Research ArticleDEGRANULATION

Tomosyn functions as a PKCδ-regulated fusion clamp in mast cell degranulation

See allHide authors and affiliations

Science Signaling  03 Jul 2018:
Vol. 11, Issue 537, eaan4350
DOI: 10.1126/scisignal.aan4350

The right partner for secretion

Basophils and mast cells are immune cells that initiate allergic reactions and participate in host defense by releasing intracellular granules that contain proteases, lipid mediators, cytokines, and histamine. Degranulation requires the fusion of secretory granules with the plasma membrane, a process that is tightly controlled by SNARE family proteins. Madera-Salcedo et al. noted that, in basophils of allergy patients, serum IgE concentration correlated with the abundance of tomosyn, an inhibitor of SNARE activity. PKCδ-dependent phosphorylation of tomosyn caused it to swap its SNARE protein binding partner. Inhibition or loss of PKC increased mast cell degranulation. The authors suggest that the IgE-correlated expression of tomosyn may function as a feedback loop to limit mast cell degranulation and allergic symptoms in patients.

Abstract

Soluble N-ethylmaleimide–sensitive factor attachment protein receptor (SNARE) family proteins mediate membrane fusion critical for vesicular transport and cellular secretion. Mast cells rely on SNARE-mediated membrane fusion for degranulation stimulated by crosslinking of immunoglobulin E (IgE) bound to the Fcε receptor (FcεRI). We investigated the mechanisms downstream of receptor activation that control degranulation. We found that the SNARE binding protein tomosyn-1 (also known as STXBP5) inhibited FcεRI-stimulated degranulation of mast cells. After mast cell activation, tomosyn-1 was phosphorylated on serine and threonine residues, dissociated from the SNARE protein syntaxin 4 (STX4), and associated with STX3. We identified PKCδ as the major kinase required for tomosyn-1 threonine phosphorylation and for regulation of the interaction with STXs. Incubation with high IgE concentrations increased tomosyn-1 abundance in cultured mast cells. Similarly, in basophils from allergic patients with high amounts of serum IgE, the abundance of tomosyn-1 was increased as compared to that in patients with normal IgE concentrations. Our findings identified tomosyn-1 as an inhibitor of mast cell degranulation that required PKCδ to switch its interaction with STX partners during fusion. We suggest that the IgE-mediated increase in tomosyn-1 abundance in allergic patients may represent a counterregulatory mechanism to limit disease development.

INTRODUCTION

Membrane fusion is an energetically demanding process catalyzed by SNARE [soluble N-ethylmaleimide–sensitive factor attachment protein (SNAP) receptor] complex formation between vesicle and target membranes that enables the merger of lipid bilayers. SNARE complexes form, through their about 60–amino acid helical SNARE motif, an energetically favored tetrameric fusion complex composed of a vesicle (v)–bound v-SNARE protein, namely, VAMP (vesicular-associated membrane protein), and two target (t) organelle t-SNARE proteins, called SNAPs (such as SNAP-23 or 25) and syntaxin (STX). Although this process can take place spontaneously (1), it is highly regulated in living cells by accessory proteins that regulate SNARE complex assembly (24). At the synapse of neuronal cells, SNARE complexes are formed very rapidly with a SNARE cycle being completed in milliseconds (5), whereas in immune cells, this process occurs more slowly, allowing capture of stimulation-dependent SNARE complex formation by classical biochemical approaches (6, 7).

Tissue mast cells participate in immune defense mechanisms against microorganisms and parasites and in venom detoxification but can also promote allergies and other inflammatory disorders (810). Upon stimulation, they release inflammatory compounds stored in cytoplasmic secretory granules, followed by de novo synthesis of many cytokines and chemokines (4). Mast cell degranulation involves secretory granule membrane and plasma membrane fusion events (11) that are triggered by cross-linking of immunoglobulin E (IgE) bound to its Fcε receptor (FcεRI) with a specific antigen (Ag) (12). These fusion events are implemented by formation of complexes between SNARE proteins present at the plasma membrane such as the t-SNAREs STX4 and SNAP-23 and those localized at the vesicular membrane such as the v-SNARE VAMP8 and STX3 (7, 1316). They are regulated by several accessory molecules that act at different steps during membrane fusion including the Sec/Munc family protein Munc18-2 (1618), various Rab family members such as Rab3 (19), Rab27b (20), Rab 5 (21), and Rab12 (22), Munc13-4 proteins (23, 24), and secretory carrier-associated membrane proteins (25).

Less is known about how these membrane fusion regulators connect to upstream signaling events, with the notable exception that elevated intracellular calcium targets the calcium fusion sensors synaptotagmin II (26), complexin (27), and Munc13-4 (28). Furthermore, phosphatidylinositol 3-kinase promotes assembly of a Rab27b-Slp3-kinesin-1 complex driving anterograde secretory granule transport (29). Tomosyn binds STX and inhibits synaptic transmission and secretion in neurosecretory cells (3032) and may also prevent inadvertent membrane fusion in mast cells. Tomosyn also prevents vesicle fusion in lower organisms (3335) and in non-neuronal cells where it restricts insulin secretion (36) and endothelial cell exocytosis (37). However, tomosyn promotes secretion in pancreatic β cells (38) and platelets (37, 39), indicating that it may have a complex role in membrane fusion. Tomosyn is encoded by two related genes, tomosyn-1 (STXBP5) and tomosyn-2 (STXBP5L), each with multiple distinct splice variants (40). Tomosyn-1 is a large (~130 kDa) protein composed of a C-terminal v-SNARE domain and an N-terminal domain with 14 tryptophan-aspartic acid (WD) 40 repeats predicted to fold into two consecutive propeller-like structures forming interaction platforms (41). Both domains are important for inhibiting vesicle membrane fusion (31, 4244). Whereas tomosyn binding to STX and SNAP-25 involves both its v-SNARE and N-terminal domain and prevents access of the cognate VAMP partner (42, 45), the N-terminal region may also inhibit fusion by promoting fusion-incompetent ternary SNARE complex oligomerization and inhibition of the calcium sensor synaptotagmin (31, 43).

