Two modes of exocytosis from synaptosomes are differentially regulated by protein phosphatase types 2A and 2B

Authors

  • Monique L. Baldwin,

    1. School of Biomedical Sciences, University of Newcastle and Clinical Neuroscience Program, Hunter Medical Research Institute, Callaghan, New South Wales, Australia
    Search for more papers by this author
  • John A. P. Rostas,

    1. School of Biomedical Sciences, University of Newcastle and Clinical Neuroscience Program, Hunter Medical Research Institute, Callaghan, New South Wales, Australia
    Search for more papers by this author
  • Alistair T. R. Sim

    1. School of Biomedical Sciences, University of Newcastle and Clinical Neuroscience Program, Hunter Medical Research Institute, Callaghan, New South Wales, Australia
    Search for more papers by this author

Address correspondence and reprint requests to Associate Professor A. T. R. Sim, School of Biomedical Sciences, University of Newcastle and Clinical Neuroscience Program, Hunter Medical Research Institute, Callaghan, NSW 2308, Australia. E-mail: Alistair.Sim@newcastle.edu.au

Abstract

The inhibitors okadaic acid (OA), fostriecin (FOS) and cyclosporin A (CsA), were used to investigate the roles of protein phosphatases in regulating exocytosis in rat brain synaptosomes by measuring glutamate release and the release of the styryl dye FM 2-10. Depolarization was induced by 30 mm KCl, or 0.3 mm or 1 mm 4-aminopyridine (4AP). OA and FOS produced a similar partial inhibition of KCl- and 0.3 mm 4AP- evoked exocytosis in both assays, but had little effect upon exocytosis evoked by 1 mm 4AP. In contrast, CsA had no effect upon KCl- and 0.3 mm 4AP-evoked exocytosis, but significantly enhanced glutamate release but not FM 2-10 dye release evoked by 1 mm 4AP. None of the phosphatase inhibitors changed calcium signals from FURA-2-loaded synaptosomes either before or after depolarization. Pretreatment with 100 nm phorbol 12-myristate 13-acetate abolished the inhibitory effect of OA on exocytosis induced by 0.3 mm 4AP. Taken together, these results show that exocytosis from synaptosomes has a phosphatase-sensitive and phosphatase-insensitive component, and that there are two modes of phosphatase-sensitive exocytosis that can be elicited by different depolarization conditions. Moreover, these two modes are differentially sensitive to phosphatase 2A and 2B.

Abbreviations used
4AP

4-aminopyridine

CsA

cyclosporin A

FOS

fostriecin

OA

okadaic acid

PKC

protein kinase C

PMA

phorbol 12-myristate 13-acetate

PP1

protein phosphatase 1

PP2A

protein phosphatase 2A

PP2B

protein phosphatase 2B

The modulation of synaptic vesicle cycling is one of the critical cellular mechanisms that controls synaptic transmission, regulating the supply of releasable transmitter at the nerve terminal. Once synaptic vesicles have fused with the plasma membrane to release neurotransmitters (exocytosis), they must be recycled so that they can be refilled and undergo the next round of release. Full synaptic vesicle fusion, whereby vesicles collapse completely into the plasma membrane, internalize by endocytosis and then recycle through the endosome (Alvarez et al. 1993; Ales et al. 1999), is the classical model of exocytosis and the principal mechanism under normal stimulation conditions (De Camilli and Takei 1996). However, an alternative mechanism has been suggested whereby vesicles fuse only transiently with the plasma membrane and are rapidly retrieved (Valtorta et al. 2001). In this mode of release, referred to as ‘transient fusion’ or ‘kiss-and-run’, vesicles fuse with the plasma membrane by forming a transient fusion pore while preserving vesicle integrity. After detachment, intact vesicles can fuse again with the plasma membrane without any preceding endosomal fusion. Kiss-and-run has been demonstrated by analysis of secretion from neuroendocrine and chromaffin cells by combining whole-cell capacitance methods and amperometry to measure vesicle fusion and catecholamine release of vesicle contents (Alvarez et al. 1993; Zhou et al. 1996). The advantage of a kiss-and-run mechanism is a rapid cycling between a fusion state and a non-fusion state, thus accelerating the turnover of the limited pool of synaptic vesicles in neurons (Cousin and Robinson 1999).

