RalA-exocyst interaction mediates GTP-dependent exocytosis.

Many secretory cells utilize a GTP-dependent pathway, in addition to the well characterized Ca2+-dependent pathway, to trigger exocytotic secretion. However, little is currently known about the mechanism by which this may occur. Here we show the key signaling pathway that mediates GTP-dependent exocytosis. Incubation of permeabilized PC12 cells with soluble RalA GTPase, but not RhoA or Rab3A GTPases, strongly inhibited GTP-dependent exocytosis. A Ral-binding fragment from Sec5, a component of the exocyst complex, showed a similar inhibition. Point mutations in both RalA (RalA(E38R)) and the Sec5 (Sec5(T11A)) fragment, which abolish RalA-Sec5 interaction also abolished the inhibition of GTP-dependent exocytosis. Moreover, transfection with wild-type RalA, but not RalA(E38R), enhanced GTP-dependent exocytosis. In contrast the RalA and the Sec5 fragment showed no inhibition of Ca2+-dependent exocytosis, but cleavage of a SNARE (soluble N-ethylmaleimide-sensitive factor attachment protein receptor) protein by Botulinum neurotoxin blocked both GTP- and Ca2+-dependent exocytosis. Our results indicate that the interaction between RalA and the exocyst complex (containing Sec5) is essential for GTP-dependent exocytosis. Furthermore, GTP- and Ca2+-dependent exocytosis use different sensors and effectors for triggering exocytosis whereas their final fusion steps are both SNARE-dependent.

monomeric GTPases, have been suggested as regulators of exocytosis. Rho family proteins have been proposed to regulate exocytosis in mast cells (10), because Rho GDP dissociation inhibitor (RhoGDI) (11) has been shown to inhibit GTP␥Sinduced exocytosis in mast cells (12). RhoGDI is known to inhibit GDP/GTP exchange in Rho family proteins including Rho, Cdc42, and Rac (11,13,14). The Rab family of proteins, particularly Rab3A, has long been a focus of exocytosis studies (15)(16)(17). Rab3A, a neuron-specific isoform localized on the secretory vesicles (18), is considered the mammalian orthologue of Sec4p GTPase in yeast (19). Sec4p is essential for secretion in yeast and the exocyst complex consisting of eight proteins (Sec3p, Sec5p, Sec6p, Sec8p, Sec10p, Sec15p, Exo70, Exo84) (20,21) was identified as an effector for Sec4p (22). In mammalian cells including neurons, a similar protein complex designated as the mammalian exocyst complex (also called Sec6-Sec8 complex), was identified (23). However, the mammalian exocyst complex is not able to interact with Rab3A, but rather interacts with RalA and RalB (24 -26), suggesting that Ral GTPase may regulate exocytosis. RalA and RalB physically bind to Sec5, a component of the exocyst complex, in a GTPdependent manner (25)(26)(27). In addition, Ral is localized on dense granules in platelets (28) and synaptic vesicles in neurons (29), supporting a possible function of Ral in exocytosis in these cells.
In yeast, exocyst components are essential for secretion (20,21). Functional analysis of exocyst proteins in higher organisms including Drosophila has only been recently initiated (30,31). The mammalian exocyst is localized at the plasma membrane of nerve terminals in neurons (23). In differentiated PC12 cells, the exocyst was found in a punctate distribution at terminals of cell processes, at or near sites of granule exocytosis (32). These results suggest the involvement of the mammalian exocyst in exocytotic processes, although its exact function in exocytosis is not yet understood.
In this study, we sought to identify a key GTPase and its effector(s) that mediate GTP-dependent exocytosis using neuroendocrine PC12 cells. Permeabilized PC12 cells have proven to be a powerful model for the molecular dissection of exocytosis (33)(34)(35)(36)(37)(38)(39)(40)(41). It was previously shown that permeabilized, cytosoldepleted PC12 cells are capable of secretion in response to nonhydrolyzable GTP analogues (42,43). Here we show that RalA-exocyst is the key sensor-effector for GTP-dependent exocytosis in PC12 cells. Because PC12 cells show robust Ca 2ϩ -dependent secretion, we also sought to explain the relationship between GTP-dependent exocytosis and Ca 2ϩ -dependent exocytosis. We suggest that GTP-dependent exocytosis and Ca 2ϩdependent exocytosis use different sensors and effectors for triggering exocytosis.