Here, we investigated the function of tomosyn-1 in mast cells and its connection to upstream signaling events. We demonstrated that tomosyn acted as a fusion clamp in mast cell degranulation. This function was dependent on threonine phosphorylation of tomosyn by protein kinase Cδ (PKCδ), which induced a switch in the interaction with cognate fusion proteins. Tomosyn abundance correlated with serum IgE titers in allergic subjects, which may represent an inhibitory mechanism to limit granulocyte secretory vesicle release.

RESULTS

Tomosyn-1 is expressed in mast cells and restricts exocytosis

A database search showed that mast cells express large amounts of tomosyn-1 (STXBP5) but not of tomosyn-2 (STXBP5L) mRNA (http://biogps.org; www.immgen.org). Reverse transcription polymerase chain reaction (RT-PCR) analysis indicated that the RBL-2H3 mast cell line and murine bone marrow–derived primary mast cells (BMMCs) expressed all tomosyn-1 splice variants including the long big (b) and the shorter medium (m) and short (s) forms (Fig. 1A). Immunoblotting for tomosyn (Fig. 1B) indicated the presence of this 130-kDa protein in BMMCs and RBL-2H3 cells. Confocal microscopy analysis showed a diffuse and punctuate distribution of tomosyn in the cytosol in resting and in stimulated BMMCs (Fig. 1C), as already reported in other cells (46, 47). These data indicated that mast cells express cytosolic tomosyn-1.

Fig. 1 Mast cells express tomosyn, which inhibits degranulation.

(A) RT-PCR analysis of STXBP5 mRNA splice variant expression in RBL-2H3 and BMMCs. The relative size of the big (b), medium (m), and short (s) isoforms are indicated. Images are representative of two independent experiments. (B) Western blot (IB) analysis for tomosyn in lysates from BMMCs and RBL-2H3 cells. Blots are representative of three independent experiments. (C) Confocal microscopy analysis of tomosyn distribution in BMMCs before and after PMA/ionomycin (ion) stimulation. Single optical section images are representative of three independent experiments. (D) Western blot analysis for tomosyn in lysates of RBL-2H3 cells transfected with control (siCtrl) and tomosyn siRNAs (si1Tomosyn, si2Tomosyn, and si3Tomosyn). Blots (top) are representative of at least three independent experiments. Tomosyn band intensities normalized to actin (bottom) are means ± SEM pooled from all experiments. (E) β-Hexosaminidase release after IgE/dinitrophenyl–human serum albumin (DNP-HSA) and PMA/ionomycin stimulation in RBL-2H3 cells transfected with control and tomosyn siRNA. Graphed data are means ± SEM from eight independent experiments per group. (F) Western blot analysis for tomosyn in lysates of BMMCs transfected with control and tomosyn siRNA. Blots (left) are representative of at least three independent experiments. Tomosyn band intensities normalized to actin (right) are means ± SEM pooled from all experiments. (G) Flow cytometry analysis of CD63 cell surface exposure on resting and IgE/DNP-HSA–stimulated BMMCs transfected with control and tomosyn siRNA. Contour plots are representative of four independent experiments. Graphed data are the mean percentage of CD63+ cells ± SEM pooled from all experiments. *P ≤ 0.05, **P ≤ 0.01, and ***P ≤ 0.001 by one-way analysis of variance (ANOVA) followed by Dunnett’s test (D and F) and two way-ANOVA followed by Bonferroni’s correction (E and G). Scale bars, 5 μm.

To assess the role of tomosyn-1 (thereafter called tomosyn) in mast cell degranulation, we used small interfering RNA (siRNA)–mediated silencing in RBL-2H3 cells. Efficient tomosyn knockdown (Fig. 1D and fig. S1A) did not affect the amount of STX3, STX4, and VAMP8 proteins (fig. S1) but enhanced IgE/Ag-induced mast cell degranulation (measured by β-hexosaminidase release) in dose-response experiments when compared to scramble siRNA, control (siCtrl)–transfected cells (Fig. 1E). Stimulation with phorbol 12-myristate 13-acetate (PMA)/ionomycin, which bypasses early FcεRI-induced signaling, similarly increased degranulation (Fig. 1E), suggesting that tomosyn acts at a late signaling step. The results were confirmed in primary BMMCs in which a robust (~70%) tomosyn knockdown (Fig. 1F) enhanced cell surface exposure of the degranulation marker CD63 upon cell stimulation (Fig. 1G). By contrast, tomosyn silencing had no effect on chemokine (C-C motif) ligand 2 (CCL2), tumor necrosis factor–α (TNF-α), and interleukin-6 (IL-6) secretion by BMMCs neither when stimulated with IgE/Ag nor with PMA/ionomycin (fig. S2). These experiments indicate that tomosyn inhibits preformed mediator release but does not affect release of de novo–synthesized cytokines/chemokines in mast cells.

