Src Kinases Regulate De Novo Actin Polymerization during Exocytosis in Neuroendocrine Chromaffin Cells

The cortical actin network is dynamically rearranged during secretory processes. Nevertheless, it is unclear how de novo actin polymerization and the disruption of the preexisting actin network control transmitter release. Here we show that in bovine adrenal chromaffin cells, both formation of new actin filaments and disruption of the preexisting cortical actin network are induced by Ca2+ concentrations that trigger exocytosis. These two processes appear to regulate different stages of exocytosis; whereas the inhibition of actin polymerization with the N-WASP inhibitor wiskostatin restricts fusion pore expansion, thus limiting the release of transmitters, the disruption of the cortical actin network with cytochalasin D increases the amount of transmitter released per event. Further, the Src kinase inhibitor PP2, and cSrc SH2 and SH3 domains also suppress Ca2+-dependent actin polymerization, and slow down fusion pore expansion without disturbing the cortical F-actin organization. Finally, the isolated SH3 domain of c-Src prevents both the disruption of the actin network and the increase in the quantal release induced by cytochalasin D. These findings support a model where a rise in the cytosolic Ca2+ triggers actin polymerization through a mechanism that involves Src kinases. The newly formed actin filaments would speed up the expansion of the initial fusion pore, whereas the preexisting actin network might control a different step of the exocytosis process.


Introduction
The subplasmalemmal actin mesh plays a pivotal function during the secretory process in neuroendocrine cells. The first studies conducted to understand the role of cortical actin filaments (F-actin) in exocytosis proposed that it constitutes a physical barrier that restricts the secretory vesicle access to the plasma membrane. Consistent with this idea, an uninterrupted cortical actin ring is observed in non-stimulated cells, while this actin ring is disrupted upon stimuli that increase cytosolic Ca 2+ concentrations and trigger exocytosis [1]. Later findings showed that F-actin is rearranged upon cell stimulation to favor the release of transmitters [2]. Indeed, the actin nucleation promoting factor (NPF) neural Wiskott-Aldrich syndrome protein (N-WASP) is recruited upon stimulation in PC12 cells, where it seems to mediate the secretory response [3]. It was also observed that Factin forms trails that favor the secretory vesicle motion to the plasma membrane [4] and that the subplasmalemmal actin network regulates the expansion of the fusion pore [5][6], an intermediate structure formed during the exocytotic process [7], and the amount of catecholamine released per individual events [6], [8]. Nevertheless, in these latter reports [5][6], [8], the role of F-actin was evaluated using pharmacological agents that indiscriminately disrupt the new F-actin formation and the preexisting cortical actin network, making it difficult to separate the differential roles of de novo actin polymerization from the preexistent actin mesh during exocytosis.
Actin polymerization and depolymerization require the concerted action of actin regulatory proteins whose activity needs to be finely tuned by protein kinases. Among them are Src kinases, a family of non-receptor tyrosine kinases that phosphorylate and regulate the activity of NPFs such as cortactin [9][10] and N-WASP [11] and severing-factors such as cofilin [12] and gelsolin [13][14].
Src-kinase family comprises nine proteins: c-Src, c-Yes, Fyn, c-FGR, Lck, Hck, Blk, Lyn and Yrk, which share a common structure that includes: a SH1 domain with tyrosine kinase activity, a SH2 domain that binds phosphotyrosines, a SH3 domain that targets proline-rich regions in different partners, and a SH4 domain containing myristylation signals that guide Src to membranes [15][16]. The SH3 domain seems to be the region responsible for the Ca 2+ -mediated regulation of some Src kinase members [17]. In this regard, a rise in the cytosolic Ca 2+ concentration inhibits c-Yes but promotes c-Src [17][18] and Fyn [19] activation. Furthermore, Ca 2+ -induced activation of c-Src reportedly mediates Ca 2+ -dependent processes such as neurite growth [20] and neurotransmitter release [21][22], suggesting that some Src kinase members may constitute key regulatory molecules of actin polymerization during exocytosis.
In the present work, we address the hypothesis that Src kinases regulate the exocytotic release of transmitters by controlling the cortical actin dynamics by using neuroendocrine adrenal chromaffin cells (ACCs). These cells constitute a widely used model to study the exocytotic release of neurotransmitters and hormones. Moreover, they reportedly express three Src family members: c-Src, c-Yes and Fyn [19], with c-Src found in both plasma membrane and secretory vesicles, and Fyn localized mainly in the plasma membrane [19], [23].
