Inhibition of Insulin-induced GLUT4 Translocation by Munc18c through Interaction with Syntaxin4 in 3T3-L1 Adipocytes*

Insulin induces the translocation of vesicles containing the glucose transporter GLUT4 from an intracellular compartment to the plasma membrane in adipocytes. SNARE proteins have been implicated in the docking and fusion of these vesicles with the cell membrane. The role of Munc18c, previously identified as an n-Sec1/Munc18 homolog in 3T3-L1 adipocytes, in insulin-regulated GLUT4 trafficking has now been investigated in 3T3-L1 adipocytes. In these cells, Munc18c was predominantly associated with syntaxin4, although it bound both syntaxin2 and syntaxin4 to similar extents in vitro. In addition, SNAP-23, an adipocyte homolog of SNAP-25, associated with both syntaxins 2 and 4 in 3T3-L1 adipocytes. Overexpression of Munc18c in 3T3-L1 adipocytes by adenovirus-mediated gene transfer resulted in inhibition of insulin-stimulated glucose transport in a virus dose-dependent manner (maximal effect, ∼50%) as well as in inhibition of sorbitol-induced glucose transport (by ∼35%), which is mediated by a pathway different from that used by insulin. In contrast, Munc18b, which is also expressed in adipocytes but which did not bind to syntaxin4, had no effect on glucose transport. Furthermore, overexpression of Munc18c resulted in inhibition of insulin-induced translocation of GLUT4, but not of that of GLUT1, to the plasma membrane. These results suggest that Munc18c is involved in the insulin-dependent trafficking of GLUT4 from the intracellular storage compartment to the plasma membrane in 3T3-L1 adipocytes by modulating the formation of a SNARE complex that includes syntaxin4.

Insulin stimulates glucose transport into muscle and adipose tissue by inducing the translocation of vesicles containing the glucose transporter GLUT4 from the intracellular compartment to the plasma membrane (1,2). This process is thought to be a major contributor to the mechanism by which insulin reduces the blood concentration of glucose. The binding of insulin to its receptor on the surface of target cells results in receptor autophosphorylation and receptor-mediated tyrosine phosphorylation of several additional proteins, including insu-lin receptor substrates 1-4 (IRS1 to IRS4). 1 The phosphorylated IRS proteins then bind other proteins, such as phosphoinositide (PI) 3-kinase, SHP-2, and GRB2, that contain SRC homology 2 (SH2) domains (3). PI 3-kinase is thought to play a role in the insulin-induced translocation of GLUT4 (4,5); however, the mechanism by which activation of PI 3-kinase results in GLUT4 translocation remains unclear (6).
The SNARE (soluble N-ethylmaleimide-sensitive fusion protein attachment protein receptor) hypothesis was initially proposed to explain the process of neurotransmitter secretion (7,8). According to this hypothesis, the docking and fusion of synaptic vesicles at the plasma membrane are initiated by the interaction of proteins, known as v-SNAREs (synaptobrevin/ VAMP), located on the vesicle surface with corresponding proteins, known as t-SNAREs (syntaxin, SNAP-25), located on the target membrane. Membrane fusion is subsequently mediated by the cytosolic proteins ␣-, ␤-, and ␥-SNAP (soluble NSF attachment protein) and NSF (N-ethylmaleimide-sensitive fusion protein), which binds SNAP and hydrolyzes ATP. This hypothesis was subsequently applied to vesicle transport from the intracellular compartment to the plasma membrane in cells other than neurons.
