Chaperoning of closed syntaxin-3 through Lys46 and Glu59 in domain 1 of Munc18 proteins is indispensable for mast cell exocytosis

Munc18-2 (also known as STXBP2) plays indispensable functions in immune cells. In humans, mutations of Munc18-2 have recently been associated with a fatal immune disorder, familial hemophagocytic lymphohistiocytosis type 5 (FHL5) (Cetica et al., 2010; Côte et al., 2009; Meeths et al., 2010; zur Stadt et al., 2009). In the immune tissue of these patients, there is a markedly reduced degranulation of lytic granules from cytotoxic T-lymphocytes and natural killer cells (Côte et al., 2009). In mice, Munc18-2 heterozygous knockdown results in significant impairment of mast cell degranulation (Kim et al., 2012). However, despite its importance, understanding the functions and its structural determinants of Munc18-2, which are necessary for immune cell exocytosis, has just begun recently (Bin et al., 2013).

In contrast, the precise roles of Munc18-1 (also known as STXBP1), a neuronal isoform of Munc18, have been studied extensively. The essential function of Munc18-1 in neuronal exocytosis (Verhage et al., 2000) is known to be mediated by its complex interactions with syntaxin-1 and/or the neuronal SNARE complex (Han et al., 2010). At least two important functions of Munc18-1 have been proposed: (1) priming of membrane fusion through direct interaction with the SNARE complex (Han et al., 2013; Han et al., 2014; Hu et al., 2010; Ma et al., 2013; Parisotto et al., 2014; Rodkey et al., 2008; Südhof and Rothman, 2009; Tareste et al., 2008; Xu et al., 2010); (2) molecular chaperone function of syntaxin-1 regulating appropriate intracellular localization and expression of syntaxin-1 (Arunachalam et al., 2008; Han et al., 2009; Malintan et al., 2009; McEwen and Kaplan, 2008; Medine et al., 2007; Rowe et al., 2001; Rowe et al., 1999; Shen et al., 2007). A similar chaperoning function (regulation of expression level as well as intracellular trafficking of its cognate syntaxin partner) of Munc18-2 has also been discovered; some FHL5 patients, who lack functional Munc18-2 protein, present with strongly reduced syntaxin-11 expression levels in their immune cells (Côte et al., 2009; zur Stadt et al., 2009). However, the underlying mechanisms governing such chaperoning function of Munc18-2 and its significance in the degranulation process remain unclear.

Therefore, two key questions remain unanswered about the Munc18-2 chaperoning function in mast cells: (1) What are the key cognate syntaxin isoform(s) that exhibit indispensable roles in mast cell degranulation and whose level and intracellular localization are regulated by Munc18-2? (2) Does the chaperoning activity of Munc18-2 directly depend on the binding of the Munc18-2 domain-1 cleft and those syntaxin(s) in the ‘closed’ conformation? We demonstrate here for the first time that loss of function in Munc18 impairs the expression level and intracellular localization of syntaxin-3. Furthermore, we show that such chaperoning activity of Munc18 is dependent on the direct binding between closed syntaxin-3 and the amino acid (aa) residues Lys46 and Glu59 in domain 1 of Munc18. Importantly, the functional contribution of these two residues in governing chaperoning activity towards syntaxin-3 is different compared with that of Munc18-1 and syntaxin-1 of neuronal exocytosis. Finally, strong and stable knockdown of syntaxin-3 results in vast deficits in mast cell degranulation. Collectively, our data invariably argue that the chaperoning activity of Munc18 towards syntaxin-3 is indispensable for the degranulation of mast cell.

We have previously found that a stable Munc18-1–Munc18-2 double-knockdown (Munc18-1/2 DKD) in RBL-2H3 mast cells causes severe defects in degranulation, accompanied with the specific downregulation of syntaxin-11 of the various syntaxin isoforms (syntaxin-2, -3, -4, -6, -7, -8 and -11) examined by us (Bin et al., 2013). However, a stable knockdown of syntaxin-11 mast cells generated by us did not exhibit significant impairment of mast cell degranulation. Thus, it is unlikely that the degranulation defects observed in our Munc18-1/2 DKD mast cells are due to the reduction of syntaxin-11 levels itself. Therefore, we hypothesized that – if one of the primary functions of Munc18 proteins (hereafter referred to as Munc18) is to mediate the chaperoning action toward the syntaxins – there must be another key cognate syntaxin isoform, whose level is regulated by Munc18. Munc18-1 and Munc18-2 are known to bind to syntaxin-2 and syntaxin-3 in addition to syntaxin-11 and neuronal syntaxin-1 (Riento et al., 1998; Tamori et al., 1998). Recent work suggested that syntaxin-3 potentially substitutes for the function of syntaxin-11 in syntaxin-11-deficient mast cells (Hackmann et al., 2013). Thus, we examined a possible change of syntaxin-3 levels in our Munc18-1/2 DKD mast cells, by using a recently generated rabbit monoclonal anti-syntaxin-3 antibody (EPR8543, Abcam). Immunoblot analysis revealed a vast reduction in the levels of syntaxin-3 within Munc18-1/2 DKD mast cells (Fig. 1A). Our quantification revealed that more than 60% of syntaxin-3 was decreased in the DKD cells when compared with control cells (P<0.05, Fig. 1B). However, when we tested again for syntaxin-3 levels with the previously used rabbit polyclonal anti-syntaxin-3 antibody (Synaptic Systems), we failed to observe any obvious change in the expression of syntaxin-3 (data not shown). We think that this may be because the latter antibody is crossreacting more to other proteins due to its polyclonal nature, which hampered us to detect the reduction of syntaxin-3 level in our previous study (Bin et al., 2013). We also confirmed our previous findings of the mild decrease in the levels of syntaxin-11 in Munc18-2 single knockdown (KD) mast cells and of a more drastic decrease in Munc18-1/2 DKD mast cells (Fig. 1A).

