Glut4 exocytosis in adipocytes uses protein machinery that is shared with other regulated secretory processes. Synapsins are phosphoproteins that regulate a `reserve pool' of vesicles clustered behind the active zone in neurons. We found that adipocytes (primary cells and the 3T3-L1 cell line) express synapsin IIb mRNA and protein. Synapsin IIb co-localizes with Glut4 in perinuclear vesicle clusters. To test whether synapsin plays a role in Glut4 traffic, a site 1 phosphorylation mutant (S10A synapsin) was expressed in 3T3-L1 adipocytes. Interestingly, expression of S10A synapsin increased basal cell surface Glut4 almost fourfold (50% maximal insulin effect). Insulin caused a further twofold translocation of Glut4 in these cells. Expression of the N-terminus of S10A synapsin (amino acids 1-118) was sufficient to inhibit basal Glut4 retention. Neither wild-type nor S10D synapsin redistributed Glut4. S10A synapsin did not elevate surface levels of the transferrin receptor in adipocytes or Glut4 in fibroblasts. Therefore, S10A synapsin is inhibiting the specialized process of basal intracellular retention of Glut4 in adipocytes, without affecting general endocytic cycling. While mutant forms of many proteins inhibit Glut4 exocytosis in response to insulin, S10A synapsin is one of only a few that specifically inhibits Glut4 retention in basal adipocytes. These data indicate that the synapsins are important regulators of membrane traffic in many cell types.

In adipocytes and muscle, glucose transport is controlled through the regulation of the total number of facilitative glucose transporters, largely the Glut4 isoform, expressed at the cell surface (Bryant et al., 2002; Foster and Klip, 2000). The increase in cell surface Glut4 occurs through a regulated secretory process. In response to insulin, Glut4 is translocated from an intracellular, vesicular pool to the plasma membrane. The large increase in glucose transport observed in adipocytes is the product of both the efficient sequestration of Glut4 under basal conditions, and the rapid redistribution of Glut4 from intracellular storage compartments to the plasma membrane after insulin stimulation.

The insulin-stimulated exocytosis of Glut4 utilizes protein machinery that is shared by other regulated secretory processes. These proteins include homologues or isoforms of synaptobrevins/VAMPS, syntaxins, SNAP-23/25, Muncs and Rabs (Bryant et al., 2002; Foster and Klip, 2000). While these proteins are necessary components of the regulated exocytic machinery, they are also components of the general membrane trafficking machinery in most cell types. The VAMPs, syntaxins and SNAP-23/25, along with NSF, α-SNAP and γ-SNAP, constitute the basic membrane fusion machinery in cells, while the Rabs and Muncs play modulatory roles. An additional set of proteins required for synaptic vesicle cycling, including synaptotagmin, synaptojanin, synaptophysin/amphiphysin and dynamin, also play basic roles in vesicle trafficking in all cell types (Sudhof, 2004).

One notable exception to the nearly ubiquitous expression of the synaptic vesicle-associated proteins was a family of highly conserved proteins termed the synapsins. Recently, however, synapsin I was found in both pancreatic β-cells (Krueger et al., 1999; Longuet et al., 2005; Matsumoto et al., 1999) and in liver epithelial cells (Bustos et al., 2001). Its function in non-neuronal cells is unknown. Genetic and biochemical knock-down experiments have shown that in neurons the synapsins maintain and organize a reserve pool of synaptic vesicles that are clustered behind the active zone (Bloom et al., 2003; Gitler et al., 2004a; Hilfiker et al., 1998; Lonart and Simsek-Duran, 2006; Pieribone et al., 1995; Rosahl et al., 1995; Ryan et al., 1996; Sun et al., 2006; Takei et al., 1995). Exactly where the synapsins function in the synaptic vesicle cycle is not fully understood (Sudhof, 2004). The synapsins are peripheral membrane proteins that coat synaptic vesicles and cluster vesicles behind the active zone within an actin-rich filamentous matrix (Bloom et al., 2003; De Camilli et al., 1983; Huttner et al., 1983). Synapsins cluster synaptic vesicles by binding to both vesicles and to each other (Brautigam et al., 2004; Gitler et al., 2004b; Hosaka and Sudhof, 1999). They were originally identified as major substrates for PKA and CaM kinases in the brain, and phosphorylation of the synapsins at multiple sites can affect their interaction with lipids and proteins. For example, phosphorylation at site 1 (serine 9/10, a site shared by all synapsin isoforms) decreases the affinity of the synapsins for liposomes in vitro (Hosaka et al., 1999) and is necessary for PKA-dependent dispersion of synapsin in neurons (Bonanomi et al., 2005). However, the precise role of synapsin phosphorylation in the regulation of synaptic vesicle trafficking remains unclear (Hosaka et al., 1999).

