Phosphorylation of SNAP-23 Regulates Exocytosis from Mast Cells*

Regulated exocytosis is a process in which a physiological trigger initiates the translocation, docking, and fusion of secretory granules with the plasma membrane. A class of proteins termed SNAREs (including SNAP-23, syntaxins, and VAMPs) are known regulators of secretory granule/plasma membrane fusion events. We have investigated the molecular mechanisms of regulated exocytosis in mast cells and find that SNAP-23 is phosphorylated when rat basophilic leukemia mast cells are triggered to degranulate. The kinetics of SNAP-23 phosphorylation mirror the kinetics of exocytosis. We have identified amino acid residues Ser95 and Ser120 as the major phosphorylation sites in SNAP-23 in rodent mast cells. Quantitative analysis revealed that ∼10% of SNAP-23 was phosphorylated when mast cell degranulation was induced. These same residues were phosphorylated when mouse platelet degranulation was induced with thrombin, demonstrating that phosphorylation of SNAP-23 Ser95 and Ser120 is not restricted to mast cells. Although triggering exocytosis did not alter the absolute amount of SNAP-23 bound to SNAREs, after stimulation essentially all of the SNAP-23 bound to the plasma membrane SNARE syntaxin 4 and the vesicle SNARE VAMP-2 was phosphorylated. Regulated exocytosis studies revealed that overexpression of SNAP-23 phosphorylation mutants inhibited exocytosis from rat basophilic leukemia mast cells, demonstrating that phosphorylation of SNAP-23 on Ser120 and Ser95 modulates regulated exocytosis by mast cells.

Regulated exocytosis is the process by which stimulation of plasma membrane receptors on secretory cells results in the release of proteins and/or peptides from intracellular stores into the extracellular space (1). One common characteristic of regulated exocytosis, whether it be from neurons, cytotoxic lymphocytes, adipocytes, or mast cells, is that the cells response to this stimulus results in pre-formed intracellular granules moving toward and fusing with the plasma membrane. Secretory granule/plasma membrane fusion is the essence of regulated exocytosis, and there is an intense effort underway to identify the molecular mechanisms regulating this process in the hopes of identifying ways to modulate exocytosis.
The RBL-2H3 1 mast cell line has been extensively studied as a model not only for mast cell biology but also as a paradigm for regulated exocytosis from non-neuronal cells (2). Stimulation of the high affinity IgE receptor, Fc⑀RI, on these cells by crosslinking initiates a signal transduction cascade that culminates in secretory granule fusion with the plasma membrane, thereby liberating a variety of inflammatory mediators (3,4). Numerous proteins necessary for the tethering, docking, and fusion steps between various membrane compartments in eukaryotic cells have been described. Among those, SNAREs (soluble NSF-attachment protein receptors) are a large family of membrane-associated proteins essential for membrane-membrane fusion (5,6). These proteins include members of the vesicle-associated synaptobrevin/VAMP family as well as members of the syntaxin and SNAP-23 families of "target" membrane SNAREs. The current model proposes that while vesicles are docked on the target membrane, SNAREs from the donor or vesicle membrane (v-SNAREs) form trans-SNARE complexes with their cognate SNARE partners on the opposing target membrane. Structurally, the exocytic SNARE complex is a trimolecular protein complex containing one member of the VAMP, syntaxin, and SNAP-23 family, each contributing to the formation of a four-helix coiled-coil bundle (7,8) whose formation is sufficient for in vitro membrane fusion (9).
Given that SNAREs play a central role in the membrane fusion process, it is likely that their function is modulated in vivo. In particular, protein kinases, which have been extensively associated with the regulation of exocytosis (10), could participate in SNARE function by phosphorylating residues essential in SNARE complex assembly or the binding of SNARE regulatory proteins (6,11,12). Members of the syntaxin (13)(14)(15) and SNAP-23/25 (14,16,17) family of proteins are substrates for various protein kinases in vitro, although precise SNARE substrates in vivo and the physiological consequences of SNARE phosphorylation are not clear.
The plasma membrane localized target SNAREs, SNAP-23 and syntaxin 4, have been shown to be important mediators of granule/plasma membrane fusion in mast cells (18 -21) and platelets (22,23). Pombo et al. (15) found that syntaxin 4 was constitutively phosphorylated in RBL mast cells and the extent of syntaxin 4 phosphorylation was not altered during secretion. By contrast, using human platelets Chung et al. (24) found that thrombin stimulation resulted in phosphorylation of this same SNARE protein. In neither of these studies were the syntaxin 4 phosphorylation sites identified or the physiological impor-tance of syntaxin 4 phosphorylation in regulated exocytosis directly examined.
In addition to identifying syntaxin 4 as a substrate for protein kinases, it has recently been reported that the plasma membrane SNARE SNAP-23 is phosphorylated during human platelet activation (17). Using mass spectroscopy these authors identified Ser 23 /Thr 24 and Ser 161 as targets of SNAP-23 phosphorylation but were unable to ascribe a functional consequence to this. In the current study we have investigated in detail the extent and kinetics of SNARE phosphorylation in mast cells. We find that SNAP-23 undergoes rapid, stimulusinduced phosphorylation and the kinetics of phosphorylation mirror that of degranulation. In contrast to the results of Polgar et al. (17), we demonstrate that the induced phosphorylation sites in rodent SNAP-23 are Ser 95 and Ser 120 . By using phosphopeptide mapping studies, site-directed mutagenesis, and phosphorylation site-specific antibodies we show that these sites are phosphorylated in vivo in RBL mast cells, bone marrow-derived mast cells, and activated mouse platelets. Overexpression of SNAP-23 phosphorylation mutants inhibits the extent of regulated exocytosis from RBL mast cells. We also find that phosphorylated SNAP-23 preferentially associated with syntaxin 4 and VAMP-2 after exocytosis was triggered. These data suggest that SNAP-23 phosphorylation is an important post-translational modification of a critical SNARE that regulates mast cell exocytosis.

EXPERIMENTAL PROCEDURES
Antibodies and Cell Culture-Rabbit antisera recognizing the SNAP-23 carboxyl terminus have been described (25). A polyclonal rabbit antisera recognizing rat syntaxin 3 (26) was the generous gift of Dr. Ann Hubbard (Johns Hopkins University, Baltimore, MD). A polyclonal rabbit antisera recognizing rat syntaxin 4 was generated by immunizing rabbits with a GST fusion protein of rat syntaxin 4 (amino acids . This antiserum showed no detectable cross-reactivity with rat syntaxin 2 or rat syntaxin 3 and was used in all immunoprecipitation and immunoblotting studies. Antisera recognizing phospho-SNAP-23-Ser 120 was generated by immunizing rabbits with the synthetic peptide Val-Ser-Lys-Gln-Pro-phospho-Ser-Arg-Ile-Thr-Ans-Gly-Gln (corresponding to amino acids 115-126 of rat SNAP-23). Antisera recognizing phospho-SNAP-23-Ser 95 was generated by immunizing rabbits with the synthetic peptide Thr-Lys-Asn-Phe-Glu-phospho-Ser-Gly-Lys-Asn-Tyr-Lys-Ala-Thr (corresponding to amino acids 90 -102 of rat SNAP-23). The VAMP-2 monoclonal antibody cl 69.1 was obtained from Synaptic Systems (Goettingen, Germany). Unless indicated, anti-DNP-IgE (clone SPE-7) was from Sigma. For the experiments shown in Figs. 8 -10, anti-DNP-IgE (clone TIB-142) was obtained from the American Type Culture Collection (Manassas, VA). Biotinylated anti-FLAG monoclonal antibody M5 was from Sigma and anti-GFP monoclonal antibody clone A.v. was from Clontech.