Mast cell activation stimulates a switch in STX4 and STX3 complex formation

Mast cell degranulation involves formation of a complex between the v-SNARE VAMP8 and the t-SNAREs STX4 and SNAP-23 (7, 15, 16). Because tomosyn is known to interact with STXs, we evaluated its binding to STX4 in resting and stimulated BMMCs using a proximity ligation assay (PLA). Numerous fluorescent spots indicated a close interaction between tomosyn and STX4 in resting cells (Fig. 2A). In agreement with the localization of STX4 at the plasma membrane, the distribution of tomosyn–STX4 interactions indicated that they were majorly located close to the plasma membrane labeled with wheat germ agglutinin both in unstimulated and stimulated cells (fig. S3). Strikingly, however, their numbers largely diminished after PMA/ionomycin stimulation, indicating significant dissociation (Fig. 2A). This was confirmed in co-immunoprecipitation (co-IP) experiments after stimulation of BMMC with IgE/Ag (Fig. 2B). Because STX3 is largely located on secretory granules but translocates to the plasma membrane upon stimulation (16, 29), we evaluated tomosyn binding to STX3. The interaction of STX3 with tomosyn was minimal in resting BMMCs both in PLA and co-IP analysis but markedly increased after stimulation with PMA/ionomycin or IgE/Ag (Fig. 2, C and D). Colabeling with wheat germ agglutinin revealed that a significant fraction of the tomosyn–STX3 complexes localized within the cytoplasm both before and after stimulation (fig. S3). Because tomosyn preferentially binds to STX–SNAP-23 binary complexes (48, 49), we noted that tomosyn–SNAP-23 complexes were also detected (fig. S3). In unstimulated cells, similar to tomosyn–STX4 complexes, tomosyn–SNAP-23 complexes were mainly found close to the plasma membrane, whereas after stimulation, they were equally distributed between the cytoplasm and the plasma membrane, resembling the distribution of STX3–tomosyn complexes. Together, these data indicate that tomosyn differentially associates with STX3 or STX4 after stimulation of BMMCs. This switch in molecular complexes formed likely involves binding to binary complexes of STX4 and SNAP-23 in unstimulated cells and STX3–SNAP-23 in stimulated cells.

Fig. 2 BMMC stimulation induces a switch in tomosyn interaction with STX4 or STX3.

(A and C) Confocal analysis of tomosyn proximity to STX4 (A) or STX3 (C) determined by PLA in BMMC at the indicated times of BMMCs after PMA/ionomycin stimulation. Three-dimensional (3D) reconstitution images (left) of the proximity of the indicated targets (red) and cellular nuclei (blue) are representative of three independent experiments. Quantified data (right) are means ± SEM of at least 45 cells per group pooled from all experiments. (B and D) Co-IP of tomosyn with STX4 (B) or STX3 (D) from lysates of BMMC stimulated for the indicated times with IgE/DNP-HSA. Blots are representative of at least three independent experiments (top). Tomosyn band intensities normalized to STX (bottom) are means ± SEM pooled from all experiments. **P ≤ 0.01 and ***P ≤ 0.001 by one-way ANOVA followed by Dunnett’s test. Scale bars, 5 μm.

PKCδ regulates tomosyn interactions with STX partners in mast cells

Tomosyn isoforms contain multiple phosphorylation sites and are substrates of PKC (50) and PKA (50, 51). Both PKA (52) and PKCβ (5355) and PKCδ (54, 55) isoforms are involved in mast cell degranulation. We therefore hypothesized that the activity of these kinases in mast cells after FcεRI engagement may control the interaction of tomosyn with STXs. In agreement, we found that preincubation with inhibitors selective for either PKA (KT5720) or PKC [bisindolylmaleimide I (Bis I)] prevented the FcεRI-mediated decrease of tomosyn–STX4 and increase of tomosyn–STX3 interactions (Fig. 3, A and B). By mass spectrometry (MS) analysis of anti-tomosyn immunoprecipitates from unstimulated and stimulated BMMC lysates, we identified nine amino acid residues (five serine, three threonine, and one tyrosine residues) within the tomosyn sequence that could be phosphorylated, six of which (three Ser and three Thr) were phosphorylated after FcεRI triggering (Fig. 4A). All of these residues were in the long N-terminal region outside the C-terminal v-SNARE domain with several of these residues located in the previously identified hypervariable loop 2 region (Ser693, Thr704, and Ser760) described to play a regulatory role (47). A few of the identified residues (Tyr1007 and Thr1036) were located in the tail region preceding the v-SNARE domain reported to have an autoinhibitory role on STX binding in yeast homologs (41). By contrast, Thr794 is not conserved in rat and human species. We probed the phosphorylation status of tomosyn in immunoprecipitates from resting and IgE/Ag-stimulated BMMCs by Western blot for p-Ser and p-Thr residues. Only blotting for p-Thr revealed a stimulation-dependent increase that reached a maximum at ~9 min, which was coincident with the end of the degranulation response (Fig. 4B and fig. S4) and with the dissociation of tomosyn from STX4 and association with STX3 (compare Figs. 2 and 4B). The PKA inhibitor had no effect on this Thr phosphorylation, whereas the PKC inhibitor abrogated its increase after stimulation (Fig. 4C).

Fig. 3 PKA and PKC control the interaction of tomosyn with STX4 and STX3.

(A and B) Co-IP of tomosyn with STX4 (A) or STX3 (B) from lysates of BMMC treated either with vehicle (Veh), PKA inhibitor (PKAi), or PKC inhibitor (PKCi) for 15 min before stimulation for the indicated times with IgE/DNP-HSA. Blots (left) are representative of five independent experiments. Tomosyn band intensities normalized to resting cells (right) are means ± SEM pooled from all experiments. *P ≤ 0.05 and **P ≤ 0.01 by one-way ANOVA followed by Dunnett’s test.

