Snapin Interacts with the Exo70 Subunit of the Exocyst and Modulates GLUT4 Trafficking*

The exocyst is a multisubunit complex that has been implicated in the transport of vesicles from the Golgi complex to the plasma membrane, possibly acting as a vesicle tether and contributing to the specificity of membrane fusion. Here we characterize a novel interaction between the Exo70 subunit of the exocyst and Snapin, a ubiquitous protein known to associate with at least two t-SNAREs, SNAP23 and SNAP25. The interaction between Exo70 and Snapin is mediated via an N-terminal coil-coil domain in Exo70 and a C-terminal helical region in Snapin. Exo70 competes with SNAP23 for Snapin binding, suggesting that Snapin does not provide a direct link between the exocyst and the SNARE complex but, rather, mediates cross-talk between the two complexes by sequential interactions. The insulin-regulated trafficking of GLUT4 to the plasma membrane serves to facilitate glucose uptake in adipocytes, and both SNAP23 and the exocyst have been implicated in this process. In this study, depletion of Snapin in adipocytes using RNA interference inhibits insulin-stimulated glucose uptake. Thus, Snapin interacts with the exocyst and plays a modulatory role in GLUT4 vesicle trafficking.

SNAREs 4 are central components of the vesicle docking and fusion machinery (1,2). In vitro liposome fusion experiments have shown that SNAREs alone are sufficient to catalyze fusion but at a much slower rate than occurs in vivo (3). In vivo, vesicle docking and fusion likely requires additional components to tightly regulate the specificity and kinetics of membrane fusion.
The seminal work of Novick and Scheckman (4) identified several genes necessary for secretion in yeast, collectively named the SEC genes. Temperature-sensitive mutations in each of these genes block secretion of invertase and accumulate membrane-bound organelles within the yeast cell. Six of the SEC genes, SEC3, SEC5, SEC6, SEC8, SEC10, and SEC15, encode subunits of an evolutionarily conserved octameric protein complex known as the exocyst. Two additional components, Exo70p and Exo84p, were subsequently biochemically purified in complex with the exocyst proteins (5)(6)(7). Similar to their yeast homologues, the mammalian exocyst proteins have been implicated in the transport of vesicles from the Golgi to the plasma membrane (8). The exocyst complex has also been implicated in the trafficking of GLUT4 vesicles to the plasma membrane (9). Among its proposed roles, the exocyst is thought to function as a vesicle tether at the plasma membrane and contribute to the specificity of membrane fusion (5). Recent studies in the ␤-cell demonstrated that the exocyst regulates the ability of insulin granules to dock with the plasma membrane (10).
Within the cell, each step of membrane trafficking is coupled to maintain the sequential nature and ensure the high fidelity of the process. Emerging evidence suggests a link between the membrane tethering and the SNARE-mediated docking/fusion steps (11,12). Genetic experiments in yeast are consistent with the SNAREs acting directly downstream of the exocyst. For example, overexpression of yeast Sec1p or the t-SNAREs Sso1p/Sso2p can bypass secretory blocks induced by mutations in the exocyst genes (13,14). In yeast, the exocyst associates through Sec15p with the vesicle-bound Rab-GTPase Sec4p (7) and acts as an effector for Sec4p during targeting of secretory vesicles to sites of exocytosis (15). In addition, the exocyst interacts with the yeast Lgl homologue Sro7p and the t-SNARE Sec9p (16,17). The above data highlight the involvement of multiple protein-protein interactions and the potential complexity of the coupling mechanism between tethering, docking, and fusion.
Exo70 is a central component of the mammalian exocyst complex, demonstrating specific interactions with almost all of the other subunits, including Sec5, Sec6, Sec8, Sec10, and Exo84 (18,19). Exo70 has recently been implicated in GLUT4 trafficking through its association with the GTP-bound form of the Rho GTPase TC10 (9). Stimulation of 3T3-L1 adipocytes with insulin was shown to cause an increase in the abundance of Exo70 at the plasma membrane. Furthermore, overexpression of an Exo70 mutant that interacts with TC10 specifically inhibited insulin-stimulated appearance of HA-GLUT4-GFP molecules at the plasma membrane, with no apparent effect on the trafficking of GLUT4 to the plasma membrane. Using RNA interference, a follow-up study showed that Exo70 is critical for insulin-regulated glucose uptake in 3T3-L1 adipocytes (20). However, in both studies, the precise mechanism of Exo70 function was not defined. More recently, Exo70 has been demonstrated to interact with the Arp2/3 complex in an epidermal growth factor-dependent manner in HeLa cells. This interaction was essential for the formation of actin-based membrane protrusions and cell migration (21). Furthermore, epidermal growth factor stimulation resulted in recruitment of Exo70 to the plasma membrane, which is consistent with the observation in adipocytes. These studies suggest that growth factor signaling regulates Exo70 function, making Exo70 an interesting candidate for hormone-regulated exocytic events.
