HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Neves, G. Articles by Lagnado, L. Search for Related Content PubMed PubMed Citation Articles by Neves, G. Articles by Lagnado, L.

Journal of Physiology (2001), 535.3, pp. 809-824
© Copyright 2001 The Physiological Society

The actions of barium and strontium on exocytosis and endocytosis in the synaptic terminal of goldfish bipolar cells

Guilherme Neves, Andreas Neef and Leon Lagnado

	ABSTRACT

Top Abstract Introduction Methods Results Discussion References

1. We investigated the properties of Ca²⁺-sensitive steps in the cycling of synaptic vesicles by comparing the actions of Ca²⁺, Ba²⁺ and Sr²⁺ in the synaptic terminal of depolarizing bipolar cells isolated from the retina of goldfish. FM1-43 fluorescence and capacitance measurements demonstrated that exocytosis, endocytosis and vesicle mobilization were maintained when external Ca²⁺ was replaced by either Ba²⁺ or Sr²⁺.
2. The rapidly releasable pool of vesicles (RRP) was equivalent to 1.5 % of the membrane surface area when measured in the presence of 2.5 mM Ca²⁺, but only 0.4 % in 2.5 mM Sr²⁺. The relative sizes of the RRP in Ca²⁺, Sr²⁺ and Ba²⁺ were 1.0, 0.28 and 0.1, respectively. We conclude that a smaller proportion of docked vesicles are available for fast exocytosis triggered by the influx of Sr²⁺ or Ba²⁺ compared to Ca²⁺.
3. The slow phase of exocytosis was not altered when Ca²⁺ was replaced by Ba²⁺, but it was accelerated 1.6-fold in Sr²⁺. The peak concentrations of Ca²⁺, Sr²⁺ and Ba²⁺ (measured using Mag-fura-5) were ~4, ~14 and ~60 µM, respectively. The order of efficiency for the stimulation of slow exocytosis was Ca²⁺ Sr²⁺ > Ba²⁺.
4. Exocytosis was prolonged after the influx of Sr²⁺ and Ba²⁺. Sr²⁺ was cleared from the synaptic terminal with the same time constant as Ca²⁺ (1.3 s), but Ba²⁺ was cleared 10-100 times more slowly. Although Ba²⁺ stimulates the slow release of a large number of vesicles, it did so less efficiently than Ca²⁺ or Sr²⁺.
5. The recovery of the membrane capacitance was equally rapid in Sr²⁺ and Ca²⁺, demonstrating that the fast mode of endocytosis could be triggered by either cation.

	INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

Ca²⁺ ions regulate the exocytosis of small vesicles at the synapse (Zucker, 1996; Neher, 1998) and large secretory granules in neuroendocrine cells (Henkel & Almers, 1996). It has also been suggested that Ca²⁺ modulates endocytosis and the intermediate steps that supply vesicles to release sites on the plasma membrane. Insights into these processes and the molecules that control them can be gained by replacing Ca²⁺ with other divalent cations, such as Sr²⁺ and Ba²⁺. For instance, rapid exocytosis synchronized with the action potential is triggered less effectively by Sr²⁺ and Ba²⁺ than it is by Ca²⁺ (Miledi, 1966; Zengel & Magleby, 1977; Silinski, 1978; Augustine & Eckert, 1984), and this has been attributed to the divalent cation selectivity of the Ca²⁺ sensor for rapid exocytosis (Goda & Stevens, 1994), possibly synaptotagmin I (Geppert et al. 1994). It has been reported that rapid endocytosis in chromaffin cells is blocked when Ca²⁺ is replaced by either Sr²⁺ or Ba²⁺ (Artalejo et al. 1995). Artalejo et al. (1996) suggested that this was because the Ca²⁺ sensor for rapid endocytosis is calmodulin, and calmodulin binds other divalent cations much less effectively than Ca²⁺ (Chao et al. 1984).

Ca²⁺ is also thought to regulate the supply of new vesicles to release sites on the plasma membrane (Dittman & Regehr, 1998; Stevens & Wasseling, 1998; Wang & Kaczmarek, 1998; Gomis et al. 1999). Bipolar cells from the retina provide an interesting preparation for studying this process because, as well as transiently supporting high rates of exocytosis (Mennerick & Matthews, 1996; Sakaba et al. 1997; Neves & Lagnado, 1999), they can support a continuous slow mode of exocytosis in response to maintained stimulation (Lagnado et al. 1996; Rouze & Schwartz, 1998). Continuous exocytosis is maintained by a pool of about one million vesicles located behind the plasma membrane (Lagnado et al. 1996; von Gersdorff et al. 1996), while the rapidly releasable pool (RRP) contains ~1500 vesicles that are docked to release sites at the active zone (Mennerick & Matthews, 1996; Neves & Lagnado, 1999). Total internal reflection microscopy has demonstrated directly that slow exocytosis in bipolar cells is maintained by the transport of vesicles to release sites (Zenisek et al. 2000).

In this study, we have characterized the divalent cation selectivity of Ca²⁺-sensitive molecules that regulate vesicle cycling at the ribbon synapse of depolarizing bipolar cells from the retina of goldfish. We focused on three processes: fast exocytosis, slow exocytosis (which involves the refilling of release sites) and endocytosis. We found that Ba²⁺ and Sr²⁺ could substitute for Ca²⁺ in all of these steps in the vesicle cycle, but that the kinetics of exocytosis were dependent upon the cation triggering release. The capacity of the RRP was significantly reduced when Ca²⁺ was replaced by either Ba²⁺ or Sr²⁺. The vesicles within the RRP therefore appeared to be heterogeneous in terms of their sensitivity to the influx of divalent cations. In addition, slow exocytosis was significantly faster in Sr²⁺ than in either Ca²⁺ or Ba²⁺, although this was not because the Ca²⁺-sensitive molecule regulating slow exocytosis had a higher affinity for Sr²⁺. Measurement of the concentrations of free Ca²⁺, Ba²⁺ and Sr²⁺ during depolarization indicated that the order of efficiency for the stimulation of slow exocytosis was Ca²⁺ Sr²⁺ > Ba²⁺. Finally, capacitance measurements demonstrated that rapid endocytosis after a brief stimulus occurred similarly in Ca²⁺ and Sr²⁺. This characterization of divalent cation selectivity should aid in the identification of the molecules that regulate these steps in the vesicle cycle.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Depolarizing bipolar cells from the goldfish retina were obtained by enzymatic dissociation (Burrone & Lagnado, 1997). Goldfish were anaesthetized using MS222, then killed by decapitation and pithed (in accordance with Home Office guidelines). Trituration of retinal pieces produced isolated terminals that were detached from the axon, as well as intact bipolar cells. The standard Ringer solution contained (mM): 120 NaCl, 2.5 CaCl₂, 2.5 KCl, 1 MgCl₂, 10 glucose and 10 Hepes (pH 7.3). The effects of Ba²⁺ and Sr²⁺ were tested by equimolar substitution of BaCl₂ and SrCl₂ for CaCl₂.

Fluorescence measurements of vesicle cycling

The styryl dye FM1-43 was used to label synaptic vesicles (Betz et al. 1996). The increase in fluorescence that occurred in response to exocytosis was monitored using either a CCD camera (described by Lagnado et al. 1996; Fig. 2 and Fig. 3) or a photomultiplier tube (described by Neves & Lagnado, 1999; Figs 1, 4 and 5). The camera was used to make measurements of low temporal resolution over long time scales. In these experiments, cells were not voltage clamped through a patch pipette, but were held at a potential of 0 mV by perfusion with a modified Ringer solution containing 50 mM KCl. The photomultiplier tube was used to make measurements with a high temporal resolution over short time scales. In these experiments, intact cells were voltage clamped using a patch pipette and maintained at a holding potential of -70 mV.

The fluorescence of FM1-43 in the plasma membrane is directly sensitive to the membrane potential, so there were two causes for changes in fluorescence in response to depolarization: a decrease in fluorescence of FM1-43 in the plasma membrane, and an increase in fluorescence associated with staining of the vesicular membrane after exocytosis. The procedure used to isolate the signal due to exocytosis is similar to that described by Neves & Lagnado (1999), and further details are given below and in Results (Fig. 4). The procedure depends upon the accurate measurement of the voltage-dependent fluorescence of the dye in the plasma membrane, so it was necessary to measure this in Ca²⁺, Ba²⁺ and Sr²⁺. Figure 1 compares the voltage dependence with 5 µM FM1-43 in the external solution, together with 2.5 mM CaCl₂, BaCl₂ or SrCl₂. Figure 1A shows the percentage change in fluorescence as a function of the change in membrane potential. These measurements were made from the terminals of cells that were voltage clamped at potentials below -45 mV, thereby preventing activation of the Ca²⁺ current and the stimulation of exocytosis. The exception is the data point corresponding to V = -70 mV, which was obtained after repolarization from a 5 s step to 0 mV (when the fluorescence signal was increasing at less than 2 % s^-1). The change in the fluorescence of FM1-43 in the plasma membrane scaled linearly with voltage for all three divalent cations. The fluorescence decreased by 9.1 ± 0.6 % per 100 mV depolarization in Ca²⁺, 8.1 ± 0.8 % per 100 mV depolarization in Ba²⁺ and 8.9 ± 0.9 % per 100 mV depolarization in Sr²⁺. On average, the fluorescence of the plasma membrane decreased by 8.7 ± 0.4 % per 100 mV depolarization (line fitted to results in Fig. 1A).