Fig. 4 BMMC activation stimulates tomosyn Thr phosphorylation dependent on PKC activity.

(A) Mass spectroscopy analysis of tomosyn immunoprecipitates from BMMC stimulated for the indicated times with IgE/DNP-HSA. Possible phosphorylation sites were localized and compiled using the average normalized abundance of three independent experiments. (B) Western blot analysis for p-Thr on tomosyn immunoprecipitates from lysates of BMMC stimulated for the indicated times with IgE/DNP-HSA. Blots are representative of three independent experiments. p-Thr tomosyn band intensities normalized to tomosyn (right) are means ± SEM pooled from all experiments. (C) Western blot analysis for p-Thr on tomosyn immunoprecipitates from lysates of BMMC treated either with control, PKA, or PKC inhibitors for 15 min before IgE/DNP-HSA stimulation for the indicated times. Blots (left) are representative of three independent experiments. p-Thr tomosyn band intensities normalized to resting cells (right) are means ± SEM pooled from all experiments. *P < 0.05 and ***P < 0.001 by one-way ANOVA followed by Dunnett’s test (B and C).

Because PKCδ has been described as a negative regulator of mast cell degranulation (55), we investigated the role of PKCδ on tomosyn Thr phosphorylation. In addition, we found that the tomosyn dissociation from STX4 and the tomosyn association with STX3 were independent of calcium (fig. S5), supporting a role for a calcium-independent PKC isoform in this process. Transfection of BMMCs with two different siRNAs decreased PKCδ expression (~90%; Fig. 5A) and enhanced CD63 surface exposure after stimulation with IgE/Ag when compared to scramble siRNA Ctrls (siCtrls; Fig. 5B). Similarly, PKCδ knockdown reduced tomosyn Thr phosphorylation after activation of BMMCs by IgE/Ag when compared to siCtrl-transfected cells (Fig. 5C). Because the calcium-dependent PKCβ isozyme is essential in mast cell degranulation (53), we also examined the role of this kinase. Despite knockdown of this highly expressed kinase of only 30 to 40%, we found that BMMC degranulation was slightly inhibited after stimulation (fig. S6, A and B), consistent with previous studies (53). However, under these conditions, we found no change in tomosyn Thr phosphorylation as compared to siCtrl-transfected cells (fig. S6C). These data suggest that PKCδ but not PKCβ phosphorylates tomosyn on Thr residues. The knockdown of PKCδ also prevented tomosyn switching STX binding partners in PLA experiments (Fig. 5, D and E). Together, our data support an important role of PKCδ in regulating dissociation from STX4 and association with STX3.

Fig. 5 PKCδ-mediated tomosyn Thr phosphorylation controls its interaction with STX partners.

(A) Western blot analysis for PKCδ in lysates from BMMCs transfected with control (siCtrl) and PKCδ siRNAs (si1PKCδ and si2PKCδ). Blots (left) are representative of three independent experiments. PKCδ band intensities normalized to actin (right) are means ± SEM pooled from all experiments. (B) Flow cytometry analysis of CD63 cell surface exposure on resting and IgE/Ag-stimulated BMMCs transfected with control and PKCδ. Contour plots (right) are representative of four independent experiments. Graphed data are the mean percentage of CD63+ cells ± SEM pooled from all experiments. (C) Western blot analysis of p-Thr on tomosyn immunoprecipitates from lysates of BMMCs transfected with control or PKCδ siRNA stimulated with IgE/DNP-HSA for the indicated times. Blots are representative of six independent experiments. p-Thr tomosyn band intensities normalized to resting cells (right) are means ± SEM pooled from all experiments. (D and E) Confocal analysis of tomosyn proximity to STX4 (D) or STX3 (E) determined by PLA in control and PKCδ siRNA knockdown BMMCs at the indicated times after PMA/ionomycin stimulation. 3D reconstitution images (top) of the proximity of the indicated targets (red) and cellular nuclei (blue) are representative of three independent experiments. Quantified data (bottom) are means ± SEM of at least 15 cells per condition in each experiment. *P < 0.05, **P < 0.01, and ***P < 0.001 by Student’s t test (A) and two-way ANOVA followed by Bonferroni’s correction (B to D). Scale bars, 5 μm. n.s., not significant.

Tomosyn protein functions in blood basophils and is regulated by IgE

Similar to mast cells, FcεRI+CD123+CCR3+CRTH2+ human blood basophils (fig. S7A) express tomosyn-1, whereas tomosyn-2 is absent (www.immgen.org). Likewise, stimulation of basophils with anti-IgE induced tomosyn dissociation from STX4 and association with STX3 (Fig. 6, A and B). Basophil activation after incubation with anti-IgE was confirmed by enhanced CD63 surface exposure using immunofluorescence (fig. S7B). Because our data in mast cells suggested that tomosyn was a negative regulator of mast cell degranulation, we explored whether there were differences in the amount of tomosyn protein in basophils from nonallergic and allergic subjects (table S1) by intracellular flow cytometric staining in resting CCR-3+CD63 human basophils (fig. S7C). Although the amount of tomosyn protein did not differ between these groups (Fig. 6C), we detected a positive correlation between tomosyn abundance in basophils and total serum IgE concentration in allergic but not in nonallergic subjects (Fig. 6D). Because allergic patients generally display higher total IgE concentration (median level, 295.5 kIU/liter versus 80.5 kIU/liter), we divided allergic patients into groups with low total IgE titers (within the same range as that seen in nonallergic individuals, <485 kIU/liter = 1.18 μg/ml) and those with elevated total IgE concentrations (>485 kIU/liter). Reanalysis indicated that tomosyn protein abundance was significantly reduced in allergic patients who had low total IgE concentration when compared to those with high total IgE concentration (Fig. 6C). Tomosyn abundance in patients with elevated total IgE concentration was comparable to that of nonallergic subjects. Together, these results suggested that high IgE titers may increase cellular tomosyn amount. When we cultured BMMCs for 7 days with IgE, we found that high (5 μg/ml) concentrations of IgE increased the intracellular abundance of tomosyn (Fig. 7A). Increasing expression of tomosyn in BMMCs markedly inhibited degranulation as measured by CD63 expression, as expected (Fig. 7B). Together, our data support a model of tomosyn inhibitory function at multiple stages of the mast cell degranulation process (fig. S8).