To further investigate the functional role of Exo70 in GLUT4 trafficking, we set out to identify new Exo70 interacting proteins that may facilitate the tethering and docking of GLUT4 vesicles at the plasma membrane. Here we show that Exo70 interacts with Snapin. This observation is exciting, because Snapin has previously been shown to interact with SNAP23 (22), a major t-SNARE regulating GLUT4 vesicle trafficking (23). Furthermore, we provide evidence for a modulatory role for Snapin in GLUT4 trafficking, possibly by coordinating vesicle tethering and docking through sequential interactions with the exocyst and SNARE complexes.
Yeast Two-hybrid Screen-Full-length rat Exo70 fused to the Gal4 DNA-binding domain in the pGBKT7 vector was a kind gift from Dr. Hugo Matern (Genentech, Inc.) and used as a bait to screen a mouse embryo Matchmaker cDNA library (Clontech). The mated yeast cells were selected on SD minimal medium supplemented with appropriate autotrophic factors except for adenine, histidine, leucine, and tryptophan (QDO plates), and putative positive clones grown on QDO were assayed for LacZ activity according to the protocol provided with the Matchmaker kit (Clontech). Plasmids from LacZ activity-positive clones were extracted and sequenced. The interac-tion of these clones was then further confirmed by individually retransforming them back into Saccharomyces cerevisiae stain AH109 with rat Exo70 in pGBKT7 and selected as described above and subjected to a ␤-galactosidase assay according to the manufacturer's protocol (Clontech).
In Vivo Binding Assays-Appropriate combinations of plasmids were transfected into 293T or COS1 cells using Lipofectamine 2000 (Invitrogen). Cells were lysed 24 h after transfection in lysis buffer (50 mM Tris, pH 7.5, 140 mM NaCl, 0.1% Triton X-100, 0.1% bovine serum albumin) for 30 min at 4°C. Lysates were spun at 13,000 rpm for 20 min in a microcentrifuge, and the supernatants were then pre-cleared for 1 h with either protein A-or protein G-agarose beads. For immunoprecipitation of Myc-or HA-tagged proteins, the precleared lysates (500 g of protein) were incubated overnight with rabbit anti-Myc or rat anti-HA and then captured by protein A or protein G-agarose beads for 2 h. Total rabbit or mouse IgG served as a control. For immunoprecipitation of FLAG-tagged proteins, precleared lysates were incubated with monoclonal M2 anti-FLAG beads for 2 h. Beads were then washed 5 times with lysis buffer.
In Vitro Binding Assays-For GST pulldown experiments, 10 g each of GST, Exo70-GST, or Exo70 ⌬100 -653-GST were first incubated with glutathione Sepharose-4B beads (PerkinElmer Life Sciences) in GST binding buffer (50 mM Tris, pH 7.5, 1 mM MgCl 2 , 1 mM dithiothreitol) for 1 h and then washed with GST binding buffer supplemented with 0.5% Nonidet P-40 twice. 10 g of His-Snapin in lysis buffer was then added to each tube. After 2-h incubation, the beads were washed 4 times in lysis buffer and twice in lysis buffer supplemented with 0.5% Triton X-100. For the in vitro competition assays, 250 nM MBP-SNAP23 was first bound to amylose beads (Biolabs) in lysis buffer for 1 h. 500 nM His-Snapin was then added together with increasing amounts either of GST, GST-Exo70, or Exo70 (cleaved from GST using a thrombin digestion kit, Novagen). After a 3-h incubation, the amylose beads were washed four times, and bound proteins were solubilized in SDS sample buffer and analyzed by SDS-PAGE and Western blot analysis. Samples were kept a 4°C during the manipulations, and buffers were supplemented with protease inhibitors (Roche Applied Science). Cellular Fractionation-Fractionation of adipocytes was carried out as described (24).
