Calcium (Ca2+)-dependent endocytosis has been linked to preferential Ca2+ entry through the L-type (α1D, CaV1.3) of voltage-dependent Ca2+ channels (VDCCs). Considering that the Ca2+-dependent exocytotic release of neurotransmitters is mostly triggered by Ca2+ entry through N-(α1B, CaV2.2) or PQ-VDCCs (α1A, CaV2.1) and that exocytosis and endocytosis are coupled, the supposition that the different channel subtypes are specialized to control different cell functions is attractive. Here we have explored this hypothesis in primary cultures of bovine adrenal chromaffin cells where PQ channels account for 50% of Ca2+ current (ICa), 30% for N channels, and 20% for L channels. We used patch-clamp and fluorescence techniques to measure the exo-endocytotic responses triggered by long depolarizing stimuli, in 1, 2, or 10 mM concentrations of extracellular Ca2+ ([Ca2+]e). Exo-endocytotic responses were little affected by ω-conotoxin GVIA (N channel blocker), whereas ω-agatoxin IVA (PQ channel blocker) caused 80% blockade of exocytosis as well as endocytosis. In contrast, nifedipine (L channel blocker) only caused 20% inhibition of exocytosis but as much as 90% inhibition of endocytosis. Conversely, FPL67146 (an activator of L VDCCs) notably augmented endocytosis. Photoreleased caged Ca2+ caused substantially smaller endocytotic responses compared with those produced by K+ depolarization. Using fluorescence antibodies, no colocalization between L, N, or PQ channels with clathrin was found; a 20–30% colocalization was found between dynamin and all three channel antibodies. This is incompatible with the view that L channels are coupled to the endocytotic machine. Data rather support a mechanism implying the different inactivation rates of L (slow-inactivating) and N/PQ channels (fast-inactivating). Thus a slow but more sustained Ca2+ entry through L channels could be a requirement to trigger endocytosis efficiently, at least in bovine chromaffin cells.
- voltage-dependent calcium channels
- endocytic proteins
neuronal communication at synapses is exerted through a Ca2+-dependent release mechanism (37), an observation that is also true for neuroendocrine secretion (18). Such process underlies the fusion of storing vesicle membranes with the plasmalemma (exocytosis) and the subsequent membrane retrieval (endocytosis). The coupling of these two processes is essential to keep an equilibrium between the amount of vesicular membrane incorporated into the plasmalemma during exocytosis and membrane retrieved during the subsequent endocytotic phase. This will preserve the size of nerve terminals and neurosecretory cells and will also warrant that a given number of secretory vesicles are available to participate in subsequent rounds of exocytosis during repetitive cell activation (11, 14, 60, 62, 63).
Calcium has been suggested to affect endocytosis and vesicle recycling in synapses and neuroendocrine cells (50, 68, 69). Much is known about Ca2+ dependence of exocytosis in these systems (19, 28, 37, 49); however, controversy exists over the manner in which membrane retrieval during endocytosis is affected by Ca2+. For instance, compensatory and excess endocytosis represent two independent Ca2+-regulated mechanisms of rapid internalization in bovine chromaffin cells (14, 21). The existence of these two different Ca2+ sensors is supported by data of Nucifora and Fox (51) in which Ca2+ and Ba2+ support excessive membrane retrieval in bovine chromaffin cells. In contrast, a previous study on the same cells showed that rapid endocytosis was supported by Ca2+ but not by Sr2+ or Ba2+ (5). There are additional studies in synapses concluding that Ca2+ was not required for endocytosis (61, 74) or behaved even as an inhibitor (59).
An additional problem emerges when considering the Ca2+ pathway that controls exo-endocytotic processes. The expression in neurons of multiple types of voltage-dependent Ca2+ channels (VDCCs) (22, 27, 53) and in adrenal medullary chromaffin cells (28) poses the interesting question of their possible specialization in controlling different functions. This seems clear in neurons, where the geographical segregation of different subtypes of VDCCs to dendrites, axon terminals, or somata facilitates their specialization for specific functions. For instance N-(α1B, CaV2.2) and PQ-types (α1A, CaV2.1) VDCCs, which are preferentially found along the length of apical dendrites and in axon terminals that synapse with dendrites (72), control the release of various neurotransmitters (73). On the other hand, L (α1D, CaV1.3)-type channels located on proximal dendrites and neuronal cell bodies (1, 70, 72) have been associated with the regulation of gene expression and enzyme activities in cortical and hippocampal neurons (6, 16, 48).
Alike neurons, adrenal medullary chromaffin cells express L-, N-, PQ-, and R-(α1E, Cav2.3)-type of VDCCs. But unlike neurons, these cells lack dendrites and axons. Thus studies on the specialization of VDCC subtypes in cultured spherical chromaffin cells have provided unclear results (28). Depending on the stimulus, i.e., long/short single depolarizing pulses (DPs), trains of short DPs or action potentials, animal species, and the methods used to measure exocytosis or catecholamine release certain specialization of the VDCCs to modulate exocytosis has been suggested. For instance, L-type channels have been preferentially associated to release responses in some studies (36, 44). Others attributed more protagonism to N channels (52) or to PQ channels (4, 41, 57, 75). Also some others concluded that exocytosis was proportional to Ca2+ entry and the density of Ca2+ current carried by each Ca2+ channel subtype in bovine (20, 45, 65) and rat chromaffin cells (10), with no specialization at all.