The data presented here show that both formation of new actin filament and disruption of preexisting cortical actin are induced by Ca 2+ concentrations that trigger exocytosis in ACCs. These two processes appear to regulate different stages of the exocytosis; while the newly formed F-actin is involved in the expansion of the fusion pore, the preexisting actin seems to control a later step. In this mechanism, Src kinases appear as key regulators of the Ca 2+dependent actin polymerization and the fusion pore expansion.

Chromaffin cells culture
Bovine ACCs were isolated and cultured as previously described [24]. Briefly, bovine suprarenal glands obtained from a local slaughterhouse, were perfused with a LOCKE/glucose solution (mM: 154 NaCl, 5.6 KCl, 3.6 NaHCO3, 10 Hepes, 100 U/mL penicillin, 100 mg/ml gentamycin at pH 7.4) and then digested in the presence of 0.25% collagenase (Roche)/0.01% trypsin inhibitor (Gibco BRL) and 0.5% bovine serum albumin (BSA). After digestion, the adrenal medulla was crumbled and filtered through a surgical gauze filter. The resulting cell suspension was subjected to a percoll gradient and the ACC enriched fraction was resuspended in a 1:1 mixture of Dulbecco's modified F-12 medium (Gibco BRL) supplemented with 10% of fetal bovine serum (Gibco), and cultured at a density of 3610 5 cells/ml in 0.01% poly-L-lysine (Sigma) treated dishes. Cultured ACCs were kept at 37uC, 5% CO 2 and 95% air for at least 24 hours prior to experiments.
Cloning, expression and purification of c-Src-SH2 and c-Src-SH3 GST-fusion proteins The pcDNA3 vector containing the chicken Src cDNA was kindly provided by Dr. Ligia Toro (California University, Los Angeles, USA). The vector was linearized using HindIII and Not I enzymes and the region encoding SH2 and SH3 domains of c-Src were amplified by PCR using specific primers (forward: 59 CTT GGA TCC CAG GCT GAA GAG TGG TAC 39 and reverse: 59 GAT GAA TTC GCA GAC GTT GGT CAG 39 for SH2 domain); (forward: 59 CTT GGA TCC GGC GGC GTC ACC ACT TTC GTG 39 and reverse: 59 GCG GAA TTC GAT GGA GTC TGA GGG CGC GAC ATA 39 for SH3 domain). The resulting PCR products were digested with BamHI and EcoRI, subcloned into the glutathione-S-transferase (GST) fusion vector pGEX4T-2, expressed in Escherichia coli BL21 cells and purified with glutathione agarose beads (Life Technologies) using standard protocols.
The GST fusion protein of the N-WASP proline-rich domain (PRD) contained into the pGEX-6P1 vector was previously described [25].

Intracellular injections and pharmacological treatments
To evaluate the role of Src kinases, cells were incubated during 20 minutes at 37uC with 10 mM of the Src kinase inhibitor PP2 or its inactive analogue PP3 (Sigma Aldrich). To acutely inhibit SH2 or SH3 domain-dependent Src kinase association to different partners, cultured ACCs were injected with 5 mM of SH2 or SH3-GST fusion proteins (or GST alone), using an InjectMan system (Eppendorf) and 0.5 mm-diameter femtotips (Eppendorf) and 30 minutes later monitored by amperometry. To suppress N-WASP activation, cultured ACCs were incubated with 5 mM of wiskostatin (Calbiochem) for 10 minutes. To disrupt actin network, cells were incubated with 4 mM of cytochalasin D (Gibco) during 10 min at 37uC.

F-actin formation and actin cytoskeleton organization in chromaffin cells
The new formation of actin fibers was assessed in cultured ACCs permeabilized during 6 minutes with 20 mM digitonin in KGEP buffer (mM: 139 K-glutamate, 20 PIPES, 5 EGTA, 2 ATP-Mg 2+ , pH 6.6), in the presence of 0.3 mM Alexa Fluor 488 (or Alexa Fluor 554) G-actin conjugate (Life Technologies) and different free Ca 2+ concentrations. After permeabilization cells were fixed for 15 minutes at 4uC with 4% paraformaldehyde (PFA) and visualized in a confocal microscope (Eclipse Nikon80i; 3 excitation lasers: 408 nm, 488 nm and 543 nm).