Confocal Microscopy-Immunofluorescence microscopy was performed on COS1 cells or lentivirus transduced 3T3-L1 adipocytes seeded on glass coverslips. Adipocytes were serumstarved for 2 h in Dulbecco's modified Eagle's medium at 37°C and 5% CO 2 and incubated in the absence or presence of 1 or 100 nM insulin for 20 min. Cells were then washed three times in cold PBS, fixed with 3% (v/v) paraformaldehyde in PBS, blocked in 2% (w/v) bovine serum albumin containing 0.1% (w/v) saponin in PBS, and labeled with rabbit anti-Snapin and mouse monoclonal anti-Myc antibodies. Cells were subsequently labeled with appropriate Alexa-labeled antibodies. COS1 cells were fixed 24 h after transfection and labeled with monoclonal anti-Myc or Rabbit anti-HA antibody and appropriate secondary antibodies. Coverslips were mounted onto glass coverslides using Immuno-fluore mounting medium (ICN Biomedicals Inc.). Images were acquired and analyzed using a Leica laser scanning confocal microscope and software (Leica Microsystems, Wetzlar, Germany).
Lentivirus Production and Transduction of 3T3-L1 Adipocytes-Concentrated Myc-Snapin and GFP control lentivirus stocks were prepared as previously described (25,26). Briefly, HEK-293FT cells were transfected at 70 -90% confluency using calcium phosphate (26). One T150 flask was transfected with 26 g of expression vector, 17.4 g of pMDLg/ pRRE (gag/pol) vector, 13.2 g of pRSV-Rev (rev) vector, and 9.2 g of pMD2.G (VSV-G) vector in 20 ml of culture medium. At 5-6 h post-transfection, medium was replaced with 12 ml of medium containing 6 mM butyric acid. Lentivirus was harvested at 24 h post-transfection. Supernatants containing the virus were collected, cleared by centrifugation (1,620 ϫ g; 10 min), and concentrated using Amicon Ultracel concentration tubes (Millipore, MA). For confocal microscopy and immunoblotting, 3T3-L1 adipocytes grown in 24-well plates and at day 6 of differentiation were incubated for 24 h with 200 l of 10-fold concentrated GFP or Snapin lentivirus containing 4 g/ml Polybrene. After 24 h, the medium was replaced with 1 ml of culture medium. Experiments were performed at 5 days post-transduction.
siRNA-mediated Gene Silencing-Snapin siRNAs at a final concentration of 10 nM were transfected into mouse NIH3T3 cells with Lipofectamine RNAiMax reagent (Invitrogen), and cells were analyzed 48 h post-transfection. For electroporation of selected siRNAs, 3T3-L1 adipocytes at day 5 of differentiation were electroporated with either 20 nmol of control nontargeting siRNA or Snapin siRNA. Briefly, three 10-cm dishes of adipocytes were trypsinized and washed twice in PBS by resuspension and centrifugation at 200 ϫ g for 5 min. Cells were resuspended in 1 ml of PBS, and this was divided between two 0.4-cm cuvettes (Bio-Rad) containing either control or Snapin siRNAs. Cells were electroporated by delivery of one 200 V/cm, 10-ms pulse using an ECM-830 electroporator (BTX Technologies, Inc., Holliston, MA). Cells were then transferred to gelatin-coated 12-well plates. At 72 h post electroporation, glucose uptake was measured, and cell lysates were prepared for Western blot analysis using rabbit anti-Snapin and as a control, rabbit anti-14-3-3 antibodies.

Identification of Snapin as an Exo70-binding Protein-
To identify new effectors of the exocyst subunit Exo70, a yeast two-hybrid screen of a mouse embryo cDNA library was performed using full-length rat Exo70 as bait. This screen identified 37 positive clones. Seven of these cDNA clones represented full-length mouse Snapin (accession number AF086838). The interaction between rat Exo70 and mouse Snapin in the yeast two-hybrid system was confirmed by the detection of X-␣-galactosidase or ␤-galactosidase activity in plate (data not shown) or liquid (Fig. 1A) assays, respectively.