Fig. 1.

Synapsin IIb mRNA and protein are expressed in adipocytes. (A) Northern blots of RNA samples (10 μg) from mouse brain (Br) or 3T3-L1 adipocytes (Ad) probed with a 185 bp pan-synapsin C domain probe, a 649 bp synapsin II C domain probe or a probe from the synapsin IIb 3′-untranslated region (UTR). (B,C) Anti-synapsin western blotting with two affinity-purified antibodies raised against the C domain of synapsin IIb (antibodies 69 and 83) from samples (10 μg) of (B) mouse brain post-nuclear supernatant or (C) 3T3-L1 adipocyte (L1), primary adipocyte (1°), or liver low density microsomes. Solid boxes indicate that the samples were run on the same gel; dashed lines indicate that the order of samples on the gel was altered in the image used in the figure.

Fig. 1.

Synapsin IIb mRNA and protein are expressed in adipocytes. (A) Northern blots of RNA samples (10 μg) from mouse brain (Br) or 3T3-L1 adipocytes (Ad) probed with a 185 bp pan-synapsin C domain probe, a 649 bp synapsin II C domain probe or a probe from the synapsin IIb 3′-untranslated region (UTR). (B,C) Anti-synapsin western blotting with two affinity-purified antibodies raised against the C domain of synapsin IIb (antibodies 69 and 83) from samples (10 μg) of (B) mouse brain post-nuclear supernatant or (C) 3T3-L1 adipocyte (L1), primary adipocyte (1°), or liver low density microsomes. Solid boxes indicate that the samples were run on the same gel; dashed lines indicate that the order of samples on the gel was altered in the image used in the figure.

Given the roles for the synapsins in the maintenance, subcellular localization, and regulation of synaptic vesicles, we postulated that a similar protein might be involved in the regulation of Glut4 exocytosis. To initially test this hypothesis, we explored whether isoforms of synapsin were also expressed in adipocytes and whether expression of functional mutants of synapsin (S10A and S10D synapsin IIb) affects Glut4 traffic.

Adipocytes express synapsin II mRNA

Expression of synapsin mRNA in adipocytes was analyzed by PCR, library screening and northern blotting. Synapsin I, II and III are expressed from three different genes, each expressing multiple splice variants (Porton et al., 1999; Sudhof et al., 1989). They are highly homologous in their N-terminal domains (designated A, B and C) but differ in their C-terminal domains. Degenerate oligonucleotide primers were used to amplify a highly conserved 185 base pair (bp) C-domain sequence shared by all synapsin isoforms (1011-1196 in synapsin IIb) from a 3T3-L1 adipocyte cDNA library. Using this product as a probe, five independent clones of synapsin II (a and b) were isolated by colony hybridization from two independently constructed 3T3-L1 (L1) adipocyte cDNA libraries (7×105 colonies were screened). Synapsin IIa and IIb are splice variants encoded by a single gene, identical in their four N-terminal domains (ABCG) and differing only in their final C-terminal domains (a:HE or b:I) and 3′-untranslated regions. Eight additional independent clones of synapsin II were isolated by RACE using a mixed 3′ oligonucleotide primer. No clones of synapsin I or III were isolated. Although no full-length clones were isolated, all were complete through the C domain, and sequences were the same as the mRNA sequences of the synapsin II expressed in the brain, including the 3′ untranslated regions.