To examine whether the observed reduction in syntaxin-3 expression levels was specifically due to the knockdown of Munc18 proteins and not the result of any off-target small hairpin RNA (shRNA) effects, we re-expressed shRNA-resistant wild-type Munc18-1 or Munc18-2 fused to emerald GFP (EmGFP) again in Munc18-1/2 DKD mast cells by using a lentivirus mediated expression system (see Materials and Methods). We succeeded in finding full restoration of reduced syntaxin-3 levels in both wild-type (WT) Munc18-1–EmGFP- and Munc18-2–EmGFP-rescued mast cells, indicating the specificity of the syntaxin-3 downregulation phenotype in our Munc18-1/2 DKD mast cells (Fig. 1C). We also rescued syntaxin-11 expression levels in our wild-type Munc18-rescued cells once more (Fig. 1C). Our results provide strong evidence that, in addition to syntaxin-11, Munc18 plays a crucial role in the regulation of syntaxin-3 expression levels in mast cells.

Severe mislocalization of syntaxin-1 is observed following the single knockdown of Munc18-1 or the double knockdown of Munc18-1 and Munc18-2 in the neuroendocrine tumor PC12 cell line (Arunachalam et al., 2008; Han et al., 2011; Han et al., 2009). Similarly, in unc-18-null C. elegans, the anterograde trafficking and synaptic localization of UNC-64, an ortholog of syntaxin-1, is significantly perturbed (McEwen and Kaplan, 2008). In contrast, we have previously found that the endosomal localization of syntaxin-11 is unchanged in our Munc18-1/2 DKD mast cells (Bin et al., 2013). Because we hypothesize that Munc18 exhibits chaperoning activity toward syntaxin-3, we next asked whether the subcellular localization of syntaxin-3 could also be perturbed, in addition to its reduced expression, in the Munc18-1/2 DKD mast cells. Although syntaxin-2 expression levels were unchanged in our Munc18-1/2 DKD mast cells (Bin et al., 2013, see Fig. 1 therein), we tested for a possible alteration in syntaxin-2 localization because Munc18 proteins are known to physically interact with both syntaxin-2 and syntaxin-3. For a direct comparison between the two syntaxin isoforms in a systematic manner, we overexpressed hemagglutinin (HA)-tagged exogenous syntaxin-2 and syntaxin-3 in our control cells and in Munc18-1/2 DKD mast cells, and examined their intracellular localization by using anti-HA monoclonal antibody. HA tags were attached in tandem to the C-terminal end of these proteins (Fig. 2A), and we stably expressed them in control and Munc18-1/2 DKD mast cells (see Materials and Methods). Immunoblotting with anti–HA antibody revealed similar levels of overexpressed syntaxin-2–HA and syntaxin-3–HA in control and Munc18-1/2 DKD cells (Fig. 2B).

Upon immunostaining, we found that the syntaxin-2–HA had specifically localized to the plasma membrane (Fig. 2C, top panels), whereas syntaxin-3–HA was shown as punctate staining, suggesting its localization to secretory lysosomal granules (Fig. 2D, top panels). In addition, we found strong expression of syntaxin-3–HA at the tips of the mast cells. Upon Munc18-1/2 double knockdown, we found that such plasmalemmal localization of syntaxin-2–HA was unchanged (Fig. 2C, bottom panels). Strikingly, however, we observed severe perturbation of secretory lysosomal localization of syntaxin-3–HA in Munc18-1/2 DKD mast cells (Fig. 2D, bottom panels). Specifically, we found vast aggregation of syntaxin-3–HA in perinuclear regions of Munc18-1/2 DKD mast cells and its expression at the tips of the mast cells was largely reduced. This observation implies that Munc18 helps syntaxin-3–HA to translocate towards the tips of the mast cells and, therefore, in their absence, syntaxin-3–HA accumulates at the perinuclear region. Thus, our results indicate that Munc18 selectively regulates the intracellular localization of syntaxin-3 but not the plasmalemmal localization of syntaxin-2, even though it physically binds to both isoforms.

We then sought to reveal whether similar subcellular localization profiles can be seen for endogenous syntaxin-3 by staining the control and the Munc18-1/2 DKD mast cells with rabbit monoclonal syntaxin-3 antibody that we used for the immunoblotting. In accordance with the syntaxin-3–HA overexpression data (Fig. 2D), we found that endogenous syntaxin-3 was shown as discrete punctate pattern with particularly strong expression at the tips of elongated mast cells (Fig. 3A, top panels). Subsequently, for the Munc18-1/2 DKD mast cells, we were able to recapitulate the aggregation of syntaxin-3 in the perinuclear region and its absence at the tips of the cells (Fig. 3A, bottom panels). Also, we found that overall intensity for syntaxin-3 staining in the Munc18-1/2 DKD cells was much less compared with control, consistent with a lower level of syntaxin-3 expression in these cells, as analyzed by immunoblotting (Fig. 1A). Quantification analysis of confocal images revealed, compared with control, an overall syntaxin-3 intensity of ∼35% in Munc18-1/2 DKD (Fig. 3B, left panel). This reduction was similar to the reduction of syntaxin-3 expression levels in our immunoblot analysis (Fig. 1B). In addition, we found that most of syntaxin-3 was expressed at the tip of the control cells (70% at the tips versus 30% in the soma). However, we found syntaxin-3 strikingly – with nearly 80% – aggregated in the soma, whereas only 20% was found at the tips of Munc18-1/2 DKD cells (Fig. 3B, right panel). This result, therefore, further confirms that the localization of syntaxin-3 is strikingly perturbed by the Munc18-1/2 double knockdown in the mast cells.