Studies using the lipophilic, but membrane impermeable, fluorescent styryl dyes in combination with the measurement of soluble neurotransmitter release have made possible the investigation of kiss-and-run in neurons (Henkel and Betz 1995; Klingauf et al. 1998). When the styryl dye FM 2-10 reversibly partitions into the outer leaflet of exposed plasma membrane its fluorescence increases. By endocytosis, the dye is able to label the luminal surface of the synaptic vesicle and, during subsequent exocytosis, is lost to the extracellular medium, accompanied by a decrease in fluorescence (Cousin and Robinson 1999). Release of neurotransmitter is detectable whether the nerve terminal uses full fusion or kiss-and-run, as neurotransmitter is soluble within the synaptic vesicle and can diffuse rapidly through an open pore. However, only the relatively slow process of full fusion would allow complete dye loss from the membrane, whereas in the relatively rapid kiss-and-run mode the synaptic vesicles are internalized before the dye can fully departition and escape from the membrane (Klingauf et al. 1998).

By comparing the relative ability of the glutamate release and styryl dye release assays to measure exocytosis, Cousin and Robinson (2000) were able to distinguish between two different modes of glutamate release from isolated nerve terminals (synaptosomes); 30 mm KCl and 0.3 mm 4-aminopyridine (4AP) evoked exocytosis by a mechanism that was readily detectable with both the glutamate and FM 2-10 release assays, and was therefore assumed to be occurring with the full fusion mode of release. However, 1 mm 4AP recruited a second mode of release that was detectable using the glutamate assay but not the FM 2-10 assay; this was proposed to be kiss-and-run. This second mode was proposed to be largely mediated by protein kinase C (PKC) (Cousin and Robinson 2000).

Phosphorylation and dephosphorylation of nerve terminal proteins is known to regulate neurotransmitter release, and experiments using membrane-permeable inhibitors of protein phosphatases suggest that serine/threonine protein phosphatases are positive regulators of this release (Sim et al. 1991; Verhage et al. 1995; Vickroy et al. 1995; Issa et al. 1999; Storchak et al. 2001). However, the relative importance of the different serine/threonine protein phosphatases in the full fusion and kiss-and-run modes of release is not known. The three major serine/threonine phosphatases are protein phosphatase 1 (PP1), 2A (PP2A) and 2B (PP2B), all three of which are found in rat brain synaptosomes (Sim et al. 1993). Okadaic acid (OA), an inhibitor of protein phosphatases PP1, PP2A and to a lesser extent PP2B (Haystead et al. 1989; Ishihara et al. 1989), inhibits depolarization-induced glutamate release from isolated nerve terminals (synaptosomes) (Verhage et al. 1995; Vickroy et al. 1995; Issa et al. 1999). However, it is impossible to ascertain which class of protein phosphatase is being affected in this attenuation of transmitter release, as OA causes partial or complete inhibition of several protein phosphatases. The use of selective inhibitors is necessary to resolve this issue.

We have used the highly selective, membrane-permeable protein phosphatase inhibitors fostriecin (FOS) and cyclosporin A (CsA) to delineate the class of protein phosphatases likely to be involved in the release of neurotransmitter from synaptosomes. Synaptosomes have been used extensively to study the mechanisms of synaptic neurotransmitter release because they remain metabolically active and preserve the ability to synthesize and release neurotransmitters. Our results support the argument for the presence of two modes of exocytosis in synaptosomes and, more importantly, show that PP2A and PP2B have distinct roles in regulating exocytosis from synaptosomes by these two modes.

Experimental procedures

Materials

OA was obtained from Alexis Biochemicals (San Diego, CA, USA); FOS and CsA were obtained from Calbiochem (San Diego, CA, USA). FM 2-10 and FURA-2 AM were from Molecular Probes (Eugene, OR, USA). All other reagents were obtained from the Sigma Chemical Co. (St Louis, MO, USA).

Glutamate release assay

Synaptosomes were prepared from rat brain cerebral cortex by centrifugation on discontinuous Percoll gradients (fractions 3 and 4 were combined) as described previously (Dunkley et al. 1986). Under the conditions used, synaptosomes remain viable for several hours and respond to multiple depolarization and repolarization signals (Dunkley et al. 1986). The glutamate release assay was performed using enzyme-linked fluorescent detection of released glutamate (Nicholls and Sihra 1986). In brief, synaptosomes were stored on ice and were diluted to 500 µg protein in 2 mL Krebs-like solution (118 mm NaCl, 5 mm KCl, 25 mm NaHCO3, 1 mm MgCl2, 10 mm Glucose, pH 7.4) at 37°C. Experiments were started after addition of 1 mm NADP+. Fifty units of glutamate dehydrogenase were added after 1 min, and the synaptosome suspension was treated after 4 min with either KCl (30 mm) or 4AP (0.3 or 1 mm). Fluorescence at 460 nm emission was measured continuously in a Perkin-Elmer (Shelton, CT, USA) LS-50B spectrofluorimeter using a four-cuvette holder that enabled four samples to be run simultaneously. A typical experiment consisted of two control conditions, and two treated conditions, each in the presence and absence of Ca2+ (1.2 mm), thus enabling each sample to be measured against its own control simultaneously. Glutamate release was quantitated by the addition of 10 nmol glutamate at the end of each run, which acted as an internal standard that allowed comparison to be made between experiments. Data are presented as the Ca2+-dependent glutamate release (expressed in nmoles per milligram protein), measured at different time points after stimulation (100, 200, 300 and 400 s), and calculated as the difference between release measured in the presence and absence of added Ca2+. In experiments in which inhibitors were used, the synaptosomes were preincubated with either OA (0.01–10 µm), FOS (0.1 or 1 µm), CsA (10 µm) or phorbol 12-myristate 13-acetate (PMA) (100 nm), for 5, 5, 15 or 7 min respectively, before depolarization with either KCl or 4AP.