Thus, SNARE proteins have been implicated in the insulininduced translocation of GLUT4-containing vesicles in adipocytes. Members of the VAMP/synaptobrevin family of proteins were first shown to localize in GLUT4 vesicles in rat adipocytes (9). Such VAMP/synaptobrevin proteins were identified as VAMP2 and cellubrevin/VAMP3 in 3T3-L1 adipocytes (10,11). Syntaxins 2, 4, and 5 are also expressed in 3T3-L1 adipocytes (12). In addition, with the use of neurotoxins that cleave VAMP2 and cellubrevin/VAMP3, these proteins were shown to function as v-SNAREs in the trafficking of GLUT4 vesicles to the plasma membrane (11,13). Syntaxin4 is also implicated in GLUT4 translocation in adipocytes. Microinjection of the cytoplasmic domain of syntaxin4 (14,15) or of antibodies to syn-taxin4 (16), or expression of the same cytoplasmic domain with a vaccinia virus vector (15), resulted in inhibition of insulininduced GLUT4 translocation to the plasma membrane in 3T3-L1 adipocytes. Furthermore, the SNARE complexes bound by recombinant NSF and ␣-SNAP in proteins solubilized from rat adipocyte membranes contained syntaxin4 but not syn-taxin2 (12). These studies thus suggest that syntaxin4 functions as a t-SNARE in the docking and fusion of GLUT4 vesi-cles at the plasma membrane in adipocytes. However, although VAMP2 and VAMP3 (v-SNAREs) as well as syntaxin4 (t-SNARE) appear to play essential roles in GLUT4 translocation in these cells, the mechanism by which insulin might regulate these SNARE components remains unknown.
Munc18b and Munc18c, the latter identified from a 3T3-L1 adipocyte cDNA library, are abundant Munc18 isoforms in these cells (31). We have now investigated the function of Munc18 proteins in the insulin-induced translocation of GLUT4 vesicles in 3T3-L1 adipocytes.

EXPERIMENTAL PROCEDURES
Antibodies and cDNAs-Polyclonal antibodies to SNAP-23, syn-taxin4, and Munc18c were generated as described previously (35). Rabbit polyclonal antibodies specific for the COOH-terminal portion of GLUT1 also were prepared as described previously (36). Mouse monoclonal antibodies to GLUT4 were kindly provided by D. E. James (University of Queensland, Australia). Rabbit polyclonal antibodies generated in response to the cytoplasmic region of syntaxin2 and to full-length Munc18b were kindly provided by M. K. Bennett (University of California, Berkeley) and V. M. Olkkonen (National Public Health Institute, Helsinki, Finland), respectively. Mouse monoclonal antibody to c-MYC (monoclonal antibody 9E10) was purchased from Oncogene Science, Inc.
Cell Culture-3T3-L1 fibroblasts were maintained in Dulbecco's modified Eagle's medium (DMEM) (Life Technologies, Inc.) supplemented with 10% fetal bovine serum (FBS). Adipogenesis was induced by treatment with insulin, dexamethasone, and isobutylmethylxanthine as described previously (37), and the cells were subjected to experiments after 8 -13 days. COS cells were also maintained in DMEM supplemented with 10% FBS.
Subcellular Fractionation of 3T3-L1 Adipocytes and Immunoprecipitation-Subcellular fractionation of 3T3-L1 adipocytes was performed as described previously (38) with minor modifications. Briefly, cells were scraped and homogenized in TES buffer (20 mM Tris-HCl (pH 7.4), 1 mM EDTA, 225 mM sucrose), and the homogenate was centrifuged at 16,000 ϫ g. The resulting pellet was layered on top of a 1.12 M sucrose cushion and centrifuged at 101,000 ϫ g, after which the plasma membrane fraction was collected from the interface of the two solutions. Immunoprecipitation of syntaxins 2 or 4 or of Munc18c from extracts of the plasma membrane fraction (prepared with a solution containing 50 mM Hepes-NaOH (pH 7.4), 150 mM NaCl, 5 mM EDTA, 1% Triton X-100, and 1 mM phenylmethylsulfonyl fluoride) was performed with the corresponding specific antibodies.