Since re-introduction of either wild-type Munc18-1 or Munc18-2 into Munc18-1/2 DKD mast cells restored the reduced expression of syntaxin-3 (Fig. 1C), we next asked whether its perturbed localization can also be rescued. For this purpose, Munc18-1/2 DKD RBL-2H3 mast cells expressing EmGFP, wild-type Munc18-1–EmGFP or wild-type Munc18-2–EmGFP were co-stained for GFP and syntaxin-3 in corresponding rescued cells by using mouse anti-GFP monoclonal and rabbit anti-syntaxin-3 monoclonal antibodies; syntaxin-3 localization was then visualized using confocal microscopy (Fig. 3C). First we noticed that, in comparison to EmGFP-rescued cells, both WT Munc18-1–EmGFP- and WT Munc18-2–EmGFP-rescued cells exhibited brighter syntaxin-3 intensity (an ∼4-fold increase in intensity compared to that of EmGFP, see Fig. 7 for quantification), consistent with their restored expression observed in the immunoblots. (Fig. 1C). In addition, we found that the strong syntaxin-3 expression in the tips re-appeared in both Munc18-1–EmGFP- and Munc18-2 WT–EmGFP-rescued cells, but not in EmGFP-rescued cells. Importantly, we found that the perinuclear aggregation of syntaxin-3 seen in EmGFP-rescued cells was largely alleviated in both Munc18-1–EmGFP- and Munc18-2 WT–EmGFP-rescued cells (Fig. 3C). Quantification analysis revealed that almost 80% of syntaxin-3 was found to be aggregated in the soma of EmGFP rescued cells, yet its strong expression pattern at the tips was restored in both Munc18-1–EmGFP- and Munc18-2–EmGFP-rescued cells with nearly 60% of syntaxin-3 in the tips versus ∼40% in the soma (Fig. 7). Collectively, our results strongly advocate that Munc18 functions as a chaperone of syntaxin-3 by regulating both its expression levels and intracellular trafficking. In addition, we illustrate here that Munc18-mediated syntaxin trafficking is not only present in neuronal cells (Arunachalam et al., 2008; Han et al., 2011; Han et al., 2009; McEwen and Kaplan, 2008) but also in mast cells.

We found that in the Munc18-1/2 DKD mast cells syntaxin-3 aggregated near the nucleus. Since we hypothesize that Munc18 acts as a molecular chaperone by mediating the trafficking of syntaxin-3, we asked whether such aggregation occurs in the Golgi complex or compartments of the endoplasmic reticulum (ER) of Munc18-1/2 DKD cells. To test for syntaxin-3 localization in the Golgi complex, we co-stained for syntaxin-3 with two different Golgi markers – GM130 for cis-Golgi and TGN38 for the trans-Golgi network – in the Munc18-1/2 DKD cells and analyzed their localizations using confocal microscopy (supplementary material Fig. S1A). We found that aggregated syntaxin-3 does not colocalize with either of the Golgi markers. Then we asked whether aggregation occurs in the ER compartment by staining cells with anti-calnexin antibody (supplementary material Fig. S1B). Since this antibody was made against rabbit calnexin, we were unable to use the anti-syntaxin-3 antibody for co-staining. However, we noticed that calnexin expression was different compared with the pattern of aggregated syntaxin-3, being more spread throughout the cells unlike the aggregated syntaxin-3 pattern in the Munc18-1/2 DKD cells. Therefore, we conclude that, although we see perinuclear syntaxin-3 aggregation, syntaxin-3 does not occur in the ER or the Golgi compartments of the Munc18-1/2 DKD mast cells.

Binding of Munc18-1 to monomeric closed syntaxin-1 is mediated by interactions of many residues, as revealed by X-ray crystallography (Misura et al., 2000). Among those residues, we have previously shown that Lys46 and Glu59 in the domain 1 of Munc18-1 play a crucial role in governing the binary interaction (interaction between Munc18 and monomeric syntaxin). We found that either of the two point mutations (i.e. K46E or E59K) significantly disrupted the binding to closed syntaxin-1, whereas the K46E/E59K double mutation resulted in complete abolishment of interaction, as seen using yeast two-hybrid assays and isothermal titration calorimetry (ITC) (Han et al., 2011; Han et al., 2009). Lys46 and Glu59 are conserved not only between Munc18-1 and Munc18-2 but also in UNC-18 of C. elegans and Rop in Drosophila melanogaster (Fig. 4A). Also, a homology model of Munc18-2 in complex with monomeric syntaxin-3 based on the crystal structure of Munc18-1–syntaxin-1A suggested that these two residues are in close proximity to Asp230 and Arg114 of closed syntaxin-3, possibly mediating electrostatic interactions (Fig. 4B). Therefore, we next investigated whether mutation of these residues in Munc18-1 and Munc18-2 would affect the binary interactions (if so, how much) with syntaxin-2 and syntaxin-3 in yeast two-hybrid assays (Fig. 4C–F).

We expressed wild-type and the binary interaction mutants (K46E, E59K and K46E/E59K) of rat Munc18 as bait and the cytoplasmic domains of rat syntaxin-2 and syntaxin-3 as prey (see Materials and Methods). When we tested the binding of the mutants to monomeric syntaxin-2 and syntaxin-3, we found significant disrupting effects of mutations K46E and E59K on the binary interactions (Fig. 4C–F). However, we found intriguing differences between syntaxin-2 and syntaxin-3. Strikingly, even the single mutation of K46E or E59K in Munc18-1 or Munc18-2 completely abolished the binding to syntaxin-3, and double mutation of K46E/E59K did not exhibit any additive effects (Fig. 4C,D). In contrast, the single binary interaction mutants of Munc18-1 and, to a lesser extent, those of Munc18-2 retained some ability to bind syntaxin-2. However, as observed in syntaxin-3, the double mutants of Munc18-1 or Munc18-2 similarly abrogated the binding to syntaxin-2 (Fig. 4E,F).

