+ and other divalent ions, as well as their potency order (Figure 5).

ECIE in quantal transmitter release

Having shown that extracellular Ca2+ inhibited photolysis-induced exocytosis in single DRG neurons and chromaffin cells (Figures 1, 2, and 3, Figure S2), we next tested whether extracellular Ca2+ also inhibited exocytosis in the rat adrenal slice. Following application of caffeine, individual quantal exocytosis of catecholamines was detected by a micro carbon fiber electrode (CFE) placed in an adrenal slice [15], [53] (Figure 6). Since extracellular Ca2+ and Mg2+ both inhibited exocytosis (Figure 5), we used Ca2+- and Mg2+-free solution to remove calcimimetic inhibition of catecholamine release. In Ca2+- and Mg2+-free solution the baseline of recording shifted upwards (Figure 6), probably due to the catecholamine release from distant cells because it was absent when the CFE holding potential was below the oxidization voltage (0 mV, [20], [38], [39]). The caffeine-induced exocytosis was inhibited in 2.5 mM [Ca2+]o (Figure 6), while the [Ca2+]i rise was similar (Figure 6). These experiments demonstrated that ECIE occurred in quantal catecholamine transmitter release in RACCs. A physiological reduction (∼34%) of extracellular Ca2+ also significantly increased caffeine-induced transmitter release in RACCs (Figure S9).

Figure 6

ECIE in quantal transmitter release.

(A) Left, upper traces: amperometric recording of typical secretory signals evoked by 30 s caffeine (20 mM) at 0 and 2.5 mM [Ca2+]o in a rat adrenal slice. Recordings were performed at 120-s intervals. Lower traces show the integrated amperometric current signals (∫Iampdt, or total catecholamine molecules detected). Right, average of normalized ∫Iampdt signals from 15 cells. Caffeine-induced exocytosis was reduced to 22±3% in 2.5 mM vs 0 mM [Ca2+]o (n = 15, p<0.001). (B) Left, caffeine-induced [Ca2+]i rise was similar when changing [Ca2+]o in an adrenal slice. [Ca2+]i rise in response to standard 20 mM caffeine at 0 and 2.5 mM [Ca2+]o was measured using fluo-4 AM. Right, average of caffeine-induced [Ca2+]i rises from 20 cells. The caffeine-induced [Ca2+]i rise was similar at 0 and 2.5 mM [Ca2+]o (101±5%, n = 20). (C) The distribution of quantal size of caffeine-induced amperometric spikes in 0 Ca2++0 Mg2+ (black traces, 423 events) and 2.5 mM Ca2++1 mM Mg2+ (gray traces, 150 events) bath solutions. Histograms show the percentage of amperometric quantal sizes. The curves show Gaussian fits of the distributions. Right inset: average events of single amperometric spikes from the top 10% of fastest events in 0 Ca2++0 Mg2+ (n = 13 cells) vs 2.5 mM Ca2++1 mM Mg2+ (n = 7) [31]. (D) Quantal analysis. Data shown are mean ± s.e.m. Left, parameters for analysis. Right, statistical table. (Q, quantal size; PA, peak amplitude; HHD, half-height duration; RT, rise-time; f, normalized frequency).

To investigate the extracellular Ca2+ effect on single vesicle release kinetics, we undertook a quantitative analysis of amperometric spikes representing individual vesicle release. Several features of the spikes are used to reflect the kinetics of vesicle fusion (Figure 6, left panel) [31], [38], [54]. Compared to caffeine-induced amperometric spikes in bath with 0 Ca2++0 Mg2+, bath containing 2.5 mM Ca2++1 mM Mg2+ inhibited the quantal size (Q) by ∼27% (Figure 6, table) and shifted the Q distribution towards smaller sizes (Figure 6). In addition, the peak amplitude was inhibited by ∼47%, while half-height duration and rise time were increased by ∼30% and ∼50%, and the normalized frequency of the amperometric spikes was reduced by ∼60% (Figure 6, table). These changes in amperometric signals are different from (but not in conflict with) a previous report where no change in quantal size was found under non-physiological high [Ca2+]o (5–90 mM) (compared to the physiological 2.5 mM) [55].

Discussion

In this work we used photolysis of a caged Ca2+ compound or caffeine to elevate [Ca2+]i without Ca2+ influx. This permitted us to determine whether extracellular Ca2+ directly regulated vesicle release. We found that ECIE occurred in the somata of DRG neurons and adrenal chromaffin cells.