		View larger version [in this window] [in a new window]
Figure 1. Voltage-dependent fluorescence of FM1-43 A, voltage-dependent changes in the fluorescence (F) of FM1-43 bound to the plasma membrane were similar using extracellular solutions containing 2.5 mM CaCl2 (; n = 17), 2.5 mM BaCl2 (; n = 11) or 2.5 mM SrCl2 (; n = 9). The change in fluorescence is expressed as a percentage of the resting fluorescence of the surface membrane. Fluorescence decreased by 9.1 ± 0.6 % per 100 mV depolarization in Ca2+, 8.1 ± 0.8 % per 100 mV depolarization in Ba2+ and 8.9 ± 0.9 % per 100 mV depolarization in Sr2+. The slope of the line fitted to the data corresponds to a decrease in fluorescence of 8.7 % per 100 mV depolarization. B, the change in FM1-43 fluorescence in response to a 5 s hyperpolarization from -50 mV to -85 mV had a similar time course in Ca2+ (21 cells), Ba2+ (17 cells) and Sr2+ (8 cells). Note that over this voltage range Ca2+ channels were closed, allowing measurement of the fluorescence of FM1-43 in the surface membrane. Traces were decimated to 50 ms per point. C, the averaged time course of the fluorescence changes shown in B. The superimposed bold line shows an empirical description in which 72 % of the change in fluorescence occurred instantaneously and the remainder with a time constant of 0.28 s. Traces were decimated to 10 ms per point.

Figure 1B compares averaged FM1-43 signals in response to a 35 mV hyperpolarization lasting 5 s. The time course of the voltage-dependent change in fluorescence of FM1-43 was similar in Ca²⁺, Ba²⁺ and Sr²⁺. The average of the responses obtained for all three divalent cations, shown in Fig. 1C, was not quite square, with 28 % of the fluorescence change occurring with a time constant of 0.28 s. There was no hysteresis in the response.

Capacitance measurements

All capacitance measurements were made in synaptic terminals that had detached from the axon, which were voltage clamped using the perforated patch technique. The solution in the patch pipette contained (mM): 110 caesium gluconate, 4 MgCl₂, 3 Na₂ATP, 1 Na₂GTP, 10 TEACl, 0.4 BAPTA and 20 Hepes, with 250 µg ml^-1 nystatin (260 mosmol l^-1, pH 7.2). Electrodes had resistances in the range 2-4 M. Capacitance measurements were only started once the series resistance was below 30 M, and the average series resistance was 20-25 M. The input resistance at -70 mV was at least 1 G. Capacitance measurements were made using the piece-wise linear technique (Neher & Marty, 1982), our implementation of which is described in another publication (Neves & Lagnado, 1999). Briefly, a sinusoidal command voltage (50 mV peak to peak, 2 kHz) was superimposed onto a holding potential of -70 mV. The output of the patch-clamp amplifier was analysed using a dual-phase lock-in amplifier, the phase of which was set using the 'capacitance dither' method, thereby allowing independent measurement of changes in membrane capacitance and conductance (Gillis, 1995). The capacitance signal was calibrated by dithering the capacitance by 100 fF at the beginning of each recording episode. The change in capacitance elicited by the stimulus (C_m) was measured by subtracting the capacitance signal obtained before the stimulus from that measured after. The 'before' measurement was averaged over a period of 100-200 ms, ending 20 ms before depolarization. The 'after' measurement was averaged over a period of 100-200 ms, beginning 20 ms after repolarization. Where Ca²⁺ currents are shown, they have been leak subtracted. Signals were typically digitized at 10 kHz. Data analysis was performed using Igor Pro software (Wavemetrics, Lake Oswego, OR, USA). Traces were smoothed by using the decimate function in Igor Pro to average contiguous sample points, thereby reducing the final sample interval.

Conversion of capacitance and FM1-43 measurements into vesicle numbers

Capacitance changes were converted to vesicle numbers by assuming that the membrane capacitance of a single vesicle was 26 aF. The uncertainties in this estimate are described by Neves & Lagnado (1999).

Increases in FM1-43 fluorescence are expressed as a percentage of the resting fluorescence measured when just the surface membrane of the terminal was stained (Lagnado et al. 1996). One of two methods was then used to convert the percentage increase in fluorescence into approximate vesicle numbers. For experiments using a photomultiplier tube (Fig. 4 and Fig. 5, Table 1), a 1 % increase in surface area was taken to represent the release of ~1200 vesicles (Neves & Lagnado, 1999). This conversion factor is based on the average capacitance of a terminal (3.1 pF), and the estimated capacitance of a single vesicle (26 aF). For experiments using a camera (Fig. 2 and Fig. 3), we calculated a conversion factor for each cell from differential interference contrast (DIC) images (Lagnado et al. 1996). We measured the diameter of the terminal (d_t), and assumed that it was shaped as a solid hemisphere (which has a surface area of 3/4d_t²). Three-dimensional reconstruction with a confocal microscope indicated that this is a good approximation to the shape of the terminal. As the membrane area of a spherical vesicle of diameter d_v is d_v², the surface area of a terminal is equivalent to about (3/4)(d_t/d_v)² vesicles.

Assaying changes in the concentration of free Ca²⁺, Ba²⁺ and Sr²⁺

Cells were loaded with the fluorescent indicator Mag-fura-5 (Molecular Probes, Portland, OR, USA) by incubation in 1 µM of the AM ester for 15 min at room temperature. This procedure led to the loading of 50-70 µM of the dye, as estimated by measuring the intensity of the fluorescence in the terminal when exciting at the isobestic wavelength for Ca²⁺ (352 nm) and comparing it with the average fluorescence measured when a sample of five terminals was loaded with 100 µM Mag-fura-5 through a patch pipette. Ratiometric measurements of changes in divalent cation concentration were made by exciting Mag-fura-5 at wavelengths of 340 and 380 nm (bandwidth 10 nm) using a monochromator-based illumination system (Cairn, Kent, UK). Fluorescence from the synaptic terminal was collected through a 475DCLP dichroic mirror and 535RDF45 emission filter using a photomultiplier tube. The concentration of free Ca²⁺ was calculated according to [Ca²⁺] = K_dQ(R - R_min)/ (R_max - R), where R is the ratio of the fluorescence emission at the two wavelengths (F₃₄₀/F₃₈₀) and R_min and R_max are the values of this ratio when the dye is completely free or fully bound to Ca²⁺. K_d is the affinity of the dye for Ca²⁺ and Q is the value of F₃₈₀ when the dye is completely free divided by its value when fully bound to Ca²⁺. All of these parameters were measured in vitro using a closed chamber on the same experimental set-up used to measure levels of divalent cations in cells. These values were also measured in the presence of Ba²⁺ and Sr²⁺, and the same expression was also used to calculate the free concentrations of these ions.

The affinity of Mag-fura-5 for Ca²⁺, Ba²⁺ and Sr²⁺ was measured using solutions containing 110 mM KCl, 1 mM MgCl₂, 10 mM EGTA and 10 mM Hepes, pH 7.2. For each divalent cation, several solutions were made containing various concentrations of the free ion. These three series of solutions were prepared using the 'pHmetric' method described by Tsien & Pozzan (1989) for making solutions of a known concentration of free Ca²⁺ by reliably controlling the ratio of Ca²⁺ and EGTA. Briefly, the free EGTA concentration in these solutions was fixed and variable amounts of Ca²⁺ added as a Ca²⁺:EGTA complex at a ratio of 1:1 to displace the desired concentration of free Ca²⁺. The Ca²⁺:EGTA mixture was prepared by adding small aliquots of 1 M CaCl₂ to the EGTA stock solution and noting the resulting change in pH until this change was reduced to less than 50 % of its original value. Solutions containing 1:1 ratios of Sr²⁺:EGTA and Ba²⁺:EGTA were made in the same way. Calculation of the final free divalent cation concentration for a given divalent cation:EGTA ratio required accurate estimates of the K_d of EGTA for the divalent cation and Mg²⁺ under our conditions of pH (7.2), ionic strength and temperature (20 °C). The values for the K_d of EGTA for Ca²⁺, Ba²⁺, Sr²⁺ and Mg²⁺ were calculated from measurements reported in Martell & Smith (1974) as 150.5 nM, 51.5 µM, 43.7 µM and 19.7 mM, respectively. Although EGTA was used to make buffered solutions containing free Ba²⁺ or Sr²⁺ concentrations of up to 2 mM, it was not appropriate for making solutions containing more than 10 µM Ca²⁺ (because EGTA has a much higher affinity for Ca²⁺). Instead, solutions containing 25-500 µM Ca²⁺ were made by simple dilution of a 1 M CaCl₂ stock solution (when EGTA was omitted from the Hepes-buffered solutions).