None of the drugs used altered basal release (in the presence or absence of added Ca2+), and none had any effect on either KCl- or 4AP-evoked Ca2+-independent release (data not shown).

Styryl dye release assay

Synaptic vesicle fusion with the plasma membrane was measured using release of the fluorescent dye FM 2-10, according to Cousin and Robinson (2000). In brief, synaptosomes were stored on ice and were diluted to 700 µg protein in 2 mL Ca2+-containing Krebs-like solution (as used in the glutamate release assay) at 37°C. FM 2-10 (100 µm) was added after 3 min, and endocytotic uptake of membrane-bound FM 2-10 was stimulated 1 min later with 30 mm KCl. After 2 min to allow internalization of the dye, synaptosomes were pelleted in a microfuge for 1 min and washed twice in Krebs-like solution containing 1 mg/mL bovine serum albumin. Synaptosomes were then resuspended in Krebs-like solution (either plus or minus 1.2 mm Ca2+) at 37°C, and release induced by the addition of either KCl (30 mm) or 4AP (0.3 or 1 mm). Release of accumulated FM 2-10 was measured as the decrease in fluorescence upon release of the dye into solution (excitation 488 nm, emission 540 nm). Data are presented as the Ca2+-dependent decrease in FM 2-10 fluorescence and calculated as the difference between release measured in the presence of 1.2 mm Ca2+ and that in the presence of 1.2 mm EGTA. Any drugs were added after the dye-loading procedure, and the synaptosomes were preincubated with OA (0.01–10 µm), FOS (0.1 µm), CsA (10 µm) or PMA (100 nm) for 5 min before depolarization with either KCl or 4AP. For experiments with PMA and OA, PMA was added first, and OA was added 30 s later.

Calcium measurements

Intrasynaptosomal calcium levels were measured using FURA-2 fluorescence according to Brent et al. (1997). FURA-2 was prepared as a 1-mm stock solution in dimethyl sulfoxide and 0.5% pluronic acid. In brief, synaptosomes (2 mg protein) were incubated with gentle shaking for 40 min at 37°C in 2 mL control buffer containing 5 µm FURA-2. The suspensions were then washed in ice-cold Krebs-like buffer and centrifuged for 2 min to remove extrasynaptosomal FURA-2. Synaptosomes were then stored on ice and diluted to 600 µg protein in 2 mL Ca2+-containing Krebs-like solution at 37°C, adjusted to give a final concentration of 1.2 mm Ca2+ in the cuvette. Measurements were made by alternating the excitation wavelengths of 340 and 380 nm (the 340 : 380 nm fluorescence ratio); fluorescent emission was monitored at 510 nm. Calibration of the fluorescence signals was performed at the end of each experiment by adding digitonin (200 µm) to obtain Fmax followed by EGTA (2 mm) to obtain Fmin (Grynkiewicz et al. 1985). Extrasynaptosomal FURA-2 was determined for each synaptosomal preparation by adding MnCl2 (10 mm) at the end of each experiment to quench the extracellular fluorescence and was stable for the duration of the experiments. Intracellular [Ca2+] was calculated by the equation of Grynkiewicz et al. (1985), using a KD of 81 nm for the Ca2+–FURA-2 complex derived under our laboratory conditions by scanning the fluorescent response of FURA-2 to different concentrations of Ca2+ (0–39.8 µm). In experiments in which drugs were used, the synaptosomes were preincubated with OA (0.1 µm), FOS (0.1 µm) or CsA (10 µm), for 5, 5 or 15 min respectively, before depolarization with either KCl or 4AP. As basal calcium levels in synaptosomes gradually increased during the 3–4 h involved in a series of experiments (86–117 nm Ca2+), the sequence of application of drugs was varied between experiments, and drug-treated samples were compared with control samples measured immediately before or after the drug-treated samples. There was no difference in drug effects associated with the age of the synaptosomes.