ECIE is supported by the following evidence: (1) for a given rise in [Ca2+]i induced by UV flash photolysis, the [Ca2+]i-induced Cm increase was enhanced by removing external Ca2+ in DRG neurons (Figure 1); (2) intracellular dialysis of GTPγS, which inhibits exocytosis-coupled endocytosis, had no effect on ECIE (Figure 2, see also [35]), indicating that the reduced ΔCm caused by external Ca2+ was not due to an enhanced dynamin-dependent endocytosis in normal external Ca2+ solution [34]; (3) direct measurements of catecholamine release using amperometry detected ECIE in rat single chromaffin cells (Figure S2) and adrenal slices (Figure 6); (4) extracellular Ca2+ and Mg2+ reduced quantal size (by 49%) and release probability (by 60%), and modulated the release kinetics of single vesicle exocytosis (Figure 6); (5) physiological reduction of [Ca2+]o (∼34%) significantly increased exocytosis in DRG neurons (Figure 4) and adrenal slices (Figure S9); (6) the exocytotic inhibition by [Ca2+]o was dose-dependent (Figure 5); and (7) ECIE was not caused by a Ca2+ gradient near the plasma membrane, which might be produced by possible Ca2+ efflux following a sudden [Ca2+]i rise, because Ca2+ efflux during [Ca2+]i-induced exocytosis would only reduce exocytosis.

ECIE was induced by physiological changes in [Ca2+]o. A 34% change (from 2.5 to 1.65 mM) [4], [5], [56], [57] of [Ca2+]o significantly increased total exocytosis (Figure 4 and Figure S9). Thus, a Ca2+ influx via Ca2+-permeable channels has dual effects on exocytosis: increasing [Ca2+]i and depleting [Ca2+]o [5], [42], [58], both of which boost exocytosis. Ca2+ may affect fusion pore/vesicle exocytosis from both sides of the plasma membrane: from the well-known intracellular site via Ca2+-sensing proteins, and from the as yet unknown extracellular site of the ECIE. To interpret less ECIE at higher [Ca2+]i, we hypothesize (1) the opening of a fusion pore is contributed to by both an intracellular force (Fin by synaptotagmins [51]) and another extracellular force (Fout of ECIE); (2) the total threshold force to open the fusion pore is Ftotal, such that Fin+Fout≥Ftotal is the condition to open fusion pore/exocytosis. This hypothesis explains that increasing Fin by higher [Ca2+]i reduces the effect of [Ca2+]o on ECIE.

ECIE does not at all contradict the well-established Ca2+ hypothesis of vesicle exocytosis [1], [59]. Instead, our data confirmed that the exocytosis was strictly dependent on [Ca2+]i. As an additional effect of [Ca2+]o, ECIE had never been examined previously under proper conditions. Most previous work used membrane depolarization/APs to trigger vesicle release [1], [3], [32], [59]. Synaptic vesicles are docked for exocytosis in active zones where Ca2+ channels are also clustered. AP-triggered vesicle release is induced by Ca2+ microdomains near the Ca2+ channel (at 20 nm, ∼100 µM; at 200 nm, ∼5–10 µM) [32], [60]. ECIE was best seen when a subthreshold [Ca2+]i (2∼3 µM) was triggered (Figure 1). A higher [Ca2+]i (∼6 µM) greatly attenuated this inhibition (Figure 3). Under physiological conditions, AP-induced secretion is achieved by mixed effects of both [Ca2+]i and [Ca2+]o. This explains why extracellular inhibition of exocytosis was overlooked by most previous work.

Regarding mechanisms of ECIE, (1) we investigated whether ECIE is mediated via the CaSR, which is a known [Ca2+]o sensor inhibiting cell secretion in parathyroid cells [43], [61]. Although Ca2+ and other extracellular calcimimetics (Mg2+, Cd2+ and G418) inhibited exocytosis, we failed to link the CaSR to ECIE (Figure 5; Figures S7, S8). An unknown Ca2+-binding protein at the extracellular side of the plasma membrane may be responsible for ECIE. (2) Generally Ca2+-dependent secretion is SNARE-dependent, although did not determine whether ECIE is SNARE dependent, because neurotoxins against somatic secretion in DRG neurons are not available. (3) Conformational changes of the Ca2+ channel induced by membrane depolarization and Ca2+ binding are reported to facilitate vesicle exocytosis in PC12 cells and chromaffin cells [62], [63]. This depolarization effect shares with ECIE the facilitation of exocytosis without Ca2+ influx. However, this effect is different from ECIE since there is no membrane depolarization to open Ca2+ channel for the Ca2+ entry. (4) It is known that extracellular Ca2+ could shifts the voltage-dependent gating of Na+ and Ca2+ channels through a surface charge effect [7], [64]. Future work may determine whether EICE is altered by the surface charge effect.