Absorption spectra and values of R were measured with 1 µM Mag-fura-5 diluted into each divalent cation solution. The values of R as a function of the free divalent cation concentration were fitted to the equation [divalent cation] = K_dQ(R - R_min)/(R_max - R). The affinity of Mag-fura-5 for Ca²⁺, Ba²⁺ and Sr²⁺ was 20, 70 and 90 µM, respectively. Similar affinities of Mag-fura-5 for Ca²⁺ and Sr²⁺ have been measured by Xu-Friedman & Regehr (1999).

All measurements are given as means ± S.E.M. Estimates of vesicle numbers were obtained by linearly scaling the percentage increase in FM1-43 fluorescence, so only the mean is given.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Ba²⁺ and Sr²⁺ supported the release, retrieval and recycling of synaptic vesicles

Experiments using the styryl dye FM1-43 to stain membranes have demonstrated that maintained depolarization can stimulate continuous exocytosis and endocytosis in the synaptic terminal of depolarizing bipolar cells (Lagnado et al. 1996; Rouze & Schwartz, 1998; Neves & Lagnado, 1999). The continuous cycling of vesicles also occurred in response to the influx of Ba²⁺ (Fig. 2) and Sr²⁺ (Fig. 3). Figure 2A shows a DIC image of a depolarizing bipolar cell, followed by a series of fluorescence images that were taken during the course of the experiment. The cell was continuously depolarized in the presence of 50 mM KCl, but initially there were no divalent cations present. Figure 2B plots the fluorescence signal from the terminal and cell body. The addition of FM1-43 caused staining of the plasma membrane (Fig. 2A, image a), but there was no further increase in fluorescence over the next 2 min because FM1-43 cannot cross membranes directly (Fig. 2A, image b). When 2.5 mM BaCl₂ was added to the medium, there was an immediate and continuous increase in fluorescence that was localized to the synaptic terminal, indicating that more membrane had come into contact with FM1-43 and that exocytosis had been stimulated (Fig. 2A, image c). The increase in fluorescence observed in the presence of FM1-43 provides a cumulative measure of the number of vesicles that have undergone exocytosis (Smith & Betz, 1996; Neves & Lagnado, 1999). Figure 2B shows that the fluorescence increase was blocked when Ba²⁺ was removed.

		View larger version [in this window] [in a new window]
Figure 2. Ba2+ influx stimulated a continuous cycle of exocytosis and endocytosis A, the top panel shows a differential interference contrast (DIC) image of the cell, and panels a-d show images of FM1-43 fluorescence. The cell was continuously depolarized in 50 mM KCl. Images a and b were obtained after adding 5 µM FM1-43 to a medium containing 1 mM EGTA and 1 mM MgCl2, but no other divalent cations. Image c was obtained 1.5 min after adding 2.5 mM BaCl2, and shows the accumulation of fluorescence in the synaptic terminal, but not in the cell body. Image d was obtained after removal of FM1-43, showing the selective retention of the dye in the synaptic terminal. Scale bar = 10 µm. B, quantification of the experiment shown in A, showing the mean fluorescence of the synaptic terminal (bold trace) and a region of the cell body (thin trace). Application of FM1-43 is indicated by the filled bar. The traces were background subtracted and are expressed as a percentage of the resting fluorescence of the terminal when only the surface membrane was stained. The letters a-d mark the timing of the corresponding images in A. Images were obtained every 10 s. The rate of exocytosis measured during exposure to 2.5 mM BaCl2 (open bar) was equivalent to 3 % of the membrane area per second, which corresponds to about 3900 vesicles s-1 for a terminal with a diameter of 13.5 µm. After removal of FM1-43 from the medium the plasma membrane rapidly destained, but the fluorescence inside the synaptic terminal was retained until exocytosis was stimulated by a second application of Ba2+.

		View larger version [in this window] [in a new window]
Figure 3. Continuous exocytosis was faster in Sr2+ than in Ca2+ A, the mean fluorescence of the synaptic terminal (bold trace) and a region of the cell body (thin trace) during an experiment similar to that shown in Fig. 2. Images were obtained at 10 s intervals. The initial rate of exocytosis on adding 2.5 mM SrCl2 (hatched bar) was equivalent to about 5300 vesicles s-1 (dashed line). After removal of FM1-43 from the medium, the fluorescence trapped inside the terminal was retained until exocytosis was stimulated by the application of 2.5 mM CaCl2 (open bar). Application of 5 µM FM1-43 is indicated by the filled bar. The diameter of the terminal was 12.2 µm. B, the rate of continuous exocytosis was faster in Sr2+ than in Ca2+. The rate of exocytosis in SrCl2 was equivalent to about 4000 vesicles s-1 (dashed line), decreasing to 1900 vesicles s-1 in CaCl2. Images were obtained at 3 s intervals. The diameter of the terminal was 11.5 µm.

The rate of vesicle cycling was estimated by normalization of the fluorescence increase to the resting fluorescence of the plasma membrane, measured before addition of the divalent cation (Lagnado et al. 1996; Neves & Lagnado, 1999). In this example, the rate of continuous exocytosis was equivalent to 3 % of the resting membrane area per second, which is equivalent to about 3900 vesicles s^-1 for a terminal of this size (13.5 µm in diameter), assuming that the diameter of a vesicle is 32.5 nm (see Methods).

Endocytosis also occurred in response to Ba²⁺ influx, as could be appreciated from the large increase in fluorescence that was observed within the terminal (Fig. 2A, image c). These labelled membranes were trapped within the terminal when FM1-43 and Ba²⁺ were removed from the external medium (Fig. 2A, image d). The fluorescence tended to accumulate close to the plasma membrane (Fig. 2A, images c and d), as has been observed when vesicle cycling is stimulated by Ca²⁺ (Lagnado et al. 1996; Job & Lagnado, 1998). How much endocytosis occurred? Figure 2B shows that when FM1-43 was removed, the rapid decrease in fluorescence caused by destaining of the plasma membrane was equivalent to the increase in fluorescence observed when the surface of the terminal was stained at the beginning of the experiment, indicating that the surface area of the terminal had remained roughly constant. So all of the increase in fluorescence that occurred during the 2 min period of stimulation was internalized, indicating that endocytosis balanced exocytosis on these time scales.

After retrieval in the presence of Ba²⁺, vesicles became available for release again. When FM1-43 was removed from the medium, reapplication of 2.5 mM Ba²⁺ caused a decrease in fluorescence (Fig. 2B), indicating that dye-loaded vesicles underwent exocytosis. Ba²⁺ therefore also supported the intermediate steps that allow a retrieved vesicle to dock to the plasma membrane and become competent for exocytosis. Similar behaviour was observed in all eight cells tested in this way. Sr²⁺ also supported continuous vesicle cycling in a manner qualitatively similar to Ba²⁺ and Ca²⁺ (four cells), and an example is shown in Fig. 3A. These results demonstrate that none of the essential steps in the release, retrieval or recycling of synaptic vesicles had an absolute requirement for Ca²⁺ influx.

An important feature of the results shown in Fig. 2B and Fig. 3A is that the rate of increase in fluorescence did not begin to fall off (reflecting exocytosis of vesicles that had already been loaded with the dye) until at least twice the resting surface area of the terminal was recycled. The total number of vesicles available for continuous exocytosis in Ba²⁺ and Sr²⁺ was therefore well in excess of 300 000. More detailed measurements made by Lagnado et al. (1996) indicate that about 750 000 vesicles are involved in maintaining the continuous vesicle cycle when it is stimulated by Ca²⁺ (Lagnado et al. 1996), and electron microscopy has shown that there are of the order of one million vesicles in a bipolar cell terminal (von Gersdorff et al. 1996). The results in Fig. 2 and Fig. 3 therefore indicate that most of the vesicles in this synaptic terminal are available for exocytosis, irrespective of whether this process is stimulated by Ca²⁺, Ba²⁺ or Sr²⁺.

Continuous exocytosis was faster in Sr²⁺

Continuous exocytosis stimulated by Sr²⁺ influx was faster than that stimulated by either Ca²⁺ or Ba²⁺. The average rate of continuous exocytosis during the first minute of stimulation in 2.5 mM Ca²⁺ was equivalent to 1.4 ± 0.2 % of the membrane surface area per second (1600 vesicles s^-1; n = 11 cells), while in 2.5 mM Ba²⁺ it was 2.2 ± 0.3 % s^-1 (2600 vesicles s^-1; n = 13 cells) and in 2.5 mM Sr²⁺ it was 3.6 ± 0.8 % s^-1 (4200 vesicles s^-1; n = 8 cells).