Confocal images were analyzed and processed using the Image-J software (NIH, USA). The cortical actin ring fluorescence (for de novo formed and pre-existing actin) was measured subtracting the mean fluorescence 1 mm under the cell periphery to the total cell fluorescence intensity.

Amperometric detection of exocytosis
Amperometric recordings were performed as previously described [26] using 5 mm-diameter carbon fibers (Thornel P-55; Amoco Corp) and a patch clamp amplifier (EPC-10 USB; HEKA Electronics). The amperometric signal was low-pass filtered at 1 KHz and digitalized at 5 Hz with the acquisition software PatchMaster (HEKA Electronics). During the recording, cells were perfused with a Krebs-Hepes solution. Exocytosis was evoked by a 10 s pressure ejection of 20 mM of the Ca 2+ -ionophore ionomycin.

Spike analysis and statistics
Amperometric spikes were analyzed using the software IGOR 4 (Wavemetrics) and macros specifically designed to filter, identify and analyze individual amperometric spikes. All used macros are available at: http://webpages.ull.es/users/rborges/. Only spikes with amplitudes .10 pA were used for analysis. The criteria to analyze foot signals was restricted to foot amplitudes .3 pA and foot durations .3 ms. Each amperometric parameter was statistically analyzed by taking the median values of the events from individual cells and then averaging these values per treatment group. For amperometric data and image analysis ''n'' refers to the number of tested cells. Statistical comparisons were performed utilizing the Kruskal-Wallis test for nonparametric data or ANOVA test for parametric data. Results are expressed as mean 6 standard error of the mean (SEM).

Pull-down assays
Plated ACCs were lysed at 4uC in lysis buffer (mM:150 NaCl, 20 Hepes, 5 EDTA, pH 7.4 and 1% Triton X-100 plus protease inhibitor cocktail) and fresh extracts were used for immunoprecipitation and pull-down experiments. For the pull-down assay, 100 ml of glutathione agarose beads (Life Technologies) were incubated with 100 mg of c-Src-SH2-GST, c-Src-SH3-GST, N-WASP-PRD-GST or GST alone for 4 hours at 4uC, blocked by 30 minutes with 10% BSA and then incubated with 800 mg of ACC extract, overnight at 4uC. Beads were washed three times in lysis buffer and bound proteins eluted with SDS sample buffer for immunoblotting analyses using specific antibodies against cortactin (Abcam), c-Src (Cell Signaling) and GST (Upstate).

Ethics statement
The present work includes the use of samples from animals (bovine adrenal glands) obtained from a local slaughterhouse, Frigorific Don Pedro, certified by the Agriculture and Livestock Service of the Chilean Government (certificate number: 04.2.03.0002) and regularly inspected by a veterinarian of the Chilean Health Service. Transport, processing and elimination of the samples were carried out in strict accordance with the Article 86 of the Sanitary Regulations of the Chilean Government (Supreme decree Nu 977/96).
The protocols described in this article were approved by a Committee of Bioethics and Biosafety of the Faculty of Science, University of Valparaíso, directed by Professor Juan Carlos Espinoza, on March 7, 2011.

Ca 2+ -dependent actin polymerization and depolymerization in adrenal chromaffin cells
In order to evaluate the contribution of Src kinases to the Ca 2+dependent actin cytoskeleton dynamics in ACCs we first analyzed how different Ca 2+ concentrations affect both F-actin polymerization and depolymerization. To accurately adjust the intracellular Ca 2+ concentrations, these experiments were performed in permeabilized ACCs. We used three different Ca 2+ concentrations: 0,1 mM that corresponds to the Ca 2+ concentration at resting conditions [27], and 1 and 10 mM free Ca 2+ , which reportedly induce exocytosis in permeabilized ACCs [28].
As shown in Figure 1A a disruption of the subplasmalemmal actin ring stained with phalloidin-rhodamine B was observed at 10 mM of free Ca 2+ (see bottom panel). However, a new formation of cortical actin filaments was also observed at 1 and 10 mM of free Ca 2+ (see top panel), suggesting that actin polymerization coexists with a disassembly of the preexisting cortical actin network at Ca 2+ concentrations that induce exocytosis in ACCs. Quantification of the cortical actin ring fluorescence intensity ( Figure 1B) showed that at 10 mM Ca 2+ the de novo cortical actin assembly increased by 50% with respect to that observed at 0.1 mM (green line), while the integrity of the cortical actin ring diminished with the higher Ca 2+ concentration (red line).