Concerning the possible specialization of VDCC subtypes in controlling endocytosis, we provided evidence that L channels were much more relevant than N or PQ channels in controlling endocytosis in voltage-clamped bovine adrenal chromaffin cells after application of single depolarizing pulses (DPs) of long duration (57). This functional coupling between L channels and endocytosis was also found at the mouse neuromuscular junction (56) and Drosophila synapse (39). The present study was planned to explore the hypothesis that the mode of Ca2+ entry through a given VDCC subtype, rather than the total amount of Ca2+ entry into the cell, is more relevant to trigger endocytosis. We have used bovine adrenal chromaffin cells because they express the same densities of VDCC subtypes than human chromaffin cells (26, 54), namely, 20% L channels, 30% N channels, and 50% PQ channels (28). Despite that L channels contribute to only 20% of total whole cell current through VDCCs (2, 24), Ca2+ entering through slow-inactivating L-type channels seems to be functionally coupled to endocytosis. This supposition arises from experiments presented here, combining patch-clamp techniques, fluorescence techniques with the probe FM1–43, photorelease of caged Ca2+, and colocalization techniques through the fluorescence labeling of VDCC subtypes and of the endocytosis proteins clathrin and dynamin.
Isolation and culture of bovine chromaffin cells.
Bovine adrenal glands were obtained from a local slaughterhouse. Chromaffin cells were isolated by digestion of the adrenal medulla with collagenase following standard methods (43) with some modifications (47). Cells were suspended in Dulbecco's modified Eagle's medium (DMEM) supplemented with 5% fetal calf serum, 10 μM cytosine arabinoside, 10 μM fluorodeoxyrudine, 50 IU/ml penicillin, and 50 μg/ml streptomycin. Cells were plated on 12-mm diameter polylysine-coated glass coverslips at a density of 5 × 104 cells/coverslip for patch-clamp experiments or 5 × 105 cells/coverslip for FM1–43 and flash-photolysis experiments. Cells were kept in a water-saturated incubator at 37°C in a 5% CO2-95% air atmosphere and used 2–5 days thereafter.
Recording of Ca2+ currents and membrane capacitance.
Calcium currents (ICa) and membrane capacitance (Cm) were measured under the perforated patch configuration of the patch-clamp technique (29, 32, 38). During recording, cells were constantly perifused with a standard control solution at pH 7.4 containing (mM) 137 NaCl; 1 MgCl2; 1, 2, or 10 CaCl2; 10 glucose; and 10 HEPES. Internal perforated-patch solution had the following composition (mM): 135 Cs-glutamate, 10 HEPES, and 9 NaCl, and pH 7.2 adjusted with CsOH. An amphotericin B stock solution was prepared every week at 50 mg/ml in DMSO, stored at 4°C, and protected from light. Fresh perforated patch pipette solution was prepared every day by addition of 10 μl stock amphotericin B to 0.5 ml Cs-glutamate internal solution; this solution was sonicated thoroughly, protected from light, and kept in ice. Patch pipettes had their tips dipped in amphotericin-free solution for 5–10 s and back-filled with freshly mixed amphotericin-containing solution. For patching the cells, pipettes of ∼2–3 MΩ resistance were pulled from borosilicate glass, partially coated with molten dental wax, and lightly fire polished. Cells were continuously perifused with external solution at a rate of 1 ml/min, and all experiments were performed at room temperature (25–28°C). Electrophysiological data were carried out using an EPC-9 amplifier under the control of Pulse software (HEKA Elektronik). Cell Cm changes were estimated by the Lindau-Neher technique (42). A 400-ms sinusoidal wave (1 kHz, 60 mV peak to peak amplitude) was then added before the protocol and a 50-s sinusoidal wave of the same characteristics after it to allow for the computation of membrane capacitance change. Membrane current was sampled at 20 kHz. Calcium channel blockers were perifused as indicated. Cells were held at −80 mV, and single depolarizing pulses to voltages where ICa peak was reached were applied at 5-min intervals. Cells were perforated to a series resistance of no more than 30 MΩ, which usually happened within 10-min after sealing. External solutions were exchanged by a fast perifusion device consisting of a modified multibarrelled pipette, the common outlet of which was positioned 50–100 μm from the cell. Control and test solutions were changed using miniature solenoid valves operated manually (25).
For experiments with FPL64176 in perforated-patch clamp configuration, cells were preincubated during 45 min with 100 μM of the cell-permeable EGTA/AM at 37°C; cells were then washed during 15 min at room temperature with the extracellular solution. Cell loading with EGTA was performed to prevent Ca2+-dependent inactivation of VDCCs by excessive intracellular Ca2+ ([Ca2+]c) rise (58).
Fluorescence imaging of exocytosis and endocytosis.