In Vitro Binding Assay-With the use of Lipofectin (Life Technologies, Inc.), COS cells were transiently transfected with full-length cDNAs encoding SNAP-23 or c-MYC epitope-tagged Munc18a, -b, or -c that had been cloned into the expression vector pcDL-SR␣ (35,39). The cells were solubilized with lysis buffer (25 mM Hepes-NaOH (pH 7.4), 150 mM NaCl, 5 mM EDTA, 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride), and the resulting extracts were incubated, with constant agitation, at 4°C for 1 h with glutathione-Sepharose 4B beads (Amersham Pharmacia Biotech) conjugated with 2 to 3 g of glutathione S-transferase (GST) fusion proteins containing the cytoplasmic portions of syntaxins 1a, 2, 3, or 4. The beads were washed three times with ice-cold lysis buffer, and proteins bound to the beads were then eluted with 20 l of Laemmli sample buffer and subjected to SDS-PAGE and immunoblot analysis with either a monoclonal antibody to c-MYC (for Munc18a, -b, or -c) or antibodies to SNAP-23.
Construction of and Infection with Adenovirus Vectors-Recombinant adenovirus vectors were generated by cloning cDNAs into pAx-CAwt (40), which contains the CAG promoter (41), and cotransfection into 293 cells with DNA-TPC, as described previously (42). Proteinencoding viruses were screened by immunoblot analysis and cloned by limiting dilution. Adenovirus vectors were propagated by a standard procedure and then purified and titrated as described previously (43). Eight to 11 days after induction of differentiation, 3T3-L1 adipocytes were infected for 2 h at the indicated multiplicity of infection (m.o.i.), in plaque-forming units (pfu) per cell, as determined by limiting dilution assay in 293 cells. The adipocytes were subjected to various experiments about 48 h after infection.
Assay of 2-Deoxy-D-glucose Transport-3T3-L1 cells were deprived of serum by incubation in 12-well plates with DMEM for 2 h. The cells were then incubated with 100 nM insulin for 20 min or 600 mM Dsorbitol for 40 min in 0.45 ml of KRH buffer (25 mM Hepes-NaOH (pH 7.4), 120 mM NaCl, 5 mM KCl, 1.2 mM MgSO 4 , 1.3 mM CaCl 2 , 1.3 mM KH 2 PO 4 ). Glucose transport was initiated by the addition of 0.05 ml of KRH buffer containing 2-deoxy-D-[1,2-3 H]glucose (final concentration, 0.05 mM; 0.25 Ci) to each well, and after 5 min, transport was terminated by washing cells three times with ice-cold KRH buffer. The cells were solubilized with 0.5% SDS, and the incorporated radioactivity was measured by liquid scintillation counting.
Plasma Membrane Lawn Assay-Translocation of GLUT1 or GLUT4 to the plasma membrane was measured by the plasma membrane lawn assay as described previously (44). In brief, 3T3-L1 cells cultured on coverslips were washed in phosphate-buffered saline (PBS) and treated with poly-L-lysine (0.5 mg/ml) in PBS. Cells were incubated in a hypotonic solution (0.33 KHMgE buffer, comprising 30 mM Hepes-NaOH (pH 7.5), 70 mM KCl, 5 mM MgCl 2 , and 3 mM EGTA) and then disrupted by placement under an ultrasonic microprobe in KHMgE buffer containing 0.1 mM phenylmethylsulfonyl fluoride and 1 mM dithiothreitol. Sonicated cells were fixed in 2% paraformaldehyde and then incubated with rabbit polyclonal antibodies to GLUT1 and mouse monoclonal antibodies to GLUT4. After washing three times with PBS, the coverslips were incubated with tetramethylrhodamine isothiocyanate-conjugated antibodies to rabbit IgG and fluorescein isothiocyanate-conjugated antibodies to mouse IgG. The cells were washed with PBS, mounted in 90% glycerol in PBS containing p-phenylenediamine (1 mg/ml), and examined with a fluorescence microscope (Axiophot; Zeiss, Jena, Germany).