EXPERIMENTAL PROCEDURES
PC12 Secretion Assay-The secretion assay followed the protocols of previously published work (33,36,37,43 Amersham Biosciences) in the presence of 0.5 mM ascorbic acid. After washing, the cells were harvested in KGlu buffer (20 mM HEPES, pH 7.2, 120 mM potassium glutamate, 20 mM potassium acetate, and 2 mM EGTA) with 0.1% bovine serum albumin (BSA), permeabilized with a ball homogenizer (33,36) or with freeze-and-thaw (37,43), and incubated for 1-3 h on ice in the presence of 10 mM EGTA to extract the cytosolic proteins. (Similar results were obtained from both methods of permeabilization, as previously shown, Ref. 43.) Permeabilized PC12 cells were washed three times with KGlu buffer containing 0.1% BSA. GTP-dependent NE secretion assays were done in KGlu buffer with 0.08% BSA containing various concentrations of GppNHp (trisodium salt, Sigma) (typically 100 M, unless otherwise indicated) and recombinant proteins for 20 -25 min at 30°C. Ca 2ϩ -dependent secretion assays were performed in KGlu buffer with 0.08% BSA, 2 mM MgATP, 1 mg/ml brain cytosol, and 1.72 mM CaCl 2 for 20 min at 30°C. 1.72 mM CaCl 2 in KGlu buffer, which contains 2 mM EGTA, produced ϳ1-3 M [Ca 2ϩ ] free (36). Secretion was terminated by chilling to 0°C, and samples were centrifuged at 4°C for 3 min. Supernatants were removed, and the pellets were solubilized in 1% Triton X-100 for liquid scintillation counting. In the Botulinum neurotoxin type E (BoNT/E)-LC experiments, permeabilized PC12 cells were washed once with KGlu buffer with 0.1% BSA, incubated with 100 nM BoNT/E-LC for 8 min at 30°C (36) and then washed twice with KGlu buffer with 0.1% BSA before inducing secretion. Concentrations of free Mg 2ϩ in the presence of EDTA were calculated based on Ref. 44, using a program written by Dr. Jochen Kleinschmidt (New York University) (45).
hGH Secretion Assay using Co-transfected PC12 Cells-PC12 cells at 70 -80% confluency in 10-cm dishes were transfected with 14.4 g of plasmid DNA and 36 l of FuGENE 6 (Roche Applied Science) (46,47). Seventy-two hours post-transfection, cells were harvested with 6 ml of KGlu buffer containing 0.1% BSA and permeabilized by freezing and thawing (37,38). Permeabilized cells were stimulated with the indicated concentrations of GppNHp for 25 min at 30°C. The amounts of hGH secreted into the medium and retained in the cells were measured by radioimmunoassay (Medicorp).
Construction of Expression Plasmids-All the expression plasmids for GST fusion proteins were generated using the parental plasmid pGex-KG (37,38). Full-length RalA, RhoA, Rab3A, RhoGDI, and the fragment (residues 2-120) of Sec5 were synthesized by PCR from a rat brain cDNA library. The point mutation in Sec5 (Sec5 T11A ) was introduced by the QuikChange TM site-directed mutagenesis kit (Stratagene), and the point mutation in RalA was introduced by PCR. For transfection, a 0.8-kb BamHI-HindIII fragment of pGex-RalA WT and pGex-RalA E38R was subcloned into pCMV9E10myc-1 generating pCMV9E10myc-RalA WT and pCMV9E10myc-RalA E38R , respectively. The hGH expression vector was described previously (46). The expression plasmids obtained by PCR were fully sequenced and verified.