RBL-2H3 mast cells were obtained from Dr. Juan Rivera (National Institutes of Health) and were grown in equal parts minimal essential medium and Iscove's medium containing 20% fetal calf serum (HyClone Laboratories), 25 mM Hepes (pH 7.5), and 50 g/ml gentamicin. Cells were maintained as subconfluent monolayers at 37°C in a humidified atmosphere containing 5% CO 2 and passaged with trypsin. Cells were sensitized with 1 g/ml anti-DNP-IgE for 24 h prior to stimulation with 100 ng/ml DNP-BSA in phenol red-free RPMI 1640 medium. Sensitized cells were mock-stimulated using 100 ng/ml BSA in the same medium. HeLa cells were grown in DMEM supplemented with 10% fetal calf serum, 20 mM Hepes, and 50 g/ml gentamicin. The cells were maintained as subconfluent monolayers at 37°C in a humidified atmosphere containing 5% CO 2 and passaged with trypsin. In some experiments transfected HeLa cells were stimulated using 10 nM phorbol myristate acetate and 1 M ionomycin.
Plasmids and Recombinant Proteins-Rat SNAP-23 was subcloned into pcDNA3 (Invitrogen) using EcoRI and XhoI sites introduced by PCR using a previously described full-length rat SNAP-23 cDNA (27). GST-rat SNAP-23 was generated by subcloning rat SNAP-23 from pcDNA3 into pGEX-4T1 (Amersham Biosciences). Ala mutants of all Ser and Thr residues present in rat SNAP-23 were generated with the QuikChange Site-directed Mutagenesis kit (Stratagene) using pGEX-4T1-rat SNAP-23 as the template according to the manufacturer's in-structions. The double mutant of S95A/S120A was generated using rat SNAP-23 S120A DNA as a template. The double mutant replacing both Ser 95 and Ser 120 with Asp residues (S95D/S120D) was generated using the QuikChange Multi-Mutagenesis kit (Stratagene) as directed by the manufacturer. The inserts of wild-type SNAP-23 and the mutants of interest were subcloned into the pCMV-Tag 2B vector (Stratagene) to generate amino-terminal FLAG-tagged proteins. The integrity of all PCR-amplified products and all mutants were confirmed by automated sequence analysis using an ABI sequencer.
Transfections-Exponentially growing RBL cells were re-suspended in serum-free and antibiotic-free DMEM and 10 ϫ 10 6 cells were transfected by electroporation (310 mV, 960 microfarads, Bio-Rad GenePulser) using 10 -20 g of empty vector, FLAG-tagged wild-type SNAP-23, or FLAGtagged mutant rat-SNAP-23 as described previously (20). Transfected cells were immediately diluted in antibiotic-free medium and allowed to adhere to plastic culture dishes for 5-6 h before adding anti-DNP IgE to the medium overnight. The next day, cells were triggered for exocytosis as described above. HeLa cells were transiently transfected using Lipofectamine (Invitrogen) according to the manufacturer's instructions.
Metabolic Labeling of Cells-RBL cells were plated at 1.5 ϫ 10 6 cells per 10-cm dish. IgE-sensitized cells were starved the next day for 15 min in phosphate-free DMEM containing 3% dialyzed calf serum (Hy-Clone Laboratories), 20 mM Hepes (pH 7.5), and gentamicin. Cells were labeled with 0.5 mCi of [ 32 P]orthophosphate (PerkinElmer Life Sciences) for 5 h in 1 ml of phosphate-free DMEM. Cells were then stimulated for secretion with DNP-BSA as described above. Cells were washed with ice-cold Hanks' balanced salt solution and tissue culture plate-bound cells were frozen until further use.
In Vitro Phosphorylation Reactions-GST and GST-SNAP-23 proteins were purified from isopropyl 1-thio-␤-D-galactopyranoside-induced BL21 Escherichia coli using standard protocols. Briefly, GST and GST-SNAP-23 were isolated from 200-l bacteria lysates with 10 l Immunoprecipitation and Electrophoresis-Immunoprecipitation of proteins from cell lysates was carried out as described previously (28). Briefly, cells were lysed for 1 h in ice-cold lysis buffer, 10 mM Tris (pH 7.4), 150 mM NaCl containing 1% Triton X-100 as well as protease inhibitors (5 mM iodoacetamide, 50 mM phenylmethylsulfonyl fluoride, and 0.1 mM N ␣ -p-tosyl-L-lysine chloromethyl ketone) and phosphatase inhibitors (5 mM EDTA, 5 mM EGTA, 50 mM NaF, 10 mM Na 4 P 2 O 7 , and 1 mM Na 3 VO 4 ). Specific immunoprecipitations were performed for 2 h at 4°C by incubating pre-cleared lysates with specific antibodies bound to protein A-Sepharose beads (Sigma). Proteins from the immunoprecipitates were separated on 12.5% SDS-PAGE gels. Radiolabeled molecules were visualized by autoradiography and/or PhosphorImager analysis. The extent of phosphorylation was quantitated by PhosphorImager analysis. For phosphopeptide mapping studies, proteins were transferred to 0.2-m nitrocellulose membranes (Schleicher & Schuell) and the location of phosphoproteins on the membrane was determined after autoradiography. For immunoblot analysis proteins were transferred to polyvinylidene difluoride membranes (Bio-Rad) and proteins were visualized by enhanced chemiluminescence using Western Lightning (PerkinElmer Life Sciences). Band intensity was determined by automatic integrating densitometry using a Amersham Biosciences densitometer. For quantitative experiments, statistical analyses were carried out by using a Student's t test. Results were considered significant when a p value of less then 0.05 was obtained.