Snapin Associates with the Exocyst Complex in Vivo-To confirm the yeast two-hybrid results and characterize the association between Snapin with Exo70 in vivo, HEK cells were transfected with Myc-tagged Snapin and HA-tagged Exo70, and their association was analyzed in coimmunoprecipitation experiments. Immunoprecipitations with Myc antibody but not with control IgG pulled down HA-Exo70 (Fig. 1C). Likewise, Myc-Snapin was present in Exo70-HA immunoprecipitates.
To explore the association of Snapin with endogenous Exo70, overexpressed Myc-Snapin was immunoprecipitated and bound endogenous Exo70 detected using anti-Exo70 antibodies. In addition to endogenous Exo70, at least two other components of the exocyst complex, Sec6 and Sec8, were coprecipitated with Myc-Snapin (Fig. 1B). The presence of other exocyst subunits in the precipitated complex was not analyzed, because antibodies to these other exocyst components were not readily available.
To further confirm the association between Snapin and Exo70 established above, we analyzed the colocalization of Myc-Snapin and HA-Exo70 overexpressed in COS-1 cells by confocal microscopy. As shown in Fig. 1D, there was substantial colocalization between Myc-Snapin and HA-Exo70 in the perinuclear region of the cells as well as in punctate structures near the plasma membrane of cellular protrusions. Collectively, these experiments demonstrate an association of Snapin with the exocyst complex by a direct interaction with the Exo70 subunit.
Exo70 Interacts with a Helical Domain in the C Terminus of Snapin-Snapin contains an N-terminal hydrophobic domain (amino acids 1-20) and two helical domains (amino acids 37-65 and 83-119), the second of which is the major binding site for SNAP23 (22). To map the Exo70 binding domain in Snapin, a series of FLAG-Snapin truncation mutants were generated ( Fig. 2A). These constructs were then coexpressed with Exo70-HA in COS-1 cells, and their interaction was monitored in coimmunoprecipitation experiments. As shown above for Myc-Snapin, FLAG-Snapin robustly interacted with Exo70-HA (Fig. 2B). N-terminal (FLAG-Snapin ⌬1-21) or C-terminal (FLAG-Snapin ⌬120 -136) deletions in Snapin did not affect binding to Exo70. A mutant containing only the C terminus (FLAG-Snapin ⌬1-83) could not be detected by immunoblotting probably due to instability of the truncated protein (data not shown and Ref. 22). Individual deletion of either the second helical domain (FLAG-Snapin ⌬84 -118) or the C terminus (FLAG-Snapin ⌬84 -136) revealed that the presence of the sec-ond helical domain is critical for binding to Exo70. Thus, the Exo70 binding domain in Snapin maps to the second helical domain and interestingly overlaps with the SNAP23 binding site (22).
The ability of the N terminus of Exo70 to directly interact with Snapin was confirmed in vitro using pulldown experiments with GST-Exo70 fusion proteins and purified His-tagged Snapin. There was a significant interaction between GST-FIGURE 1. Snapin interacts with Exo70. A, yeast-two hybrid ␤-galactosidase assay. Liquid culture ␤-galactosidase assay using yeast carrying pGADT7ϩpGBKT7 (negative control; bar 1), pGADT7-TϩpGBKT7-p53 (positive control; bar 2), and pGADT7-SnapinϩpGBKT7-Exo70 (bar 3) were carried out using o-nitrophenyl-␤-D-galactopyranoside as a substrate. Absorbance at 600 nm is shown in arbitrary units (n ϭ 3); error bars, ϮS.D. B, Myc-Snapin coprecipitates with endogenous exocyst components. Equal protein amounts of Myc-Snapin-transfected 293T cell lysates were immunoprecipitated with 2 or 10 g of rabbit IgG (control) or rabbit anti-Myc antibodies, and immunoprecipitated proteins were analyzed by SDS-PAGE and Western blot analysis using antibodies to endogenous Exo70, Sec6, Sec8, and a mouse anti-Myc antibody. As a control, total cell lysate (50 g) was also probed. IgG heavy chains from the antibodies used to immunoprecipitate were also detected and are indicated. C, Myc-Snapin and HA-tagged Exo70 interact in vivo. Lysates of cells coexpressing Myc-Snapin and HA-Exo70 (left panel) or Myc-Snapin and Exo70-HA were immunoprecipitated and bound HA-tagged Exo70 or Myc-Snapin, respectively, detected by Western blot analysis. Rabbit or rat IgG served as a control and total cell lysate was analyzed to monitor expression of the different proteins. Note that when immunoprecipitated with anti-HA antibodies, the antibodies bound to the N-terminal HA tag interfered with the association of HA-Exo70 with Myc-Snapin. D, Myc-Snapin and HA-Exo70 colocalize in COS1 cells. COS1 cells coexpressing Myc-Snapin and HA-Exo70 were detected by immunolabeling using mouse anti-Myc and rabbit anti-HA antibodies, followed by fluorescently labeled secondary goat antibodies and visualized using confocal microscopy. Yellow labeling in the merged image indicates colocalization. Bar ϭ 5 m.