Association of Munc18 Isoforms and SNAP-23 with Syntaxins in Vitro-
To investigate the functional roles of Munc18 isoforms, syntaxins, and SNAP-23 in GLUT4 translocation in 3T3-L1 adipocytes, we examined the affinity of Munc18 isoforms and SNAP-23 for GST fusion proteins containing the cytoplasmic portions of syntaxins 1a, 2, 3, or 4. Interaction with syntaxin5 was not assessed because this protein appears to function as a t-SNARE in transport from the endoplasmic reticulum to the Golgi (45). SNAP-23 as well as Munc18a, -b, and -c tagged at their COOH termini with a human c-MYC epitope were transiently expressed in COS cells, and the corresponding cell extracts were incubated with glutathione-Sepharose beads containing immobilized GST-syntaxin fusion proteins. After washing, proteins eluted from the beads were analyzed by SDS-PAGE and immunoblotting with antibodies to c-MYC (for the three Munc18 isoforms) or to SNAP-23.
Munc18a, which is identical to n-Sec1 and Munc18 -1, bound to the GST fusion proteins containing syntaxins 1a, 2, and 3, but there was no detectable binding of Munc18a to syntaxin4 (Fig. 1). Munc18b showed a pattern of binding similar to that of Munc18a. In contrast, Munc18c showed a marked interaction with syntaxins 2 and 4, interacted to a much lesser extent with syntaxin1, and exhibited no detectable binding to syntaxin3. Although it is not clear whether the Munc18 isoforms expressed in COS cells bind directly or indirectly to the recombinant GST-syntaxin proteins, these results show that Munc18c is the only Munc18 isoform that interacts substantially with syntaxin4 in vitro. The observation that SNAP-23, an adipocyte homolog of SNAP-25 (35), interacted markedly with syntaxins 1a and 4 and slightly with syntaxin2 is consistent with our previous results obtained with the yeast two-hybrid system (35). Given that syntaxins 2 and 4 (8) and Munc18b and -c (31) are the major syntaxin and Munc18 isoforms in 3T3-L1 adipocytes, it is likely that Munc18b interacts with syntaxin2 and that Munc18c interacts with syntaxins 2 or 4 in these cells.
Interactions of Munc18c, Syntaxins 2 and 4, and SNAP-23 in 3T3-L1 Adipocytes-We next examined the physiological interactions among Munc18b or -c, syntaxins 2 or 4, and SNAP-23 in 3T3-L1 adipocytes. Our polyclonal antibodies generated in response to a synthetic COOH-terminal peptide of Munc18c are highly specific and do not recognize Munc18a or Munc18b overexpressed in COS cells, as assessed by immunoblot analysis or immunoprecipitation (data not shown). These antibodies detected a major protein of 67 kDa, corresponding to the predicted size of Munc18c, on immunoblot analysis of a detergent extract of the plasma membrane fraction of 3T3-L1 adipocytes ( Fig. 2A). In contrast, we did not detect Munc18b in 3T3-L1 adipocytes by immunoblot analysis with antibodies specific for this Munc18 isoform (data not shown). Previous studies have detected Munc18b mRNA by Northern blot analysis (31) and Munc18b protein by immunoblot analysis (16) in these cells. This discrepancy is probably attributable to a difference in the efficacy of the antibodies used in the present and previous (16) studies; we were able to detect Munc18b overexpressed in COS cells.