Results

Experimental approach

Exocytosis from isolated synaptosomes was measured using two complementary approaches. Results from the measurement of glutamate and FM 2-10 release in response to three different stimuli are shown in Fig. 1. Although the time course of exocytosis is different for the two assays, this is a reflection of the nature of the assays used rather than an inherent difference in the mechanisms of exocytosis of glutamate and FM 2-10. Specifically, the exocytosis of glutamate is slower than that of FM 2-10 as a consequence of the dilution of released glutamate into a large volume of buffer. This results in submaximal conditions for the glutamate dehydrogenase used in the assay. The signal generated over minutes by the assay therefore represents enzymatic amplification of the glutamate released within seconds. FM dye is released more slowly than glutamate owing to the requirement for it to departition from the membrane before a signal is generated. However, it is measured directly, producing a signal in seconds, and therefore has a shorter time course than measurement of glutamate.

Figure 1.

Ca2+-dependent glutamate (a) and FM 2-10 (b) release evoked by 30 mm KCl, 0.3 mm 4AP and 1 mm 4AP. Arrowhead shows when depolarizing agents were added. Each point is the mean ± SEM of 3–4 independent experiments.

Measurement of glutamate release clearly restricts the measurement of exocytosis to glutamatergic synaptosomes, although these do represent a large proportion of the synaptosomes in the preparation. Measurement of FM 2-10 dye release, on the other hand, is independent of the neurotransmitter, and represents exocytosis from all synaptosomes. In order to restrict analysis to exocytosis rather than signals generated by non-specific leaching or bleaching of the dye, results are reported as only the Ca2+-dependent release.

Two alternate modes of depolarization were studied (KCl and 4AP) to investigate whether any effects of the phosphatase inhibitors on glutamate release were due to actions on events associated with exocytosis, as opposed to the depolarization-activated steps that trigger exocytosis. KCl produces a clamped depolarization, whereas 4AP, at the concentrations used, leads to propagation of repetitive action potentials by blocking both A type and delayed rectifier potassium channels, thereby preventing correction of spontaneous depolarizations that occur from ion diffusion (Nicholls 1993). Levels of Ca2+-dependent glutamate release evoked by 30 mm KCl and 0.3 mm 4AP were similar, whereas depolarization with 1 mm 4AP evoked more release (Fig. 1a). However, the depolarization-stimulated FM 2-10 release was similar for all depolarizing stimuli (Fig. 1b). These results are consistent with the findings of Cousin and Robinson (2000) that the slow rate of diffusion of FM 2-10 dye limits its capacity to detect short-duration modes of exocytosis, suggesting that the increased exocytosis promoted by higher 4AP concentrations represents a ‘kiss-and-run’-like mode.

Regression analysis of FM 2-10 release indicated that the data were consistent with two phases of release. The biphasic nature of exocytosis from synaptosomes has been well documented although the interpretation of this remains unclear. Analysis showed that the initial rate constant for each of the three depolarization conditions was similar (τ = 4–6 s). Importantly, these kinetics of FM dye release derived from a population of synaptosomes are within the same order of magnitude as that observed in a single presynaptic bouton from an intact neuron (Ryan et al. 1993).

Effects of OA and FOS: role of PP1 and PP2A in exocytosis

The protein phosphatase inhibitor OA has previously been shown to inhibit KCl-evoked glutamate release in synaptosomes (Verhage et al. 1995). In this study OA inhibited this release by a mean ± SEM of 58 ± 5% at 100 s (Fig. 2a). Investigation of the concentration dependence of this inhibition of glutamate release showed that inhibition was maximal at 0.1 µm; increasing the OA concentration 100-fold to 10 µm did not significantly change this inhibition (data not shown). This also suggests that PP2B, which is inhibited by OA at 10 µm but not at 0.1 µm (Cohen 1989), has little role in regulating release under these conditions. However, it cannot be determined whether the major inhibitory effect on release seen at lower OA concentrations results from inhibition of PP2A and/or PP1. FOS (0.1 µm), a potent and selective inhibitor of PP2A (Evans and Simon 2001), inhibited glutamate release to a similar extent as OA (Fig. 2b). Increasing the concentration of FOS to 1 µm did not change the extent of inhibition (data not shown). There was no significant difference (p > 0.24, Student t-test) between the level of inhibition of Ca2+-dependent release by OA and FOS at any time after depolarization. The effect of OA and FOS on glutamate release evoked by 0.3 mm 4AP (Figs 2c and d) followed the same pattern as that observed following KCl-evoked release; OA (0.1 µm) and FOS (0.1 µm) inhibited Ca2+-dependent release to a similar extent. These results suggest a major role for PP2A in the regulation of KCl- and 0.3 mm 4AP-evoked glutamate release. In contrast, the effect of phosphatase inhibition on release evoked by 1 mm 4AP was markedly different: OA and FOS (Figs 2e and f respectively) had little or no effect on Ca2+-dependent glutamate release.