In summary, our study demonstrated that extracellular Ca2+ directly regulated exocytosis. It is known that under pathological conditions, such as brain ischemia, extracellular Ca2+ and Mg2+ decrease greatly [6], [65]. This may directly enhance synaptic transmission according to our finding. ECIE was most effective at moderate levels (≤6 µM), but less effective in high [Ca2+]i. Thus, ECIE may affect neurotransmitter release more in neural somata than in synapses, where synchronized exocytosis is triggered by Ca2+ microdomains (10∼100 µM) following an AP [32]. ECIE could be a new extension of the classic Ca2+ hypothesis of exocytotic transmitter release [1], [60].

Facilitation of photolysis-induced exocytosis by cAMP elevation with forskolin in DRG neurons. (A) ΔCm and [Ca2+]i signals in response to Ca2+ release by trains of UV flashes (4 flashes at 0.5 Hz). The neuron was perfused with standard bath solutions containing 0 mM Ca2+, 0 mM Ca2++100 µM forskolin, 2.5 mM Ca2+, and 2.5 mM Ca2++100 µM forskolin, respectively, at 120 s intervals. UV flashes produced similar [Ca2+]i changes, while the corresponding exocytosis was facilitated by forskolin both in the presence and absence of external Ca2+. The initial value of Cm is noted. [Ca2+]i was monitored by Fura-6F measurements. (B) Comparison of averaged ΔCm traces with forskolin in 0 Ca2+ (left, n = 4) and 2.5 mM Ca2+ (right, n = 4). (C–D) Average results from 4 cells showing that forskolin enhanced exocytosis in both 0 and 2.5 mM Ca2+. Exocytosis in 0 and 2.5 mM Ca2+ in the absence of forskolin was 70±8% and 22±9% of that with forskolin treatment. The corresponding [Ca2+]i values were similar in the presence and absence of forskolin. Compared to the [Ca2+]i rise with forskolin application, the [Ca2+]i rise was 88±6% in 0 [Ca2+]o and 103±5% in 2.5 mM [Ca2+]o solutions.

(DOC)

Click here for additional data file.

ECIE measured by combined membrane capacitance and amperometry in cultured rat adrenal chromaffin cells. (A) Representative ΔCm and amperometric current (Iamp) responses to [Ca2+]i rises induced by UV flashes in the presence (2.5 mM) or absence (0 mM) of external Ca2+. The arrowheads indicate the time points of UV flashes. A [Ca2+]i increases and secretion signals (ΔCm and Iamp or ∫Iampdt, the oxidization charge proportional to the number of oxidized catecholamines) were first induced by a train of UV flashes in 2.5 mM extracellular Ca2+ solution (2.5 Ca2+) (left). Subsequently, another similar [Ca2+]i increase induced by the second UV train produced much larger secretion signals in Ca2+-free solution (0 Ca2+) (middle). Finally, a similar [Ca2+]i rise induced by the third UV train triggered secretion signals similar to that by the first UV train when extracellular solution was changed back to 2.5 mM (right). (B) Statistics. Following UV flashes, the secretion signals were significantly smaller in 0 vs 2.5 mM Ca2+ (ΔCm, spike numbers and ∫Iampdt were 69±5, 68±6 and 59±9% of controls), while the [Ca2+]i values were similar (107±4% of control).

(DOC)

Click here for additional data file.

Extracellular Ca Increase of [Ca2+]i to ∼8 µM almost abolished the extracellular Ca2+ inhibition of photolysis-induced exocytosis in DRG neurons.

(DOC)

Click here for additional data file.

Positive control of specific reagents (calhex and calindol) against CaSR. (A) Typical trace of [Ca2+]i measurement in CaSR-transfected HEK 293 cells. Cells initially bathed in 10 µM [Ca2+]o solution. [Ca2+]o at 1 mM induced a [Ca2+]i rise, which was smaller in the presence of calhex (1 µM) (n = 4). [Ca2+]i was monitored using Fura-2 AM. (B) Cells bathed in 0.25 mM [Ca2+]o solution. Calindol (1 µM) induced a [Ca2+]i rise (n = 5).

(DOC)

Click here for additional data file.