Exocytosis and endocytosis were monitored using the membrane fluorescence indicator FM1–43 (63). The optical measurement of FM1–43 was performed as described previously (55). Before experiments, cells attached to the coverslip were washed twice for 5 min with standard external solution. Thereafter, the coverslip was mounted in the stage of a Zeiss Axiovert 100 S inverted microscope using a ×100 Plan-Neofluor objective and incubated in the standard external solution that contained 3 μM FM1–43 during 10 min. Fluorescence of FM1–43 was excited with 480 nm light and then collected at 535 nm. Images were acquired every second. Total cellular FM1–43 fluorescence was measured from individual cells, and exo-endocytosis was monitored using the protocol described in Fig. 7. Background fluorescence was measured in the same way from regions containing no cells and was subtracted from total cellular fluorescence. Total cellular FM1–43 fluorescence was normalized with respect the total basal cell fluorescence after 10 min of FM1–43 exposition.
Flash photolysis experiments.
Flash photolysis experiments were performed as described (30). Photorelease of cytosolic caged Ca2+ was achieved using a 25 μM concentration of ο-nitrophenyl EGTA/AM in Krebs-HEPES buffer with the following composition (in mM): 145 NaCl, 5.6 KCl, 1.2 MgCl2, 2 CaCl2, 10 glucose, and 15 HEPES, pH adjusted to 7.4 using an NaOH solution (15). To facilitate cell loading with the indicator, stock solutions of NP-EGTA.AM were dissolved in DMSO with 10% wt/vol Pluronic F-127. After 45-min incubation, three washes were performed, and the glass coverslip containing the cells was mounted in the stage of a Zeiss Axiovert 100 S inverted microscope using a ×100 Plan-Neofluor objective. In this system a dual-port condenser allowed excitation of the specimen with a monochromator Polychrome IV (8% incoming light, Till-Photonics, Munich, Germany) and simultaneous application of a 5-ms UV flash using a pulsed Xenon arc lamp system (92% incoming light, UV Flash II, Till-Photonics), therefore allowing [Ca2+]c changes measurement using Fura-2 excitation at 340/380 nm (for this purpose, 4 μM Fura-2/AM was incubated simultaneously with the NP-EGTA/AM) (30). Fluorescence emission from a 20 × 20-μm restricted area (Viewfinder III, Till-Photonics) was detected in a photomultiplier tube (Hamamatsu, Japan). Control of excitation light and acquisition were performed with PCLAMP 8.0 software (Axon Instruments, Foster City, CA) on a PC. FM1–43 was present at a concentration of 3 μM in the Krebs-HEPES buffer bathing the cells; in this manner, the rapid release of vesicles caused by the fast Ca2+ elevation could be detected by the increase in FM1–43 fluorescence following vesicle membrane incorporation by exocytosis (63). FM1–43 fluorescence after excitation at 475 nm was determined and integrated at 50-ms intervals using a fluorescein emission filter set.
Colocalization studies using fluorescence-coupled antibodies.
For immunofluorescence and confocal microscopy, cells were fixed using a 2% paraformaldehyde (PFA) in phosphate-buffered saline solution (PBS) for 15 min at room temperature. Then cells were permeabilized with 0.1% Triton X-100 during 1 min incubation and washed twice with PBS within 10 min.
To detect clathrin, the cells were incubated for 45 min with a mouse monoclonal anti-clathrin (X22) antibody (7) (Abcam, Cambridge, UK) diluted 1:100 in PBS containing 1% goat serum. To detect dynamin, mouse anti-dynamin [clone 41; an antibody that detects all dynamin isoforms (23)] at a 1:100 dilution in PBS (containing 1% goat serum) was used as the primary antibody. After extensive washes, secondary goat anti-mouse antibodies coupled to Alexa-488 (1:1,000 dilution in PBS, Invitrogen) were added, and after 45 min unbound material was washed away using PBS.
Calcium channel subtypes were labeled using rabbit polyclonal antibodies (Alomone Labs, Jerusalem, Israel). Anti-L-type (CaV1.3) channels were raised against a highly purified peptide corresponding to the residues 859–875 of the α1D subunit of rat brain voltage-dependent calcium channels (35) and affinity purified using a column immobilizing the same peptide. Anti-P/Q (CaV2.1) channels were produced against residues 865–881 of α1A subunit (64) of rat brain voltage-dependent calcium channels. Similarly, anti-N-type (CaV2.2) channel antibodies were produced against residues 851–867 of α1B subunit from the same biological source (71). Immunolabeling was performed as described above using a 1:100 dilution of the antibodies and the nonspecific labeling investigated using antibody solutions preincubated with 3 μg/ml of the corresponding peptides during 1 h at room temperature. Goat anti-rabbit secondary antibodies coupled to Alexa-546 (1:1,000 dilution in PBS) were used to visualize epitope distribution.