FIGURE 2. A coil-coil domain in Snapin mediates the interaction with Exo70.
A, schematic diagram of mouse Snapin and its deletion mutants. Molecular modeling predicted the N-terminal first 20 amino acids to from a hydrophobic domain and residues 84 -118 to form a coil-coil domain. B, coimmunoprecipitation of Snapin and its deletion mutants with Exo70. Lysates of COS1 coexpressing Exo70-HA and the indicated FLAG-tagged Snapin constructs were immunoprecipitated with M2 anti-FLAG beads, and the precipitates were analyzed by Western blot using anti-HA antibodies to detect bound Exo-70 or anti-FLAG antibodies to confirm binding of the FLAG-Snapin protein to the M2 beads. Cell lysates were probed with the two antibodies to monitor expression levels. Untransfected COS1 cells served as a control. The data indicate that the coil-coil domain of Snapin (amino acids 84 -118) is required for binding to Exo70.
Exo70 and His-Snapin (Fig. 3C), whereas no binding of His-Snapin was observed using GST alone. The GST-Exo70 ⌬100 -653 truncation mutant interacted with His-Snapin with a similar efficiency as full-length Exo70, confirming the importance of the N terminus of Exo70 for the interaction.
These experiments confirm a direct interaction between Snapin and Exo70 and map the interaction domain in Exo70 to the N-terminal 99 residues. Interestingly, computational modeling of the Exo70 amino acid sequence revealed a putative coiled-coil domain within this region of the Exo70 N terminus.
Exo70 and SNAP23 Compete for Binding to Snapin-Because Exo70 and SNAP23 apparently interact with an overlapping region in Snapin (see above, supplemental data, and Ref. 22), we determined if Exo70 and SNAP23 simultaneously bind to Snapin, or if their association with Snapin is mutually exclusive. For this purpose, equal amounts of MBP-SNAP23 immobilized on amylose beads were incubated with His-Snapin in the presence of increasing concentrations of GST-Exo70, GST alone, or Exo70 cleaved from the GST moiety. As shown in Fig. 4 (A and  B), the interaction between Snapin and SNAP23 was significantly reduced in the presence of increasing concentrations of either GST-Exo70 or cleaved Exo70 but not GST alone. We have been unable to detect any significant interaction between Exo70 and SNAP23 (data not shown), consistent with the notion that Exo70 and SNAP23 compete for the same binding site in Snapin.
Subcellular Distribution of Snapin-The Snapin-binding proteins Exo70 and SNAP23 have been shown to play a role in the final stages of GLUT4 trafficking (9,28). Snapin itself has been previously implicated in regulated exocytic events such as calcium-induced neurosecretion (29,30). To investigate the functional role of Snapin in vesicle docking/fusion we next studied the role of Snapin in GLUT4 trafficking in adipocytes.