The higher rate of continuous exocytosis observed in the presence of Sr²⁺ was confirmed by the experiment shown in Fig. 3B, which was designed to compare the efficacy of Sr²⁺ and Ca²⁺ in the same cell. The addition of 2.5 mM SrCl₂ for 30 s evoked exocytosis at a constant rate of 4.3 % s^-1. Switching to a solution containing 2.5 mM CaCl₂ immediately reduced the rate of exocytosis to 2 % s^-1. A comparison with Fig. 3A indicates that this sudden deceleration in the FM1-43 signal occurred long before the gradual deceleration caused by the accumulation of dye-labelled vesicles (note the different time scales). The rate of exocytosis in Sr²⁺ was always higher than in Ca²⁺, and this was also evident as an acceleration of exocytosis when switching from CaCl₂ to SrCl₂. Experiments over these time scales indicate that continuous exocytosis in Sr²⁺ was faster than in Ca²⁺, by a factor of 1.5 ± 0.2 (five cells). The rate of continuous exocytosis in Ba²⁺ was not significantly different to that measured in Ca²⁺.

Measuring exocytosis with higher temporal resolution

To measure fast modes of exocytosis, we used voltage-clamped cells and recorded FM1-43 fluorescence using a photomultiplier tube (Neves & Lagnado, 1999). It was then necessary to correct records using the procedure shown in Fig. 4. The upper part of Fig. 4A shows the fluorescence change elicited by a 5 s depolarization from -70 mV to 0 mV, delivered in 2.5 mM Ca²⁺. The trace shown is the response averaged from 19 cells, and the signal reflects both the stimulation of exocytosis and the direct voltage-dependent fluorescence of the dye in the plasma membrane (see Methods). The latter component is shown by the smooth trace, which is the empirical description of the averaged response to a 35 mV hyperpolarization (bold line in Fig. 1C), scaled to match the expected amplitude of the response to a 70 mV depolarization (Fig. 1A). The lower part of Fig. 4A shows the time course of exocytosis obtained after subtraction of the signal originating from the plasma membrane. Figure 4B and C shows the application of this subtraction procedure to responses in Sr²⁺ (9 cells) and Ba²⁺ (11 cells). The kinetics of exocytosis in Ca²⁺ , Ba²⁺ and Sr²⁺ are described below.

		View larger version [in this window] [in a new window]
Figure 4. Correction of records of FM1-43 fluorescence to measure exocytosis on short time scales A, top, the fluorescence change evoked by a 5 s depolarization from -70 mV to 0 mV in 2.5 mM CaCl2. The trace is an average of 19 stimuli, each in a different cell. The change in fluorescence had two components: the direct voltage-dependent change in the fluorescence of FM1-43 in the plasma membrane, and the increase in the amount of membrane stained by FM1-43 due to exocytosis. The smooth trace shows the estimate of the latter component, obtained from measurements of the purely voltage-dependent change in fluorescence (Fig. 1 and Methods). Note that in each case the expected amplitude of the response to the 70 mV depolarization was obtained from measurements in solutions of the same divalent cation. The bottom panel shows the result of subtracting this purely voltage-dependent component. The resulting signal is thought to measure the time course of exocytosis. Traces were decimated to 20 ms per point. B and C show the same processing applied to the averaged fluorescence signals obtained in Sr2+ (9 cells) and Ba2+ (11 cells).

The capacity of the rapidly releasable pool of vesicles was reduced in Sr²⁺ and Ba²⁺

When the membrane potential was stepped from -70 mV to 0 mV in the presence of 2.5 mM Ca²⁺, there were two rapid but transient components of release followed by a third, continuous component (bold lines in Fig. 5B and E). The first phase of exocytosis was complete within about 20 ms, reflecting exocytosis of the RRP (Fig. 5B and C). The capacity of the RRP, measured by extrapolating the intermediate phase back to the beginning of the stimulus, was equivalent to 1.5 ± 0.2 % of the resting membrane area, or ~1800 vesicles. Similar estimates of the size of the RRP have been made using the capacitance technique (Mennerick & Matthews, 1996; Sakaba et al. 1997; Gomis et al. 1999; Neves & Lagnado, 1999), and roughly the same number of vesicles are morphologically docked to the plasma membrane (von Gersdorff et al. 1996).

Three phases of exocytosis were also observed in response to Sr²⁺ influx, but the most rapid component was significantly smaller than that observed in Ca²⁺ (Fig. 5B and C). By extrapolating the first linear phase of the fluorescence increase back to time zero, the amplitude of the most rapid component of exocytosis was found to be equivalent to 0.4 ± 0.2 % of the resting membrane area (Fig. 5B and C), or ~500 vesicles. In Ba²⁺ only two phases of exocytosis were apparent: extrapolating the first linear component of the fluorescence increase back to time zero did not reveal a faster phase of release (Fig. 5D-F). The different components of exocytosis observed in response to the influx of Ca²⁺, Ba²⁺ and Sr²⁺ are compared in Table 1.

		View larger version [in this window] [in a new window]
Figure 5. Rapid exocytosis stimulated by Ca2+, Sr2+ and Ba2+ A, a comparison of the increases in FM1-43 fluorescence generated by the stimulation of exocytosis in 2.5 mM Ca2+ (black trace) and 2.5 mM Sr2+ (grey trace). The traces are from Fig. 4, after subtraction of the voltage-dependent change in FM1-43 fluorescence in the plasma membrane. The timing of the stimulus (a 5 s depolarization to 0 mV) is shown below the averaged currents. The initial amplitude of the current evoked by depolarization was similar in the two solutions, but inactivation was slower in SrCl2. B, the average time course of exocytosis evoked by Sr2+ and Ca2+ influx on an expanded time scale. The superimposed bold lines describing the time course of exocytosis in both Ca2+ and Sr2+ had three components, as described in the text. C, the initial 300 ms of stimulation is shown on an expanded time scale. Sr2+ was much less effective than Ca2+ at triggering exocytosis from the RRP. Error bars show S.E.M. D, a comparison of the increases in FM1-43 fluorescence generated by the stimulation of exocytosis in 2.5 mM Ca2+ (black trace) and 2.5 mM Ba2+ (grey trace). Exocytosis continues for tens of seconds in Ba2+ after closure of the Ca2+ channels. The initial amplitude of the current evoked by depolarization was similar in the two solutions, but inactivation was slower in BaCl2. E, the average time course of exocytosis evoked by Ba2+ and Ca2+ influx on an expanded time scale. The superimposed bold line describing the response in Ba2+ had only two distinct components. F, the initial 300 ms of stimulation is shown on an expanded time scale. When Ba2+ acted as a charge carrier, no exocytosis could be detected above the noise level in the first 100 ms, nor could any be inferred by extrapolation of the first linear component of the fluorescence increase. Traces in A, B, D and F were decimated to 20 ms per point. Traces in C and F were decimated to 10 ms per point.

The resolution of the FM1-43 technique was at the limit where exocytosis of the RRP could be detected. We therefore used the capacitance technique to test further how the amount of rapid exocytosis depended upon the divalent cation triggering release. Figure 6A shows capacitance responses to 10 ms depolarizations to 0 mV - a stimulus that is sufficient to release the whole of the RRP when delivered in the presence of 2.5 mM Ca²⁺ (Mennerick & Matthews, 1996; Sakaba et al. 1997; Gomis et al. 1999; Neves & Lagnado, 1999). The capacitance response was first measured in Ca²⁺, and then the terminal was held at -70 mV for at least 1 min to allow the RRP to refill (Gomis et al. 1999). A second stimulus was delivered within seconds of replacing Ca²⁺ with Sr²⁺ (bold trace in Fig. 6A). On average, the capacitance response to a 10 ms depolarization in Sr²⁺ was 28 ± 6 % of that in Ca²⁺ (n = 15; Fig. 6A and C), which was similar to the decrease in the amount of rapid exocytosis measured using FM1-43 (Fig. 5C). The capacitance response to a 10 ms depolarization in Ba²⁺ was 11 ± 5 % of that in Ca²⁺ (n = 9; Fig. 6B and C), indicating that Ba²⁺ could support a small amount of rapid exocytosis that could not be detected using FM1-43. When Sr²⁺ or Ba²⁺ was replaced by Ca²⁺, the size of the RRP recovered (Fig. 6A and B).

		View larger version [in this window] [in a new window]
Figure 6. Capacitance measurements of the most rapid phase of exocytosis stimulated by Ca2+, Sr2+ and Ba2+ A, capacitance responses (Cm) to a 10 ms depolarization from -70 mV to 0 mV. The bold trace shows the response in 2.5 mM Sr2+, and the thin traces show two responses in 2.5 mM Ca2+ (one before and one after perfusion with SrCl2). The inset shows the currents activated by the stimulus. B, a similar experiment to that shown in A, comparing responses in Ca2+ and Ba2+. C, averaged results from experiments shown in A and B. The capacitance increase in Ba2+ was 11 ± 5 % of the increase observed in the same cell in Ca2+ (; n = 9), and in Sr2+ it was 28 ± 6 % (; n = 15). The integral of the depolarization-evoked current in Ba2+ was 94 ± 6 % of that in Ca2+. In Sr2+ it was 87 ± 2 % of that in Ca2+.