GST Pull-down Experiments with Recombinant RalA Proteins-One frozen rat brain was homogenized with 10 ml of homogenization buffer (10 mM HEPES-NaOH, pH 7.4, 320 mM sucrose) and centrifuged at 800 ϫ g for 10 min. The supernatant was centrifuged at 12,000 ϫ g for 20 min, and the pellet was resuspended with 5 ml of KGlu buffer containing 0.3% Triton X-100 and 1 mM MgCl 2 . The resuspended material was subsequently centrifuged at 100,000 ϫ g for 30 min, and the supernatant (brain homogenate) was used for binding. For loading guanine nucleotides, GSH agarose containing ϳ20 g of GST-RalA WT or GST-RalA E38R was incubated with KGlu buffer containing 1 mM EDTA and 200 M GDP or GppNHp for 30 min at 30°C. The buffer was adjusted to 5 mM MgCl 2 at the end of reaction. The GSH-agarose was recovered by centrifugation and was incubated with 1 ml of brain homogenate at 4°C overnight and washed three times with KGlu buffer containing 0.3% Triton X-100 and 1 mM MgCl 2 . Samples were subjected to 10% SDS-PAGE and transferred to nitrocellulose membrane. Sec6 and Sec8 were probed with monoclonal antibodies (StressGen) and detected by chemiluminescence (Renaissance, PerkinElmer Life Sciences).
GST Pull-down Experiment with Recombinant Sec5 Fragments-Brain homogenate was prepared as above except that the pellet was resuspended with 5 ml of KGlu buffer containing 0.3% Triton X-100 and 1 mM EDTA. Brain homogenate was incubated with 200 M GDP or GppNHp for 30 min at 30°C and was adjusted to 3 mM MgCl 2 at the end of the reaction. GSH-agarose containing ϳ20 g of GST-Sec5-(2-120) WT or GST-Sec5-(2-120) T11A was incubated with 1 ml of brain homogenate loaded with either GDP␤S or GppNHp at 4°C overnight and was washed five times with KGlu buffer containing 0.3% Triton X-100 and 1 mM MgCl 2 . The presence of RalA and Rab3A was determined by Western blotting using monoclonal antibodies (for anti-RalA, purchase from BD Biosciences, for anti-Rab3A, Cl.42.2, a kind gift from Dr. Reinhard Jahn).
Immunocytochemistry and Confocal Laser-scanning Microscopy-PC12 cells transfected with pCMV9E10myc-RalA WT or pCMV9E10myc-RalA E38R were grown on polylysine-coated glass coverslips, washed with phosphate-buffered saline (PBS), fixed for 15 min with PBS containing 4% paraformaldehyde and permeabilized with PBS containing 0.2% Triton X-100 and 0.3% bovine serum albumin for 5 min. Nonspecific sites were blocked for 1 h at room temperature in PBS containing 0.3% bovine serum albumin. Primary antibodies against Myc (mouse 9E10 ascites diluted 1:1000, Covance) and secretogranin II (rabbit polyclonal antiserum diluted 1:1000, GED Bioscience) were diluted in blocking buffer and applied to the permeabilized cells for 1 h at room temperature. Following three washes in blocking buffer, goat fluorescein-conjugated anti mouse (diluted 1:1000) and goat rhodamine red-xconjugated anti-rabbit antibodies (diluted 1:3000) (Jackson Immuno-Research Laboratories) were diluted in blocking buffer and applied for 1 h at room temperature. Samples were washed three times in blocking buffer and mounted in FluorSave TM reagent (Calbiochem). Immunofluorescence staining was monitored with a Zeiss laser confocal scanning microscope (LSM 410) with a water immersion objective lens (ϫ63).