Figure 2.

Effect of OA and FOS on Ca2+-dependent release of glutamate in synaptosomes depolarized by 30 mm KCl (a and b), 0.3 mm 4AP (c and d), and 1 mm 4AP (e and f). Synaptosomes were preincubated with either 0.1 µm OA (a, c and e; dotted line) or 0.1 µm FOS (b, d and f; dashed line) before depolarization. Each point is the mean ± SEM of 3–4 independent experiments.

To investigate whether the effects of OA and FOS were specific for glutamate release or more generally applicable to the mechanisms of synaptic exocytosis, we next investigated FM 2-10 release and compared the results with the profile of glutamate release under the same conditions. The effects of the phosphatase inhibitors OA and FOS on depolarization-stimulated FM 2-10 release are shown in Fig. 3. OA and FOS both inhibited FM 2-10 release in the same manner as that observed for glutamate release. Inhibition was greatest when release was evoked by KCl or 0.3 mm 4AP (Figs 3a–d): 31 ± 4 and 28 ± 2% inhibition of KCl-evoked release, and 30 ± 3 and 31 ± 3% inhibition of 0.3 mm 4AP-evoked release respectively (3 min after depolarization). Kinetic analysis showed that the rate constants (τ) of release were not affected by the inhibitors (not shown); rather, as indicated by the figures, the effect was on the extent of the first phase. In contrast to release measured in response to 30 mm KCl or 0.3 mm 4AP, neither OA nor FOS significantly affected release evoked by 1 mm 4AP (Figs 3e and f).

Figure 3.

Effect of OA and FOS on Ca2+-dependent release of FM 2-10 in synaptosomes depolarized by 30 mm KCl (a and b), 0.3 mm 4AP (c and d), and 1 mm 4AP (e and f). Arrowhead shows when depolarizing agent was added. Synaptosomes were preincubated with either 0.1 µm OA (a, c and e) or 0.1 µm FOS (b, d and f) before depolarization. C, control. Each point is the mean ± SEM of 3–4 independent experiments.

It has been shown that PMA greatly increases glutamate release stimulated by 0.3 mm 4AP, but not FM 2-10 release; this has been interpreted as resulting from a switch from full fusion to kiss-and-run release (Cousin and Robinson 2000). Although PMA has a number of targets, this effect of PMA is assumed to be related to activation of PKC as it is blocked by inhibition of PKC (Cousin and Robinson 2000). Consistent with these results, we found that PMA increased 0.3 mm 4AP-evoked Ca2+-dependent glutamate release by 103 ± 8% but had no effect on FM 2-10 release (Fig. 4). In addition, OA (0.1 µm) produced no change in 0.3 mm 4AP-evoked release in the presence of PMA for both assays, whereas it reduced 0.3 mm 4AP-evoked release of glutamate and FM 2-10 by 64 ± 7% and 30 ± 3% respectively in the absence of PMA (Fig. 2c and Fig. 3c respectively). Taken together, these results are consistent with the hypothesis that PMA induces a switch from full fusion to kiss-and-run as the predominant mode of release (Cousin and Robinson 2000).

Figure 4.

Effect of OA on 0.3 mm 4AP-evoked glutamate release (a) and FM 2-10 release (b) in the presence of PMA. Arrowhead shows when depolarizing agent was added. Each point is the mean ± SEM of 3–4 independent experiments.

Effects of CsA: role of PP2B in exocytosis

To further investigate the role of different phosphatases in exocytosis we investigated the effect of the PP2B-selective inhibitor CsA (Liu et al. 1991) on both glutamate and FM2-10 release. Using this inhibitor, we observed no significant change in Ca2+-dependent glutamate release followingKCl- or 0.3 mm 4AP-induced depolarization (Figs 5a and b respectively). This is consistent with our findings that 10 µm OA (which also inhibits PP2B) had no greater effect on glutamate release than 0.1 µm OA, suggesting that PP2B plays little role in regulating the release of glutamate evoked by KCl or 0.3 mm 4AP. In contrast, CsA significantly increased glutamate release evoked by 1 mm 4AP, and the degree of stimulation increased with time reaching 59 ± 6% above control values by 400 s (Fig. 5c).

Figure 5.

Effect of selective inhibition of PP2B on Ca2+-dependent release of glutamate in synaptosomes depolarized by 30 mm KCl (a), 0.3 mm 4AP (b) and 1 mm 4AP (c). Synaptosomes were preincubated with 10 µm CsA (dotted line) before depolarization. Each point is the mean ± SEM of 3–4 independent experiments.