If all the vesicles comprising the RRP had the same sensitivity to divalent cations, replacing Ca²⁺ with either Ba²⁺ or Sr²⁺ would be expected to release the same total number of vesicles, albeit at different rates. In fact, the results in Fig. 5 and Fig. 6 show that Ba²⁺ and Sr²⁺ caused a large reduction in the total number of vesicles that could undergo rapid exocytosis. Why are Ba²⁺ and Sr²⁺ only capable of releasing a small fraction of the RRP? The reduced amount of rapid exocytosis could not be attributed to reduced influx of these cations, because the currents they carried were similar, both in the FM1-43 measurements (Fig. 5A and D) and in the capacitance measurements (insets to Fig. 6A and B, and 6C). We also think it is unlikely that the reduced size of the RRP was attributable to the accumulation of Ba²⁺ or Sr²⁺ causing an increase in the rate of spontaneous vesicle fusion (Zengel & Magleby, 1977; Goda & Stevens, 1994; Xu-Friedmann & Regehr, 1999). All of the measurements shown in Fig. 5 and Fig. 6 were made by applying Ba²⁺ or Sr²⁺ at a membrane potential where Ca²⁺ channels were closed (-70 mV), and then delivering the depolarizing stimulus within seconds. Furthermore, the capacitance responses recovered fully after re-applying external Ca²⁺ (Fig. 6). These results therefore indicate that Ba²⁺ and Sr²⁺ could only trigger the rapid exocytosis of a small fraction of the docked vesicles comprising the RRP.

The continuous mode of exocytosis was faster in Sr²⁺

The slower continuous phase of exocytosis began about 1 s into a depolarizing step (Fig. 5B and E). The initial rate was 1.0 ± 0.1 % s^-1 in Ca²⁺ and 0.9 ± 0.2 % s^-1 in Ba²⁺ (Fig. 5E), but it was significantly faster in Sr²⁺, occurring at a rate of 1.6 ± 0.2 % s^-1 (Fig. 5B). These measurements confirm that the rate of continuous exocytosis was dependent on the cation triggering release. It should be noted that the initial rate of continuous exocytosis measured in Fig. 5B and E was slower than the rate averaged over 1-2 min (Fig. 2 and Fig. 3). Recent experiments indicate that this is because continuous exocytosis accelerates after a delay of 10-20 s (M. Holt & L. Lagnado, unpublished observations).

What process determines the rate of exocytosis during maintained stimulation? Figure 5B shows that the acceleration of continuous exocytosis by Sr²⁺ was apparent immediately after the two more rapidly released pools of vesicles were exhausted. These two pools were equivalent to ~5 % of the surface area of the terminal, or ~6000 vesicles. The continuous phase of exocytosis therefore began after the release of only ~1 % of the available vesicles. These measurements indicate that the rate of continuous exocytosis was not limited by depletion of vesicles, nor by the rate of replenishment by endocytosis. The majority of vesicles in the bipolar cell terminal are located some distance from the plasma membrane or synaptic ribbons (Lagnado et al. 1996; von Gersdorff et al. 1996), so it seems likely that the rate-limiting process during the slow phase of exocytosis was either the transfer of these vesicles to release sites on the plasma membrane or their conversion into a release-ready state once they were at those sites. Total internal reflection microscopy has directly demonstrated that slow exocytosis in bipolar cells involves the transport of vesicles to the plasma membrane (Zenisek et al. 2000). Furthermore, vesicle mobilization is sensitive to Ca²⁺, because refilling of the release sites can be inhibited by the introduction of the Ca²⁺ buffer EGTA (Gomis et al. 1999).

Asynchronous release was prolonged by Ba²⁺and Sr²⁺

Measurements using FM1-43 and the capacitance technique both indicate that the introduction of a large Ca²⁺ load into the bipolar cell terminal can lead to the stimulation of exocytosis for some time after the closure of Ca²⁺ channels (Neves & Lagnado, 1999). This form of 'asynchronous release' was potentiated by both Sr²⁺ and Ba²⁺. Figure 5A shows that after a 5 s depolarization in Sr²⁺, the rate of exocytosis was maintained for ~3 s before gradually stopping. The effect of Ba²⁺ was even more pronounced: exocytosis continued at a fixed rate for ~15 s after the introduction of a Ba²⁺ load (Fig. 5D). The rate of continuous exocytosis was not immediately affected by the closure of Ca²⁺ channels, from which we draw two conclusions. First, asynchronous release was simply a continuation of the slow mode of exocytosis. Second, slow exocytosis was driven by the global rise in divalent cation concentration rather than the concentration close to an open Ca²⁺ channel. The dramatic potentiation of asynchronous release in Sr²⁺ and Ba²⁺ may arise, at least in part, because inactivation of Ca²⁺ channels was reduced, leading to the introduction of larger Ba²⁺ or Sr²⁺ loads (Fig. 5A and D). Two other factors that will determine the time course of asynchronous release are the relative efficiencies with which these ions trigger slow exocytosis and the rates at which they are removed from the cytoplasm of the synaptic terminal.

A comparison of Ca²⁺, Sr²⁺ and Ba²⁺ signals in the synaptic terminal

To understand how Sr²⁺ and Ba²⁺ are handled in the synaptic terminal of bipolar cells, we measured their free concentrations using the fluorescent indicator Mag-fura-5. Figure 7A compares increases in Ca²⁺ and Ba²⁺ concentrations occurring during a 5 s depolarization to -10 mV, measured in the same terminal. This stimulus was of the same duration as that used in Fig. 5 to measure the time course of exocytosis using FM1-43. As shown in Fig. 7A, the concentration of free Ca²⁺ peaked at about 6 µM after ~3 s. In contrast, the concentration of free Ba²⁺ continued to rise throughout the depolarizing step, peaking at ~80 µM. In Fig. 7B these traces have been normalized to the divalent cation concentration at the end of the stimulus to allow direct comparison of the rate at which these ions were removed from the cytoplasm. Ca²⁺ was cleared with a time constant () of 1.6 s, while Ba²⁺ was cleared with = 12 s. On average, the concentration of free Ca²⁺ during a 5 s depolarization peaked at 3.6 ± 0.7 µM, and Ca²⁺ was cleared with = 1.4 ± 0.2 s (n = 13). The average peak amplitude of the Ca²⁺ current in this sample of measurements was 100 ± 9 pA. The concentration of free Ba²⁺ observed during a 5 s depolarization peaked at 57 ± 13 µM, and Ba²⁺ was cleared with = 46 ± 28 s (range 12-143 s; n = 5). The average peak amplitude of the Ba²⁺ current in this sample was 61 ± 5 pA. These results demonstrate that the prolonged secretory response triggered by the influx of Ba²⁺ occurred because it was cleared from the cytoplasm slowly.

		View larger version [in this window] [in a new window]
Figure 7. Concentrations of free Ca2+, Ba2+ and Sr2+ during and after depolarization A, changes in [Ca2+] (black trace) and [Ba2+] (grey trace) evoked by a 5 s depolarization to -10 mV. Measurements were made from the same terminal in the presence of 2.5 mM Ca2+ or 2.5 mM Ba2+. The lower panel shows the currents measured during the depolarization and also marks the timing of depolarization. The concentration of Ba2+ rises much higher than that of Ca2+. B, traces shown in A were scaled to the level at the end of the depolarization to allow comparison of the rates at which Ca2+ and Ba2+ were removed from the cytoplasm. Ba2+ was cleared with a time constant of 12 s, while Ca2+ was cleared with a time constant of 1.3 s, as shown by the fitted traces. C, changes in [Ca2+] (black trace) and [Sr2+] (grey trace) during a 5 s depolarization to -10 mV, measured in the same terminal. The concentration of Sr2+ plateaued at 12 µM during the depolarization, while that of Ca2+ only reached 3 µM. D, traces from C scaled in the same manner as in B. Ca2+ was cleared with a time constant of 1.3 s and Sr2+ with a time constant of 1.5 s.

Asynchronous release after the influx of Sr²⁺ was not as prolonged as that caused by the influx of Ba²⁺ (Fig. 5) because Sr²⁺ ions were removed from the cytoplasm much more rapidly. Figure 7C compares increases in the concentration of free Ca²⁺ and free Sr²⁺ measured in the same terminal. In this example the concentration of free Sr²⁺ reached a plateau of ~12 µM. Sr²⁺ and Ca²⁺ were cleared from the terminal at similar rates (Fig. 7D). On average, the concentration of free Sr²⁺ during a 5 s depolarization peaked at 14 ± 2 µM, and Sr²⁺ was cleared with = 1.3 ± 0.3 s (n = 4). The average peak amplitude of the Sr²⁺ current in this sample of measurements was 95 ± 8 pA.