The expression of synapsin II mRNA was verified by northern blotting (Fig. 1A). Synapsin transcripts in total RNA prepared from mouse brain or L1 adipocytes were detected using three different probes: the 185 bp probe conserved in synapsins I and II, a 649 bp synapsin-II-specific C-domain probe, and a synapsin-IIb-specific 3′-untranslated region (UTR) probe. Three transcripts were detected in the brain samples with the 185 bp probe (3.9, 3.4 and 2.9 kb) corresponding to synapsins IIb, Ia/b and IIa (expected sizes: 4.2, 3.3 and 3.0 kb) (Sudhof et al., 1989). The 649 bp synapsin II probe detected the two synapsin II transcripts, but not the synapsin I transcripts. The synapsin IIb 3′-UTR probe detected only the synapsin IIb transcript. Two transcripts were detected in the adipocyte samples, synapsin IIb (3.9 kb) and IIa (2.7 kb). Consistent with the cloning data, no synapsin I transcripts were detected in the adipocyte samples.

Adipocytes express synapsin II protein

Two independent affinity-purified polyclonal antibodies against the C domain of synapsin II were made. Both antibodies (69 and 83) recognized synapsins Ia, Ib, IIa and IIb (86, 80, 74 and 55 kDa) (Sudhof et al., 1989) in brain samples (Fig. 1B). Both antibodies recognized a protein of the approximate molecular mass of synapsin IIb in samples from L1 and primary adipocytes (Fig. 1C, Fig. 2A). Synapsin IIa was not detected. It may be expressed, but at levels beneath the sensitivity of detection. Synapsin I was not detected in samples from adipocytes, but was detected in samples prepared from liver, as has been previously reported (Bustos et al., 2001). Synapsin IIb is also expressed in liver.

To determine the subcellular localization of synapsin IIb in adipocytes, fractions from both primary and L1 adipocytes were analyzed (Fig. 2). Glut4 was found primarily in the low density microsome (LDM) fraction of primary adipocytes under basal conditions (Fig. 2A). In response to insulin, the Glut4 in this fraction decreased by 50% and redistributed to the plasma membrane (PM) fraction. Like Glut4, the synapsin IIb is also enriched in the LDM fraction in basal cells. Unlike Glut4, however, synapsin IIb is not translocated out of the LDM fraction in response to insulin. Both Glut4 and synapsin IIb were also enriched in the 150,000 g LDM pellet fraction isolated from L1 adipocytes (Fig. 2B). The distribution of synapsin IIb is similar to the distribution of the SCAMPs, known protein components of Glut4 vesicles in adipocytes (Cain et al., 1992; Laurie et al., 1993). The SCAMPS are highly enriched in the LDM fraction in Glut4-containing vesicles but do not translocate with Glut4 to the PM fraction in response to insulin.

Fig. 2.

Distribution of synapsin IIb protein in subcellular fractions. (A) Fractions from basal and insulin stimulated primary adipocytes; M/N, mitochondria/nuclei; PM, plasma membrane, HDM, high density microsomes; LDM, low density microsomes; cyt, cytosol. (B) Soluble (S) and particulate (P) fractions from basal L1 adipocytes (Ad) prepared by differential centrifugation (1 K=1000 g). (C) Subcellular fractions from L1 fibroblasts (Fib). Equal protein from each fraction was analyzed (10 μg for synapsin and 1 μg for Glut4) and proteins were detected using antibodies specific for Glut4 or synapsin IIb (Syn; two antibodies used, 83 and 69). In the synapsin blots using antibody 69, an additional non-specific band was detected that was also observed when the blots were probed with secondary antibody alone (2°).

Fig. 2.

Distribution of synapsin IIb protein in subcellular fractions. (A) Fractions from basal and insulin stimulated primary adipocytes; M/N, mitochondria/nuclei; PM, plasma membrane, HDM, high density microsomes; LDM, low density microsomes; cyt, cytosol. (B) Soluble (S) and particulate (P) fractions from basal L1 adipocytes (Ad) prepared by differential centrifugation (1 K=1000 g). (C) Subcellular fractions from L1 fibroblasts (Fib). Equal protein from each fraction was analyzed (10 μg for synapsin and 1 μg for Glut4) and proteins were detected using antibodies specific for Glut4 or synapsin IIb (Syn; two antibodies used, 83 and 69). In the synapsin blots using antibody 69, an additional non-specific band was detected that was also observed when the blots were probed with secondary antibody alone (2°).