Tryptic Phosphopeptide Mapping-Nitrocellulose membranes containing excised SNAP-23 were incubated for 1 h at 37°C in 1% polyvinylpyrrolidone-40 (Sigma), washed twice with 0.4% NH 4 HCO 3 (pH 8.0), and incubated overnight with 0.5 mg/ml trypsin at 37°C in the same buffer. The released peptides were dried in a Speed-vac, washed with H 2 O several times, and resuspended in 5 l of H 2 O. Peptides were spotted at the bottom of a 20 ϫ 20-cm thin layer chromatography plate (#5577, E. M. Science), dried, and subjected to electrophoresis for 75 min at 1000 V in a buffer of 2.5% formic acid, 7.5% acetic acid. After electrophoresis, the plates were dried for at least 1 h and subjected to analysis in a second dimension of thin layer chromatography using a solvent containing 62.5% isobutyric acid, 4.8% pyridine, 1.9% butanol-1, 2.9% acetic acid until the solvent front migrated to within 1 cm of the top of the TLC plate (generally overnight). The plates were dried and analyzed with a PhosphorImager. Subcellular Fractionation-RBL cells were harvested by trypsinization and washed in phosphate-buffered saline. The cells were re-suspended in hypotonic buffer (10 mM Tris, 10 mM KCl, 1 mM EGTA, 0.5 mM MgCl 2 , pH 7.4) and were disrupted by repeated passage of cells through a 25-gauge syringe. Nuclei and unbroken cells were removed by centrifugation at 1000 ϫ g and the post-nuclear supernatant was subjected to centrifugation at 100,000 ϫ g for 1 h at 4°C to isolate membranes (pellet) and cytosol (supernatant). In some experiments RBL cells were transfected with pEGFP (Clontech) to allow expression of the cytosolic marker protein GFP. The membrane pellet and cytosolic supernatant were brought to the same volume in hypotonic buffer and each was adjusted to a final concentration of 1% Triton X-100. Equal portions of each fraction were analyzed by SDS-PAGE and immunoblotting.
Mast Cell Exocytosis Assay-The degranulation in RBL-2H3 cells was monitored by measuring the ␤-hexosaminidase activity released from cells grown in 6-or 12-well plates. IgE-sensitized cells were washed twice with RPMI and mock-stimulated or stimulated by the addition of DNP-BSA. The cell supernatant was collected and the cells were lysed in phenol red-free RPMI containing 0.2% Triton X-100 to determine the total enzyme content. A colorimetric assay with p-nitrophenyl-N-acetyl-␤-D-glucosaminide (Sigma) as the substrate was used to measure the amount of ␤-hexosaminidase released into the medium and remaining in the cells as described previously (20). The ␤-hexosaminidase activity released was expressed as a percentage of the activity released into the medium relative to the total activity (released plus cell-associated).
In some experiments, RBL cells were transfected by electroporation with an expression vector of human growth hormone (2 g) together with empty pCMV vector or pCMV-FLAG-SNAP-23 (20 g). After overnight culture, the cells were sensitized with IgE and mock-stimulated or stimulated with DNP-BSA as described above. The amount of human growth hormone released into the medium or remaining cell associated was determined using a human growth hormone enzyme-linked immunosorbent assay (Roche Diagnostics Corp.) as described previously (29). For quantitative experiments, statistical analyses were carried out by using a Student's t test. Results were considered significant when a p value of less than 0.05 was obtained.
Platelet Preparation and Measurement of Thrombin-induced Secretion-Mouse platelets were isolated as described previously (30). Briefly, blood was obtained from the right ventricle of the heart of sacrificed C57BL/6 mice. The blood was mixed with 1.8% sodium citrate (pH 7.4) to a final concentration of 0.18%. Platelet-rich plasma was prepared by centrifugation at 100 ϫ g for 10 min. Care was taken to harvest only the platelet-containing upper layer to avoid the buffy coat layer of nucleated cells above the red blood cells.
Platelets were assayed for dense core release of [ 3 H]5-hydroxytryptamine, lysosomal release of ␤-hexosaminidase, and ␣-granule release of platelet factor IV, as described previously (31). Platelets were activated by the addition of thrombin (0.5 units/ml; Chronolog) in a Hepes/Tyrode's release buffer containing 0.7 mM CaCl 2 at 25°C for the indicated times. The data were tabulated as the percent release compared with the total present in each reaction. The background release of each marker in unstimulated platelets was subtracted from that obtained after thrombin stimulation to yield net granule marker release. A parallel set of reactions was solubilized in SDS-PAGE sample buffer for analysis by immunoblotting.

SNAP-23 Is Phosphorylated in Stimulated RBL Cells-RBL
cell exocytosis can be triggered by cross-linking of surface Fc⑀RI receptors for IgE by the appropriate antigen (3,4). To determine whether SNAREs were phosphorylated during physiological stimulation in mast cells, RBL-2H3 cells were metabolically labeled with [ 32 P]orthophosphate prior to cross-linking surface Fc receptors for DNP-specific IgE using DNP-BSA. The cells were then mock-stimulated (using unconjugated BSA) or stimulated for 20 min with DNP-BSA and the phosphorylation status of the SNARE proteins was analyzed by immunoprecipitation, SDS-PAGE, and PhosphorImager analysis (Fig. 1A). In agreement with published results (15), syntaxin 4 was phosphorylated in mock-treated RBL cells and phosphorylation was not altered by stimulation. Whereas syntaxin 3 is clearly present in RBL cells, syntaxin 3 was not detectably phosphorylated either before or after stimulation. By contrast, SNAP-23, which was also phosphorylated to a small extent in the resting state, underwent a stimulation-dependent increase in phosphorylation after stimulation with DNP-BSA. Under these conditions we observed a 2.5-fold increase in SNAP-23 phosphorylation after 20 min of stimulation as compared with mock-treated cells.
Quantitative immunoblot analysis of the SNAP-23 immunoprecipitate showed no changes in the total amount of SNAP-23 after stimulation of RBL cells, however, the SNAP-23 band became heterogeneous and higher molecular weight isoforms of SNAP-23 appeared after stimulation. (The more rapidly migrating band in the SNAP-23 immunoblot is an amino-terminal degradation product of SNAP-23 that is variably observed in rat tissues.) Because altered electrophoretic mobility, or "fuzziness," is often associated with protein phosphorylation, we examined the mobility of SNAP-23 isolated from stimulated RBL cells by immunoprecipitation and treatment with alkaline phosphatase. Incubation of the SNAP-23 immunoprecipitate with phosphatase resulted in the removal of the more slowly migrating electrophoretic species and SNAP-23 now migrated with a mobility that was indistinguishable from that from unstimulated cells (Fig. 1B). These data therefore demonstrated that the electrophoretic mobility of SNAP-23 is indeed altered in SDS-PAGE by phosphorylation.