Fig. 6 Stimulation of human basophils oppositely regulates the interaction of tomosyn with STX partners and effect of total IgE concentrations on tomosyn amount.

(A and B) Confocal analysis of tomosyn proximity to STX4 (A) or STX3 (B) determined by PLA in resting and anti-IgE–stimulated human basophils. 3D reconstitution images (left) of the proximity of the indicated targets (red), cellular nuclei (blue), and FcεRI (green) are representative of three independent experiments. Quantified data (right) are means ± SEM of at least 42 cells per group pooled from all experiments. (C) Flow cytometry analysis of tomosyn abundance in human basophils from allergic and nonallergic subjects. Data are the individual geometric mean fluorescence intensity (gMFI) of tomosyn normalized to isotype controls from 10 independent experiments. The median + interquartile range is indicated for healthy nonallergic, allergic, and allergic patients classified by serum IgE concentration. (D) Correlation of tomosyn abundance data in (C) with serum total IgE concentration in nonallergic and allergic subjects. The best-fit line and Spearman’s correlation analysis are indicated. **P < 0.01 by Mann-Whitney test. Scale bars, 2 μm.

Fig. 7 IgE increases tomosyn protein in mast cells, and transfection of tomosyn decreases mast cell degranulation.

(A) Flow cytometry analysis of tomosyn abundance in mast cells after culture for 7 days with indicated concentrations of IgE. Data are means ± SEM from at least three independent experiments. (B) Flow cytometry analysis of CD63 cell surface exposure on green fluorescent protein–positive (GFP+) resting and IgE/DNP-HSA–stimulated BMMCs transfected with GFP-control and GFP-tomosyn. Contour plots (left) are representative of four independent experiments. Graphed data (right) are the mean percentage of CD63+ cells ± SEM pooled from all experiments. *P < 0.05 by one way-ANOVA followed by Dunnett’s test (A) or two-way ANOVA followed by Bonferroni’s correction (B).

DISCUSSION

Tomosyn and tomosyn homologs in different organisms act as a fusion clamp of synaptic transmission, as membrane fusion events are enhanced in their absence (30, 3335). In endothelial cells and insulin-secreting cells (INS-1), tomosyn similarly acts as fusion clamp (36, 37, 39) by inhibiting binding of cognate SNARE proteins and generating fusion-incompetent ternary SNARE complexes (31, 32, 45). However, tomosyn also promotes exocytosis in an insulin-secreting cell line (38) and in platelets (37, 39), suggesting that tomosyn has complex functions. Although the mechanism of this positive action is presently unclear, tomosyn may interact with essential cytoskeletal elements or effectors of secretion such as Sec4, a Rab3 homolog (32, 39, 56, 57).

In mast cells, which express high amounts of STXBP5 mRNA, our present study indicates that tomosyn inhibits secretory granule secretion. Reduced amounts of tomosyn in mast cells enhanced degranulation, whereas its overexpression decreased degranulation. In contrast, the secretion of newly synthesized chemokines and cytokines was not altered by changes in tomosyn abundance. However, neosynthesized proteins use different vesicular carriers than prestored mediators in various immune cells, albeit little is known about the regulation of release from these compartments (4, 58). We also found that tomosyn dissociates from STX4 and associates with STX3 after mast cell stimulation. This reciprocal association was surprising because both STX4 and STX3 contribute to the membrane fusion process in mast cells (14, 16, 59, 60). STX3 in particular, because of its location on secretory granules, may play a role in secretory granule fusion, which is relevant for the multivesicular/compound mode of exocytosis in mast cells (11). Increased association with STX3 appears to contradict a simple fusion clamp model where SNARE-bound tomosyn prevents formation of fusogenic SNARE complexes, and stimulation relieves inhibition (30, 31, 37, 43, 48, 49).

We found that tomosyn–STX4 complexes were located mostly near the plasma membrane, whereas tomosyn–STX3 complexes were widely distributed between the cytoplasm and the plasma membrane. The subcellular distribution of tomosyn may be controlled by the distribution of its binding partners. Resting mast cells may lack granule-localized tomosyn–STX3 complexes because of the absence of SNAP-23 in cytoplasmic granules (13, 14), because it represents a crucial interaction partner in inhibitory SNARE complexes. Although distinct STXs can bind tomosyn alone, complex formation is strongly enhanced in the presence of SNAP-23 or SNAP-25 in adipocytes and adrenal cells (48, 49). Thus, although tomosyn may readily bind to binary complexes of SNAP-23 and STX4 in resting cells at the plasma membrane, it may only encounter binary STX3–SNAP-23 complexes after mast cell activation, which stimulates redistribution of STX3 (16, 29) and SNAP-23 (13). In support of binary complex binding, we demonstrated that tomosyn also interacted closely with SNAP-23. Similar to our observations with STX3, an increase in tomosyn–STX1A complexes after stimulation also occurs in bovine adrenal chromaffin cells (46). Formation of these tomosyn–STX1A complexes accompanies the relocation of tomosyn to the plasma membrane, where it forms core complexes with SNAP-25. However, the biological significance of enhanced complex formation of tomosyn with SNARE proteins remains unsolved. We propose that, in mast cells, the STX binding switch may function as a negative feedback mechanism that gradually blocks membrane fusion. This may ultimately limit mast cell degranulation, which can be life-threatening during an anaphylactic reaction. It is an intriguing possibility that the switch in tomosyn binding partners could, in a narrow time frame, both allow degranulation by alleviating plasma membrane fusion block and limit its extent by generating another intracellular fusion block slightly shifted in time that would prevent uncontrolled degranulation (fig. S8).