To evaluate whether the newly formed F-actin associates with secretory vesicles, we performed a colocalization analysis between de novo polymerized F-actin and chromaffin granules immunola-beled with a specific antibody against the intragranular protein chromogranin A. As observed in Figure 1C, chromaffin granules colocalized in some specific sites with newly formed actin filaments. The mean score for the Pearson Coefficient between chromogranin A and F-actin at these sites was 0.8960.03 (n = 12), an indicative value of a statistically significant correlation. These data suggest a role of the new actin polymerization during the exocytosis of secretory granules in ACCs. Src kinases control the de novo cortical actin assembly Given that Src kinases are involved in both actin polymerization [9][10] and Ca 2+ -dependent transmitter release [21][22], our next step was to analyze the role of Src kinases on actin dynamics in both permeabilized and intact AACs. Permeabilized cells allow us to adjust the Ca 2+ concentration, introduce fluorescent-tagged-Gactin, and separately evaluate the de novo F-actin formation from the F-actin network organization. In this condition, we did not observe a loss of Src kinases as compared with cortactin, a protein of a higher molecular weight ( Figure S1), but we can not discard the leakage of smaller molecules involved in actin polymerization/ depolymerization dynamics. On the other hand, intact cells allow us to visualize the general organization of the cortical F-actin network under resting or stimulated conditions. As observed in Figure 2A, 10 mM of PP2 significantly reduced the new F-actin formation in cells permeabilized in the presence of 10 mM free Ca 2+ , as compared with the treatment with its inactive isomer PP3 (Figure 2A-B). However, it had no effects on the Ca 2+induced disruption of the cortical F-actin stained with phalloidinrhodamine B ( Figure 2C-D).
We also analyzed the effects of the Src kinase inhibitor PP2 on the integrity of the cortical actin network in intact cells maintained in resting conditions or stimulated with 20 mM of the Ca 2+ ionophore ionomycin during 10 s. Reportedly the cortical F-actin ring in chromaffin cells is visualized as an uninterrupted ring in resting conditions, but in the presence of a stimulus that increased cytosolic Ca 2+ concentrations, such as ionomycin this F-actin ring is disrupted [5]. As shown in Figure 2E-H, this change in the cortical F-actin organization was also observed in cells treated with PP3 (10 mM for 20 min). PP2 did also not affect the cortical Factin network under either resting or stimulated conditions. Given that in ACCs SH2 and SH3 domains of Src-kinases mediate Src association to partners required for actin polymerization such as N-WASP and cortactin (see Figure S2), we next evaluated the impact of an acute disruption of these associations on the formation of F-actin. ACCs were permeabilized in the presence of 10 mM free Ca 2+ and growing concentrations of chicken c-Src SH2-or SH3-GST fusion proteins. Amino acid sequences of chicken and bovine c-Src SH2 or SH3 domains are 100% identical, while sequence homology between chicken c-Src SH2 or SH3 domain and the respective domain of bovine Fyn or c-Yes are between 71 and 79% identical (see Table S1).