RESULTS
Pharmacological Characterization of GppNHp-induced Exocytosis-We first confirmed a previous finding that GppNHp together with Mg 2ϩ induces NE secretion from permeabilized PC12 cells (43). As previously reported (43), the dose-response curve was bell-shaped, and with high concentrations of Mg-GppNHp, NE release was suppressed (Fig. 1A). To clarify the relative role of Mg 2ϩ for this suppression, we tested GppNHpinduced NE release in the absence of Mg 2ϩ . Low concentrations (0.1-10 M) of GppNHp and Mg-GppNHp induced similar levels of exocytosis. However, at high concentrations (100 M to 1 mM), GppNHp stimulated increased levels of exocytosis relative to Mg-GppNHp (Fig. 1A). This result suggests that suppression at high doses is primarily because of high concentrations of Mg 2ϩ . To further validate our findings we also examined the effects of Mg 2ϩ concentrations on GppNHp-induced exocytosis by varying the concentration of free Mg 2ϩ ([Mg 2ϩ ] free ) independent of GppNHp by using a Mg 2ϩ /EDTA buffer. When Mg 2ϩ was chelated by 1.8 mM EDTA (0 M ([Mg 2ϩ ] free ) GppNHpinduced secretion was completely abolished, suggesting an essential role for Mg 2ϩ in GTP-dependent exocytosis. In contrast 0.7 M [Mg 2ϩ ] free rescued GppNHp-induced exocytosis (Fig.  1B). However, at 6 -45 M [Mg 2ϩ ] free , GppNHp-induced exocytosis was less efficient than at 0.7 M [Mg 2ϩ ] free (Fig. 1B). Thus, we conclude that while submicromolar to micromolar [Mg 2ϩ ] free is required for GppNHp-induced exocytosis, more than 10 M [Mg 2ϩ ] free inhibits GppNHp-induced exocytosis.
The complex action of Mg 2ϩ on GppNHp-induced exocytosis is reminiscent of the interaction between Mg 2ϩ and GTPases. Thus, we hypothesized that GppNHp-induced exocytosis is mediated by activated GTPases and that the inhibition of exocytosis by high concentrations of Mg 2ϩ is due to the stabilizing effects of Mg 2ϩ on the nucleotide-bound form of GTPases (48,49). After permeabilization and washes, GTPases in PC12 cells are mostly in a GDP-bound form and Mg 2ϩ prevents GppNHp from binding to the target GTPase by drastically decreasing the off-rate of GDP (48,49). Therefore, we predicted that GppNHpinduced exocytosis would be inhibited by increased concentrations of GDP. Indeed, GDP inhibited GppNHp (100 M)-induced exocytosis in a concentration-dependent manner (Fig.  1C), supporting a role for activated GTPases in GppNHpinduced exocytosis.
Inhibition of GTP-dependent Exocytosis by Soluble RalA, but Not RhoA or Rab3A-We sought to identify the GTPase involved in GppNHp-induced exocytosis. Three small GTPases were considered as possible candidates: RhoA, Rab3A, and RalA. We hypothesized that if one of these GTPase proteins was responsible for GTP-dependent exocytosis, it would inhibit exocytosis by functioning as a dominant negative protein. This is a realistic possibility because the recombinant proteins generated in E. coli lack prenylation for proper membrane localization. On incubation of permeabilized PC12 cells with soluble recombinant proteins, we found that the strong inhibition was mediated by recombinant RalA, but not by RhoA or Rab3A (Fig.  2B). This effect was not caused by variable purities among GST fusion proteins as determined by SDS-PAGE and Coomassie Blue staining ( Fig. 2A). Strong inhibition by RalA was not caused by competition for binding of GppNHp to native GT-Pases (buffering effect) because the concentration of recombinant proteins used in Fig. 2B was ϳ1 M, 100 times lower than the concentrations of GppNHp (100 M). A dose-response curve indicated that the inhibition by RalA became evident at concentrations of 0.1 M (Fig. 2C). Furthermore, higher concentrations of GppNHp (e.g. 1 mM) did not reverse the inhibition by RalA (Fig. 2D), thus further arguing against a potential buffering of available GppNHp. Therefore, we reasoned that recombinant RalA inhibits exocytosis as a dominant negative by binding to the target of endogenous GTPase(s).