CsA did not affect FM 2-10 release under any of the depolarization conditions used (Fig. 6). The lack of effect on KCl- and 0.3 mm 4AP-stimulated release correlates well with results for glutamate release. However, CsA greatly increased Ca2+-dependent glutamate release stimulated with 1 mm 4AP, whereas FM 2-10 release was virtually unaffected. This is consistent with CsA affecting a second mode of exocytosis that is less apparent in the FM 2-10 assay.

Figure 6.

Effect of selective inhibition of PP2B on Ca2+-dependent release of FM 2-10 in synaptosomes depolarized by 30 mm KCl (a), 0.3 mm 4AP (b) and 1 mm 4AP (c). Arrowhead shows when depolarizing agent was added. Synaptosomes were preincubated with 10 µm CsA before depolarization. C, control. Each point is the mean ± SEM of 3–4 independent experiments.

Effects of phosphatase inhibition on calcium levels

To test the possibility that the effects of protein phosphatase inhibition on exocytosis were via modulation of calcium influx, mean calcium levels were measured using the fluorescent dye, FURA-2. Figure 7 shows that, under the conditions used, none of the inhibitors changed synaptosomal calcium levels under basal conditions or after depolarization with either KCl or 4AP (0.3 mm and 1 mm). The increase in levels of calcium induced by depolarization with 1 mm 4AP, the stimulus that induces the kiss-and-run-like mode of exocytosis, was significantly greater than that induced by depolarization with 0.3 mm 4AP, the stimulus that primarily induces the full fusion mode of exocytosis (p < 0.0001, Student's t-test). There was no significant difference between the calcium levels induced by depolarization with 0.3 mm 4AP or KCl, both of which induce exocytosis by full fusion.

Figure 7.

Effect of protein phosphatase inhibitors on calcium levels in synaptosomes depolarized with 30 mm KCl (a, b and c), 0.3 mm 4AP (d, e and f) or 1 mm 4AP (g, h and i). Each panel contains average traces for 3–5 experiments normalized to a common starting calcium level. Control (filled circle) and drug treatment (open square) traces are both plotted but are virtually coincident. Addition of drug and depolarizing agent are indicated by the downward and upward arrowheads respectively.

Discussion

Isolated synaptosomes provide a useful model with which to study synaptic exocytosis. Although the measurement of exocytosis and endocytosis in all in vitro preparations does not fully match the millisecond timescale that occurs in vivo, the molecular machinery underlying exocytosis is the same and all in vitro approaches, including electrophysiological ones, take advantage of artificially reduced rates to probe these molecular mechanisms. Interestingly, the rate constant of diffusion of FM dye in the present study, which measures a population of nerve terminals, is similar to that seen by direct imaging of individual hippocampal neurons (Ryan et al. 1993).

The inhibition of KCl-evoked release by OA has been observed in several cell types (Verhage et al. 1995; Vickroy et al. 1995; Issa et al. 1999), but no previous release studies have examined the effects of specifically inhibiting PP2A. We have now been able to do this using FOS, a PP2A inhibitor that is 10 000-fold more selective for PP2A than for PP1 (Walsh et al. 1997). Our finding that FOS inhibited KCl- and 0.3 mm 4AP-evoked glutamate release to a similar extent as OA shows that the principal phosphatase regulating release under these conditions is PP2A, although a role for PP1 cannot be ruled out. The results with FM 2-10 strengthen this interpretation, depicting a similar pattern of effect of OA and FOS on release evoked by KCl and 0.3 mm 4AP. The lack of effect of PP2B inhibition using the specific inhibitor CsA (Liu et al. 1991) confirmed that PP2B plays little role in release under these conditions. In view of the well recognized role of PP2B in endocytosis (Liu et al. 1994), one might have expected to see some effect of PP2B inhibition on the release of glutamate. The fact that PP2B inhibition had no observable effect on release indicates that glutamate release measurements under these experimental conditions primarily reflect steps in the synaptic vesicle fusion cycle associated with recruitment of release-ready vesicles and fusion with the membrane, rather than the whole synaptic vesicle cycle. It is important to note that maximal inhibition of PP2A evoked by KCl- and 0.3 mm 4AP evoked release does not completely inhibit release. Therefore, exocytosis from synaptosomes has both phosphatase-sensitive and phosphatase-insensitive components.

Cousin and Robinson (2000) reported no increase in FM 2-10 release as the concentration of 4AP was increased from 0.3 to 1 mm, despite an increase in glutamate release under the same conditions. Our results follow the same pattern and are therefore consistent with the interpretation that a different mechanism of release is recruited at higher concentrations of 4AP. Our results with phosphatase inhibitors reveal that under these different depolarization conditions exocytosis also shows a differential pharmacological sensitivity. This further strengthens the hypothesis that different mechanisms of release occur under different depolarization conditions. Specifically, OA and FOS inhibited exocytosis evoked by KCl and 0.3 mm 4AP, but not that evoked by 1 mm 4AP, whereas CsA increased glutamate release evoked by 1 mm 4AP but had no effect on release evoked by 0.3 mm 4AP and KCl.