Based on our homology model illustrating Munc18-2 bound to closed syntaxin-3 (Fig. 4B), we speculated that the wild-type syntaxin-3 adopts a closed conformation upon binding to Munc18. To verify that, indeed, wild-type syntaxin-3 adopts a closed conformation and that Munc18-2 primarily binds to this closed conformation, we examined whether binding between Munc18-2 and syntaxin-3 is significantly reduced when syntaxin-3 adopts an open conformation. In a previous NMR study of syntaxin-1A (Dulubova et al., 1999), mutations (L165A/E166A) in a linker region between its N-terminal Habc domain and its SNARE domain were found to cause this protein to open. These residues are highly conserved from nematode to human; similar mutations (L165A/E166A) in the syntaxin-1 ortholog UNC-64 also cause this protein to adopt an open conformation (Richmond et al., 2001). Therefore, we generated a syntaxin-3 mutant that has an open conformation (L165E/E166A, hereafter referred to as LE open mutant) and then tested its interaction with both isoforms of Munc18 in yeast two-hybrid assays. We found that, compared to wild-type syntaxin-3, the syntaxin-3 open mutant merely exhibited ∼10% of binding to either Munc18-1 or Munc18-2 isoforms (Fig. 4G). This indicates that the interaction of wild-type syntaxin-3 with Munc18 is mainly governed by syntaxin-3 in a closed conformation. Importantly, the fact that K46E and E59K alone completely abolished the interaction with wild-type syntaxin-3 indicates that these two chaperoning mutants play an essential role in mediating binary interaction with closed syntaxin-3 (Fig. 4C,D). In addition, although the binding mode of Munc18-1 and Munc18-2 towards closed syntaxin-2 seems similar to the binding mode of Munc18-1 and closed syntaxin-1 in neuronal exocytosis, as previously reported (Han et al., 2011; Han et al., 2009), their interactions with closed syntaxin-3 are substantially different, in that even a single mutation is enough to disrupt such interaction.

To test the functional outcome of these binary interaction mutations with regard to the regulation of syntaxin-3 expression levels, localization and mast cell degranulation, we stably expressed the EmGFP-tagged mutants of Munc18-2 (K46E and E59K single mutants, and K46E/E59K double mutant) in Munc18-1/2 DKD mast cells together with wild-type and examined their ability to restore compromised syntaxin-3 expression level, subcellular localization as well as defective degranulation. Immunoblot analysis illustrated that all mutants were expressed at levels similar to that of wild-type Munc18-2, implying that there were no major folding issues caused by these point mutations (Fig. 5A). However, we found that none of the binary interaction mutants of Munc18-2 was able to restore the reduced syntaxin-3 expression level of the Munc18-1/2 DKD mast cells (Fig. 5A). We then examined whether the mislocalization of syntaxin-3 seen in the Munc18-1/2 DKD mast cells can be rescued by the mutants (Fig. 5B). Consistent with the reduced expression of syntaxin-3, we found that overall staining intensities were substantially less in the each of the mutant-rescued cell in comparison to the wild-type-rescued cells. In fact, quantification of signal intensities showed similarly less intense signals in Munc18-2 EmGFP-tagged mutant-rescued cells and EmGFP-rescued cells (Fig. 7A). Strikingly, we observed expression of syntaxin-3 in the binary-interaction-mutant-rescued cells in an aggregated manner and failed to see syntaxin-3 expression at the tips of these cells (Fig. 5B, Fig. 7B for quantification). Finally, we tested the ability of these mutants to rescue the defective degranulation phenotype of Munc18-1/2 DKD mast cells by measuring release of β-hexosaminidase (Fig. 5C). This protein resides inside of secretory granules and/or lysosomes of mast cells and we triggered its release using immunoglobulin E (IgE) and 2,4-dinitrophenyl hapten conjugated to human serum albumin (DNP-HSA) or ionomycin as two different stimulation methods (Bin et al., 2013). IgE-DNP antibody is recognized by the high-affinity IgE receptor (FcεRI), a cell surface receptor of mast cells. Upon cross-linking its antigen DNP-HSA, signaling pathways eventually lead to the increase in intracellular Ca²⁺ levels, resulting in degranulation. Ionomycin is a potent Ca²⁺ ionophore that directly induces Ca²⁺ influx, thus triggering the degranulation. Strikingly, we found that none of the Munc18-2 binary mutants were able to rescue β-hexosaminidase secretion in response to either stimulation (Fig. 5C), indicating that the degranulation defects lie within the fusion machineries of mast cells. In accordance with the results of our yeast two-hybrid assays (Fig. 4D), we did not see any additive effects of the double mutant (K46E/E59K) over the single mutants (K46E or E59K), suggesting that even a single mutation is detrimental, i.e. abolishes the functional outcome of the binary interaction towards syntaxin-3 in the mast cells.

We then performed additional experiments to investigate the effect of the same mutations in Munc18-1 on the chaperoning of syntaxin-3 and the mast cell degranulation (Fig. 6). We have previously found that the ability of Munc18-1 to rescue the phenotype of Munc18-1/2 DKD PC12 cells is correlated to their ability to bind syntaxin-1. We observed that the K46E and E59K single mutants retained some binding to syntaxin-1 and, thus were able to restore β-hexosaminidase secretion to 70% compared to the wild-type rescue, whereas the double mutant lost its ability to bind to syntaxin-1 and, therefore, rescued secretion at the level of merely 10% compared to wild-type (Han et al., 2011). In addition, synaptic transmission of hippocampal neurons that lack Munc18-1 was relatively well rescued by the E59K single mutant, which exhibited EPSC size that was comparable to that of wild-type (Meijer et al., 2012). However, in the rescue of the Munc18-1/2 DKD mast cells, we found that – as for Munc18-2 – even the single mutations of K46E or E59K of Munc18-1 were unable to restore reduced syntaxin-3 expression levels, its mislocalization or the impaired degranulation (Fig. 6 and Fig. 7C,D for quantification). Thus, in the case of the mast cells, the ability of Munc18-1 and Munc18-2 to rescue the loss of syntaxin-3 chaperoning and defective degranulation in Munc18-1/2 DKD cells was strikingly different from the results of Munc18-1 and syntaxin-1 interaction in neuronal secretion. Also, this functional phenotype of the binary interaction mutants was well reflected on their ability to directly bind to syntaxin-3, which we saw in our yeast two-hybrid assays (Fig. 4C).