SNAP-23 Phosphorylation and Mast Cell Degranulation Occur with Similar Kinetics-To establish whether the kinetics of SNAP-23 phosphorylation correlate with the kinetics of mast The cells were then mock-triggered or stimulated for degranulation using DNP-BSA for 20 min. The cells were washed, lysed in Triton X-100, and immunoprecipitations were performed using SNAP-23, syntaxin 4, and syntaxin 3 antisera. The immunoprecipitates were analyzed by SDS-PAGE and analyzed by Phos-phorImager (to visualize 32 P-labeled proteins) or by immunoblot analysis using the indicated antibodies. The more rapidly migrating band in the SNAP-23 immunoblot is a degradation product of SNAP-23 that is only seen in rat tissues. B, RBL cells were coated with DNPspecific IgE and incubated with BSA (mock-triggered) or DNP-BSA (triggered) for 20 min. The cells were lysed in Triton X-100 and SNAP-23 was isolated by immunoprecipitation (I.P.) using a SNAP-23 antiserum. The immunoprecipitates were incubated in alkaline phosphatase buffer in the absence (buffer alone) or presence (phosphatase) of alkaline phosphatase. The samples were then analyzed by immunoblot analysis using a SNAP-23 antibody. cell degranulation, supernatants were collected at various times after stimulation and the extent of degranulation was determined by analyzing the amount of the granule enzyme ␤-hexosaminidase released from the cells. RBL mast cell exocytosis was essentially complete within 15 min of stimulation ( Fig. 2A). In a parallel series of experiments RBL cells were pre-loaded with [ 32 P]orthophosphate and stimulated by Fc⑀RI cross-linking for various times before harvesting cells and analyzing the extent of SNAP-23 phosphorylation. SNAP-23 was transiently phosphorylated, with a peak of phosphorylation after 15 min of stimulation that declined over the next 30 min down to the resting state level (Fig. 2B). Analysis of the SNAP-23 immunoprecipitates by immunoblotting confirmed that stimulation did not alter the amount of SNAP-23 present in the cells. Detailed kinetic analysis of SNAP-23 phosphorylation after short times of stimulation showed that significant SNAP-23 phosphorylation was observed after only 5 min of stimulation (a 4-fold increase in SNAP-23 phosphorylation at 5 min as compared with mock-treated cells; Fig. 2C). Because the kinetics of mast cell degranulation correlate with the kinetics of SNAP-23 phosphorylation, these data raise the possibility that these two cellular processes are linked.
SNAP-23 Is Phosphorylated by Protein Kinase C-Protein kinase C (PKC) is a key kinase involved in mast cell signaling and is an important regulator of exocytosis (10). To determine whether PKC is involved in SNAP-23 phosphorylation, RBL cells were treated with the PKC inhibitor bisindolylmaleimide (I) for 10 min before and during the stimulation. Bisindolylmaleimide (I) inhibits a wide variety of PKC isoforms, including both conventional and non-conventional PKCs (32). Both stimulated exocytosis and SNAP-23 phosphorylation were inhibited by treatment of RBL cells with bisindolylmaleimide (I), revealing that SNAP-23 phosphorylation requires PKC activity in RBL cells (Fig. 3A). We next set out to determine whether SNAP-23 could serve as a substrate for PKC. In vitro phosphorylation experiments using purified PKC and recombinant GST-SNAP-23 fusion proteins revealed that the latter was phosphorylated by PKC under conditions in which GST alone was not (Fig. 3B). Whereas such an analysis does not prove that PKC is responsible for SNAP-23 phosphorylation in vivo, it prompted us to identify the PKC phosphorylation sites in SNAP-23.
To establish more clearly if PKC was involved in SNAP-23 phosphorylation in vivo, we performed two-dimensional phosphopeptide analyses of endogenous phospho-SNAP-23 isolated from resting RBL cells, stimulated RBL cells, and GST-SNAP-23 phosphorylated by purified PKC in vitro. In the mock-stimulated state, one SNAP-23 phosphopeptide is observed (indicated by the arrowhead in Fig. 3C). Upon stimulation the intensity of this basal spot is not altered, demonstrating that phosphorylation of this phosphopeptide is not augmented or diminished by mast cell triggering for exocytosis. Cross-linking of surface Fc⑀RI with DNP-BSA leads to the appearance of three major phosphopeptides (circled in Fig. 3C) and quantitative analysis of multiple maps revealed that these three spots represent 61 Ϯ 6% of the total phosphorylation signal and 75 Ϯ 6% of the induced phosphorylation signal. Comparing the phosphopeptide map of endogenous SNAP-23 from RBL cells to that of in vitro phosphorylated GST-SNAP-23 demonstrated that the "basal" phosphorylation site as well as the three, induced phosphopeptides were generated by PKC treatment, further suggesting that PKC phosphorylates SNAP-23 in vivo. In addition to the highlighted phosphopeptides, PKC directly phosphorylates additional residues on GST-SNAP-23 that are not present in endogenous SNAP-23 isolated from mast cells. In vitro phosphorylation artifacts that are not relevant to in vivo phosphorylation are not uncommon. In this case we did not choose to characterize these "irrelevant" phosphopeptides further.
Identification of SNAP-23 Phosphorylation Sites-Having identified SNAP-23 as a protein that is inducibly phosphorylated by triggering mast cell exocytosis, we set out to identify the phosphorylation sites on the protein. Because phosphorylation of GST-SNAP-23 with PKC resulted in phosphopeptide maps that contained most of the spots present in the maps of endogenous SNAP-23 phosphorylated in RBL cells (Fig. 3B), we individually mutated each of the 29 serine and threonine residues in GST-SNAP-23 to alanine, phosphorylated the recombinant proteins with PKC in vitro, and analyzed the tryptic phosphopeptides on 40% alkaline PAGE gels (not shown). Mutants that appeared to lack one or more prominent phosphopeptide(s) were then analyzed by two-dimensional phosphopeptide mapping. Analysis of these maps revealed that mutation of Ser 95 resulted in the disappearance of one phosphopeptide, whereas mutation of Ser 120 resulted in the disappearance of two distinct phosphopeptides (Fig. 4). The disappearance of these two spots with a single mutation most likely reflects partial tryptic proteolysis around Ser 120 , although it is theoretically possible that phosphorylation of Ser 120 regulates the An aliquot was also analyzed using preimmune serum to demonstrate the specificity of the immunoprecipitation reaction. The immunoprecipitates were analyzed by SDS-PAGE and analyzed by PhosphorImager (to visualize 32 P-labeled SNAP-23) or by immunoblot analysis using SNAP-23 antibodies. The intensity of the phospho-SNAP-23 signal was quantitated using PhosphorImager and is shown for the 60-min secretion time course in B and the 20-min secretion time course in C.
phosphorylation of an additional residue that was not identified in our initial screen of mutants.
Additional mutagenesis studies revealed that the phosphopeptide indicated by an asterisk in Fig. 4 represents Ser 161 (as identified by Polgar et al. (17)). This minor phosphopeptide was not reproducibly observed in vivo, accounted for only 1.1 Ϯ 2% of the induced phosphorylation signal, and was therefore not analyzed further. The phosphopeptide migrating just below and to the left of the basal phosphorylation site represented only 22 Ϯ 10% of the induced phosphorylation signal, however, because this phosphopeptide was not generated by in vitro phosphorylation of GST-SNAP-23 with PKC the nature of this residue was not determined.
To unambiguously demonstrate that SNAP-23 Ser 95 and Ser 120 are phosphorylated upon mast cell stimulation, we introduced wild-type FLAG-tagged SNAP-23 or its Ser 95 and Ser 120 mutants into RBL cells and triggered the cells for degranulation. The addition of the FLAG-epitope tag allowed us to distinguish endogenous rat SNAP-23 from the transfected rat SNAP-23 mutants. Phosphopeptide mapping studies confirmed that FLAG-SNAP-23 was inducibly phosphorylated like endogenous SNAP-23 by Fc⑀RI cross-linking (Fig. 4). Most importantly, mutation of Ser 95 and Ser 120 to Ala also resulted in the disappearance of the three induced phosphopeptide species observed in endogenous SNAP-23. These data demonstrate that Ser 95 and Ser 120 are the predominant phosphorylation sites of SNAP-23 in RBL cells.