Fluorescence was investigated using a Leica TCS SP2 confocal microscope with a ×100 objective (spatial resolution based in this objective and the scanner characteristics was estimated in 60–80 nm for 2–3 pixel separations). This system allows for z-axis reconstruction with theoretical z slice of about 0.5 μm thick and sequential mode studies in double labeling experiments. Images were processed using the LAS-AF (Leica Application Suite Advance Fluorescence; Leica Microsystems) program, and colocalization analysis were done using images with the same magnification.
Materials and solutions.
The materials used were the following: collagenase type I, nifedipine, FPL64176, and amphotericin B were from Sigma (Madrid, Spain); Dulbecco's modified Eagle's medium (DMEM), bovine serum albumin fraction V, fetal calf serum, and antibiotics were from GIBCO (Madrid, Spain). Agatoxin-IVA was from Peptide Institute (Sandhausen, Germany). ω-Conotoxin GVIA and ω-conotoxin MVIIC were from Bachem Feinchemikalien (Bubendorf, Switzerland). EGTA/AM was from Calbiochem (Barcelona, Spain). FM1–43, ο-nitrophenyl EGTA-AM (NP-EGTA/AM), and Alexa-labeled secondary antibodies were from Invitrogen (Eugene, OR). ω-Conotoxin GVIA and ω-conotoxin MVIIC were dissolved in distilled water and stored frozen in aliquots at 0.1 mM. Nifedipine (10 mM), NP-EGTA/AM (2.5 mM), and Fura-2/AM (2 mM) were prepared in dimethylsulfoxide (DMSO); care was taken to protect solutions from light. FPL64176 at 1 mM was prepared in ethanol, kept at −20°C in aliquots, and protected from light. Final concentrations of drugs were obtained by diluting the stock solution directly into the extracellular solution. At these dilutions, solvents had no effect on the parameters studied. FM1–43 was dissolved in distilled water at a concentration of 1 mM.
The whole cell inward ICa was analyzed after the initial 10 ms of each depolarizing pulse to get rid of the Na+ current (INa). In electrophysiology experiments, exo- and endocytosis were measured by monitoring changes in cell capacitance. In the experiments to study the effect of FPL on exo-endocytosis, QCa was measured until 3 ms after the end of the depolarizing pulse to calculate the Ca2+ influx, which takes place during the inward and tail current. Exocytosis peak was measured by subtracting the basal mean Cm obtained 400-ms previous to depolarization to 50-ms after the end of the depolarizing pulse to avoid a possible Na+ channel gating artifact (34). After the exocytotic peak, Cm changes were measured during the ensuing 50-s period; endocytosis was calculated as the difference in Cm at the beginning and the end of such 50-s period. For imaging experiments with FM1–43 and flash photolysis, the initial size of the fluorescence signal after 10-min of FM1–43 perifusion was used to normalize subsequent fluorescence changes triggered by exocytosis. Endocytosis was calculated as the difference between fluorescence after wash and fluorescence before beginning the experiment (55). Comparisons between means of group data were performed by one-way analysis of variance (ANOVA) followed by Duncan post hoc test when appropriate. A P value equal or smaller than 0.05 was taken as the limit of significance.
Calcium gradients and the relationship among Ca2+ entry, exocytosis, and endocytosis.
All subtypes of VDCCs undergo a pronounced Ca2+-dependent inactivation in voltage-clamped bovine chromaffin cells. However, such inactivation is substantially smaller for L channels compared with N or PQ channels (33). It was therefore expected that the rate and amount of Ca2+ entry through each channel type could undergo visible changes in ICa, QCa, ΔCm (exocytosis), and Cm decay (endocytosis) upon application of DPs at different extracellular Ca2+ ([Ca2+]e). We therefore performed experiments to explore the variations of those four parameters under conditions of cell perifusion with external solutions containing 1 mM Ca2+ (1Ca2+), 2 mM Ca2+ (2Ca2+), or 10 mM Ca2+ (10Ca2+). Cells were voltage clamped at −80 mV, and DPs were applied to generate ICa and the corresponding Cm changes. Current-voltage curves revealed that maximum peak ICa was achieved upon application of test pulses to 0 mV at 1 and 2 mM [Ca2+]e and to +10 mV at 10 mM [Ca2+]e. Thus at the beginning of each experiment the test pulse giving maximum ICa was determined and used throughout the experiment. Concerning the length of the test pulse used here, we studied ICa and Cm changes generated by pulses of 50 to 2,000 ms duration; however, a clear endocytotic response was visible from 200 ms onwards. Hence, we decided to use single DPs of 500 ms duration. With the perforated-patch configuration we learned in a previous report that the sequential application of 500 ms DPs to a cell at 5-min intervals generated reproducible ICa and Cm changes (57). Thus we applied this protocol here.