Glut4 is not expressed in fibroblasts and is upregulated with adipocyte differentiation (Fig. 2B). In contrast to Glut4, synapsin IIb expression did not increase upon differentiation and its subcellular distribution in the fibroblasts was the same as that observed in the adipocytes (Fig. 2C). Many of the protein components necessary for Glut4 traffic are expressed in both cell types because they are also required for general membrane traffic (Bryant et al., 2002). Therefore, synapsin IIb might play a role in general membrane trafficking as well as in regulated exocytosis.

Our fractionation data is consistent with co-localization of Glut4 and synapsin. However, we could not use immunoisolation of vesicles to determine whether the synapsin in the LDM was directly associated with Glut4-containing vesicles, as we have done previously for a number of other proteins (Cain et al., 1992; Laurie et al., 1993; Mastick et al., 1994; Mastick and Falick, 1997). Synapsins are released from synaptic vesicles by treatment with as little as 40 mM NaCl and are insoluble in Triton-X 100 (TX100) in the absence of salt (Huttner et al., 1983). Synapsin IIb in the LDM fractions prepared from adipocytes (primary or L1) is also released into the soluble fraction by salt treatment (supplementary material Fig. S1). However, in the absence of salt, both the Glut4 and the synapsin in the LDM fractions non-specifically binds to the affinity matrices (data not shown). We therefore turned to fluorescence microscopy and functional studies to examine the interaction of synapsin IIb and Glut4-containing compartments.

Synapsin co-localizes with Glut4 in perinuclear clusters

To examine the co-localization of synapsin with Glut4 vesicles, differentiated L1 adipocytes were infected with recombinant adenoviruses that express HA-Glut4-GFP and co-express Flag-epitope tagged synapsin IIb from a single virus. The Flag epitope was added because the synapsin antibodies used for western blotting are not suitable for microscopy or immunoprecipitation experiments. The trafficking of this Glut4 reporter has been well characterized and it behaves like the endogenous Glut4 in adipocytes (Dawson et al., 2001; Zeigerer et al., 2002). To examine the functional significance of synapsin expression in adipocytes, mutant forms of synapsin IIb in which Ser10 was substituted with alanine (S10A) or aspartic acid (S10D) were also expressed. Ser10 is the only site of phosphorylation reported in synapsin IIb and these mutations affect synapsin function in a dominant fashion when expressed in neurons (Bonanomi et al., 2005; Chi et al., 2001; Chi et al., 2003; Kao et al., 2002). Cells infected with the viruses expressed similar levels of HA-Glut4-GFP (supplementary material Fig. S2A,B). The total Glut4 expressed in infected L1 adipocytes was not significantly elevated relative to uninfected cells, indicating that the HA-Glut4-GFP is expressed at relatively low levels (supplementary material Fig. S2C,D). At this low level of expression the expressed HA-Glut4-GFP does not saturate the basal retention machinery. There is no increase in the basal surface-to-total ratio observed as the amount expressed increases over the entire range of expression observed in these experiments (supplementary material Table S1). The Flag-synapsins were also expressed similarly, at levels approximating the expression level in the brain (supplementary material Fig. S2E). The tandem expression system ensures that all cells expressing the GFP-tagged Glut4 also express Flag-synapsin (verified by flow cytometry; supplementary material Fig. S2F).

The co-localization of Flag-synapsin with Glut4 was examined in infected L1 adipocytes (Fig. 3). Flag-epitope was detected in cells co-expressing both HA-Glut4-GFP and Flagsynapsin, but not in cells expressing Glut4 alone. WT-synapsin was found both in the cytosol (diffuse labeling) and enriched in perinuclear clusters where it co-localized with Glut4. Both S10A and S10D synapsin were also enriched in the perinuclear region in clusters co-localized with Glut4. The co-localization of synapsin and Glut4 can be seen in multiple adjacent Z-stack images (supplementary material Fig. S3). In all cells there were additional clusters of Glut4 vesicles that were not enriched in synapsin. As observed in the cell fractionation experiments, stimulation with insulin did not cause redistribution of the synapsin (data not shown). The co-localization of synapsin with a population of clustered Glut4 vesicles in adipocytes is similar to the co-localization of synapsin with a clustered pool of synaptic vesicles (the reserve pool) observed in nerve terminals (Bloom et al., 2003).