Polyclonal antibodies generated in response to GST fusion proteins containing the cytoplasmic portions of rat syntaxins 2 or 4 or to a GST fusion protein containing mouse SNAP-23 detected proteins of the predicted molecular sizes in the detergent extracts of plasma membranes from 3T3-L1 adipocytes (Fig. 2, B-D). Next, we confirmed that both syntaxins 2 and 4 were immunoprecipitated from detergent extracts of the plasma membrane fraction of 3T3-L1 adipocytes with the corresponding specific antibodies (Fig. 2, B and C). Despite the interaction of Munc18c with both syntaxins 2 and 4 in the in vitro binding assay (Fig. 1), Munc18c was coimmunoprecipitated with syntaxin4 but not with syntaxin2 ( Fig. 2A). In contrast, SNAP-23 was coimmunoprecipitated with both syn-taxin2 and syntaxin4 (Fig. 2D), consistent with the results of the in vitro binding assay (Fig. 1). When the immunoprecipitation was performed with antibodies to Munc18c, syntaxin4 (Fig. 2C), but not syntaxin2 (Fig. 2B) or SNAP-23 (Fig. 2D), was coprecipitated with Munc18c. In summary, Munc18c was detected in the plasma membrane of 3T3-L1 adipocytes and was associated with syntaxin4 but not with syntaxin2. SNAP-23 associated with both syntaxins 2 and 4 in the plasma membrane of 3T3-L1 adipocytes, although it was not associated with syntaxin4 complexed with Munc18c.
Insulin-or Sorbitol-stimulated Glucose Transport in 3T3-L1 Adipocytes Overexpressing Munc18b or Munc18c-To examine the function of Munc18b and Munc18c in 3T3-L1 adipocytes, we prepared recombinant adenoviruses that were expressed with an efficiency of Ͼ95% in these cells as assessed by ␤-galactosidase staining (46). Both c-MYC epitope-tagged Munc18b and -c were overexpressed in 3T3-L1 adipocytes with the use of this adenovirus-mediated gene transfer system. Both proteins were expressed to similar extents in an m.o.i.-dependent manner (Fig. 3, A and C). Moreover, the amount of overexpressed Munc18c bound to syntaxin4 also increased in an m.o.i.-dependent manner and was maximal at an m.o.i. of 15-30 pfu/cell (Fig. 3C); overexpressed Munc18b did not bind syntaxin4 (Fig.  3A).
Overexpression of Munc18b had no significant effect on glucose transport, compared with that in noninfected cells, in the presence of 100 nM insulin, although it showed a slight increase (15.7% at an m.o.i. of 30) on unstimulated glucose transport (Fig. 3B). In contrast, overexpression of Munc18c inhibited insulin-stimulated glucose transport (with a maximal inhibition of 50.6% at an m.o.i. of 30) in proportion to the amount of overexpressed protein bound to syntaxin4 in addition to unstimulated glucose transport to the extent of 63.0% at an m.o.i. of 30 (Fig. 3, C and D). Moreover, Munc18c overexpression also inhibited glucose uptake induced by hyperosmolarity (600 mM D-sorbitol) which promotes glucose transport through a path-way different from that for insulin (46,47) to a little lesser extent (33.9% inhibition at an m.o.i. of 30) than that observed for insulin; Munc18b overexpression had no significant effect on sorbitol-induced glucose transport (Fig. 4).
GLUT1 and GLUT4 Translocation in 3T3-L1 Adipocytes Overexpressing Munc18b or Munc18c-In addition to GLUT4, 3T3-L1 adipocytes express another facilitative glucose transporter, GLUT1, which undergoes translocation to the plasma membrane in response to insulin to a lesser extent than does GLUT4. We performed plasma membrane lawn assays (44) with polyclonal antibodies generated in response to a synthetic COOH-terminal peptide of GLUT1 and with a monoclonal antibody to GLUT4 to investigate which transporter isoform was responsible for the reduced stimulation of glucose transport in 3T3-L1 adipocytes overexpressing Munc18c. Insulin induced a marked increase in immunoreactivity of both GLUT1 and GLUT4 in the plasma membrane of noninfected 3T3-L1 adipocytes (Fig. 5). Adenovirus-mediated overexpression of Munc18c inhibited the insulin-induced increase in GLUT4 immunoreactivity in the plasma membrane, whereas Munc18b