Generation of Phospho-SNAP-23-specific Antibodies-Having identified Ser 120 and Ser 95 as the major phosphorylation sites in SNAP-23 in RBL cells we generated rabbit antisera against SNAP-23 peptides containing phospho-Ser 120 or phospho-Ser 95 (Fig. 5A). To confirm the specificity of these antisera, HeLa cells were transfected with either empty vector or vector containing wild-type, S95A, or S120A mutants of FLAG-SNAP-23. After stimulation of PKC with phorbol myristate acetate and ionomycin, the anti-phospho-SNAP-23-Ser 120 antibody recognized wild-type FLAG-SNAP-23 and the SNAP-23 S95A mutant but, as anticipated, did not recognize the SNAP-23 S120A mutant (Fig. 5B). Conversely, the anti-phospho-SNAP-23-Ser 95 antibody recognized wild-type FLAG-SNAP-23 and the SNAP-23 S120A mutant but not the SNAP-23 S95A mutant (Fig. 5B). Immunoblot analysis with a SNAP-23 antibody confirmed the presence of endogenous human SNAP-23 and transfected FLAG-SNAP-23 in the cells. Note that neither phospho-SNAP-23-specific antibody recognized endogenous HeLa SNAP-23, as Ser 120 is not present in human SNAP-23 and the sequence surrounding Ser 95 is not conserved between human SNAP-23 and rodent SNAP-23.
The anti-phospho-SNAP-23 antibodies were then used to examine endogenous SNAP-23 phosphorylation in RBL cells and bone marrow-derived mouse mast cells. In unstimulated RBL cells the antisera recognized a very weak phospho-SNAP-23 band (Fig. 5C), likely representing SNAP-23 that is phosphorylated during the low basal degranulation observed in mast cells. In contrast, when cells are stimulated with DNP-BSA for 15 min, very strong phospho-SNAP-23-Ser 120 and phospho-SNAP-23-Ser 95 signals were observed in RBL cells (Fig. 5C). In addition, when bone marrow-derived mast cells were triggered we observed robust phosphorylation of SNAP-23 using the phospho-SNAP-23-Ser 120 antibody ( Fig. 5C; Ser 95

FIG. 5. Phosphorylation site-specific antibodies recognize phospho-SNAP-23 in rat and mouse mast cells.
A, a cartoon of the structure of SNAP-23 highlighting the location of Ser 95 and Ser 120 in relation to the cysteine-rich linker region of SNAP-23 and the aminoand carboxyl-terminal coiled-coil domains. B, HeLa cells were transiently transfected with cDNAs encoding empty expression vector (mock) or FLAG-tagged wild-type SNAP-23, SNAP-23 S95A, and SNAP-23 S120A as indicated. The cells were stimulated with phorbol myristate acetate and ionomycin for 20 min and aliquots of the cell lysates were analyzed by immunoblotting with a phospho-SNAP-23-Ser 120 antiserum (upper panel), a phospho-SNAP-23-Ser 95 antiserum (middle panel), or an antibody that recognizes total SNAP-23 (lower panel). Note that FLAG-tagged rat SNAP-23 migrates with a slower electrophoretic mobility than endogenous human SNAP-23 in HeLa cells. C, RBL mast cells and mouse bone marrow-derived mast cells were coated with DNP-specific IgE and mock-stimulated or stimulated with DNP-BSA for 15 min. Equal aliquots of the cell lysates were analyzed by immunoblotting with a phospho-SNAP-23-Ser 120 antiserum, a phospho-SNAP-23-Ser 95 antiserum, or an antibody that recognizes total SNAP-23. phosphorylation was not examined). These data demonstrate that endogenous SNAP-23 in both mast cell types is phosphorylated on Ser 120 upon stimulation of exocytosis. These data highlight the specificity of these antibodies for phospho-SNAP-23 and demonstrate that SNAP-23 phosphorylation is not only induced in RBL cells but in bone marrow-derived mast cells.
Phosphorylated SNAP-23 Is Bound to Membranes-Both Ser 95 and Ser 120 are adjacent to the cysteine-rich region known to be involved in the anchoring of SNAP-23 to the plasma membrane by palmitoylation (33). To investigate the possibility that SNAP-23 phosphorylation regulates membrane attachment of SNAP-23, we isolated membrane and cytosol fractions from resting and stimulated RBL cells. Both before and after DNP-BSA stimulation, essentially all phospho-SNAP-23-Ser 120 , like the total pool of SNAP-23 and syntaxin 4, was associated with the membranes (Fig. 6). The membrane fraction was essentially devoid of the cytosolic marker protein GFP, confirming the ability of our subcellular fractionation protocol to separate cellular membranes from cytosol. These data demonstrate that phosphorylation of SNAP-23 does not interfere with membrane attachment of SNAP-23.

SNAP-23 Is Phosphorylated in Stimulated Mouse
Platelets-It has recently been reported that triggering of platelet degranulation results in SNAP-23 phosphorylation on Ser 23 / Thr 24 and Ser 161 (17). Given our failure to identify these phosphorylation sites in stimulated mast cells together with our identification of Ser 95 and Ser 120 as the major induced phosphorylation sites in SNAP-23, we sought to examine SNAP-23 phosphorylation in stimulated mouse platelets. Thrombin activates platelets and induced secretion from each of the three platelet granule stores: ␣-granules, dense core granule, and lysosomes. Thrombin treatment leads to a time-dependent increase in SNAP-23 phosphorylation as revealed by immunoblotting using phospho-SNAP-23-Ser 120 and phospho-SNAP-23-Ser 95 antibodies (Fig. 7). Note that as in mast cells, the characteristic fuzziness of platelet SNAP-23 in the SNAP-23 immunoblot indicates phosphorylation. In this same experiment the extent of release from each of the granules were as follows: 6 and 38% for ␣-granules, 52 and 73% for dense core granules, and 4 and 27% for lysosomes at the 15-and 120-s time points, respectively. These data demonstrate that as in rodent mast cells, SNAP-23 in mouse platelets is phosphoryl-ated on Ser 95 and Ser 120 following thrombin stimulation.