Endocytosis is a mechanism that depends on the prior exocytosis, and both are known to be Ca2+-dependent processes. The supposition that variations of extracellular/intracellular Ca2+ gradients could distinctly affect ΔCm and Cm decay was tested in cells being perifused with extracellular solutions containing 1Ca2+, 2Ca2+, or 10Ca2+. The example ICa and Cm traces shown in Fig. 1A were obtained with different perifused extracellular solutions containing 1Ca2+, 2Ca2+, and 10Ca2+. Current traces show an initial current peak generated by Na+ entry (INa) that inactivated quickly, giving rise to a slow inactivating ICa. Both ICa and ΔCm were substantially higher in 10Ca2+ compared with 1Ca2+ and/or 2Ca2+. This is better seen in Fig. 1, B–D, where quantitative averaged results from 15 cells were plotted. ICa peak rose almost threefold when 1Ca2+ was switched to 10Ca2+ solution. This increase was even higher for QCa that rose near fivefold on switching the extracellular solution from 1Ca2+ to 10Ca2+. Exocytotic and endocytotic responses rose as the Ca2+ gradient increased. For instance, exocytosis augmented 2.2-fold in 2Ca2+ and 6-fold in 10Ca2+, with respect to 1Ca2+. This increase was even more drastic for endocytosis that rose 4-fold in 2Ca2+ and near 10-fold in 10Ca2+, with respect to 1Ca2+. However, ratios between endocytosis and exocytosis were similar at the three [Ca2+]e studied (Fig. 1D).
A quite close correlation between QCa and Cm changes was found. This is better seen in Fig. 2 where regression plots on exocytosis/QCa (A), endocytosis/QCa (B), and endocytosis/exocytosis (C) are shown. There is a reasonable correlation between Ca2+ entry, exocytosis, and endocytosis, indicating that the three events are acting in a coordinated manner.
Blockade of VDCC subtypes does not always lead to a parallel blockade of exocytosis and endocytosis.
As described above, a positive correlation among an increased Ca2+ entry gradient, exocytosis, and endocytosis was apparent. Because in bovine chromaffin cells the VDCC of the L-, N-, and PQ-subtype contribute differently to exocytosis during stimulation (28), the question arose as to whether the Ca2+ entering through each channel type contributes similarly or differently to the control of endocytotic responses. To explore this question, selective blockers of those channels were used. We and many others have widely used these blockers in bovine chromaffin cells and found that 3 μM nifedipine, 1 μM ω-conotoxin GVIA (GVIA), or 1 μM ω-agatoxin IVA (Aga-IVA) fully block L, N, or PQ channels, respectively (reviewed in Ref. 25). These were the concentrations used in this study. The experiments were performed using 2 and 10 mM external Ca2+ as charge carrier and two initial control DPs were applied, followed by another two DPs given in the presence of the blocker that was perifused since 1–3 min before and during the two DPs. Figures 3 and 4 and Table 1 display the results obtained with the three blockers.
In 10 mM Ca2+, nifedipine caused 25% and 50% reduction of ICa and QCa. Greater reduction of QCa was likely due to higher Ca2+-dependent inactivation of N and PQ channels (33, 58). This led to a 22% reduction of ΔCm and to as much as 93% reduction of Cm decay (Fig. 3; Table 1). In 2 mM Ca2+ nifedipine caused similar effects on the four parameters, namely 31%, 53%, 26%, and 92% reduction for, respectively, ICa, QCa, ΔCm, and Cm decay (Table 1).
GVIA blocked ICa to an extent similar to nifedipine, i.e., 29% in 2Ca2+ and 33% in 10Ca2+. However, QCa blockade was significantly lower than that of nifedipine, namely 23% in 2Ca2+ and 35% in 10Ca2+. This could be explained by greater Ca2+ entry along the 500-ms pulse via slow-inactivating L channels. Exocytosis was unaffected and endocytosis was reduced 33% in 2Ca2+ and 40% in 10Ca2+ (Fig. 3, Table 1).
At 1 μM, Aga-IVA reduced ICa by as much as 68% in 2Ca2+ and 51% in 10Ca2+. QCa was reduced to a similar extent, 69% in 2Ca2+, and 64% in 10Ca2+. Exocytosis was markedly reduced, i.e., 82% in 2Ca2+ and 66% in 10Ca2+, and endocytosis was also largely reduced, i.e., 68% in 2Ca2+ and 90% in 10Ca2+ (Fig. 3, Table 1).
Thus it seemed that a defined blockade of a subcomponent of ICa and QCa did not always lead to parallel inhibitions of ΔCm and Cm decay. This is clearer when looking at the ratios between endocytotic and exocytotic responses (Endo/Exo; Fig. 3). At 10 mM Ca2+, the Endo/Exo ratio in nifedipine-treated cells was 0.19, meaning that the exocytotic response was much more pronounced than the endocytotic response and that nifedipine was markedly blocking endocytosis with scarce exocytosis blockade. In the case of GVIA-treated cells, such ratio was 0.77, indicating the presence of a healthy endocytotic response. Aga-IVA-treated cells showed an intermediate behavior: both exo and endocytotic responses were markedly blocked and thus the Endo/Exo ratio was 0.54.