overexpression had no substantial effect on this action of insulin. Overexpression of either Munc18b or Munc18c had no effect on the insulin-induced increase in GLUT1 immunoreactivity in the plasma membrane. We also confirmed by immunoblot analysis with the specific antibodies to GLUT1 or GLUT4 that overexpression of Munc18c but not Munc18b inhibited the insulinstimulated translocation of GLUT4 to the plasma membrane, although overexpression of either Munc18b or Munc18c had no remarkable effects on the insulin-stimulated translocation of GLUT1 (Fig. 6). Thus, the inhibitory effect of Munc18c on insulin stimulation of glucose transport in 3T3-L1 adipocytes appears to be attributable to inhibition of the translocation of GLUT4 to the plasma membrane. DISCUSSION We have shown that Munc18c is expressed in the plasma membrane of 3T3-L1 adipocytes in association with syntaxin4. Furthermore, overexpression of Munc18c in these cells inhibited insulin-induced translocation of GLUT4, but not that of GLUT1, to the plasma membrane. Consequently, Munc18c overexpression also reduced the extent of insulin-stimulated glucose transport. These inhibitory effects of Munc18c were not reproduced by overexpression of Munc18b, which is also expressed naturally in 3T3-L1 adipocytes (31) but which does not bind to syntaxin4 either in these cells or in vitro. Overexpression of Munc18c, but not of Munc18b, also inhibited stimulation of glucose transport by sorbitol, which also induces GLUT4 translocation in 3T3-L1 adipocytes but by a pathway different from that triggered by insulin (46,47). These data suggest that Munc18c plays an inhibitory role in the translocation of GLUT4-containing vesicles in 3T3-L1 adipocytes by binding to syntaxin4. The inhibitory effects of Munc18c overexpression are consistent with the observation that overexpression of ROP, the Drosophila homolog of n-Sec1/Munc18, in third-instar larvae reduced neurotransmitter release (34). On the other hand, ROP was recently reported to be a rate-limiting regulator of exocytosis that performs not only inhibitory but positive functions in neurotransmission (48). Thus in neurotransmission, it is suggested that ROP may play more complicated roles than expected.
Syntaxin4 plays a crucial role as a t-SNARE in insulinstimulated GLUT4 translocation in both 3T3-L1 adipocytes (14 -16) and rat adipocytes (12). Moreover, we have previously shown that SNAP-23, a homolog of SNAP-25, is expressed predominantly in the plasma membrane of 3T3-L1 adipocytes (35). Our present observation that SNAP-23 formed a complex with syntaxin4 in 3T3-L1 adipocytes suggests that this molecule might function together with syntaxin4 as a t-SNARE in GLUT4 translocation in 3T3-L1 adipocytes. The observation that botulinum E toxin, which cleaves SNAP-25 but not SNAP-23, showed no inhibitory effect on insulin-stimulated GLUT4 translocation in 3T3-L1 adipocytes (49) supports the possibility that SNAP-23 functions as a t-SNARE in association with syntaxin4 in these cells (35). In rat brain tissue, n-Sec1 (Munc18a) binds to syntaxin1 with high affinity and is absent from 7 S particles containing syntaxin, SNAP-25, and VAMP or 20 S fusion particles consisting of NSF and ␣-SNAP in addition to syntaxin, SNAP-25, and VAMP (33). Our observation that SNAP-23 was not present in the complexes immunoprecipitated with antibodies to Munc18c, whereas syntaxin4 was coimmunoprecipitated with these antibodies, may also suggest that Munc18c binds to syntaxin4 in a manner exclusive of SNAP-23 in adipocytes. We have previously shown that Munc18c inhibits the binding of SNAP-23 to syntaxin4 in vitro (35). Furthermore, Munc18c inhibits the interaction of syn-taxin4 with VAMP2 and VAMP3 in vitro (16). On the basis of these various in vitro data, we propose that Munc18c plays a negative role in the docking and fusion of GLUT4 vesicles at the plasma membrane of 3T3-L1 adipocytes by inhibiting the association of syntaxin4 with SNAP-23 (t-SNARE) and VAMP2 or -3 (v-SNAREs).