Phosphorylation of SNARE and accessory proteins represents an important mechanism to regulate membrane fusion (51, 6064). Our proteomic analysis of tomosyn immunoprecipitates identified nine different phosphorylation sites, six of which (three Ser and three Thr) were possibly enhanced after cell stimulation. We confirmed tomosyn Thr phosphorylation reaching its maximum at about the same time as the degranulation response, suggesting a possible mechanistic relationship (Fig. 4B and fig. S4). Our analysis using PKC and PKA inhibitors indicated that phosphorylation events are involved in tomosyn function because both inhibitors affected the switch in tomosyn binding to STX isoforms. However, only the PKC inhibitor affected tomosyn Thr phosphorylation, implying that PKA may phosphorylate tomosyn on Ser residues. In rat neuroblastoma cells, PKA-induced phosphorylation of tomosyn-1 on Ser724 reduces its binding to STX1, enabling cognate SNARE complex formation (51). Likewise, Akt phosphorylated tomosyn-1 on Ser783, which disabled binding to STX4 (63). Conversely, phosphorylation of STX1 by ROCK increased its affinity for tomosyn, thereby inhibiting formation of fusogenic SNARE complexes (46, 65). Whereas tomosyn-2 (also known as STXBP5L) is phosphorylated by both PKA and PKC, a phosphomimetic Ser-Asp mutant within the 11 identified Ser phosphorylation sites was unable to inhibit insulin secretion when overexpressed in INS-1. This was attributed to an enhanced tomosyn turnover and degradation (50). In neurons, cyclin-dependent kinase 5–dependent phosphorylation of tomosyn regulates the distribution of various vesicular pools, an activity that also involved the interaction with Rab3A and synapsin (66). Together with our own data, these studies show direct phosphorylation of tomosyn by multiple kinases reduces its inhibitory function.

Because blocking Thr phosphorylation by PKC reduces phosphorylation of tomosyn (5255), we analyzed which PKC isozyme was involved, focusing on the calcium-independent PKCδ, because both tomosyn dissociation from STX4 and tomosyn association with STX3 were calcium-independent. In contrast, calcium-dependent tomosyn-initiated blockade of synaptotagmin I limits neurotransmitter release (43). In mast cells, PKCδ is rapidly tyrosine-phosphorylated after stimulation (67) and plays both positive and negative regulatory roles in mast cell and platelet secretion (54, 55, 68, 69). This complexity is in agreement with the large number of possible substrates and functions of PKCδ (70). We found that the knockdown of PKCδ partially blocked Thr phosphorylation of tomosyn enhancing degranulation, as may have been predicted (55). Conversely, it also strongly affected the stimulation-dependent switch in STX binding, supporting a positive regulatory role in this function. Similar opposing functions have also been described for PKCβ, which overall is necessary for mast cell degranulation to occur (53) but which can also restrict fusion by phosphorylation of VAMP8 (64). However, PKCβ did not seem to be implicated in tomosyn phosphorylation.

Our data identified that PKCδ-mediated phosphorylation of tomosyn inhibits membrane fusion in mast cells. Mutation of tomosyn phosphorylation sites indicates that there are various consequences of tomosyn posttranslational modifications including promotion of both STX dissociation (51, 63) and association (46, 65), regulation of vesicular pool sizes (66), and decreasing tomosyn stability (50). Our present results in live mast cells indicated that inhibition of tomosyn phosphorylation or the absence of PKCδ both promoted dissociation from STX4 and association with STX3. Although the precise link between phosphorylation and tomosyn action remains to be fully resolved, our data suggest that posttranslational modification of tomosyn also alters its subcellular localization and inhibitory function. All identified tomosyn phosphorylation sites are located N-terminally from the v-SNARE domain, which contains the WD40 domains necessary for tomosyn inhibitory activity (31, 7173). Because STX4 is involved in the SNARE-mediated fusion process during mast cell degranulation (14, 16, 59), we propose that STX4-bound tomosyn acts as an initial fusion clamp to prevent formation of fusogenic SNARE complexes. The clamp may be relieved after stimulation by phosphorylation of tomosyn by PKCδ and PKA before gradual installment of a new fusion block that depends on tomosyn redistribution to STX3.

Our data indicate that activation of human basophils, which are closely related to mast cells (74), also promotes tomosyn dissociation from STX4 and association with STX3. In allergic subjects, the abundance of tomosyn in basophils positively correlated with total IgE titers, known to be elevated in many allergic patients (75). Prolonged incubation of cultured mast cells with high concentrations of IgE increased tomosyn abundance. IgE stimulation of mast cells can have statistically significant biological activities (76, 77). High IgE titers also protect against the severity of sting reactions in Hymenoptera venom allergy (78). Thus, the IgE-stimulated increase in tomosyn abundance may represent one mechanism that limits the deleterious effects of uncontrolled mast cell activation in allergic patients.