A Significant Pool of SNAP-23 Is Phosphorylated Upon Stimulation-Whereas we were able to unambiguously determine the induced SNAP-23 phosphorylation sites by phosphopeptide mapping and site-directed mutagenesis studies, such techniques do not readily allow one to determine the extent of protein phosphorylation on each residue. To examine the significance of SNAP-23 phosphorylation, we took advantage of our finding that the phospho-SNAP-23-Ser 120 antiserum was capable of immunoprecipitating phospho-SNAP-23 from cell extracts. (Whereas the phospho-SNAP-23-Ser 95 antiserum was effective for immunoblot analysis, this antiserum was not effective in removing all phospho-SNAP-23-Ser 95 from cell extracts.) RBL mast cells were triggered with DNP-BSA for 10 min (or not), cell extracts were prepared, and immunoprecipitations of equal aliquots of cell extract were carried out using control rabbit antiserum, an antiserum that recognizes total SNAP-23, or phospho-SNAP-23-Ser 120 antiserum. Analysis of both the immunoprecipitated material as well as aliquots of the remaining supernatant from each reaction confirmed that the control rabbit antiserum did not nonspecifically bind SNAP-23, and that the pan-SNAP-23 or phospho-SNAP-23-Ser 120 antisera removed the majority of SNAP-23 or phospho-SNAP-23-Ser 120 from each reaction, respectively (Fig. 8A). Quantitative analysis revealed that the phospho-SNAP-23-Ser 120 antiserum bound 2% of the total pool of SNAP-23 in unstimulated cells, and that after stimulation the phospho-SNAP-23-Ser 120 antiserum bound more than 10% of the total pool of SNAP-23 (Fig.  8B). To eliminate the possibility these values were skewed by co-precipitation of non-phosphorylated SNAP-23 with phospho-SNAP-23, cell lysates were boiled in SDS prior to immunoprecipitation to ensure dissociation of all SNARE proteins. After this treatment the percent of phospho-SNAP-23 present in the SNAP-23 immunoprecipitate was similar to that in the unboiled samples, increasing from 0.2% before stimulation to 8.8% after stimulation. These data reveal that stimulation of mast cell degranulation leads to phosphorylation of a significant fraction of the total pool of SNAP-23.
Phosphorylation of SNAP-23 Modulates Regulated Exocytosis from RBL Mast Cells-Given the striking similarity in kinetics of mast cell degranulation and regulated phosphorylation of SNAP-23, we introduced a SNAP-23 phosphorylation mutant into RBL cells to determine whether overexpression of this protein interfered with mast cell degranulation. For these studies RBL cells were transfected with empty FLAG vector alone, FLAG-tagged wild-type SNAP-23, the FLAG-tagged SNAP-23 S95A/S120A double mutant (that is not subject to induced phosphorylation), or the FLAG-tagged SNAP-23 S95D/ FIG. 6. Phospho-SNAP-23-Ser 120 resides exclusively on membranes. RBL mast cells were sensitized with DNP-specific IgE and mock-stimulated or stimulated with DNP-BSA for 15 min. The cells were fractionated into cytosolic and membrane fractions, and equivalent aliquots of each fraction were analyzed by immunoblotting with phospho-SNAP-23-Ser 120 antiserum (upper panel), an antibody that recognizes total SNAP-23 (middle panel), or an antibody that recognizes syntaxin 4 (lower panel). In a parallel experiment, RBL cells expressing the cytosolic control protein GFP were fractionated and the cytosolic and membrane fractions were analyzed by immunoblotting with a GFP antibody.
FIG. 7. Phosphorylation site-specific antibodies recognize phospho-SNAP-23 in mouse platelets. Platelets were kept resting or stimulated for the indicated times with 0.5 units/ml thrombin. The reactions were stopped by the addition of hirudin and SDS-PAGE sample buffer to each sample. The SNAP-23 present in each sample was analyzed by immunoblotting with a phospho-SNAP-Ser 120 antiserum, a phospho-SNAP-23-Ser 95 antiserum, or an antibody that recognizes total SNAP-23. S120D double mutant (that is a phosphomimetic mutant). Each construct was transiently transfected together with a trace amount of human growth hormone secretion reporter as described previously (29). To determine the effect of the mutations on regulated exocytosis, we calculated the percent of secretion from cells expressing each construct (Fig. 9A). Overexpression of wild-type SNAP-23 had no effect on the extent of regulated exocytosis from RBL cells (data no shown). By contrast, overexpression of both SNAP-23 phosphorylation mutants resulted in statistically significant alterations in exocytosis as compared with cells overexpressing wild-type SNAP-23 (measured 15 min after triggering the cells with DNP-BSA). Whereas cells expressing the SNAP-23 Ser/Ala mutants showed a modest inhibition of exocytosis (23% inhibition relative to wild-type SNAP-23), the cells expressing the phosphomimetic Ser/Asp mutant showed a much more profound inhibition of exocytosis (35% inhibition relative to wild-type SNAP-23). Immunoblot analysis of the cells confirmed that each form of FLAG-SNAP-23 was expressed in the cells and the level of expression of each of these FLAG-tagged forms of SNAP-23 was approximately twice that of endogenous SNAP-23 (Fig. 9B). Co-immunoprecipitation studies confirmed that each mutant protein was able to interact with syntaxin 4 and no obvious differences in syntaxin binding between FLAG-SNAP-23 and these mutants were detected (data not shown). Because expression of either a SNAP-23 mutant that cannot be phosphorylated or a SNAP-23 mutant that appears constitutively phosphorylated inhibits exocytosis even in cells that constitutively express large amounts of endogenous SNAP-23, these data demonstrate that regulation of SNAP-23 phosphorylation is an important regulator of stimulated mast cell exocytosis.
Syntaxin 4 and VAMP-2 Are Preferentially Bound to Phospho-SNAP-23-SNAP-23 is known to bind to both syntaxin 4 and VAMP-2 and all three proteins are important for mast cell degranulation (18 -21). We have observed that only a small proportion of SNAP-23 is bound to syntaxin 4 and VAMP-2 in RBL mast cells, and we set out to determine whether triggering exocytosis altered syntaxin 4 or VAMP-2 binding to SNAP-23.
Triggering exocytosis for 10 min did not lead to any detectable change in the extent of SNAP-23 binding to syntaxin 4 or VAMP-2, with ϳ5% of all SNAP-23 binding to syntaxin 4 and VAMP-2 both before and after stimulation. Curiously, we found that after stimulation the majority of SNAP-23 bound to both syntaxin 4 and VAMP-2 was phosphorylated, because very little non-phosphorylated SNAP-23 could be detected in complex with either syntaxin 4 or VAMP-2 (Fig. 10). It should be pointed out, however, that even under these conditions the majority of phospho-SNAP-23 is not bound to syntaxin 4 (and vice versa). These data demonstrate that triggering mast cell degranulation promotes SNAP-23 phosphorylation and leads to the preferential association of syntaxin 4 and VAMP-2 with phospho-SNAP-23. Like SNAP-25 (34), SNAP-23 is unable to efficiently bind to VAMP in the absence of syntaxin family members, 2 suggesting that the preferred, but not exclusive, substrate for SNAP-23 phosphorylation may be the SNAP-23 present in the ternary SNARE complex. DISCUSSION Secretory granule fusion with the plasma membrane is the final step in a complex series of biochemical events that lead to the release of inflammatory mediators from mast cells. As in membrane/membrane fusion events in other cell types, members of the SNARE fusion machinery are thought to mediate regulated exocytosis in mast cells. SNAP-23 and syntaxin 4 have been proposed to be essential plasma membrane SNAREs (18 -20), whereas on the secretory granules themselves VAMP-2, VAMP-7, and VAMP-8 have been implicated in the granule/plasma membrane fusion process (18,19,21,35). In this study we have investigated changes in SNARE phosphorylation during mast cell exocytosis and studied the biochemical and functional consequences of SNARE phosphorylation in vivo. We find that SNAP-23 is phosphorylated in the RBL mast cell line and in bone marrow-derived mast cells upon Fc⑀RI receptor cross-linking, the physiological trigger for mast cell 2 N. Puri and P. A. Roche, manuscript in preparation.