The release of glutamate is detectable whether the nerve terminal uses the full fusion or kiss-and-run-like mode of exocytosis because glutamate is soluble within the synaptic vesicle and able to diffuse rapidly into the extracellular space, even if the vesicle remains fused for only a short period (as happens during kiss-and-run). However, the loss of FM 2-10 styryl dye from the luminal surface of the vesicle membrane is slow, so that only the relatively long duration of full fusion would allow substantial dye loss from the membrane. In the relatively short duration of the kiss-and-run mode of exocytosis the synaptic vesicles are internalized before most of the dye can escape from the membrane. It is therefore proposed that the mode of release seen at high concentrations of 4AP represents a kiss-and-run-type mechanism because it is observed with the glutamate release assay but not with the FM 2-10 release assay. Furthermore, we propose that PP2A is a positive regulator of the full fusion mode of exocytosis, shown by the ability of PP2A inhibitors (OA and FOS) to inhibit both glutamate and FM 2-10 release stimulated by KCl or 0.3 mm 4AP. However, PP2A appears to have little role to play in the kiss-and-run-like mode of exocytosis, as indicated by the lack of effect of OA and FOS on glutamate and FM 2-10 release evoked by 1 mm 4AP. The differential ability of the two measures of exocytosis to detect the effects of PP2B inhibition on 1 mm 4AP-evoked release suggests that PP2B is involved in regulating the kiss-and-run-like mode of exocytosis, in addition to its known role in endocytosis (Liu et al. 1994).

The differential effect of PP2B inhibition on glutamate release induced by 30 mm KCl and 1 mm 4AP was first noted by Nichols et al. (1994) but, in that study, the difference was assumed to be due to differences in the mode of depolarization. Our results show that this is not the case: changing the mode of depolarization does not change the sensitivity to the phosphatase inhibitors, as long as the same mechanism of release is activated. Using two different depolarization stimuli, KCl and 0.3 mm 4AP, each of which induces full fusion release, no sensitivity to the PP2B inhibitor CsA was observed. However, sensitivity to the PP2B inhibitor was seen when the same depolarizing agent (4AP) was used at a higher concentration (1 mm) because a different mode of exocytosis that is sensitive to PP2B action was induced.

Valtorta et al. (2001) discussed the possibility of both modes of release operating in synapses, with a switch to kiss-and-run occurring as vesicles acquire the ‘competence’ to release neurotransmitter in this way. By using amperometry, Ales et al. (1999) showed that, in response to an increase in extracellular Ca2+, chromaffin cells shifted from the full fusion mode to the kiss-and-run mechanism. During kiss-and-run events the fusion pore appeared to expand briefly to a large size, allowing for rapid and complete transmitter release. There was no size difference between vesicles that underwent full fusion and those that released transmitter by the kiss-and-run mechanism. This shows that vesicle cycling can be consistent with both full fusion and transient fusion. The rise in intracellular Ca2+ can be induced by increasing the concentration of Ca2+ in the incubation medium or by applying robust electrical stimulation paradigms (von Gersdorff and Matthews 1994; Ales et al. 1999; Pyle et al. 2000; Beutner et al. 2001), but also after the application of phorbol esters (Cousin and Robinson 2000). Phorbol esters, which activate PKC, increase the release of neurotransmitter in many neuronal systems (Malenka et al. 1986; Gerber et al. 1989; Barrie et al. 1991; Coffey et al. 1993; Capogna et al. 1995, 1999; Stevens and Sullivan 1998; Chen et al. 1999; Cousin and Robinson 1999; Yawo 1999), including synaptosomes. Although PMA has a number of targets, the enhancement of exocytosis is blocked by inhibitors of PKC (Cousin and Robinson 2000). Activation of PKC in chromaffin cells accelerated fusion pore expansion and subsequent pore closure and vesicle retrieval (Graham et al. 2000).

Is the additional mode of exocytosis induced in synaptosomes at high 4AP in addition to full fusion, or does the predominant mode of release switch from one mode to another? Our studies with phosphatase inhibitors add support to the latter conclusion. As FOS or OA inhibited release evoked by 0.3 mm but had no effect on release evoked by 1 mm 4AP, the simplest interpretation is that, by increasing the concentration of 4AP, a switch has occurred in the phosphatase-sensitive component of release from full fusion to the kiss-and-run-like mode. Cousin and Robinson (2000) reported that PMA induced a switch in the predominant mode of release from full fusion to kiss-and-run when synaptosomes were stimulated by 0.3 mm 4AP. We have confirmed these results and now show that, in the presence of PMA, OA produced no change in 0.3 mm 4AP-evoked release; however, in the absence of PMA, OA inhibited 0.3 mm 4AP-evoked release. These results are further evidence for the interpretation that a switch occurs in the predominant mode of release under these conditions, at least for the phosphatase-sensitive component of release.