MATERIALS AND METHODS

Antibodies and reagents

Specific antibodies (Abs) used in these studies are listed in table S2. The PKC inhibitor Bis I was purchased from Cell Signaling Technology. The PKA inhibitor KT5720, PMA, and ionomycin were from Calbiochem. Murine IL-3 and stem cell factor (SCF) were from PeproTech. Inhibitor stock solutions were prepared in dimethyl sulfoxide and used at 10 μM for KT5720 and 4 μM for Bis I. The DNP-HSA used for stimulation of anti–DNP-IgE–sensitized cells was prepared as a stock solution in phosphate-buffered saline (PBS) at 10 μg/ml.

Cell culture

BMMCs from C57BL/6J mice (8 to 12 weeks old) were generated and maintained as described (7). FcεRI/c-kit–positive (>98%) cultures were used between 4 and 8 weeks. The rat RBL-2H3 mast cell line was maintained as described (79). Cells were either stimulated with PMA (20 nM) plus ionomycin (1 μM) or IgE-sensitized before addition of DNP-HSA (10 ng/ml) for cells for up to 9 min.

RT-PCR analysis

To analyze the presence of tomosyn splice variants in RBL-2H3 cells and in BMMC, the primers previously described (38) were used yielding products of 354, 246, and 195 base pairs corresponding to b–, m–, and s–tomosyn-1, respectively. Briefly, total RNA was extracted and reverse-transcribed using TRIzol (Thermo Fisher Scientific), and RT was carried out using the QuantiTect Reverse Transcription Kit (Qiagen). Standard PCR amplification was performed using Taq polymerase (Thermo Fisher Scientific) and the above oligonucleotides (38). Products were visualized after gel electrophoresis.

Gene modifications

Usually, four specific siRNAs were designed and assayed, together with a universal scramble siRNA. Sequence-specific siRNAs were used to knockdown tomosyn-1, PKCβ, and PKCδ (table S3). RBL-2H3 cells were transfected with annealed siRNAs (20 nM) by electroporation as described (16). For BMMCs, 1 × 106 cells in 0.1 ml were transfected with 50 or 100 nM siRNA in free-serum Iscove’s Modified Dulbecco’s Medium using a NEPA21 (NEPA GENE) electroporator (poring pulse: 275 V, 3-ms length, 50-ms interval, no. 2 with a 10% rate and + polarity; and transfer pulse: 20 V, 50-ms length, 50-ms intervals, no. 5 with a 40% rate and +/− polarity). After transfection, BMMCs were cultured in IL-3/SCF-containing medium and were assayed after 48 hours. Effective silencing was confirmed by Western blot. To express tomosyn in BMMC, the human tomosyn (73) was cloned in frame into the pEGFP-C1 (Clontech) to yield N-terminal enhanced GFP (eGFP)–tagged tomosyn. The eGFP-C1 vector was used as transfection control. One million cells in 0.1 ml of free-serum medium were transfected with plasmid (10 μg/ml) using a NEPA21 apparatus as above. Cells were assayed for CD63 surface amount after 48 hours.

Degranulation assays

Release of secretory granule content was determined by measuring β-hexosaminidase release (14). BMMC degranulation was evaluated by measuring CD63 surface abundance using flow cytometry as described (16).

Immunofluorescence and PLA

Immunofluorescence analysis of RBL-2H3 cells and BMMC was performed as described (80). PLA was performed using a Duolink in situ kit (Sigma-Aldrich). Cultured BMMCs (100,000 cells) were plated on coverslips coated with fibronectin (10 μg/ml; Sigma-Aldrich) for 45 min in MnCl2 (1 mM)–containing medium and maintained at 37°C in a humidified atmosphere. After PMA/ionomycin (20 nM:1 μM) stimulation for indicated time points, cells were fixed, quenched, permeabilized (16), and blocked for 1 hour at 37°C before incubation with primary Abs for 2 hours at room temperature. PLA was then performed according to the manufacturer’s instructions. Complexes were analyzed using an LSM 780 Zeiss confocal microscope (Carl Zeiss) at 63× steps of a 0.38-μm microscope. Images were processed with Imaris 8.1 software (Bitplane, Oxford Instruments), and the spots per nucleus were quantified. To evaluate membrane or cytoplasmic localization, cells were additionally labeled with Alexa Fluor 488–wheat germ agglutinin (Thermo Fisher Scientific) before the permeabilization procedure, and dots were evaluated for membrane-proximal or cytoplasmic localization using Imaris 8.1 software counting at least 30 cellular sections.

Immunoprecipitation

Ten million BMMCs at 1 × 106/ml per condition were IgE-sensitized overnight. Cells were washed and stimulated with DNP-HSA (10 ng/ml) for the indicated times. In some experiments, stimulation was performed in the absence of extracellular calcium after addition of EGTA (5 mM) 5 min before stimulation. Reaction was stopped by adding cold PBS followed by incubation with 0.25 mM N-ethylmaleimide for 15 min on ice. Cells were washed twice with PBS and lysed with 50 mM Hepes (pH 7.2), 1% Triton X-100, 150 mM NaCl, 5 mM KCl, 1 mM MgCl2, 1 mM sodium orthovanadate, and proteases inhibitors. IPs and immunoblot analysis with indicated Abs were performed as described (16).