FIG. 8. Approximately 10% of all SNAP-23 is inducibly phosphorylated in stimulated mast cells.
A, RBL mast cells were sensitized with DNP-specific IgE and mock-stimulated or stimulated with DNP-BSA for 10 min. The cells were lysed in Triton X-100 and immunoprecipitations were performed using control (preimmune) rabbit antiserum, an antiserum that recognizes total SNAP-23, or a phospho-SNAP-23-Ser 120 antiserum. Aliquots of each immunoprecipitate and the remaining supernatants of each immunoprecipitation reaction were analyzed by SDS-PAGE and immunoblotting using an antibody recognizing total SNAP-23 (upper and middle panels) or a phospho-SNAP-23-Ser 120 antiserum (lower panel). B, the relative band intensities of total SNAP-23 and phospho-SNAP-23-Ser 120 were quantified by densitometry and the amount of SNAP-23 phosphorylated was calculated as a percent of total amount of SNAP-23 present in each sample. The result is a mean Ϯ S.D. of four independent experiments and asterisks indicate statistically significant differences in the percentage of SNAP-23 phosphorylated in resting and DNP-BSA-stimulated RBL cells (***, p Ͻ 0.0005).
degranulation. Using site-directed mutagenesis and phosphopeptide mapping from cells labeled in vivo we have identified Ser 95 and Ser 120 as the two major phosphorylation sites of SNAP-23 in RBL mast cells, representing 75% of the induced SNAP-23 phosphorylation. These residues are located in the palmitoylated "linker" region separating the amino-and carboxyl-terminal coiled-coil domains of SNAP-23. Whereas Ser 95 is conserved in both rodent and human SNAP-23, Ser 120 is present in rodent, but not human, SNAP-23. Like syntaxin 4 in platelets (24) and RBL cells (Ref. 15 and this study), SNAP-23 is constitutively phosphorylated in unstimulated cells. Phosphopeptide mapping studies using SNAP-23 phosphorylation mutants confirmed that the basal phosphorylation site is not Ser 95 or Ser 120 , therefore unambiguous determination of the basal phosphorylation site(s) in SNAP-23 will require additional studies.
In our attempt to reveal a biological consequence of SNAP-23 phosphorylation, we overexpressed SNAP-23 phosphorylation mutants into RBL mast cells together with a human growth hormone-regulated exocytosis reporter (29). Expression of SNAP-23 S95A/S120A and SNAP-23 S95D/S120D phosphorylation mutants inhibited exocytosis from mast cells, although expression of the phosphomimetic SNAP-23 S95D/S120D phosphorylation mutant inhibited exocytosis to an even greater extent. These data clearly demonstrate that phosphorylation of SNAP-23 regulates the efficiency of regulated exocytosis, although the precise molecular mechanism governing this remains to be determined (as discussed below).
Most studies using in vitro phosphorylated SNAREs have all suggested that phosphorylation inhibits SNARE interactions (13)(14)(15)(16)(17). By contrast, we find that after stimulation essentially all of the SNAP-23 bound to syntaxin 4 and VAMP-2 is phosphorylated, clearly demonstrating that in vivo phosphorylation of SNAP-23 does not prevent its binding to these SNAREs. Curiously, we find that despite the preferential association of phospho-SNAP-23 with syntaxin 4 and VAMP-2, stimulation does not alter the absolute amount of SNAP-23 bound to these proteins. Indeed, co-immunoprecipitation experiments revealed that a large proportion of phospho-SNAP-23 is not bound to syntaxin 4, a finding that is consistent with our failure to detect changes in the total amount of SNAP-23 bound to syntaxin 4 after stimulation. Quantitative analyses revealed that in RBL mast cells only 5% of all SNAP-23 is bound to syntaxin 4 or VAMP-2 and less than 2% of syntaxin 4 is bound to SNAP-23 both before and after stimulation, demonstrating that the vast majority of SNAP-23, syntaxin 4, and VAMP-2 are either "free" or bound to other proteins (such as Munc18c (21)). We have found that, like SNAP-25 (34), VAMP-2 is unable to efficiently bind to SNAP-23 in the absence of syntaxin. 2 This is consistent with the idea that the SNAP-23 present in the ternary SNARE complex is the substrate for phosphorylation, however, additional studies will be required to examine this in greater detail.
During the course of our investigation of SNAP-23 phosphorylation in rodent mast cells Polgar et al. (17) published a study showing that SNAP-23 Ser 23 /Thr 24 and Ser 160 were phosphorylated when human platelets were activated. We found that while rat SNAP-23 Ser 161 was very efficiently phosphorylated by protein kinase C in vitro, quantitative analysis of multiple phosphopeptide maps revealed that phosphorylation of SNAP-23 Ser 161 represented only 1% of induced SNAP-23 phosphorylation. Furthermore, we did not observe differences in phosphorylation of a FLAG-tagged SNAP-23 Ser 23 /Thr 24 mutant relative to wild-type FLAG-tagged-SNAP-23 by onedimensional phosphopeptide mapping. We did examine the phosphorylation of activated mouse platelets using our phosphorylation state-specific antibodies and these studies clearly showed that SNAP-23 Ser 95 and Ser 120 are phosphorylated when mouse platelets are activated with thrombin. It is interesting to note that Polgar et al. (17) also observed SNAP-23 FIG. 9. Overexpression of SNAP-23 phosphorylation mutants inhibits exocytosis from RBL mast cells. A, RBL cells expressing FLAG-tagged wild-type SNAP-23, SNAP-23 S95A/S120A, or SNAP-23 S95D/S120D were co-transfected with a trace amount of human growth hormone secretion reporter plasmid as described in the text. The cells were sensitized with DNP-specific IgE and mock-stimulated or stimulated with DNP-BSA. Cell supernatants were harvested after 15 min and the extent of degranulation was determined by measuring the amount of human growth hormone released from the cells. The amount of secretion from cells in each experimental condition was expressed as a percentage of the total amount of human growth hormone present in the cells before stimulation. Data are from three independent experiments, and asterisks indicate statistically significant differences in the percentage of human growth hormone released from each mutant as compared to FLAG-tagged wild-type SNAP-23 (*, p Ͻ 0.05 and **, p Ͻ 0.0005). B, equivalent amounts of cell lysate from mock-transfected RBL mast cells or RBL cells expressing FLAG-tagged wild-type SNAP-23, SNAP-23 S95A/S120A, or SNAP-23 S95D/S120D that were mock-stimulated or stimulated with DNP-BSA were analyzed by immunoblot analysis using a SNAP-23 carboxyl terminus antibody. This antibody detects endogenous SNAP-23 and FLAG-SNAP-23 equally well. The mobility of endogenous rat SNAP-23 and transfected FLAG-SNAP-23 are indicated by arrows. A representative blot is shown. Ser 95 phosphorylation in vitro but did not identify the in vivo phosphopeptide by mass spectrometry. Whether the differences between the phosphorylation sites identified by us and by Polgar et al. (17) represent quantitative differences in the extent of phosphorylation of particular residues or differences between rodent platelets and human platelets remains to be resolved. Nevertheless, it is clear from our data that SNAP-23 Ser 95 and Ser 120 are the dominant inducible SNAP-23 phosphorylation sites in rodent mast cells and platelets.