Snapin was readily detected in 3T3-L1 adipocytes (Fig. 5B) and was predominantly targeted to the plasma membrane (Fig.  5, A and B). In contrast to SNAP23, but similar to Exo70, Snapin was present in both low and high density membrane fractions (Fig. 5A) in adipocytes. Stimulation of 3T3-L1 adipocytes with Lysates of COS1 coexpressing FLAG-Snapin and the indicated HA-tagged Exo70 constructs were immunoprecipitated with M2 anti-FLAG beads, and the precipitates were analyzed by Western blot using anti-HA antibodies to detect bound HA-Exo70 or deletions, or anti-FLAG antibodies to confirm binding of the FLAG-Snapin to the M2 beads. Cell lysates were probed with the two antibodies to monitor expression levels. Untransfected COS1 cells served as a control. An asterisk shows the position of the Exo70 ⌬385-653 band. The IgG heavy (HC) and light chains (LC) are labeled. The data show that the N terminus of Exo70, including the coil-coil domain (amino acids 5-36), is required for binding to Snapin. C, GST pulldown assay to demonstrate a direct interaction between the N-terminal domain of Exo70 and Snapin. GST alone or the indicated GST-Exo-70 proteins coupled to beads were incubated with purified His-tagged Snapin. Bound Snapin was detected by Western blot using rabbit anti-Snapin antibody. Deletion of the C terminus of Exo70 (amino acids 100 -653) did not abolish its binding with Snapin, confirming a direct role of the N-terminal coil-coil domain. Quantity and quality of the GST and His fusion proteins used in the assay were monitored by SDS-PAGE and Coomassie Blue staining. An asterisk shows the position of the His-Snapin band. . Snapin bound to the MBP-SNAP23 beads was monitored by Western blot. As a control, the amount of MBP-SNAP23 present in each reaction was also determined. The amount of His-Snapin associated with MBP-SNAP23 decreased with increasing concentrations of GST-Exo70 or Exo70, suggesting that Snapin and Exo70 compete for binding to SNAP23.
insulin, which results in the translocation of GLUT4 to the plasma membrane (Fig. 5A), did not affect the subcellular distribution of Snapin. As previously reported for the exocyst (9), an enrichment of Exo70 in the plasma membrane fraction was observed after insulin stimulation of 3T3-L1 adipocytes.
Snapin Modulates Insulin-stimulated GLUT4 Trafficking-To explore a possible functional role for Snapin in insulin-regulated GLUT4 vesicle trafficking, we next studied the effect of modulating Snapin expression on insulin action in adipocytes. 3T3-L1 adipocytes were infected with lentivirus expressing full-length Myc-Snapin. Despite some variability in Snapin expression between individual cells, immunofluorescence microscopy revealed that Ͼ90% of the cells in the culture were expressing Myc-Snapin. Moreover, immunoblotting revealed a protein product of the expected size for Myc-Snapin in adipocytes (data not shown). The localization of Myc-Snapin was indistinguishable from the endogenous protein indicating that the protein was not vastly overexpressed using this approach (Fig. 5B). Cells overexpressing Myc-Snapin did not exhibit any impairment in insulin-stimulated glucose transport (data not shown). We also expressed a range of Snapin mutants in adipocytes using this same approach, but again we were unable to detect any impairment in insulin action (data not shown).
We next examined the effects of reduced Snapin expression on insulin action using RNA interference. As shown in Fig. 6A, we identified four different siRNAs that efficiently reduced Snapin protein levels when transfected either individually or as a pool into either NIH3T3 or 3T3-L1 cells. The expression of control proteins, either glyceraldehyde-3-phosphate dehydrogenase or 14-3-3 was unaffected under these conditions. As shown in Fig. 6B, reduced Snapin expression had no effect on 2-deoxyglucose uptake in either the absence of insulin or in the presence of a maximum dose of insulin. However, at submaximal concentrations of insulin, where we observed an ϳ2-fold effect of insulin on 2-deoxyglucose uptake in control cells, reduced Snapin expression resulted in a blunted response. These observations implicate a role for Snapin in insulin-stimulated glucose transport in adipocytes likely by modulating the tethering or docking/fusion of GLUT4 containing vesicles with the plasma membrane.

DISCUSSION
Snapin was originally identified in a human brain yeast twohybrid screen as an SNAP25 interacting protein (29). This initial study suggested Snapin was a neuronal-specific protein, Basal uptake was independent of insulin stimulation and unaffected by siRNAs. Submaximal stimulation (1 nM insulin) increased 2-deoxyglucose uptake by ϳ2-fold, and this was significantly inhibited by the Snapin siRNAs as compared with a control siRNA. The inhibitory effect was no longer significant at maximal insulin stimulation (100 nM; n ϭ 3); error bars, ϮS.E. C, silencing of Snapin in differentiated 3T3-L1 adipocytes. In parallel to the 2-deoxyglucose uptake experiment, one set of cells treated with control or Snapin siRNA was used to monitor Snapin expression by Western blot. In differentiated adipocytes, Snapin was silenced to a similar extent as in NIH3T3 cells, but the specific siRNA did not affect expression of 14-3-3, which served as a control.
because a Snapin-specific antibody was unable to detect Snapin in other tissues. It was later demonstrated by Northern blotting and subsequent Western blotting, that Snapin is ubiquitously expressed and in fact also can bind the SNAP25 homologue SNAP23 (22). However, the latter study did not address the functional significance of the SNAP23 and Snapin interaction.