A Ral Binding Fragment from Sec5, but Not RhoGDI, Inhibits GTP-dependent Exocytosis-The exocyst has been recently identified as an effector complex for Ral (24 -26). Ral physically binds to Sec5, one of the eight proteins comprising the exocyst complex, in a GTP-dependent manner. We hypothesized that if the exocyst functions as the effector for Ral in GTP-dependent exocytosis, a fragment of Sec5 (residues 2-120) that binds to Ral (25)(26)(27) should inhibit exocytosis by preventing endogenous Ral from binding to the exocyst complex. Indeed we found that this Sec5 fragment inhibited GTP-dependent exocytosis in a concentration-dependent manner (Fig. 3, A and B). We also examined the effects of RhoGDI (11), which inhibits signaling pathways mediated by Rho family proteins, including Rho, Rac, and Cdc42 (11,13,14), on exocytosis. If Rho family proteins are critical for GTP-dependent exocytosis, it is expected that RhoGDI will inhibit the exocytosis. In our preparation, we found no inhibition by RhoGDI (Fig. 3A). Thus, the RalAexocyst complex, which contains Sec5, and not the Rho signal-ing pathway, is important for GTP-dependent exocytosis. However, maximal inhibition by soluble RalA and the Sec5 fragment was ϳ70% of GTP-dependent exocytosis. Thus our results do not exclude the possibility that another GTPase may mediate GppNHp-induced exocytosis. Point Mutations in RalA (RalA E38R ) and the Sec5 (Sec5 T11A ) Fragment, Which Abolish RalA-Sec5 Interaction, also Abolish the Inhibition of GTP-dependent Exocytosis-We next examined the specificity of the inhibition by soluble RalA protein and the Sec5 fragment. A recent structural study of the RalA-Sec5 interaction identified potentially important residues for the interaction (27). Point mutants for these residues (RalA E38R , Sec5 T11A ) abolished the interaction between these two recombinant proteins, as measured by isothermal titration calorimetry (27). We expressed both mutant RalA E38R and wild-type RalA as GST fusion proteins and showed that the purity of these two fusion proteins was similar (Fig. 4A). Then, we examined recombinant RalA binding to the exocyst complex in brain homogenates using a GST pull-down assay in which the presence of the exocyst was examined with antibodies against Sec6 and Sec8, two major components of the exocyst complex. We found that wild-type RalA bound to the exocyst in a GTPdependent manner, whereas binding of RalA E38R to the exocyst was at background levels in the presence or absence of GppNHp or GDP (Fig. 4B). Thus, the mutated residue in RalA E38R is sufficient to abolish the interactions with the native exocyst. Similarly, we expressed a mutant Sec5 fragment (Sec5 T11A ) (residues 2-120) as a GST fusion protein (Fig. 4A) and observed that this mutation abolished the GTP-dependent interaction between Sec5 and native RalA in the brain (Fig. 4C,  upper panel). As a control we demonstrated that neither the wild-type nor the mutant Sec5 fragment bound Rab3A in brain homogenate (Fig. 4C, lower panel). We then introduced these mutant proteins into permeabilized PC12 cells and examined whether they would mimic the inhibitory effects of recombinant wild-type RalA and the Sec5 fragment. These point mutations almost completely abolished the inhibition (Fig. 4D). Thus, the inhibition by recombinant wild-type RalA and the Sec5 fragment appears to be caused by their interactions with the native exocyst complex and RalA, respectively. Therefore, we conclude that the GTP-dependent interaction between RalA and the exocyst is essential for GTP-dependent exocytosis.