It has been suggested that Sr²⁺ and Ba²⁺ potentiate asynchronous release because the Ca²⁺-sensitive molecule controlling this mode of exocytosis has a higher affinity for these ions (Zengel & Magleby, 1977; Goda & Stevens, 1994). Our measurements demonstrate that Ba²⁺ stimulates the slow mode of exocytosis considerably less effectively than does Ca²⁺. Thus, the continuous phase of exocytosis occurred at approximately the same rate during the influx of Ca²⁺ and Ba²⁺ (Fig. 5), even though the concentration of free Ba²⁺ rose to levels that were 10-20 times higher than the concentration of free Ca²⁺ (Fig. 7). The failure of these high levels of Ba²⁺ to accelerate the slow mode of exocytosis could not be explained by saturation of this process, because continuous exocytosis was accelerated 1.6-fold during depolarization in Sr²⁺.

The 1.6-fold acceleration of continuous exocytosis when depolarizing in Sr²⁺ must have been due, at least in part, to the concentration of free Sr²⁺ rising to levels 3-4 times higher than the concentration of free Ca²⁺ (Fig. 7). Nonetheless, the efficiency with which Sr²⁺ stimulated the slow mode of exocytosis was roughly comparable to that of Ca²⁺. The order of efficiency for the stimulation of slow exocytosis was therefore Ca²⁺ Sr²⁺ > Ba²⁺. Sr²⁺ was cleared at the same rate as Ca²⁺, indicating that the prolongation of asynchronous release was due primarily to the large rise in [Sr²⁺].

Rapid endocytosis occurred in Sr²⁺

FM1-43 measurements made over long time scales indicated that endocytosis occurred in the presence of all three divalent cations at rates sufficient to balance continuous exocytosis (Fig. 2 and Fig. 3). However, endocytosis at the bipolar cell synapse can occur by both fast and slow mechanisms, the rates of which differ by a factor of about 10 (Neves & Lagnado, 1999). The faster route of membrane retrieval occurs with a time constant of about 1 s but only predominates after brief stimuli (von Gersdorff & Matthews, 1994; Neves & Lagnado, 1999). We used the capacitance technique to test how fast endocytosis might be affected when Ca²⁺ was replaced by Sr²⁺ or Ba²⁺. The capacitance responses to 10 ms stimuli were too small to do this reliably (Fig. 6), so we used stimuli lasting 500 ms. Figure 8A shows two capacitance responses to depolarizing steps from -70 mV to 0 mV, one delivered in Ca²⁺ (thin trace) and the second in Sr²⁺ (bold trace). The time course over which the membrane was retrieved was compared by scaling the responses to the same maximum (Fig. 8B). The kinetics of endocytosis were identical in Ca²⁺ and Sr²⁺. In this example, the surface area recovered to pre-stimulus levels in about 3 s, indicating that all of the membrane was retrieved by the faster mechanism.

		View larger version [in this window] [in a new window]
Figure 8. Recovery of membrane capacitance was similar after stimulation in Ca2+ and Sr2+ A, capacitance responses to a 500 ms depolarization from -70 mV to 0 mV, delivered in 2.5 mM Ca2+ (thin trace) and 2.5 mM Sr2+ (bold trace). The inset shows the currents activated by the stimulus (scale bars: 50 pA and 250 ms). B, the responses in A scaled to the same maximum to allow comparison of the kinetics of membrane retrieval. The membrane capacitance fell with a single time constant of 1 s. C, a second example of the experiment shown in A. In this terminal the recovery in membrane capacitance occurred with fast and slow phases. Note the different time scale. The inset shows the currents activated by the stimulus (scale bars: 150 pA and 250 ms). D, the responses in C scaled to the same maximum.

A second example of capacitance responses in Ca²⁺ and Sr²⁺ is shown in Fig. 8C. In this terminal, there were both fast and slow phases of retrieval after a 500 ms stimulus (note the different time scale). The difference between the capacitance responses in Fig. 8A and C is probably related to the difference in the Ca²⁺ currents: the introduction of larger Ca²⁺ loads tends to increase the proportion of membrane retrieved by the slower route (Neves & Lagnado, 1999). When responses were scaled to the same maximum (Fig. 8D), it could again be seen that the kinetics of membrane retrieval were very similar in Ca²⁺ and Sr²⁺. The same proportion of membrane was retrieved by the fast and slow mechanisms. Similar observations were made in all six synaptic terminals that were tested in this way. We conclude that neither the fast nor slow mechanisms of membrane retrieval at this synapse have an absolute requirement for Ca²⁺. In contrast, rapid endocytosis in chromaffin cells has been reported to be completely blocked when Ca²⁺ is replaced with Sr²⁺ (Artalejo et al. 1995, 1996).

The kinetics of endocytosis in Ba²⁺ were more difficult to assess. Figure 9A and B compares capacitance responses to 500 ms depolarizations in Ca²⁺ and Ba²⁺. In Ca²⁺, the capacitance began to fall immediately after the stimulus, indicating that the rate of endocytosis exceeded the rate of exocytosis. In contrast, the capacitance response in Ba²⁺ continued to increase for about 4 s, showing that the rate of exocytosis exceeded endocytosis for at least this period. The capacitance then plateaued, reflecting a balance between the rates of exocytosis and endocytosis. We cannot assess these rates from capacitance measurements alone, but the FM1-43 measurements shown in Fig. 5 demonstrate that the introduction of a Ba²⁺ load can stimulate asynchronous exocytosis for prolonged periods. In four of the nine terminals tested in this way, no recovery in membrane capacitance was observed within 20 s. In the remaining five experiments, there was a clear but slow fall in capacitance that began within seconds, confirming that endocytosis occurred at an appreciable rate in Ba²⁺. An example of such an experiment is shown in Fig. 9C and D. Although these results confirm that a slow form of endocytosis occurs after stimulation in Ba²⁺, it remains to be seen whether the fast mechanism of membrane retrieval is also supported.

		View larger version [in this window] [in a new window]
Figure 9. Recovery of membrane capacitance after stimulation in Ca2+ and Ba2+ A, capacitance responses to a 500 ms depolarization from -70 mV to 0 mV, delivered in 2.5 mM Ca2+ (thin trace) and 2.5 mM Ba2+ (bold trace). The inset shows the currents activated by the stimulus (scale bars: 50 pA and 250 ms). B, the responses in A scaled to the same maximum to allow comparison of the kinetics of membrane retrieval. In this example the membrane capacitance continued to rise for several seconds after stimulation in Ba2+. C, a second example of the experiment shown in A. In this terminal the recovery in membrane capacitance began seconds after stimulation in Ba2+. The inset shows the currents activated by the stimulus (scale bars: 100 pA and 250 ms). D, the responses in C scaled to the same maximum.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

It has long been known that Sr²⁺ and Ba²⁺ can substitute for Ca²⁺ in triggering exocytosis in neurones and endocrine cells. At a number of synapses, Sr²⁺ and Ba²⁺ reduce the rapid component of exocytosis synchronized with the action potential but increase the slower 'asynchronous' component of release that follows (Zengel & Magleby, 1977; Goda & Stevens, 1994). It has been suggested that this difference reflects the control of fast and slow exocytosis by two distinct Ca²⁺ sensors that have different divalent cation selectivities. In this study we characterized the effects of Sr²⁺ and Ba²⁺ on defined kinetic components of exocytosis and endocytosis in retinal bipolar cells. Our results run contrary to two commonly held views. First, we found that Sr²⁺ and Ba²⁺ could trigger fast exocytosis at rates comparable to that triggered by Ca²⁺. The reduced response to a brief stimulus in Sr²⁺ or Ba²⁺ occurred because fewer vesicles were available for rapid release by these cations. Second, we found that Ba²⁺ stimulated slow exocytosis less efficiently than either Ca²⁺ or Sr²⁺. Direct measurements of the intracellular concentration of free Ba²⁺ demonstrated that this ion potentiated exocytosis because it was cleared from the cytoplasm much more slowly than either Ca²⁺ or Sr²⁺.