It seemed interesting to compare the behavior of cells treated with nifedipine with others treated with Aga-IVA but at a concentration that blocked ICa and QCa to an extent similar to that produced by nifedipine. In doing so, it was plausible that cell endocytotic responses differed provided the supposition was true that Ca2+ entry through L channels was more prone to trigger endocytosis compared with Ca2+ entering through PQ channels. At 10 mM Ca2+ (Fig. 4 and Table 1), 200 nM Aga-IVA caused 45% reduction of ICa and 59% reduction of QCa; these values were comparable with those of nifedipine. Despite this, 200 nM Aga-IVA reduced endocytosis only by 37%, whereas nifedipine caused 93% inhibition. Conversely, nifedipine caused 22% inhibition of exocytosis, whereas 200 nM AgaIVA caused 41% blockade (**P < 0.01). Thus it seems that more relevant that the total Ca2+ entry into the cell during the 500-ms depolarizing pulse is the pathway used by Ca2+ to trigger endocytosis, i.e., the slow-inactivating L type of VDCCs.
Enhanced Ca2+ entry through L-type VDCCs selectively augments endocytosis.
If restriction of Ca2+ entry via L channels aborts endocytosis, then augmentation of Ca2+ entry through this pathway should enhance this response. This supposition was tested using the L-type Ca2+ channel activator FPL64176 (FPL). To unmask FPL-elicited augmentation of ICa under the perforated patch configuration of the patch-clamp technique, we recoursed to cell loading with EGTA/AM before doing the experiments. In this manner, we could prevent Ca2+-dependent inactivation of VDCCs and see augmentation of ICa (58).
Figure 5A shows two pairs of ICa traces generated by 500-ms DPs to −10 mV (left) and 0 mV (right). In the presence of 1 μM FPL, ICa inactivated more slowly and deactivated also at a slower rate. Cm traces shown in Fig. 5B had a similar ΔCm; however, the trace obtained in the presence of FPL showed a drastic increment of endocytosis compared with the trace obtained before the compound. Figure 5, C and D, is averaged results showing that FPL caused around 11% and 18% increase of ICa and QCa, respectively. Augmentation of Ca2+ entry gave rise to an almost threefold increase of the Endo/Exo ratio.
Although this experiment supported the supposition of enhanced endocytosis by augmented Ca2+ entry via L channels, an additional experiment with isolation of L from N/PQ channels was performed. To block N/PQ channels, 2 μM ω-conotoxin MVIIC (MVIIC) was used (28). Figure 6A shows three Cm recordings obtained upon challenging a cell with three 500-ms DPs given before (control), during MVIIC treatment, and after combined MVIIC plus FPL treatment. MVIIC reduced the endocytotic response; when given on top of MVIIC, FPL augmented endocytosis well above the control level. Averaged data are plotted in Fig. 6, B–F. MVIIC reduced ICa by 51% and QCa by 59%. Added on top of MVIIC, FPL restored ICa to 80% of the initial control and QCa to 75% of the initial control. This enhanced Ca2+ entry, exclusively through the L channel, doubled the Endo/Exo ratio, indicating a selective augmentation of endocytosis.
We also explored the possibility that more sustained Ca2+ entry through non-L (N/PQ) VDCCs could also activate endocytosis. To this aim we used roscovitine that at 100 μM is known to delay N/PQ channel deactivation, thereby prolonging channel opening time (8, 9, 17, 76). In nine cells under control conditions, DPs produced an endocytotic response of 95 ± 13 fF. In the presence of nifedipine, endocytosis was reduced to 31 ± 5 fF. When roscovitine was given, still in the presence of nifedipine, endocytosis resumed and reached 110 ± 20 fF.
Monitoring exo-endocytotic responses with FM1–43 dye: effects of VDCC blockers.
To get further insight into the supposition that the various VDCC pathways contribute differently to endocytotic response, chromaffin cells were challenged with a high-K+ solution in the presence of FM1–43, a styryl hydrophilic dye widely used to study vesicle recycling. In response to long-lasting depolarization, the dye inserts into the outer leaflet on the plasma membrane and is internalized during vesicle retrieval (63). Since the remaining extracellular dye is quickly washed away, the internalized fluorescence persisting within the cell is considered a good marker of endocytosis.
Upon addition of FM1–43, the dye stained the cell plasmalemma; fluorescence increased during the initial 10-min equilibration period to reach a steady-state plateau (Fig. 7, Aa and Ba). Upon challenging the cell with a solution containing 59 mM K+ and 2 mM Ca2+ (Fig. 7, Ab and Bb), fluorescence rapidly rose to reach a plateau in about a minute of K+ exposure, an indication that vesicles were undergoing exocytosis and more membrane that was being incorporated into the plasmalemma, entered in contact with FM1–43; such increase was around 30% of fluorescence after 10 min loading. After the K+ DP, cells were washed; fluorescence due to extracellular dye fades off to a new plateau above basal. The height of such plateau was due to FM1–43 trapped within the membrane vesicle retrieved, an indication of endocytosis (Fig. 7, Ac and Bc). Fluorescence tended to accumulate near the plasmalemma, a behavior indicating Ca2+-dependent vesicle cycling (40). In this example cell, endocytosis was 30% of total fluorescence. Nonspecific cell loading with FM1–43 was tested in nonstimulated cells; under these conditions, dye fluorescence indicating endocytosis did not occur (Fig. 7C, open circles).