One potential implication of a switch in the predominant mode of exocytosis is that release evoked by 0.3 mm 4AP plus PMA might be increased by PP2B inhibition with CsA (as seen for 1 mm 4AP). However, for this to be true, the CsA-sensitive step must be rate limiting even in the presence of PMA. We found that CsA had no effect on synaptosomes pretreated with 100 nm PMA (maximal stimulation of glutamate release) and depolarized with 0.3 mm 4AP (results not shown). Attempts to repeat the experiment with submaximal concentrations of PMA were unsuccessful because of the extremely steep dose–response curve of PMA (50 nm had no discernible effect on glutamate release).

At which point in the synaptic vesicle cycle are the serine/threonine phosphatases acting to differentially regulate the two modes of release? These enzymes are likely to have many substrate proteins involved in many aspects of the synaptic vesicle cycle such that no single protein is likely to be the target. One possibility is that the protein phosphatase inhibitors modulate exocytosis through modulation of calcium levels. However, experiments using FURA-2 did not show an effect of any phosphatase inhibitor on calcium levels under basal or stimulated conditions (Fig. 7). This suggests that the major role for PP2A and PP2B in regulating exocytosis is downstream of calcium influx, but their targets remain unknown. Interestingly, CsA also increased exocytosis from synaptosomes induced by ionomycin (Jovanovic et al. 2001), which is consistent with a role for PP2B downstream of calcium entry. Our studies also suggest that the mode of exocytosis induced by ionomycin is also of the kiss-and-run type.

Although the molecular mechanisms underlying the two modes of exocytosis have not been delineated, the emerging concept is that the two modes involve the same molecular components (Valtorta et al. 2001). Thus after a vesicle has fused with the membrane, it reaches a point in the common pathway when it must decide whether to proceed via the full fusion or the kiss-and-run mode; this decision is made by regulatory molecular mechanisms that alter the rates of the two pathways in response to different stimulatory conditions. This is interesting in view of recent mathematical modeling which suggests that the predominant role of protein phosphatases in the control of intracellular processes is to regulate the rate and duration of signaling (Heinrich et al. 2002). The simplest interpretation of our results is that PP2A promotes the step(s) that commit the vesicle to full fusion and, under normal depolarization conditions, the mechanism for retrieval of the vesicle by the kiss-and-run-like mode is not activated. Upon stimulation with a high 4AP concentration, the alternative pathway for the retrieval of the vesicle by the kiss-and-run-like mode is activated and the retrieval of vesicles is held in check by PP2B activity. Alternatively, PP2A and/or PP2B might be influencing the delivery of synaptic vesicles to the readily releasable pool or the plasma membrane. This is consistent with the recent finding that in response to 1 mm 4AP there is a rapid dephosphorylation of PP2B-specific sites on synapsin I, a protein involved in synaptic vesicle trafficking from reserve pools (Jovanovic et al. 2001). Given that PP2A influences only full fusion and PP2B influences only the kiss-and-run-like mode of exocytosis, this interpretation of our data would, however, imply that each mode uses different molecular machinery for the delivery of synaptic vesicles to the plasma membrane.

We have focused on the phosphatase-sensitive component of exocytosis and argued that increasing stimulation induces a switch in the predominant mode of exocytosis. This leaves open the question as to whether a similar switch occurs in the phosphatase-insensitive component. Chromaffin cells can show an increase in kiss-and-run from 20 to 80% of the release mechanism under conditions of high stimulation (Ales et al. 1999). As we do not know how many kiss-and-run cycles occur during a typical assay or what proportion of the vesicular styryl dye is released during a single fusion event, we cannot determine the relative proportions of full fusion and kiss-and-run involved in the phosphatase-insensitive modes of release from synaptosomes using these techniques. The mechanisms responsible for the switch between the modes of release, and the identity of the protein phosphatase-sensitive control points, remain to be determined.

Acknowledgements

This research was supported by grants from the National Health Medical Research Council of Australia, a University of Newcastle postgraduate scholarship to MLB, and infrastructure funding from NSW Department of Health through the Hunter Medical Research Institute. We are grateful to Dr Derek Laver for assistance with linear regression analysis of FM 2-10 release and we also thank Professor Peter Dunkley and Dr Martín Cammarota for helpful discussions.

Ancillary