Transfection with Wild-type RalA, but Not with RalA E38R , Enhances GTP-dependent Exocytosis-To obtain independent evidence that RalA may function as a GTP sensor in GTP-dependent exocytosis, we conducted a co-transfection assay in which hGH functions as a marker of transfection. PC12 cells secrete transfected hGH when stimulated by KCl depolarization (46,47). We found that similar to GppNHp-induced NE release, GppNHp triggers hGH secretion from permeabilized PC12 cells in a concentration-dependent manner. To test if RalA acts as a sensor for the GTP analogue in exocytosis, we co-transfected PC12 cells with hGH and wild-type RalA (RalA WT ), mutant RalA (RalA E38R ), or a control plasmid. RalA WT and RalA E38R were expressed as Myc fusion proteins to facilitate their detection in PC12 cells (see below). Transfection with Myc-RalA WT caused a major shift in the GppNHp response curve to lower GppNHp concentrations, making the cells responsive to 0.1 M GppNHp. In addition, Myc-RalA WT enhanced maximal secretion in response to GppNHp (Fig.  5A). Thus, transfected Myc-RalA WT sensitized GppNHpinduced hGH secretion, indicating that RalA functions as a sensor for the GTP analogue in GTP-dependent exocytosis. In contrast, Myc-RalA E38R did not effect GppNHp-induced hGH secretion.
We verified that the lack of effect by RalA E38R was not caused by the mislocalization of this protein, because transfected Myc-RalA WT and Myc-RalA E38R were both largely colocalized with Secretogranin II (SgII), a marker protein of secretory granules in PC12 cells (Fig. 5B). The localization of transfected Myc-RalAs on secretory granules is consistent with a previous finding of native Ral on dense granules in platelets (28). Therefore, we conclude that the RalA-exocyst interaction is essential for transfected RalA to act as a sensor for the GTP analogue in GTP-dependent exocytosis.
GDP, RalA, and the Sec5 Fragment Show No Inhibition of Ca 2ϩ -dependent Exocytosis-We also examined the relationship between GTP-dependent and Ca 2ϩ -dependent exocytosis. In PC12 cells, both GppNHp and Ca 2ϩ triggers NE release from large dense core vesicles, although Ca 2ϩ with MgATP and cytosolic factors can induce a more robust exocytosis response (33,36,37,38,43). If the same pathway utilized in GTP-dependent exocytosis contributes to Ca 2ϩ -dependent exocytosis, it would be expected that GDP, Ral and the Sec5 fragment, all strong inhibitors of GTP-dependent exocytosis, will at least partially inhibit Ca 2ϩ -dependent exocytosis. However, we found that none of these agents inhibited Ca 2ϩ -dependent exocytosis (Fig. 6, A and B). Thus we conclude that independent pathways mediate GTP-dependent and Ca 2ϩ -dependent exocytosis. These results also emphasize the specificity of the inhi-  bition of GTP-dependent exocytosis by recombinant RalA and Sec5 fragments.
Cleavage of Soluble N-ethylmaleimide-sensitive Factor Attachment Protein Receptor (SNARE) Inhibits both GTP-and Ca 2ϩ -dependent Exocytosis-The SNARE complex, consisting of SNAP-25, syntaxin, and synaptobrevin (also called VAMP) (50), is essential for Ca 2ϩ -dependent exocytosis (reviewed in Ref. 51). We examined the role of the SNAREs in GTP-dependent exocytosis by utilizing BoNT/E-LC, which specifically cleaves SNAP-25.