Variations in the [Ca2+]c were tested in cells loaded with Fura-2, subjected to a similar protocol to that used for the FM1–43 experiment described above. The ratio 340/380 augmented 1.6-fold (Fig. 7D). This K+-evoked [Ca2+]c transient was evidently sufficient to trigger the exo-endocytotic responses measured with FM1–43.
The effect of VDCC blockers on the FM1–43 exo-endocytotic responses after 3-min K+ challenging were tested in another series of experiments. The Endo/Exo ratio relationship was not affected by GVIA, a similar result to that obtained with Cm change monitoring. However, MVIIC reduced by 40% the Endo/Exo ratio. In the case of nifedipine, a 53% reduction of Endo/Exo ratio was observed; this effect was lower compared with the nifedipine action on endocytosis measured with Cm change monitoring that practically abolished the endocytotic response (Fig. 3E). This difference could be explained on the basis of the two different types of stimuli used, i.e., 500 ms DPs under voltage clamp in Fig. 3 and 3 min K+ stimulation here. FPL did not change the Endo/Exo ratio when given alone; however, the compound reversed the decrease of the Endo/Exo when applied on top of MVIIC (Fig. 7D). Although less clearly than data derived from patch-clamp experiments, the FM1–43 data however support the view that Ca2+ entering through L-type VDCCs seemed to be preferentially coupled to endocytosis.
Exo-endocytotic responses triggered by photoreleased Ca2+ monitored with FM1–43 fluorescence.
Experiments so far described explored the exo-endocytotic responses triggered by Ca2+ entry through VDCCs. Under this frame, the question arose on whether Ca2+ directly released into the cytosol without the intervention of those channels could trigger responses similar to those activated by DPs. To this purpose, cells were loaded with NP.EGTA/AM and Fura-2. A 5-ms UV flash caused the rapid Ca2+ transient shown in the Fig. 8B. This transient elicited the exo- and endocytotic responses defined by the double-head arrows of Fig. 8A. Differences between variations in the [Ca2+]c are shown in Fig. 8B. In Fig. 8C, the time courses of FM1–43 fluorescence changes evoked by photoreleased Ca2+ and K+ were superimposed; this clearly shows that while the exocytotic responses had similar time courses, endocytosis was smaller in the case of cell stimulation with photoreleased Ca2+ compared with K+ (double-head arrows to the right of Fig. 8B). This is quantitatively shown in Fig. 8D where the Endo/Exo ratios are plotted; this ratio was 2.5-fold higher for K+ compared with photoreleased Ca2+.
Colocalization of calcium channels with dynamin or clathrin.
Confocal microscopy was used to identify the possible colocalization of the different calcium channel subtypes with the endocytosis-related proteins dynamin and/or clathrin. Figure 9 shows a representative experiment in which a chromaffin cell was simultaneously labeled with an anti-L-type (CaV1.3) channel antibody (A) and anti-dynamin antibody (B). Figure 9C shows a merged image that corresponds to colocalization of both pixels labeled with both antibodies in the same cell. Figure 9E shows averaged data on colocalization rate for the different combinations of dynamin and clathrin with the three VDCC subtypes studied. The colocalization of clathrin was practically negligible with the three calcium channel subtypes (CaV1.3, CaV2.1, and CaV2.2). A mild colocalization rate (about 20–30%) was observed for VDCCs and dynamin; however, significant differences between L-, N-, and PQ-type VDCCs were not found (Fig. 9E).
Upon stimulation with DPs, chromaffin cells respond with an exocytotic response that is followed by an endocytotic response. Both processes are activated by a [Ca2+]c elevation and are terminated when the cytosolic Ca2+ transient is cleared by the mitochondrion, the endoplasmic reticulum, and the plasmalemmal Ca2+ transporters (28). This exo-endocytotic coupling by Ca2+ is clearly shown in the experiments of Figs. 1 and 2 where both processes increase in parallel with augmentation of the Ca2+ gradient between the extracellular and intracellular compartments. A priori, these experiments support the view that exo-endocytosis is a simple function of the amount of Ca2+ (QCa) entering the cell, whichever the pathways used by the cation, i.e., L, N, or PQ VDCCs. However, these channel subtypes are expressed with different densities and inactivate at different rates when they are recruited by long DPs as those used here, namely 500 ms in the perforated-patch recording (Figs. 1⇑⇑⇑–5) and 3 min in the fluorescence studies using FM1.43 to monitor the exo-endocytotic responses triggered by K+ (Figs. 7 and 8). L channels are more resistant to [Ca2+]c-dependent and voltage-dependent inactivation, whereas N and PQ channels are more prone to inactivation (33, 67).