Calhex and calindol had no effect on [Ca (A) [Ca2+]i rise induced by caffeine was monitored by Fura-2 measurements. Left traces show the [Ca2+]i rise induced by 20 mM caffeine in the presence of 2.5 mM [Ca2+]o and 1 µM calhex. Right traces show the [Ca2+]i rise induced by caffeine in the presence of 0 mM [Ca2+]o and 1 µM calindol. [Ca2+]i was monitored by Fura-2 measurements. There was a 90-s interval between continuous recordings. (B) Statistics of the caffeine-induced [Ca2+]i rise. Compared to signals in 0 mM [Ca2+]o and 2.5 mM [Ca2+]o, the [Ca2+]i rise was similar in calindol (97±4%, n = 5) or calhex (100±2%, n = 5).

(DOC)

Click here for additional data file.

NSCCs did not exist on DRG neurons. DRG neurons were on-cell patched. The upper four traces show consecutive current recordings at 60-s intervals in different perfusion solutions (#1, 2.5 mM Ca2+ and 1 mM Mg2+; #2, 0.1 mM Ca2+ and 0 mM Mg2+; #3, 0 mM Ca2+ and 0 mM Mg2+; #4, 2.5 mM Ca2+ and 1 mM Mg2+). The lowest trace shows the stimulation pulse. DRG neurons were stimulated with a step depolarization from −40 mV to +150 mV, followed by a hyperpolarization to −100 mV. Voltage-induced membrane currents in different external solutions were the same and this excludes the existence of NSCCs in the somata of DRG neurons [9], [50].

(DOC)

Click here for additional data file.

CaSR is not linked to SNARE complex. (A) Immunoblot of exocytosis-related proteins in DRG extracts with antibodies against syntaxin 1, synaptotagmin 1, SNAP 25, and complexin 1 and 2. GAPDH was used as a control. For the complexin doublet, the upper band is complexin 2 and the lower band is complexin 1. (B) Immunoprecipitation using a polyclonal antibody against aa 45–81 in complexin 2 with or without 3.5 mM Ca2+. Bound proteins were detected with monoclonal antibodies to syntaxin 1, SNAP 25, and CaSR. The upper strip shows that complexin antibody immunoprecipitated syntaxin 1 and SNAP 25. The lower strip shows no CaSR in the complexin immunocomplex.

(DOC)

Click here for additional data file.

CaSR knockdown had no effect on ECIE. (A) Western blot of CaSR and β-actin in shRNA 1 (sh 1), shRNA 2 (sh 2) and control shRNA-treated HEK 293 cells. Note that the lower band (CaSR) is missing in sh 1 and sh 2-treated cells. The upper band labeled with an asterisk (*) is a non-specific band recognized by the antibody. (B) Electrophysiology of ECIE in control shRNA, sh 1 and sh 2-treated DRG neurons. Experiments were done 4–5 days after transfection. (C) Statistics of normalized exocytosis and [Ca2+]i rise. The caffeine (20 mM)-triggered exocytosis in 2.5 mM Ca2+ was reduced to 41±8% (n = 6), 40±5% (n = 6) and 50±6% (n = 6) in control shRNA, sh 1 and sh 2, respectively. Extracellular Ca2+ inhibition was unaffected in sh 1 and sh 2-treated DRG neurons compared to that of control. [Ca2+]i was monitored by Fura-2 measurements. Compared to the [Ca2+]i rise in 0 mM Ca2+, the caffeine-induced Ca2+ rise was greater in 2.5 mM Ca2+ in the control cells (135±11%, p = 0.03, n = 5). In sh 1 and sh 2-transfected neurons. the caffeine-induced [Ca2+]i rise was similar to that in 0 mM Ca2+ and 2.5 mM Ca2+ (sh 1-tranfected neurons, 112±18%, n = 5; sh 2-transfected neurons, 115±11%, n = 6).

(DOC)

Click here for additional data file.

Physiological levels of extracellular Ca (A) Left, amperometric recording from an adrenal slice. A 34% reduction of [Ca2+]o increased the 20 mM caffeine-induced amperometric signal. Upper traces show amperometric recordings, lower traces show the corresponding ∫Iampdt signals. Right, the caffeine-induced [Ca2+]i rise was similar when changing [Ca2+]o in another rat adrenal slice. [Ca2+]i was measured using Fura-2 AM. (B) Statistics of normalized exocytosis and [Ca2+]i rise signals. A 34% [Ca2+]o reduction increased caffeine-induced exocytosis signals by 68±12% in chromaffin cells (n = 7, p<0.001). The caffeine (20 mM)-induced [Ca2+]i rise was similar at [Ca2+]o of 1.65 mM and 2.5 mM in chromaffin cells (107±7%, n = 15).

(DOC)

Click here for additional data file.