As shown in Figures 3A-B, 0.1 mM of the c-Src SH2 or SH3 domain was able to significantly disrupt the new formation of actin filaments as compared with 5 mM GST alone. At such concentration, the inhibitory effect of the SH3 domain was larger than that of the SH2 domain (59.3% of inhibition with the SH3 domain v/s 28.5% of inhibition with the SH2 domain). These data indicate that Src kinase associations via SH2 or SH3 domain play an important role in promoting actin polymerization in ACCs. Src kinases regulate fusion pore expansion and the kinetics of exocytotic events in chromaffin cells Since cortical actin actively participates during exocytosis adjusting the characteristics of the release events in neuroendocrine cells [5][6], [8] we investigated how Src kinase function influences exocytosis. The release of catecholamines was measured using amperometry with carbon fiber microelectrodes, a technique that has a good time resolution, and therefore yields information of the different steps of exocytosis [6,24]. Amperometric spike parameters such as spike amplitude (Imax), spike charge (Q), spike half-width (t 1/2 ) and rise time (tP) were analyzed. All these parameters provide information about the characteristics of the exocytotic events [5]. The number of exocytotic events and the current that precedes the spike, usually called foot, were also analyzed. Given that Src kinases influence both nicotinic receptor [29] and voltage-dependent Ca 2+ channel activity [30], we induced exocytosis with the Ca 2+ ionophore ionomycin (20 mM during 10 s). Src kinase activation was disrupted by the incubation of ACCs with 10 mM PP2 for 10 min before the stimulus. Figure 4A shows a typical amperometric recording in a nontreated control cell. In these cells a 10 s pulse of 20 mM ionomycin induced 65.067.1 amperometric spikes in 100 s (n = 35), in which average values of Imax, Q, t 1/2 and tP were 98.765.0 pA, 1.760.1 pC, 12.960.9 ms and 5.260.3 ms, respectively. With the exception of the Imax that was reduced by 20%, the treatment with 10 mM PP3 did not modify other amperometric parameters (Figures 4 and 5). Conversely, Src kinase inhibition with 10 mM PP2 drastically changed the characteristics of the amperometric spikes ( Figures 5B-C). Indeed, as compared with PP3, PP2 significantly reduced Imax and Q by 45% and 40% respectively, indicating that Src kinases control the amount of catecholamines released during individual events. We also found out that PP2 dramatically increased t 1/2 and tP ( Figure 5). Given that t 1/2 reflects the duration of the exocytotic events and tP has been related with the kinetics of the fusion pore expansion [31][32], it is possible that the inhibition of Src kinases delays fusion pore expansion, giving rise to events with slow kinetics, and incomplete release. Therefore, we next analyzed the amperometric foot, which reflects the release of catecholamines through the initial fusion pore [33]. Here we observed that PP2 significantly increased the frequency of foot events, and prolonged the foot duration ( Figure 5C), which reflects the stability of the fusion pore [7]. Conversely, PP2 did not influence the foot amplitude (Table  S2) that correlates with the conductance of the fusion pore [34]. Thus, these latter findings confirm that Src kinases control fusion pore expansion. On the other hand, PP2 also significantly reduced the number of exocytotic events ( Figure 4B). As we will discuss later this effect could be a consequence of the role of c-Src on the secretory vesicle trafficking [35]. Table S2 shows the values of the amperometric parameters in these latter conditions. Src kinases usually bind to their partners via SH3 or SH2 domains [36]. Therefore, we studied the effects of the intracellular injection of the chicken c-Src SH2 or SH3 domain fused to GST on the characteristics of the exocytotic events. These fusion proteins were injected 30 min before the stimulus at a concentration of 5 mM, and their effects were compared with the injection of GST alone. As compared with non-injected cells (control), the injection of GST alone reduced Imax by 20% and the frequency of foot signals by 48% without affecting any other amperometric parameters (Table S2). Conversely, the injection of the c-Src SH2 or SH3 domain reduced Imax values by 55% and increased over twice t 1/2 and tP, as compared with non-injected or GST-injected cells ( Figure 5C). Also in comparison with GST, neither c-Src SH2 nor SH3 affected the number of exocytotic events ( Figure 4B), but they increased the frequency of foot signals ( Figure 5C). On the other hand, the injection of the c-Src-SH3, but not of the cSrc-SH2, significantly reduced Q and prolonged foot duration ( Figure 5C). Table S2 shows the values of the amperometric parameters in these latter conditions.
In order to determine whether c-Src-SH3 regulates exocytosis through an actin-mediated pathway, we analyzed the effects of the injection of c-Src-SH3 in the presence of the actin-disrupting drug Cytochalasine D (CytoD). As shown in Figure 6, in cells injected with GST alone, CytoD significantly increased tP, foot duration and foot frequency. These same effects were observed in cells injected with cSrc-SH3 and treated with DMSO or CytoD. Regarding the fusion pore conductance, neither the injection of cSrc-SH3 nor the incubation with CytoD affected the foot amplitude; however, the combination of both treatments significantly decreased this foot parameter. On the other hand, contrary to what we observed with the injection of cSrc-SH3, CytoD significantly increased Imax and Q (Fig. 6), but such effects were not observed in cells injected with cSrc-SH3 and treated with CytoD. In the latter condition Imax importantly decreased, while Q was not affected. Table S3 shows the values of the amperometric parameters in these latter conditions.