In the present studies, Snapin was identified in a yeast twohybrid screen as an Exo70-binding protein. Snapin was shown to associate, through Exo70, with additional exocyst components, suggesting that it interacts with the native exocyst complex. Buxton and colleagues (22) have previously shown that Snapin can also form a complex with SNAP23 and Syntaxin4 in vitro. This finding is interesting, because Snapin may represent a prime candidate for mediating or regulating exocyst and SNARE complex interactions. Because SNAP23 and Exo70 appear to bind Snapin in the same C-terminal portion, it is interesting to speculate that crosstalk between exocyst and SNARE complexes may be transferred by sequential binding to Snapin.
In vivo binding experiments demonstrated that the Snapin interaction is mediated through the N-terminal 99 amino acids of Exo70. The crystal structure of yeast Exo70p was recently published (31), and sequence alignments suggest that the mammalian orthologues have similar structural domains. Exo70p consists of multiple helical bundles, which are divided into three domains to form an elongated structure. Interestingly, the first 99 amino acids are predominantly surface-exposed residues, which is consistent with them contributing to a protein interaction site. The N terminus of Exo70 also serves as the binding site for TC10, and overexpression of the N terminus of Exo70 (Exo70-N: amino acid 1-384) in 3T3-L1 adipocytes was shown to inhibit fusion of GLUT4 vesicles with the plasma membrane (9). The effect of this mutant is thought to be due to competitive displacement of endogenous Exo70 for TC10 binding sites. Based on our findings, an alternative explanation is that the Exo70-N mutant competes for binding of Snapin on the plasma membrane.
Experiments in chromaffin cells suggest a positive regulatory role for Snapin in exocytosis (30). This study showed that Snapin is phosphorylated by protein kinase A, resulting in its enhanced affinity for SNAP25 and an increased association of synaptotagmin with SNARE complexes. Overexpression of a phosphomimetic Snapin mutant resulted in increased exocytosis of large dense core vesicles. Earlier studies demonstrated that overexpression of the C terminus of Snapin or injection of Snapin peptides corresponding to the SNAP25 binding site have a slight inhibitory effect on synaptic transmission (29). In this study, knockdown of Snapin caused a reduction in insulinstimulated glucose transport. This effect was only observed at submaximal insulin concentrations likely, because the knockdown of endogenous Snapin was incomplete. Snapin knockout mice have recently been described (32), and consistent with our findings, Snapin was shown to have a modulatory role in neurosecretion, because the absence of Snapin resulted in significantly reduced exocytosis in embryonic chromaffin cells. We speculate that reduction of Snapin may impair the dialogue between the exocyst and SNARE complex assembly.
We did not observe any consequence of over expression of Snapin or C-terminal truncation mutants on insulin-stimu-lated glucose transport in adipocytes (data not shown). This may denote that these proteins were not expressed at sufficiently high levels to observe a phenotype. This interpretation is consistent with a more recent publication re-investigating the role of Snapin in neurosecretion (33). In contrast to the two prior studies, Vites and co-authors were unable to observe defects in synaptic transmission induced by Snapin mutant overexpression. The overexpression studies performed in the present studies, however, do not exclude a role for Snapin in GLUT4 trafficking, because it cannot be confirmed that overexpression of Snapin wt and/or mutants would have a dominant negative effect. The presence of the endogenous Snapin may be sufficient to allow normal GLUT4-trafficking events. Our data thus suggest that, although Snapin may not ultimately be required, it plays a positive role, possibly by regulating the fidelity of GLUT4 vesicle docking. Future work will elucidate the precise mechanism of action of Snapin in targeting, docking, or fusion of GLUT4 vesicles and explore the physiological role of Snapin in glucose homeostasis.
The present studies have thus identified a novel interaction of Snapin with the exocyst and a novel modulatory role for Snapin in insulin-stimulated GLUT4 vesicle transport in adipocytes.