If the chaperoning of syntaxin-3 by Munc18-1 and Munc18-2 is functionally important in mast cell degranulation, we would anticipate that this syntaxin isoform plays an important independent role in exocytosis. To test this hypothesis, we generated two stable RBL-2H3 mast cell lines whose endogenous syntaxin-3 was knocked down (syntaxin-3–5 KD and syntaxin-3–8 KD, see Materials and Methods). Immunoblotting with rabbit monoclonal anti-syntaxin-3 antibody revealed some reduction in the level of syntaxin-3 in both cell-lines, but syntaxin-3–8 KD resulted in a stronger knockdown than syntaxin-3–5 KD (Fig. 8A,C). Also, we did not observe any obvious changes in the expression of syntaxin-2 and syntaxin-11 in cells subjected to the stronger syntaxin-3 knockdown, indicating that these two isoforms were not upregulated (Fig. 8A). When performing degranulation assays using IgE- and ionomycin-dependent stimulations, we found significant degranulation defects in both syntaxin-3 KD mast cells (Fig. 8B,D). However, the degree of degranulation defect was correlated to the level of syntaxin-3 reduction seen in two different syntaxin-3 KD cell lines. We saw modest, yet statistically significant, reduction in degranulation in syntaxin-3–5 KD cells. However, we found a reduction greater than 4-fold (∼46% versus ∼11%) in IgE-induced degranulation, whereas ionomycin-induced degranulation showed a 2–3-fold reduction (∼35% versus ∼12% for 0.5 µM and ∼45% versus ∼10% for 2.5 µM) in syntaxin-3–8 KD mast cells compared to the control. This result suggests that syntaxin-3 plays an important independent role in mast cell degranulation, and the severe secretion defects seen in the Munc18-1/2 DKD mast cells might be attributed to the lack of syntaxin-3 chaperoning.

In the present study, we examined the role of binary interaction between domain 1 of Munc18 and its cognate syntaxin isoforms in the mast cell exocytosis. This interaction mode has been regarded as of importance in neuronal secretion, as the mutations in Munc18-1 that disrupt such interaction have demonstrated lack of chaperoning activity towards syntaxin-1 thus resulting in compromised neuronal exocytosis (Han et al., 2011; Han et al., 2009). However, no attempts have been made to study the binary interaction mode in the mast cell exocytosis, despite the fact that similar Munc18-regulated secretion processes take place. We demonstrated, for the first time, the crucial role of the binary interaction between Munc18 and syntaxin-3 in mast cell degranulation.

Importantly, we have shown that Munc18 performs an essential role in chaperoning (i.e. regulating the expression levels as well as intracellular trafficking) of its cognate partner syntaxin-3. We saw that loss of Munc18 function in mast cells impairs the expression levels of syntaxin-3 (Fig. 1A,B). These decreased syntaxin-3 levels were completely restored upon re-expression of Munc18-1 or Munc18-2, illustrating the specificity of such phenotype (Fig. 1C). Furthermore, our finding of mislocalized syntaxin-3 in the perinuclear region of Munc18-1/2 DKD mast cells, which was rescued by re-introduction of wild-type Munc18 (Figs 2, 3), provides additional strong evidence that Munc18 plays a crucial role in chaperoning syntaxin-3. In an attempt to investigate an independent role of syntaxin-3 in mast cell degranulation, we generated stable syntaxin-3 KD mast cells, which exhibited strong impairments in both IgE- and ionomycin-dependent degranulation (Fig. 8). This is different compared with our previous attempt to see the role of syntaxin-11 in the mast cell exocytosis, where strong knockdown of endogenous syntaxin-11 did not result in any defect of degranulation (Bin et al., 2013). Our new findings are in accordance with the previous studies suggesting the importance of Munc18-2 and syntaxin-3 in mast cell degranulation (Brochetta et al., 2014; Tadokoro et al., 2007). Thus, we now state that, although Munc18 proteins regulate both syntaxin-3 and syntaxin-11 isoforms in mast cells, the former interactions provide more functional contributions to the mast cell exocytosis.

Recent analysis of Munc18-2 and syntaxin-3 in RBL-2H3 cells illustrates the significant function of these two proteins in the mast cell degranulation (Brochetta et al., 2014). In that study, the authors separately knocked down Munc18-2 and syntaxin-3 RBL-2H3 by using small interfering RNA (siRNA) and found substantial reduction of degranulation within both these KD cell lines. However, the downregulation of syntaxin-3 expression levels or perturbation of its intracellular localization was not found in Munc18-2 KD cells, leading to the conclusion that Munc18-2 and syntaxin-3 participate in distinct steps of mast cell degranulation that function independently of each other. Our results, however, clearly indicate that, in Munc18-1/2 DKD RBL-2H3 cells, levels of syntaxin-3 are profoundly reduced – accompanied with its severe mislocalization – in addition to a defective degranulation phenotype. Moreover, the fact that this phenotype was completely rescued following the re-introduction of wild-type Munc18 but not of its binary interaction mutants provide strong evidence that Munc18 participates in the process of mast cell exocytosis by regulating syntaxin-3 expression levels and trafficking. One possible difference between the study by Brochetta et al. and our current study could be that Munc18-2 single KD cells were used in the former study, whereas we used Munc18-1/2 DKD mast cells. The fact that wild-type Munc18-1 can rescue the phenotype of Munc18-1/2 DKD cells indicates that Munc18-1 and Munc18-2 are functionally redundant (Figs 1C, 3, 5, 6). Therefore, in Munc18-2 single KD cells, the remaining Munc18-1 is likely to overtake the role of missing Munc18-2 and is, thus, still providing adequate chaperoning activity toward syntaxin-3.