The disrupting effects of CytoD on the formation of new actin filaments and on the organization of the actin network in intact cells are shown in Figure 7. Similar to that observed with Src kinase inhibition, CytoD suppressed the new F-actin formation ( Figure 7A-B). However, contrary to what we observed with Src inhibition, CytoD disrupted the cortical actin network in resting conditions ( Figure 7C-D) and induced a further F-actin disruption in ionomycin-stimulated cells ( Figure 7E-F). Therefore, we analyzed how the combined treatment of CytoD plus cSrc-SH3 affects the cortical actin network in the presence of high Ca 2+ concentrations. In order to adjust cSrc-SH3 concentrations, these experiments were performed in cells permeabilized in the presence of 10 mM free Ca 2+ . In this condition we observed that cSrc-SH3 at both concentrations (0.1 and 5 mM), significantly diminished the F-actin disruption observed in the presence of CytoD (Figure 8).

Inhibition of N-WASP suppresses cortical actin polymerization and reduces the quantal release
Given that N-WASP contributes to the formation of actin filaments during exocytosis in chromaffin cells [3], we evaluated the effect of wiskostatin (Wsk), an agent that prevents N-WASP activation by stabilizing its autoinhibited conformation [37]. As shown in Figure 9A  We also evaluated the effects of 5 mM Wsk on the characteristics of exocytosis and found out that Wsk significantly reduced the number of individual exocytotic events ( Figure 9H) and reduced Imax and Q by almost 70% (Figure 9I-J). However, t 1/2 and tP values were not affected by the treatment with the N-WASP inhibitor (Table S4). As shown in Table S4, only 26.860.5 of spikes exhibited foot signals, which had significantly smaller amplitude, as compared with cells treated with the vehicle DMSO (Table S4).

Discussion
The cortical actin network constitutes a dynamic structure that controls and directs the motion of secretory vesicles to the plasma membrane [38]. As demonstrated here, both actin polymerization and depolymerization are induced by Ca 2+ in ACCs (Figure 1). Pioneering studies of Trifaró and collaborators [1] showed that high cytosolic Ca 2+ levels activate the actin filament-severing protein scinderin, inducing disassembly of the cortical actin network and allowing secretory vesicles to reach the plasma  16) or injected with SH3 treated with DMSO (n = 24) or CytoD (n = 14). All amperometric parameter values correspond to the median values of the events from individual cells, which were subsequently averaged per treatment group. *p,0.05 compared with cells injected with GST and treated with DMSO. Note that SH3 and CytoD have common effects in tP, foot duration and foot frequency, but show dissimilar effects on Imax and Q. In the latter parameters, the effects of c-Src-SH3 prevail over those of CytoD. doi:10.1371/journal.pone.0099001.g006 membrane [39]. As later reported, a localized F-actin assembly at the exocytotic sites appears to take place in response to stimulation of neuroendocrine cells [3]. However, how a stimulation-induced increase in the cytosolic Ca 2+ levels activates F-actin assembly is still unclear. As it is discussed here, one possibility is that Src kinases orchestrate this Ca 2+ -dependent actin polymerization and thereby modulate different stages of exocytosis.
In the present paper, we found that in ACCs Src kinases regulate actin polymerization ( Figure 2) and the fusion pore expansion ( Figure 5). These Src kinase actions seem to be dependent on Src kinase catalytic activity as well as on the associations to other proteins through its SH2 and SH3 domains (Figure 3-5). As shown here, these two domains are also involved in the association of c-Src to the NPFs N-WASP and cortactin in ACCs ( Figures S2). Both N-WASP and cortactin are known Src kinase substrates [11], [40]. Furthermore, Src kinases regulate N-WASP/cortactin interaction [25], [41], which acts synergistically during actin polymerization [42][43][44].