Mass spectrometry

BMMC lysates from resting or IgE-stimulated cells were immunoprecipitated in triplicate with mouse anti-tomosyn before MS analysis using the Proteomic/Mass spectrometry Platform of the Institut Jacques-Monod. Spectra were converted to text files, searched against a target-decoy Mus musculus database, and downloaded from Uniprot using Mascot scores. Proteome Discoverer 2.1 (Thermo Fisher Scientific) was used to filter peptides with e >0.001.

Human blood samples

Studies were performed on samples from patients referred to the Medical School of Marseille for a history of airborne, food-induced, or drug-induced symptoms, which underwent allergy diagnostic or follow-up procedures according to current guidelines (8284). Nonallergic and allergic samples from 78 patients (table S1) were identified, retrospectively. Under French law, ethical approval and patient consent were unnecessary for this type of noninterventional study, although patients were informed and retained the right to oppose the use of their anonymized medical data for research purposes. This study was approved by the local ethics committee (Aix Marseille University, APHM).

Human basophil degranulation

Basophil enrichment from the blood of healthy human volunteers was performed using an EasyStep human basophil negative selection kit (StemCell Technologies). Purity was evaluated by determining the percentage of basophils expressing FcεRI, CRTH2, CCR3, and CD123. Cells (50,000) were then plated on coverslips and stimulated or not with anti-human IgE (0.5 μg/ml) for 9 min (80, 81). Basophils were fixed, quenched, permeabilized, blocked, and incubated with anti-CD203c and anti-CD63 for 2 hours at room temperature followed by a 3-min incubation with 4′,6-diamidino-2-phenylindole to stain nuclei. CD203c (basophil marker) and CD63 (degranulation marker) amount was evaluated by confocal microscopy using an LSM 780 Zeiss confocal microscope as detailed before.

Tomosyn abundance in human basophils

Basophil identification in EDTA-treated whole-blood samples drawn from nonallergic and allergic subjects was performed with anti–CCR3-phycoerythrin and anti–CD63–fluorescein isothiocyanate using Flow CAST reagents (Bühlmann Laboratories AG). For intracellular tomosyn staining, cells were permeabilized using a Cytofix/Cytoperm kit (BD Biosciences). Whole-blood basophils were incubated with primary rabbit anti-tomosyn or isotype control for 30 min at 4°C, followed by the second Ab. Acquisition was performed with a Canto II cytometer and a FACSDiva software (BD Biosciences), and data were analyzed using FlowJo software.

Serum total IgE quantification

Circulating levels of total IgE (kIU/liter, 1 IU = 2.44 ng/ml) were measured with the ImmunoCAP method (Thermo Fisher Scientific) according to the manufacturer’s instructions and current recommendations for certified clinical laboratories (85). Allergic subjects with low (in the range of nonallergics) and high total IgE titers taking IgE (485 kIU/liter) as cutoff (the mean of nonallergic IgE concentration +2 SD).

Statistical analysis

GraphPad Prism software (version 6) was used for statistical analyses.

SUPPLEMENTARY MATERIALS

www.sciencesignaling.org/cgi/content/full/11/537/eaan4350/DC1

Fig. S1. Tomosyn silencing in RBL-2H3 cells does not modify expression of SNARE proteins.

Fig. S2. Tomosyn silencing in BMMCs does not modulate IL-6, TNF-α, and CCL2 secretion.

Fig. S3. Plasma membrane and cytosolic localization of tomosyn complexes with STX4, STX3, and SNAP-23.

Fig. S4. Kinetic of β-hexosaminidase release in BMMCs.

Fig. S5. Tomosyn/STX4 dissociation and tomosyn/STX3 association do not depend on calcium mobilization.

Fig. S6. PKCβ is required for mast cell degranulation but does not phosphorylate tomosyn.

Fig. S7. Basophil gating strategy by flow cytometry and basophil activation by immunofluorescence.

Fig. S8. Proposed model for tomosyn action in mast cells.

Table S1. Patient characteristics.

Table S2. List of antibodies used in the study.

Table S3. List of siRNAs used in the study.

REFERENCES AND NOTES

Acknowledgments: We thank the Proteomics Core Facility at the Institut Jacques-Monod (T. Léger and C. Garcia) for the liquid chromatography–tandem MS (LC-MS/MS) experiments and the Region Ile-de-France (SESAME), the Paris-Diderot University (Agence Régionale de Santé), and CNRS for funding part of the LC-MS/MS equipment. Funding: This research project has been supported by the Investissements d’Avenir program (ANR-11-IDEX-0005-02), Sorbonne Paris Cite, and Laboratoire d’Excellence INFLAMEX and by an International Collaboration Grant between Agence Nationale de la Recherche (ANR) France (ANR-12-ISV3-0006-01) to U.B. and Conacyt Mexico (Conacyt-ANR 188565) to C.G.-E. and M.M.-S. The research project has also received funding from the Société Française d’Allergologie 2014 (to U.B.) and 2016 (to G.M.). Author contributions: I.K.M.-S., N.T., M.M.-S., C.G.-E., J.V., and U.B. conceived the project. I.K.M.-S., L.D., N.T., B.D., E.P., S.V., G.M., M.B., J.B., C.A., V.L., C.K., N.C., and J.V. performed the experiments. J.B., C.A., V.L., and C.K. provided resources. I.K.M.-S., M.B., J.V. and U.B. wrote the manuscript. Competing interests: The authors declare that they have no competing interests. Data and materials availability: The MS proteomics data have been deposited to the PRoteomics IDEntifications (PRIDE) database (accession number PXD010047). All other data needed to evaluate the conclusions in the paper are present in the paper or the Supplementary Materials.
View Abstract

Navigate This Article