Although we found that the syntaxin-3 knockdown caused strong impairments in the mast cell degranulation (Fig. 8), we cannot fully attribute the almost complete abolishment of degranulation found Munc18-1/2 DKD mast cells to the reduction in syntaxin-3 levels alone. Moreover, we have previously observed that even Munc18-2 single KD mast cells exhibit severe degranulation defects (Bin et al., 2013). But, in our present study, we show that the level of syntaxin-3 was unchanged in Munc18-2 single KD mast cells (Fig. 1A). Thus, we speculate that there are other possibilities through which Munc18 participates in the mast cell degranulation in addition to the chaperoning of syntaxin-3. For example, Munc18 might regulate other syntaxin isoforms including syntaxin-11 and/or might directly interact with the SNARE complex, thereby controlling the priming step of the exocytosis. Further analyses to test the veracity of these possibilities will provide a clearer view on how Munc18 is involved in mast cell exocytosis.

In contrast to neuronal exocytosis, where fusions occur predominantly between the plasma membrane and the secretory vesicle, mast cells have been shown to undergo compound fusions (multivesicular fusions) allowing substantial release of contents in response to minimal stimulation (Blank and Rivera, 2004). To see the possible role of Munc18 in the compound granule fusions, we performed total internal reflection fluorescence microscopy (TIRFM) with control and Munc18-1/2 DKD RBL-2H3 cells expressing neuropeptide Y (NPY) tagged to the pH-sensitive GFP pHluorin (NPY–pHluorin) (Zhu et al., 2007). In this setting, we can only see the fluorescent signals from the fusion of granules and, by analyzing the intensity as well as radius of fluorescent signals, we can distinguish primary fusion (single granules fusing with the plasma membrane) and compound fusion. After simultaneous stimulation with ionomycin and PMA (phorbol ester analogue), we found rapidly brightening fluorescent signals as a result of fusions of granules in the control cells. We saw that most fusion events were primary fusions, less than 10% of total fusion events were compound fusions. When we performed the same experiments on Munc18-1/2 DKD cells, we found much fewer granule fusions than in the control cells (data not shown). Therefore, we conclude that primary fusion is the main type of fusion event in the RBL-2H3 cells, which was greatly inhibited in the Munc18-1/2 DKD cells.

In neuronal exocytosis, a single binary interaction mutation is not sufficient to disrupt binding between Munc18-1 and monomeric syntaxin-1 (Han et al., 2011; Malintan et al., 2009). However, our present study illustrates that a single binary interaction mutant (K46E or E59K) of either Munc18-1 or Munc18-2 is as detrimental as the double mutant (K46E/E59K) in both binding closed syntaxin-3 (Fig. 4) and rescuing the phenotype of the Munc18-1/2 DKD mast cells (Figs 5, 6). These results are strikingly different from the binary interaction between Munc18-1 and syntaxin-1, and their functional implications. In neuronal exocytosis, compared to wild-type, single binary interaction mutants exhibit only marginally reduced ability to rescue the defective secretion phenotype of Munc18-1/2 DKD PC12 cells and the synaptic transmission of hippocampal neurons that lack Munc18-1 (Han et al., 2011; Han et al., 2009; Meijer et al., 2012). Therefore, our results clearly indicate an important difference in the functional contribution residues Lys46 and Glu59 have on their own in mediating the binary interaction in immune cell compared with neuronal exocytosis. We speculate that, although the importance of the binary interaction between Munc18 and cognate syntaxin partners is well-conserved in both neuronal and immune cell exocytosis, the binding affinity and, thus, its functional outcome are different between the two types of cell. Since even a single mutant was able to fully abrogate the binary interaction between Munc18 and syntaxin-3, we hypothesize that the binary interaction between Munc18 and cognate syntaxin in immune cells is weaker than that in neuronal cells. Therefore, if even one single residue is altered, the whole binary interaction would be abolished because binding by the other residue is not strong enough to sustain the interaction on its own. This theory fits well with our previous finding that the interaction between the N-peptide of syntaxin and the hydrophobic pocket of Munc18 is also functionally important in mast cell degranulation (Bin et al., 2013). Because the binary interaction itself is weak, binding between the Munc18 hydrophobic pocket and the syntaxin N-peptide may work in concert to maintain interaction of Munc18 with its syntaxin partners in mast cells. Thus, unlike neuronal exocytosis, where the latter interaction mode is regarded to be of limited function (Malintan et al., 2009), both modes of interactions are equally valuable molecular actions between Munc18 and syntaxin that are essential for mast cell exocytosis.

A number of point mutations in Munc18-2 were recently identified to be implicated with FHL5, a lethal immune disease during which affected patients' innate immune players, such as natural killer cells and cytotoxic T-lymphocytes are compromised in their degranulation capability (Cetica et al., 2010; Côte et al., 2009; Meeths et al., 2010; zur Stadt et al., 2009). Based on our results illustrating the importance of the chaperoning function of Munc18 toward syntaxin-3 in degranulation, we speculate that those Munc18-2 mutants that are associated with FHL5 are incompetent in the regulation of its chaperoning protein, such as syntaxin-3, to govern correct expression levels as well as intracellular trafficking. Thus, collectively these Munc18-2 mutants might contribute to the lack of proper exocytosis in immune cells of FHL5 patients. As such, our study sheds light on the fact that not only Munc18-2 itself but also the syntaxin protein it chaperones are potentially important therapeutic targets in treating patients with such a devastating immune disorder.