Probably as a consequence of its actions on actin polymerization, Src kinases regulate the fusion pore expansion. In fact, the inhibition of Src kinase activity with PP2, or the disruption of its association to other proteins through its SH3 or SH2 domain has some effects that are similar to those produced by the actindisrupting drug CytoD. Both treatments inhibited de novo actin polymerization (Figures 2, 3 and 7), and slowed down fusion pore expansion: they increased tP, foot duration and foot frequency ( Fig. 5 and 6; see also [6]). However, differently to that observed with CytoD, Src kinase inhibition did not disturb the cortical Factin organization under resting or stimulated conditions (compare Figure 2E-H with Figures 7C-F), neither increased the amperometric parameters Imax and Q (Figures 5 and 6). These findings suggest to us that the de novo actin polymerization and the preexisting F-actin network regulate different steps of exocytosis; while the new F-actin formation appears to promote the expansion of the initial fusion pore, the preexisting F-actin would control a later step of the exocytosis. Corey Smith and collaborators [8] proposed that the disruption of the cortical F-actin network (observed in cells subjected to a strong stimulus or in the presence of CytoD) favors the full collapse of the secretory vesicles into the plasma membrane, while an uninterrupted actin network (that is maintained in cells subjected to low frequency stimulation) would prevent the vesicle collapse and thereby favoring a kiss-and-run mechanism [8]. Since CytoD does not affect the vesicle volume in ACCs [5,6], Smith's hypothesis could explain the increase in the quantal size induced by CytoD observed here ( Figure 6) and also in our previous work [8]. This idea is depicted in the Figure 10C.
The combined treatment of CytoD with the cSrc-SH3 domain also supports the idea that cSrc kinase and the de novo actin polymerization regulate the expansion of the initial fusion pore, and not later steps of exocytosis. As shown in Figures 6 and 9, the cSrc-SH3 domain diminished the disruption of the cortical actin network induced by CytoD, and prevented the increase of Imax and Q also induced by this actin-disrupting agent. How cSrc-SH3 domain prevents the actin disruption is not clear to us at this moment. One possibility is that this action is mediated by the GTPase dynamin. In fact, dynamin is a Src substrate involved in regulating actin dynamics in several cellular processes [45][46]. Furthermore, Src kinase regulates the association of dynamin to cortactin [47] and the complex dynamin/cortactin stabilizes Factin bundles [48]. In ACCs, c-Src binds dynamin via its SH3 domain (see Figure S2C) and as we previously demonstrated dynamin regulates the organization of the cortical actin [6].
In order to understand better the role of the new F-actin formation on the dynamics of the fusion pore, we analyzed the effects of Wsk on the cortical actin dynamics and the amperometric spikes. As observed with CytoD, this N-WASP inhibitor suppressed the Ca 2+ -dependent actin polymerization ( Figure 9A-B). However, it slightly disturbed the actin network in resting condition ( Figure 9C-D), and did not affect the ionomycininduced cortical F-actin disruption ( Figure 9E-F). Interestingly, the effects of Wsk on the exocytosis were totally different to those of CytoD: Wsk significantly reduced Imax and Q. Similar effects were observed by Corey Smith and collaborators using Wsk during high frequency stimulation [49]. We interpret these effects as an inhibition of the fusion pore expansion that gives rise to ''stand-alone foot'' events [34], [50][51]. In this type of event only a small fraction of the intravesicular catecholamines is released [34], [51].
Based on the similar effects of Wsk and PP2, as well as the effects of cSrc-SH3, on the spike amplitude and quantal size (compare Figure 5 and Figure 9), we propose that N-WASP and Src kinases operate through a common mechanism in exocytosis, which relies on new actin polymerization (see our model in Figure 10). This mechanism seems to be perturbed differently by Wsk or PP2. While Wsk appears to prevent the expansion of the initial fusion pore, giving rise to ''stand-alone foot'' events, PP2 only decelerates the pore expansion, generating amperometric events with a long foot signal, slow rise time and small quantal size. They affect the fusion pore dynamics differently, probably because N-WASP is directly involved in actin nucleation, while Src kinases regulate actin dynamics by acting on different substrates and modulating their associations. Besides this, Wsk reportedly perturbs membrane transport, decreasing cellular ATP levels [52] and Ca 2+ influx [49]. These actions of Wsk could further contribute to the failure of the fusion pore expansion, and also to the dramatic reduction in the number of exocytotic events. Indeed, these two types of processes are highly dependent on Ca 2+ levels [53].