The effects of divalent cations on the fast mode of exocytosis

The most rapid form of exocytosis in depolarizing bipolar cells was inhibited by replacing Ca²⁺ with either Sr²⁺ or Ba²⁺. The divalent cation selectivity of rapid exocytosis at this ribbon synapse was therefore similar to that observed in a variety of other synapses, including the neuromuscular junction (Zengel & Magleby, 1977), squid giant synapse (Augustine & Eckert, 1984), hippocampal neurones in culture (Goda & Stevens, 1994) and cerebellar granule cells (Xu-Friedman & Regehr, 1999, 2000). It has been suggested that this is because the Ca²⁺ sensor for rapid exocytosis binds Sr²⁺ and Ba²⁺ with lower affinity, leading to a lower rate constant of release (Geppert et al. 1994). A second mechanism that would inhibit rapid exocytosis, over and above any change in the rate constant of release, would be a reduction in the size of the RRP. In fact, this was the most obvious effect we observed when replacing Ca²⁺ with Sr²⁺ or Ba²⁺. For instance, Fig. 5C shows that in Sr²⁺, about 30 % of the RRP was released at a rate comparable to that observed in Ca²⁺, while the remaining vesicles were released at least 100 times more slowly. The fraction of the RRP that could be rapidly released in Ba²⁺ was even smaller (Table 1). It may be that Sr²⁺ or Ba²⁺ also caused a reduction in the rate constant of exocytosis, but our capacitance measurements do not have the resolution to test this idea. It would be interesting to know whether replacing Ca²⁺ with Sr²⁺ or Ba²⁺ also caused a reduction in the size of the RRP at other synapses.

Why does the RRP appear to be smaller when exocytosis is triggered by Sr²⁺ or Ba²⁺? It is difficult to answer this question because we do not know the concentrations of Ca²⁺, Ba²⁺ and Sr²⁺ that vesicles within this pool experience when Ca²⁺ channels open. If all vesicles experience a similar rise in divalent cation concentration and a single Ca²⁺ sensor is involved in triggering exocytosis, the RRP would be expected to have a fixed capacity. However, it may be that vesicles within the RRP do not all experience the same Ca²⁺ signal, and that vesicles further away from the Ca²⁺ channels are released more slowly (Burrone & Lagnado, 2000; Sakaba & Neher, 2001). The size of the RRP in Sr²⁺ and Ba²⁺ might then be reduced because only a subset of docked vesicles experience concentrations of these divalent cations sufficient to trigger measurable release.

A second cause of heterogeneity within the RRP might be that different vesicles may possess different Ca²⁺ sensors, only some of which bind Sr²⁺ or Ba²⁺ effectively. Blank et al. (1998) found that Ca²⁺ concentrations above threshold caused the fusion of only a sub-population of secretory vesicles in sea-urchin eggs, and proposed that individual vesicles have different sensitivities to Ca²⁺. However, existing evidence indicates that synaptic vesicles in bipolar cells have a uniform sensitivity to Ca²⁺, since all vesicles within the RRP are released with a single time constant, irrespective of the concentration of Ca²⁺ (Heidelberger et al. 1994).

Clearly it is important to establish the identity of the Ca²⁺ sensor(s) that regulates vesicle fusion. Synaptotagmin I is a Ca²⁺-binding protein that plays a key role in rapid exocytosis at the synapse (Nonet et al. 1993; Littleton et al. 1993; Geppert et al. 1994; Fernandez-Chacon et al. 2001), and it has a lower affinity for Sr²⁺ and Ba²⁺ than it does for Ca²⁺ (Li et al. 1995), leading to the suggestion that it underlies the divalent cation selectivity of rapid exocytosis (Geppert et al. 1994). Synaptotagmin I is present at all ribbon synapses in the mammalian retina (reviewed in Morgans, 2000).

The effects of divalent cations on the slow mode of exocytosis

The most rapid component of exocytosis at the bipolar cell synapse involves the release of vesicles docked at the plasma membrane (Mennerick & Matthews, 1996; Zenisek et al. 2000), but continuous exocytosis must involve vesicle movements from the bulk cytoplasm, because the numerical equivalent of the RRP is released every second (Lagnado et al. 1996; Neves & Lagnado, 1999). The movement of vesicles to the membrane during the slow phase of exocytosis in bipolar cells has been visualized using total internal reflection microscopy (Zenisek et al. 2000). It therefore seems likely that the rate-limiting processes controlled by Ca²⁺ are different during the transient and continuous phases of exocytosis. Support for this idea comes from the Ca²⁺ dependence of these processes. Whereas release from the RRP requires a Ca²⁺ concentration of at least 20 µM (Heidelberger et al. 1994), continuous exocytosis can occur at a Ca²⁺ concentration below 1 µM (Lagnado et al. 1996; Rouze & Schwartz, 1998). The idea that provision of new vesicles competent for fusion is rate limiting during continuous exocytosis is also supported by the demonstration that refilling of the RRP is accelerated by Ca²⁺ (Gomis et al. 1999).

Given that the rates of fast and slow exocytosis are likely to be limited by different processes, might they be differentiated by their relative sensitivities to Ca²⁺, Sr²⁺ and Ba²⁺? Our results do not provide any clear evidence for such a difference. Thus Ba²⁺, which supported only a small amount of fast exocytosis, also stimulated the slow mode of exocytosis less efficiently than Ca²⁺. Sr²⁺ accelerated the slow mode of exocytosis by a factor of 1.6, but only under conditions expected to make the Sr²⁺ transient 3-4 times larger than the Ca²⁺ transient. Xu-Friedman & Regehr (1999, 2000) measured Ca²⁺ and Sr²⁺ transients at the synapse of mouse cerebellar granule cells, and found that Sr²⁺ stimulated asynchronous release 5-10 times less efficiently than Ca²⁺. Taken together, these results suggest that the prolongation of asynchronous release by Ba²⁺ and Sr²⁺ does not represent the existence of a molecule that has a higher affinity for these divalent cations than it does for Ca²⁺.

There is, however, an important difference in the mechanisms by which Sr²⁺ prolongs asynchronous release in mouse cerebellar synapses compared to goldfish bipolar cells. Xu-Friedman & Regehr (1999) found that the Sr²⁺ transient in cerebellar synapses decayed about 6 times more slowly than the Ca²⁺ transient, whilst we found that Sr²⁺ and Ca²⁺ are cleared from bipolar cell terminals at the same rate (Fig. 5). Zenisek & Matthews (2000) found that a plasma membrane Ca²⁺-ATPase was the major mechanism by which Ca²⁺ was cleared from the synaptic terminal of bipolar cells. It therefore appears that this pump extrudes Sr²⁺ and Ca²⁺ at similar rates, but removes Ba²⁺ much more slowly (if at all).

The role of divalent cations in endocytosis and recycling

Endocytosis in bipolar cells did not have an absolute requirement for Ca²⁺, because it also occurred in Ba²⁺ and Sr²⁺ at rates sufficient to balance exocytosis during maintained stimulation (Fig. 2 and Fig. 3). Furthermore, vesicles retrieved in the presence of Ba²⁺ or Sr²⁺ became available for a second round of release, indicating that they were also able to pass through the intermediate steps by which vesicles are recycled.

Neves & Lagnado (1999) found that there were both fast and slow forms of endocytosis at the bipolar cell synapse, with fast endocytosis only predominating after a brief stimulus. Here we show that fast endocytosis in bipolar cells does not have an absolute requirement for Ca²⁺, because it occurred equally rapidly in Sr²⁺ (Fig. 8). In contrast, it has been reported that rapid endocytosis in chromaffin cells is completely blocked by replacing Ca²⁺ with either Sr²⁺ or Ba²⁺ (Artalejo et al. 1995, 1996). It is unclear whether this represents a real difference in the mechanisms of endocytosis operating at the synapse and chromaffin cells, because Nucifora & Fox (1998) found that rapid endocytosis in chromaffin cells was maintained after replacing Ca²⁺ with Ba²⁺.