Such differences in inactivation rates may explain that blockade by nifedipine of L channels caused 31% blockade of peak ICa and 53% blockade of QCa, whereas Aga-IVA caused similar blockade of peak ICa and QCa near 70% in 2 mM [Ca2+]e (Table 1). Considering that channel densities are 20% for L channels and 50% for PQ channels in bovine chromaffin cells (2, 24), it seems that over 50% blockade of QCa by nifedipine is explained by a slower but sustained Ca2+ entry through L channels along a DP compared with a faster and greater initial Ca2+ entry followed by rapid decay when Ca2+ enters through PQ channels. These two modes of [Ca2+]c increases are compatible with the idea 1) that fast exocytosis requires a subplasmalemmal rapid increment of Ca2+ in the tenths micromolar levels (28, 49), 2) that may be better achieved when Ca2+ enters the cell through PQ channels; endocytosis, however, is a slower process that could require lower but more sustained elevations of [Ca2+]c, 3) a signal that is better provided when Ca2+ enters the cell through slow-inactivating L channels.
Our results are compatible with the above supposition since L channel blockade (that caused little inhibition of exocytosis) suppressed the endocytotic response when triggered by electrical DPs (Fig. 3D) and caused a drastic inhibition of endocytosis when FM1.43 and K+ pulses were used (Fig. 7E). Conversely, enhanced activation of L channels by FPL (that caused a mild enhancement of exocytosis) elicited a drastic augmentation of endocytosis, either under the perforated-patch technique (Figs. 5 and 6) or when using the FM1.43 dye (Fig. 7E). This differed considerably from the changes afforded by Aga-IVA; suppression of Ca2+ entry through PQ channels led to a parallel drastic inhibition of exocytotic and endocytotic responses (Fig. 3); this was also true for MVIIC, a blocker of N and PQ channels (Fig. 6). However, partial blockade of PQ channels by a low Aga-IVA concentration (to an extent similar to that produced by 3 μM nifedipine causing a full L channel blockade) caused 40% and 20% inhibition of exocytosis, respectively, for the former and the latter. The opposite was found with endocytosis that was inhibited 37% by Aga-IVA and by as much as 93% by nifedipine (Fig. 4, Table 1).
The present study reports a quite wide set of different experiments providing results that are consistent with the supposition that L VDCCs are functionally coupled more tightly to the endocytotic machinery while N or PQ channels exhibit a lesser degree of coupling. However, immunofluorescence experiments showed no colocalization of VDCC subtypes with clathrin and only about 20–30% coupling with dynamin, two proteins of the endocytotic machinery (Fig. 9). This is not against the concept that different mode of Ca2+ entry may determine the preferential role of slow-inactivating L channels in controlling endocytosis compared with more densely expressed and fast-inactivating N/PQ channels. We will try to explain these differences in the context of the functional triad, a hypothesis arising from experiments performed in bovine chromaffin cells (3, 28, 46).
Using mitochondria-targeted aequorins with different Ca2+ affinities, we found that about 50% of mitochondria located near VDCCs at subplasmalemmal sites were taking up vast amounts of the Ca2+ entering through VDCCs. Considering that the mitochondrial Ca2+ uniporter has low affinity for Ca2+, this rapid Ca2+ uptake that allowed to reach near millimolar [Ca2+] in the mitochondrial matrix could be possible only near the VDCCs where large Ca2+ changes in the tenths micromolar concentrations are reached during cell activation (49). The functional consequences of such mitochondrial Ca2+ sink were unmasked by interrupting the mitochondrial Ca2+ uptake by protonophores that obviously elicited a large potentiation of exocytosis (12, 31, 46). Having a KD for Ca2+ of 6 μM (66) the mitochondrial uniporter will see the large local Ca2+ transients generated by PQ channels, which account for about 50% of the whole population of VDCCs expressed by bovine chromaffin cells (2, 24); this will help to dissipate the Ca2+ transient and to quickly terminate the exocytotic response. In contrast, Ca2+ entering via L channels that accounts for only 15–20% will however generate lower but more sustained [Ca2+]c elevations that will be only slightly affected by subplasmalemmal mitochondria; this may serve to modulate the slower endocytotic process, as discussed above. This interpretation is supported by the observation that N/PQ channels are pharmacologically converted by roscovitine in slow-inactivating channels and thus Ca2+ entry through them triggers an endocytotic response.
In conclusion, data reported here are compatible with the supposition that the different modes of Ca2+ entry through inactivating PQ-type of VDCCs and slow-inactivating L-type of VDCCs may preferentially control exocytosis and endocytosis, respectively. This is consistent with the idea of the specialization of VDCCs subtypes in bovine chromaffin cells.
This work was partially supported from the following grants from Spanish institutions (to A. G. Garcia): SAF2006-03589 and SAF2010-21795, Ministerio de Ciencia e Innovación, Spain; NDE 07/09, Agencia Laín Entralgo, Comunidad de Madrid; PI016/09, Fundación C.I.E.N., Instituto de Salud Carlos III; RD 06/0026 RETICS, Instituto de Salud Carlos III; S-SAL-0275-2006, Comunidad de Madrid. Also by Grant SAF2007-65181 and SAL-2010-18837, Ministerio de Ciencia e Innovación, Spain (to L. Gandia).
We are not aware of financial conflict(s) with the subject matter or materials discussed in this manuscript with any of the authors, or any of the author's academic institutions or employers.
We thank Fundación Teófilo Hernando for continued support.
- Copyright © 2011 the American Physiological Society