Parental pLKO-puro plasmid for lentivirus-mediated knockdown was purchased from Sigma-Aldrich (Oakville, ON, Canada). pLVX-IRES-puro plasmid for lentivirus-mediated expression was purchased from Clontech Laboratories (Mountain View, CA). psPAX2 was purchased from Addgene (Cambridge, MA). We obtained rabbit monoclonal anti-syntaxin-3 antibody (clone EPR8543) from Abcam (Cambridge, UK), mouse monoclonal antibodies against GM130 (clone 35), Munc18-1 (clone 31), TGN38 (clone 2) from BD Biosciences (Mississauga, ON, Canada); GAPDH (clone 6C5) from Millipore (Billerica, MA); HA.11 (clone 16B21) from Covance (Princeton, NJ); GFP (clone 9F9.F9) from Novus Biologicals (Lilleton, CO), rabbit polyclonal antibodies against syntaxin-3 from Synaptic Systems (Göttingen, Germany) and syntaxin-11 from Proteintech (Chicago, IL). Rabbit polyclonal anti-Munc18-2 antibody was a kind gift from Dr Vesa Olkkonen.

Detailed methods for generating rat Munc18-1 and Munc18-2 expression plasmids have been previously described (Bin et al., 2013). We made binary interaction point mutations (K46E, E59K and K46E/E59K) containing silent nucleotide mutations (SNM) using site-directed mutagenesis. Munc18-1 and Munc18-2 inserts (wild-type and mutants, all without stop codons), which had been digested from pCMV5 vector with EcoRI/XbaI, were subcloned into pLVX–EmGFP–IRES-blast. Sequences of all created plasmids were verified by DNA sequencing. Expression plasmids were co-transfected with psPAX2 and pCMV-VSVG into HEK-293FT cells to generate recombinant lentivirus.

pLVX–HA–IRES-blast was generated by subcloning the hemagglutinin (HA) sequence in tandem into pLVX-IRES-blast using BamHI/XbaI sites. Rat syntaxin-2 and syntaxin-3 without stop codon inserts were digested from pCMV5 vector with EcoRI/XbaI for subcloning into the same sites of pLVX–HA–IRES-blast. Sequences of all created plasmids were verified by DNA sequencing. Expression plasmids were co-transfected with psPAX2 and pCMV-VSVG into HEK-293FT cells in order to generate recombinant lentivirus.

Munc18-1/2 DKD RBL-2H3 mast cells (Bin et al., 2013) were maintained in DMEM containing 10% calf serum (HyClone Laboratories, Logan, UT), penicillin (100 U/ml)/streptomycin (0.1 mg/ml) (Sigma-Aldrich) and 250 ng/ml amphotericin B (Sigma-Aldrich). Lentivirus expressing EmGFP (control), EmGFP-tagged Munc18-1 or Munc18-2 wild-type and binary interaction mutants (K46E, E59K, K46E/E59K) were applied to the Munc18-1/2 DKD RBL-2H3 cells in the presence of polybrene (8 µg/ml) (Sigma-Aldrich) for 2 days. For the isolation of cells that were successfully infected with lentivirus, we selected the cells with blasticidin (20 µg/ml) (Invivogen, San Diego, CA). For overexpression of syntaxin-2–HA and syntaxin-3–HA, lentivirus expressing HA-tagged syntaxin-2 or syntaxin-3 were applied to control and Munc18-1/2 DKD RBL-2H3 cells in the presence of polybrene (8 µg/ml) (Sigma-Aldrich) for 2 days; blasticidin (20 µg/ml) (Invivogen) was used for selection.

A detailed protocol has been described previously (Bin et al., 2013). Briefly, RBL-2H3 cells were plated in 24-well plates; 2–3 days after plating, cells that were to be stimulated through an IgE-dependent pathway were sensitized with 0.01 µg/ml DNP-IgE Spe7 antibody (Sigma-Aldrich) overnight. The next day, cells were washed twice with physiological saline solution (PSS) containing 145 mM NaCl, 5.6 mM KCl, 2.2 mM CaCl₂, 0.5 mM MgCl₂, 5.6 mM glucose, 15 mM HEPES, and 0.1% BSA at pH 7.4. Degranulation was stimulated for 1 hour by 50 or 100 ng/ml DNP-HSA (Sigma-Aldrich), 0.5 µM ionomycin (Sigma-Aldrich) or 2.5 µM ionomycin. After transferring the supernatant, cells were solubilized in 0.5% Triton X-100 to obtain the pellet portion. To evaluate release of β-hexosaminidase, 1 mM p-nitrophenyl N-acetyl-β-D-glucosaminide (Sigma-Aldrich) was added and incubated for 1 hour. Reaction was quenched by adding 0.05 µM sodium carbonate buffer and absorbance at 405 nm was measured by using a spectrophotometer; (OD₄₀₅ of supernatant/OD₄₀₅ of total)×100% was calculated to be the release of β-hexosaminidase in percent.

Rat full-length wild-type Munc18-1 or Munc18-2 and one mutant (K46E, E59K or K46E/E59K) with SNM (Bin et al., 2013) were subcloned into the EcoRI/PstI sites of a bait vector, pLexN. The cytoplasmic part of rat syntaxin-2 and syntaxin-3 was subcloned into the EcoRI/BglII sites of prey vector pVP16-3 (Okamoto and Südhof, 1997). The syntaxin-3 open mutant (L165E/E166A) (Dulubova et al., 1999) was generated by site-directed mutagenesis on pVP16-3-rat syntaxin-3. Yeast strain L40 (Vojtek et al., 1993) was transfected with bait and prey vectors by using the lithium-acetate method (Schiestl and Gietz, 1989). Transformants were plated on selection plates lacking uracil, tryptophan and leucine. After 2 days of incubation at 30°C, colonies were inoculated into supplemented minimal medium lacking uracil, tryptophan and leucine, and placed in a shaking incubator at 30°C for 2 days. β-galactosidase assays were performed as described previously (Han et al., 2011).