In 3T3-L1 adipocytes, the population of GLUT1-containing vesicles is largely distinct from the GLUT4 vesicle population, although there is some overlap in these two compartments (50,51). Moreover, in these cells, GLUT1 and cellubrevin/VAMP3 are predominantly targeted to recycling endosomes, whereas GLUT4, VAMP2, and an aminopeptidase (VP165) are excluded from this compartment (52)(53)(54). Insulin may thus stimulate the translocation of GLUT1 and GLUT4 vesicles from two distinct intracellular compartments and by two different pathways in 3T3-L1 adipocytes. Our observation that overexpression of Munc18c inhibited translocation of GLUT4 but not that of GLUT1 suggests that the syntaxin4-Munc18c complex is in- volved in the translocation of GLUT4-containing vesicles but not that of GLUT1-containing vesicles targeted to recycling endosomes. Consistent with our data, introduction of the cytoplasmic domain of syntaxin4 into 3T3-L1 adipocytes resulted in inhibition of insulin-stimulated GLUT4 translocation, but not in that of GLUT1 translocation, to the plasma membrane (15). It has been shown that not only GLUT1 but GLUT4 constantly recycles between intracellular compartments and the plasma membrane both in the absence and presence of insulin (55). Since overexpression of Munc18c inhibits insulin-induced translocation of GLUT4, but not that of GLUT1, overexpression of Munc18c is speculated to reduce the unstimulated glucose transport by inhibiting the constant recycling of GLUT4 to the plasma membrane in the absence of insulin. In this case, we cannot exclude the possibility that overexpression of Munc18c reduced the unstimulated glucose transport by affecting the constant recycling of GLUT1 vesicles at the basal state. Further studies are necessary to identify the t-SNARE and v-SNARE proteins that mediate the docking and fusion of GLUT1 vesicles.
In the absence of insulin, GLUT4-containing vesicles are sequestered in an intracellular compartment, from which, on exposure of cells to insulin, they move to the plasma membrane (1,2). They subsequently dock and fuse with the cell membrane through the action of SNARE proteins (11)(12)(13)(14)(15)(16)35). Although the molecular mechanism by which insulin regulates the trafficking of GLUT4-containing vesicles is not known, it is likely that it targets several steps in this process. It was shown that introduction of a synthetic peptide corresponding to the COOHterminal, cytoplasmic domain of GLUT4 into rat adipocytes or cytoplasmic domain of insulin-responsive aminopeptidase into 3T3-L1 adipocytes results in insulin-like stimulation of glucose transport or GLUT4 recruitment (56,57). These observations suggest the existence of a regulatory protein in these cells that interacts with the cytoplasmic domain of GLUT4 and restricts it to the intracellular storage compartment. The insulin signal might modulate this regulatory protein, resulting in the release of the GLUT4 vesicles anchored to the intracellular compartment. Our data also suggest that insulin might also regulate the docking and fusion of GLUT4 vesicles at the plasma membrane through the action of Munc18c. Phosphorylation of Munc18a by protein kinase C blocks its interaction with syntaxin1a in vitro (58). In addition, the DOC2 protein, whose cDNA was cloned from a human brain library (59), regulates the formation of Munc18a-syntaxin1 complexes by binding to Munc18a (60). These observations suggest the possibility that insulin might induce either the phosphorylation of Munc18c or its association with some other protein. Furthermore, since overexpression of Munc18c also decreased sorbitol-induced glucose transport and GLUT4 translocation, 2 this step might be regulated by other signals than insulin which induce GLUT4 tranlocation.
In summary, we propose that Munc18c affects the insulindependent trafficking of GLUT4 to the plasma membrane of 3T3-L1 adipocytes by binding to syntaxin4, and thereby inhibiting the formation of a putative SNARE complex (consisting of syntaxin4, SNAP-23, and VAMP2 or -3) that mediates the docking and fusion of GLUT4-containing vesicles.