However, it must be taken into account that the fusion pore expansion is not the only step of the secretory pathway that could be affected by the perturbation of the actin dynamics. Since actin dynamics is critical in maintaining a proper Golgi-complex architecture and function [54], we cannot rule out that an actindisrupting treatment influences early stages of the secretory pathway and thus modifies vesicle size, vesicle content and/or number of exocytotic events.  Regarding the mechanism by which the different treatments affect the amount of exocytotic events, our data show that Wsk and PP2 decreased the number of amperometric spikes by 52% and 38%, respectively, while CytoD increased it by 38% (see Tables S2, S3 and S4). As aforementioned, the preexisting cortical actin network constitutes a barrier that limits the access of secretory vesicles to plasma membrane [1], therefore its disruption with CytoD increases the number of exocytotic events. On the other hand, F-actin appears to form trails that direct the motion of vesicles to the exocytotic sites [4]. Src kinase function could modulate this last process since it phosphorylates the small insert isoform of myosin VI making it able to bind secretory vesicles in a Ca 2+ -dependent manner, and then recruit them to the cortical actin network [35].
Therefore, the preexisting cortical actin network and the newly formed F-actin appear to regulate differently the access of secretory vesicles to the plasma membrane, the expansion of the fusion pore and the mode of exocytosis. Both the disruption of the preexisting actin and the formation of new actin filaments rely on cytosolic Ca 2+ levels, as well as on the concerted action of proteins, among them Src kinases that play a pivotal role in regulating Ca 2+dependent actin polymerization. Figure 10 depicts our proposed model of how this mechanism works. Table S2 Amperometric parameters of exocytotic events induced by 20 mM ionomycin in cells treated with the Src kinase inhibitor PP2 or its inactive analogue PP3, or injected with c-Src SH2-GST, c-Src SH3-GST or GST alone. Exocytosis was induced with 20 mM ionomycin and monitored by amperometry. Cells were incubated with 20 mM PP2 or its inactive analog PP3 20 min before the exocytosis induction. These agents were kept during the recording. GST, c-Src SH2-GST or c-Src SH3-GST were injected 30 min before cell stimulation. Control corresponds to non-treated cells. Data are means 6 SEM of averages. &p,0.05 compared with control; *p, 0.05 compared with cells treated with PP3; {p,0.05 compared with cells injected with GST. (DOC)

Supporting Information
Table S3 Amperometric parameters of exocytotic events cells injected with c-Src SH3-GST and treated with cytochalasin D. Exocytosis was induced with 20 mM ionomycin and monitored by amperometry. Cells injected with 5 mM GST or c-Src SH3-GST (SH3) were incubated with 4 mM cytochalasin D (CytoD) or its vehicle DMSO during 10 min at Figure 10. Influence of the different treatments on actin dynamics. (A) A stimulus that induces an increase in the cytosolic Ca 2+ hastens SNARE-mediated fusion of secretory vesicles with the plasma membrane, but it also promotes both disruption of the preexisting actin network (green rosary beads) and the formation of new actin filaments (yellow rosary beads). This actin dynamics appears to favor the expansion of the fusion pore, but also prevents the collapse of the vesicle in the plasma membrane. This mechanism allows the fast release of soluble catecholamines (see the purple spike at the right). (B) The N-WASP inhibitor wiskostatin (Wks) strongly inhibits the new actin polymerization, while it slightly disturbs the preexisting cortical F-actin network ( Figure 9). Such effect could be due to a reduction of the slowrate of actin polymerization/depolymerization in resting conditions. Wsk also gives rise to incomplete release events (see the small green spike), suggesting that the lack of actin polymerization hinders the expansion of the fusion pore. Wsk could also perturb membrane transport, decreasing cellular ATP levels and Ca 2+ entry [52]. These effects could also contribute to the failure of the fusion pore expansion. (C) Cytochalasin D (CytoD) greatly disrupts the preexisting F-actin network, as well as the new F-actin formation (Figure 7). CytoD also slows down the fusion pore expansion and increases the quantal size of the release events (see the big brown spike with a long foot). Probably the loss of actin meshwork favors the collapse of the vesicle in the plasma membrane, as previously proposed by Doreian et al. [8]. (D) The inhibition of Src kinases with PP2 suppresses the de novo F-actin formation, but it does not disturb the pre-existing cortical actin network (Fig. 2). PP2 also slows down the fusion pore expansion, and reduces the quantal size of the release events (note the small blue spike with a long foot). Similar effects were observed with the microinjections of cSrc-SH2 or -SH3 domains. These findings support the idea that the new F-actin formation favors the expansion of the fusion pore.