Sterilized circular glass coverslips (0.25 mm in width, 1.8 cm in diameter) were placed in 2.2-cm wells within 12-well cell-culture plates. The coverslips were then coated for 1 hour with poly-D-lysine (0.1 mg/ml). RBL-2H3 cells were allowed to adhere to the coverslips for 3 days. For cells overexpressing HA-tagged syntaxin-2 or syntaxin-3, the cells were washed with phosphate-buffered saline (PBS) and fixed for 15 min with PBS containing 4% paraformaldehyde (PFA). Cells were rinsed three times (10 minutes each time) with 1 ml of PBS per well. The fixed cells were permeabilized with PBS containing 0.2% Triton X-100 and 0.3% BSA for 5 minutes, followed by three washes with PBS. Nonspecific sites were blocked for 1 hour at room temperature in PBS containing 0.3% BSA. The mouse monoclonal HA antibody was applied as the primary antibody (1:1000) for 1 hour. After three washes in blocking buffer, Alexa-Fluor-488–conjugated goat anti-mouse secondary antibody (diluted 1:1000; Invitrogen, Carlsbad, CA) was applied for 1 hour. Samples were washed again three times in blocking buffer and 300 nM DAPI was applied for 30 minutes. The slides were mounted in Fluoromount-G reagent (Southern Biotechnology, Birmingham, AL).

For all other stainings, cells were washed once with 0.1% PBS+Tween-20 and fixed for 20 minutes with PBS containing 2% paraformaldehyde. Then, the cells were rinsed twice (10 minutes each time) with 1 ml of 0.1% PBS+Tween-20 followed by permeabilization with 0.1% Triton X-100 for 5 min. Nonspecific sites were blocked for 1 hour at room temperature in PBS containing 10% normal goat serum (Gibco, Burlington, ON, Canada). Then, the primary antibodies were applied at 1:50 for rabbit monoclonal syntaxin-3 antibody, 1:250 for mouse monoclonal GFP antibody, 1:1000 for mouse monoclonal TGN38 and GM130, and 1:1000 for rabbit polyclonal calnexin antibody overnight at room temperature. After three washes in blocking buffer, either Alexa-Fluor-488 or Rhodamine-x-conjugated goat anti-mouse or anti-rabbit secondary antibodies were applied as 1:1000 for 1 hour at room temperature. Samples were washed again three times in blocking buffer and 300 nM DAPI was applied for 30 minutes. Samples were mounted using Fluoromount-G reagent. Immunofluorescence staining of cells overexpressing HA-tagged syntaxin-2 or syntaxin-3 was recorded with a laser confocal scanning microscope (LSM510; Carl Zeiss, Jena, Germany), all other staining was recorded using an LSM710 confocal microscope (Carl Zeiss, Jena, Germany) with an oil immersion objective lens (63×).

The quantification of syntaxin-3 expressions was performed using ImageJ. First, all raw files of confocal images were exported to TIFF then each image was opened on ImageJ. For syntaxin-3 overall intensity, an entire RBL-2H3 cell was manually selected using Freehand Selection, then mean intensity was calculated using the ‘Analyze→Measure’ option. For soma versus tip analysis of syntaxin-3 expression, either soma or tip of the RBL-2H3 cell was selected using Freehand Selection. Then, the intensity was calculated using the ‘Analyze→Measure’ option. The total intensity of syntaxin-3 expression was calculated as above. Then, the value of intensity at the soma or tip was normalized to the total intensity of syntaxin-3 and given in percent. At least 20 cells were analyzed for each knockdown or rescue experiment and the mean values were presented.

To knock down the rat syntaxin-3 gene, we generated two different constructs targeting (1) the 21-nucleotide sequence 5′-GCTCGAAAGAAATTGATAATT-3′ (named syntaxin-3–8 KD) and, (2) the 21-nucleotide sequence 5′-GTTTGTGGAGGTGATGACAAA-3′ (named syntaxin-3–5 KD) in rat syntaxin-3. The hexameric sequence CTCGAG was used as a linker sequence in both constructs. These oligonucleotides, containing sense and antisense of the target sequences, were annealed, phosphorylated and ligated into AgeI/EcoRI sites of pLKO-puro (Sigma-Aldrich) for syntaxin-3–8 KD and pLKO-neo (Sigma-Aldrich) for syntaxin-3–5 KD, generating the syntaxin-3 knockdown plasmids. These knockdown plasmids were co-transfected with psPAX2 and pMD.G into HEK-293FT cells to generate recombinant lentivirus. Then the lentivirus, expressing shRNA against rat syntaxin-3 and an empty vector control, was applied to RBL-2H3 wild-type cells (ATCC, Manassas, VA) in the presence of polybrene (8 µg/ml) (Sigma-Aldrich) for 2 days, and infected cells were isolated by using puromycin (10 µg/ml) (Bioshop, Burlington, ON, Canada).

The homology model of rat Munc18-2 (GenBank accession no. U20283.1) and rat syntaxin-3 (GenBank accession no. L20820.1) was constructed based on a template crystal structure of neuronal Sec1–syntaxin-1A complex (PDB 3C98) (Misura et al., 2000). The aa sequences of Munc18-2 and syntaxin-3 were inputted into Zdock Server software and the predicted structure of the protein complex was visualized using Jmol software.

The control and Munc18-1/2 DKD RBL-2H3 cells were transiently expressed with of NPY–pHluorin by electroporating 15 µg of pcDNA3.1-NPY–pHluorin (Zhu et al., 2007). After 2 days, cells were plated on the 25 mm circular glass coverslip coated with poly-D-lysine (0.1 mg/ml) and were allowed to attach for 2 days. Total internal reflection fluorescence microscopy (TIRFM) setup, image acquisition and data analysis were performed as described here (Xie et al., 2013). Stimulation of exocytosis was triggered by adding 2.5 µM of ionomycin and 0.1 µM of PMA.

We thank Dr Vesa Olkkonen (Minerva Foundation Institute for Medical Research, Helsinki, Finland) for reagents used in this study.