	REFERENCES

Top Abstract Introduction Methods Results Discussion References

ARTALEJO C. R., ELHAMDANI, A. & PALFREY, H. C. (1996). Calmodulin is the divalent cation receptor for rapid endocytosis, but not exocytosis, in adrenal chromaffin cells. Neuron 16, 195-205	[Medline]
ARTALEJO C. R., HENLEY, J. R., MCNIVEN, M. A. & PALFREY, H. C. (1995). Rapid endocytosis coupled to exocytosis in adrenal chromaffin cells involves Ca2+, GTP and dynamin but not clathrin. Proceedings of the National Academy of Sciences of the USA 92, 8328-8332	[Abstract]
AUGUSTINE G. J. & ECKERT, R. (1984). Divalent cations differentially support transmitter release at the squid giant synapse. Journal of Physiology 346, 257-271	[Abstract]
BETZ W. J., MAO, F. & SMITH, C. B. (1996). Imaging exocytosis and endocytosis. Current Opinion in Neurobiology 6, 365-371	[Medline]
BLANK P. S., CHO, M. S., VOGEL, S. S., KAPLAN, D., KANG, A., MALLAY, J. & ZIMMERBERG, J. (1998). Submaximal responses in calcium-triggered exocytosis are explained by differences in the calcium sensitivity of individual secretory vesicles. Journal of General Physiology 112, 559-567	[Abstract/Full Text]
BURRONE J. & LAGNADO, L. (1997). Electrical resonance and Ca2+ influx in the synaptic terminal of depolarizing bipolar cells from the goldfish retina. Journal of Physiology 505, 571-584	[Abstract]
BURRONE J. & LAGNADO, L. (2000). Synaptic depression and the kinetics of exocytosis in retinal bipolar cells. Journal of Neuroscience 20, 568-578	[Abstract/Full Text]
CHAO S.-H., SUZUKI, Y., ZYSK, J. R. & CHEUNG, W. Y. (1984). Activation of calmodulin by various metal cations as a function of ionic radius. Molecular Pharmacology 26, 75-82	[Abstract]
DITTMAN J. S. & REGEHR, W. G. (1998). Calcium dependence and recovery kinetics of presynaptic depression at the climbing fiber to Purkinje cell synapse. Journal of Neuroscience 18, 6147-6162	[Abstract/Full Text]
FERNANDEZ-CHACON R., KONIGSTORFER, A., GERBER, S. H., GARCIA, J., MATOS, M. F., STEVENS, C. F., BROSE, N., RIZO, J., ROSENMUND, C. & SUDHOF, T. C. (2001). Synaptotagmin I functions as a calcium regulator of release probability. Nature 410, 41-49	[Medline]
GEPPERT M., GODA, Y., HAMMER, R. E., LI, C., ROSAHL, T. W., STEVENS, C. F. & SÜDHOF, T. C. (1994). Synaptotagmin I: a major Ca2+-sensor for transmitter release at a central synapse. Neuron 79, 717-727
GILLIS K. D. (1995). Techniques for membrane capacitance measurements. In Single Channel Recording, ed. SAKMANN, B. & NEHER, E., pp. 155-198. Plenum, New York
GODA Y. & STEVENS, C. F. (1994). Two components of transmitter release at a central synapse. Proceedings of the National Academy of Sciences of the USA 91, 12942-12946	[Abstract]
GOMIS A., BURRONE, J. & LAGNADO, L. (1999). Two actions of calcium regulate the supply of releasable vesicles at the ribbon synapse of retinal bipolar cells. Journal of Neuroscience 19, 6309-6317	[Abstract/Full Text]
HEIDELBERGER R., HEINEMANN, C., NEHER, E. & MATTHEWS, G. (1994). Calcium dependence of the rate of exocytosis in a synaptic terminal. Nature 371, 513-515	[Medline]
HENKEL A. W. & ALMERS, W. (1996). Fast steps in exocytosis and endocytosis studied by capacitance measurements in endocrine cells. Current Opinion in Neurobiology 6, 350-357	[Medline]
JOB C. & LAGNADO, L. (1998). Calcium and protein kinase C regulate the actin cytoskeleton in the synaptic terminal of retinal bipolar cells. Journal of Cell Biology 143, 1661-1672	[Abstract/Full Text]
LAGNADO L., GOMIS, A. & JOB, C. (1996). Continuous vesicle cycling in the synaptic terminal of retinal bipolar cells. Neuron 17, 957-967	[Medline]
LI C., DAVLETOV, B. A. & SÜDHOF, T. C. (1995). Distinct Ca2+ and Sr2+ binding properties of synaptotagmins: definition of candidate Ca2+-sensors for the fast and slow components of neurotransmitter release. Journal of Biological Chemistry 270, 24898-24902	[Abstract/Full Text]
LITTLETON J., STERN, M., SCHULTZE, K., PERIN, M. & BELLEN, H. J. (1993). Mutational analysis of Drosophila synaptotagmin demonstrates its essential role in Ca2+-activated neurotransmitter release. Cell 74, 1125-1134	[Medline]
MARTELL A. E. & SMITH, R. M. (1974). Critical Stability Constants, vol. 1. Plenum, New York
MENNERICK S. & MATTHEWS, G. (1996). Ultrafast exocytosis elicited by calcium current in synaptic terminals of retinal bipolar neurons. Neuron 17, 1241-1249	[Medline]
MILEDI R. (1966). Strontium as a substitute for calcium in the process of transmitter release at the neuromuscular junction. Nature 212, 1233-1234
MORGANS C. W. (2000). Presynaptic proteins of ribbon synapses in the retina. Microscopy Research and Technique 50, 141-150	[Medline]
NEHER E. (1998). Vesicle pools and Ca2+ microdomains: new tools for understanding their roles in neurotransmitter release. Neuron 20, 389-399	[Medline]
NEHER E. & MARTY, A. (1982). Discrete changes of cell membrane capacity observed under conditions of enhanced secretion in bovine adrenal chromaffin cells. Proceedings of the National Academy of Sciences of the USA 79, 6712-6716	[Medline]
NEVES G. & LAGNADO, L. (1999). The kinetics of exocytosis and endocytosis in the synaptic terminal of goldfish retinal bipolar cells. Journal of Physiology 515, 181-202	[Abstract/Full Text]
NONET M. L., GRUNDAHL, K., MEYER, B. J. & RAND, J. B. (1993). Synaptic function is impaired but not eliminated in C. elegans mutants lacking synaptotagmin. Cell 73, 1291-1305	[Medline]
NUCIFORA P. G. & FOX, A. P. (1998). Barium triggers rapid endocytosis in calf adrenal chromaffin cells. Journal of Physiology 508, 483-494	[Abstract/Full Text]
ROUZE N. C. & SCHWARTZ, E. A. (1998). Continuous and transient vesicle cycling at a ribbon synapse. Journal of Neuroscience 18, 8614-8624	[Abstract/Full Text]
SAKABA T. & NEHER, E. (2001). Quantitative relationship between transmitter release and calcium current at the calyx of Held synapse. Journal of Neuroscience 21, 462-476	[Abstract/Full Text]
SAKABA T., TACHIBANA, M., MATSUI, K. & MINAMI, N. (1997). Two components of transmitter release in retinal bipolar cells: exocytosis and mobilization of synaptic vesicles. Neuroscience Research 27, 357-370	[Medline]
SILINSKI E. M. (1978). On the role of barium in supporting the asynchronous release of acetylcholine quanta by motor nerve impulses. Journal of Physiology 274, 157-171	[Abstract]
SMITH C. B. & BETZ, W. J. (1996). Simultaneous independent measurements of endocytosis and exocytosis. Nature 380, 531-534	[Medline]
STEVENS C. F. & WESSELING, J. F. (1998). Activity-dependent modulation of the rate at which synaptic vesicles become available to undergo exocytosis. Neuron 21, 415-424	[Medline]
TSIEN R.Y & POZZAN, T. (1989). Measurement of cytosolic free Ca2+with Quin-2. Methods in Enzymology 172, 230-262	[Medline]
VON GERSDORFF H. & MATTHEWS, G. (1994). Dynamics of synaptic vesicle fusion and membrane retrieval in synaptic terminals. Nature 367, 735-739	[Medline]
VON GERSDORFF H., VARDI, E., MATTHEWS, G. & STERLING, P. (1996). Evidence that vesicles on the synaptic ribbon of retinal bipolar neurons can be rapidly released. Neuron 16, 1221-1227	[Medline]
WANG L. Y. & KACZMAREK, L. K. (1998). High-frequency firing helps replenish the readily releasable pool of synaptic vesicles. Nature 394, 384-388	[Medline]
WU L. G. & BETZ, W. J. (1996). Nerve activity but not intracellular calcium determines the time-course of endocytosis at the frog neuromuscular junction. Neuron 17, 769-779	[Medline]
XU-FRIEDMAN M. A. & REGEHR, W. G. (1999). Presynaptic strontium dynamics and synaptic transmission. Biophysical Journal 76, 2029-2042	[Abstract/Full Text]
XU-FRIEDMAN M. A. & REGEHR, W. G. (2000). Probing fundamental aspects of synaptic transmission with strontium. Journal of Neuroscience 20, 4414-4422	[Abstract/Full Text]
ZENGEL J. E. & MAGLEBY, K. L. (1977). Transmitter release during repetitive stimulation: selective changes produced by Sr2+ and Ba2+. Science 19, 67-69
ZENISEK D., STEYER, J. A. & ALMERS, W. (2000). Transport, capture and exocytosis of single synaptic vesicles at active zones. Nature 406, 849-854	[Medline]
ZENISEK D. & MATTHEWS, G. (2000). The role of mitochondria in presynaptic calcium handling at a ribbon synapse. Neuron 25, 229-237	[Medline]
ZUCKER R. S. (1996). Exocytosis: a molecular and physiological perspective. Neuron 17, 1049-1055	[Medline]

Acknowledgements

G.N. was a student of the Gulbenkian PhD Program in Biology and Medicine and was supported by a grant from Fundação para a Ciência e Tecnologia. We thank Juan Burrone, Ana Gomis, Anne Cooke and Matthew Holt for useful discussions.

Corresponding author

L. Lagnado: MRC Laboratory of Molecular Biology, Hills Road, Cambridge CB2 2QH, UK.

Email: ll1{at}mrc-lmb.cam.ac.uk

This article has been cited by other articles:

T. J. Searl and E. M. Silinsky Mechanisms of Neuromodulation as Dissected Using Sr2+ at Motor Nerve Endings J Neurophysiol,