BoNT/E-LC almost completely blocked both Ca 2ϩ -and GppNHpinduced exocytosis (Fig. 7), which is in agreement with previous results that used native SNARE-cleaving toxins including BoNT/E (42). Thus, SNARE function is essential for both Ca 2ϩand GTP-dependent exocytosis. DISCUSSION In this study, we have discovered the long sought after signaling pathway that mediates GTP-dependent exocytosis. In various mammalian cell types, sustained activation of one or more GTPases by nonhydrolyzable GTP induces exocytotic secretion (1)(2)(3)(4)(5)(6)(7)(8)(9). Therefore, the discovery of a mediating GTPase for this process is important for the cell biology of secretion. We showed that the interaction between RalA GTPase and the exocyst is critical for GTP-dependent exocytosis. The involvement of other GTPases (Rab3A and Rho) in GTP-dependent exocytosis has been previously suggested (12,15). The experiments using the Rab3A fragment (effector loop peptide) in mast cells (15) are now being questioned because the fragment induces Ca 2ϩ transients in mast cells, raising the possibility that its mode of action is nonspecific (52). In our preparation, we did not observe inhibitory effects by recombinant Rab3A, RhoA, or RhoGDI on GTP-dependent exocytosis. Because RhoGDI inhibits GTP␥S-induced exocytosis in mast cells (12), further experiments are required to investigate whether the Ral-exocyst interaction is essential for GTP-dependent exocytosis in other secretory cells such as mast cells. In addition, we do not exclude involvement of other GTPase signaling pathways in GTP-dependent exocytosis in PC12 cells, because the maximal inhibition exerted by soluble RalA and the Sec5 fragment is ϳ70%.
In contrast to GTP-dependent exocytosis, we found no role for Ral-exocyst interaction in Ca 2ϩ -dependent exocytosis. Based on this finding, we suggest that Ca 2ϩ -dependent and GTP-dependent exocytosis utilize independent pathways to trigger exocytosis. However, this finding appears to contradict the results of Moskalenko et al. (25), which suggested that overexpression of RalA inhibited high K ϩ -induced, Ca 2ϩ -dependent exocytosis. Although their work suggests that RalA is a negative regulator of Ca 2ϩ -dependent exocytosis, it is important, however, to note that overexpression of other GTPases, including Rab3 (53) and Rab11b (54), exhibited even stronger inhibition of high K ϩ -induced exocytosis. Thus, we suggest that the observed inhibition of high K ϩ -induced exocytosis by several GTPases may lack specificity.
The exocyst complex was first characterized in yeast (20,21). In yeast, exocyst components are essential for secretion and the entire complex functions as an effector for Sec4 GTPase. Little is currently known of the function of the exocyst complex in higher organisms. In Drosophila, mutations of Sec5, a central component of the exocyst complex, resulted in embryonic lethality and general defects in membrane trafficking (31). Interestingly, however, Ca 2ϩ -dependent neurotransmitter release from synaptic vesicles persisted in this mutant (31). RNAi-mediated knock-down of Sec10 expression, another component of the exocyst, failed to show defects in neurotransmission, but knock-down of Sec10 expression did cause defects in hormonal secretion from endocrine cells (30). Our results suggest that the function of the exocyst is critical for GTP-depend-ent but not for Ca 2ϩ -dependent exocytosis in neuroendocrine PC12 cells. In transgenic mice expressing a dominant inhibitory form of RalA, Ca 2ϩ -dependent glutamate release was normal, but refilling of the readily releasable pool was suppressed (29). Thus, the GTP-dependent interaction of RalA and the exocyst may also contribute to modulation of the readily releasable pool of synaptic vesicles in neurons.
We found that both GTP-dependent exocytosis and Ca 2ϩ -dependent exocytosis are SNARE-dependent. Even in the well studied Ca 2ϩ -dependent exocytosis, how Ca 2ϩ -sensors, such as synaptotagmins (38,39,41), transmit the signal to the SNAREs remains controversial. At least two scenarios may explain the communication between signaling sensors and the SNAREs. In the first scenario, either the Ral-exocyst or synaptotagmin-phospholipid/phosphatidylinositol 4,5-bisphosphate (PIP 2 ) interaction (38,39,41) apposes the vesicular and plasma membranes, which drives the formation of the SNARE complex and membrane fusion (55). In the second scenario, either the Ral-exocyst or synaptotagmin-phospholipid/PIP 2 physically interacts with the SNAREs to trigger the fusion (41). The identification of the mechanisms by which GTP-dependent Ralexocyst interaction communicates with the SNAREs to trigger exocytosis/membrane fusion is an area for future study.