It is interesting to compare and contrast the phosphorylation of SNAP-23 with that of its neuronal homolog SNAP-25. SNAP-25 is phosphorylated on Ser 187 when neuroendocrine cells (or pancreatic beta cells) are directly stimulated with phorbol myristate acetate (16,36,37), and phosphorylation of SNAP-25 alters binding to syntaxin in vitro (16). However, Ser 187 is located in one of the coiled-coil domains known to regulate SNAP-25/syntaxin association, so in this respect an effect on SNARE binding is not unexpected. In marked contrast to what we and others have observed with respect to SNAP-23 phosphorylation (this study and Ref. 17), the kinetics of SNAP-25 phosphorylation do not correlate with the kinetics of exocytosis either in regulated secretion from PC12 cells (37) or pancreatic beta cells (36). We find that inhibiting PKC activity with bisindolylmaleimide (I) prevents regulated SNAP-23 phosphorylation and mast cell exocytosis, whereas bisindolylmaleimide (I) prevented stimulus-induced SNAP-25 phosphorylation but had no effect on triggered exocytosis from either PC12 cells (37) or pancreatic beta cells (36). Such data leave the importance of SNAP-25 phosphorylation in regulated exocytosis an open question. Whereas SNAP-25 phosphorylation has clearly been observed when PC12 cells are cultured in nerve growth factor (38), when seizure activity and LTP are induced in hippocampal neurons (39), and when mice are chronically administered morphine (40), there is no direct evidence that these changes are a consequence of stimulated exocytosis. Indeed, we were unable to demonstrate phosphorylation of SNAP-25 in PC12 cells that were stimulated using physiological triggers for degranulation (41).
Kinetic analyses revealed a correlation between stimulusinduced mast cell degranulation and SNAP-23 phosphorylation, although whether phosphorylation occurs immediately prior to membrane fusion, concomitant with fusion, or immediately following membrane fusion could not be resolved by the techniques used here. Although we are unable to precisely identify the molecular mechanism by which SNAP-23 phosphorylation regulates exocytosis, we currently favor the view that phosphorylation primarily affects a post-fusion step in exocytosis, such as SNARE recycling. This is based primarily on our assumption that if phosphorylation were a prerequisite for exocytosis, then preventing phosphorylation (by overexpressing the SNAP-23 S95A/S120A phosphorylation mutant) would inhibit exocytosis more than simply introducing a phosphomimetic mutant (which could even augment exocytosis). This was not the case, suggesting that the two classes of phosphorylation mutants inhibited exocytosis by different mechanisms. Phosphorylation of SNAP-23 Ser 95 /Ser 120 is clearly important for exocytosis, as demonstrated by our results using the SNAP-23 S95A/S120A mutant. This is consistent with a vast literature on the importance of protein phosphorylation on function in a variety of systems. On the other hand, we have shown that after an initial burst of phosphorylation, SNAP-23 becomes dephosphorylated. Dephosphorylation could also have a specific role in exocytosis. If SNAP-23 dephosphorylation enhances SNARE complex disassembly, then overexpression of the phosphomimetic SNAP-23 S95D/S120D mutant could limit the recycling of SNAREs, thereby inhibiting exocytosis. There is clear evidence that efficient SNARE disassembly/reuse is important for exocytosis, as revealed in the phenotype of the temperaturesensitive mutations of NSF in Drosophila (42) and by introducing ATPase-deficient forms of NSF into mast cells (29). In both cases, inactivation of the SNARE disassembly enzyme, NSF, leads to an accumulation of SNARE complexes and to a cessation of regulated exocytosis. Consistently, our preliminary studies revealed that the absolute amount of SNARE complexes present before or after stimulation was very slightly higher in RBL cells overexpressing the SNAP-23 phosphomimetic mutant, a result that is consistent with this hypothesis (data not shown).
Whereas we did not observe dramatic alterations in the absolute amount of SNAP-23 bound to either syntaxin 4 or VAMP-2 when either SNAP-23 phosphorylation mutant was introduced into RBL mast cells, we did observe the selective inclusion of phospho-SNAP-23 into SNAP-23⅐syntaxin 4⅐VAMP-2 complexes after stimulation. While we favor a model in which triggering exocytosis stimulates the preferential phosphorylation of SNARE complex-associated SNAP-23, we cannot formally rule out the possibility that free SNAP-23 becomes phosphorylated and then displaces non-phosphorylated SNAP-23. In fact, although after stimulation most SNAP-23 bound to syntaxin 4 and VAMP-2 is phosphorylated, most phospho- FIG. 10. After mast cell stimulation most syntaxin 4-and VAMP-2-associated SNAP-23 is phosphorylated. A, RBL mast cells were sensitized with DNP-specific IgE and mock-stimulated or stimulated with DNP-BSA for 10 min. The cells were lysed in Triton X-100 and equal portions of each lysate were analyzed by immunoprecipitation using a control (preimmune) antiserum, an antiserum that recognizes total SNAP-23, or an antiserum that recognizes phospho-SNAP-23-Ser 120 . Aliquots of each immunoprecipitate were analyzed by SDS-PAGE and immunoblotting using antibodies that recognize rat syntaxin 4 (upper panels) or VAMP-2 (lower panels). B, the percent of all SNAP-23⅐syntaxin 4 complexes containing phospho-SNAP-23-Ser 120 was calculated by expressing the amount of syntaxin 4 in a phospho-SNAP-23-Ser 120 immunoprecipitate relative to the total amount of syntaxin 4 bound to all SNAP-23 in the sample. Each data point represents a mean Ϯ S.D. of four independent experiments. The percent of all SNAP-23⅐VAMP-2 complexes containing phospho-SNAP-23-Ser 120 was calculated by expressing the amount of VAMP-2 in a phospho-SNAP-23-Ser 120 immunoprecipitate relative to the total amount of VAMP-2 bound to all SNAP-23 in the sample. Each data point represents a mean Ϯ S.D. of two independent experiments. Analysis of the supernatant after each immunoprecipitation confirmed that the SNAP-23 antibody and phospho-SNAP-23-Ser 120 antibody removed essentially all SNAP-23 and phospho-SNAP-23-Ser 120 from the sample, respectively. SNAP-23 is not necessarily bound to syntaxin 4 or VAMP-2. Such an observation highlights the difficulty in identifying a specific molecular mechanism to explain the inhibition of exocytosis observed here, because only a small amount of SNAP-23, syntaxin, or VAMP are actually present in SNARE complexes at any given time. Although further analysis will be required to determine the precise mechanistic rationale, it is clear that phosphorylation of SNAP-23 is important for efficient regulated exocytosis from mast cells.