S10A synapsin inhibits Glut4 intracellular retention in basal adipocytes

Consistent with a role in Glut4 traffic, expression of S10A synapsin caused a redistribution of Glut4 in basal adipocytes (Fig. 4). In control basal adipocytes, the Glut4-GFP appeared punctate and distributed throughout the cell, with some clustering in the perinuclear region and very little at the cell surface (Fig. 4A). Insulin caused a loss of Glut4 from the small punctate structures and Glut4 was now detected at the plasma membrane (diffuse and rim fluorescence). Expression of WTsynapsin did not change the distribution of Glut4 in either basal or insulin-stimulated cells. By contrast, expression of S10A synapsin caused a clear redistribution of Glut4. S10A synapsin reduced the number of small punctate Glut4-containing vesicles and increased diffuse plasma membrane staining (also seen as rim fluorescence in z-stack images, supplementary material Fig. S3). Images of randomly selected infected adipocytes (∼60 for each condition) were scored for cell surface Glut4 (rim fluorescence and diffuse fluorescence; similar to G4, Insulin) or intracellular Glut4 (dispersed, punctate fluorescence; similar to G4, Basal). Glut4 was redistributed to the plasma membrane and perinuclear clusters in approximately 90% of the basal cells expressing S10A synapsin (Fig. 4B). In all cell types, Glut4 was localized at the cell surface and perinuclear clusters in greater than 90% of cells after insulin stimulation.

Fig. 3.

Synapsin IIb and Glut4 are co-localized. L1 cells were differentiated on coverslips, then infected with virus expressing HA-Glut4-GFP alone (G4), or coexpressed with WT-Flag-synapsin IIb (G4/S), S10A Flag-synapsin IIb (G4/S10A), or S10D Flag-synapsin IIb (G4/S10D). Two days post-infection they were fixed, permeabilized, and labeled with anti-Flag antibody. Representative confocal images of Glut4 (GFP: green) and Flag-synapsin (Cy3: red) are shown. Images are maximum projections of 10-0.5 μm slices through the center of the cell (montages of the individual slices are shown in supplementary material Fig. S3).

Fig. 3.

Synapsin IIb and Glut4 are co-localized. L1 cells were differentiated on coverslips, then infected with virus expressing HA-Glut4-GFP alone (G4), or coexpressed with WT-Flag-synapsin IIb (G4/S), S10A Flag-synapsin IIb (G4/S10A), or S10D Flag-synapsin IIb (G4/S10D). Two days post-infection they were fixed, permeabilized, and labeled with anti-Flag antibody. Representative confocal images of Glut4 (GFP: green) and Flag-synapsin (Cy3: red) are shown. Images are maximum projections of 10-0.5 μm slices through the center of the cell (montages of the individual slices are shown in supplementary material Fig. S3).

Expression of S10A synapsin also increased the apparent size of the perinuclear clusters in basal adipocytes. This could be scored by the computer using pixel intensity threshold protocols to analyze the distribution of Glut4-GFP in the 60 images. The threshold was set to eliminate the diffuse cell surface staining, leaving only the brightest clusters. The area of the base of each intensity peak was determined as an approximate measure of object size (Fig. 4D; the position of the cell in the image is indicated in the x-y plane and pixel intensity is plotted on the z-axis). Multiple objects were found in each cell (∼10 in cells expressing HA-Glut4-GFP alone or co-expressing WT-synapsin, and five in cells co-expressing S10A synapsin). The Glut4-containing structures were on average three times larger in basal S10A synapsin expressing cells than in basal cells expressing HA-Glut4-GFP alone or co-expressing WT-synapsin (Fig. 4C). If only the largest clusters are analyzed (greater than 100 pixels; 1-2 objects per cell), the objects are still twice as big in basal S10A synapsin expressing cells than in the other two cell types. After insulin stimulation there were fewer objects detected in all cell types (∼2-4 per cell). The average size of the clusters in the S10A synapsin expressing cells decreased after stimulation.

To determine whether the Glut4 translocated to the cell periphery in S10A synapsin expressing cells was inserted into the plasma membrane, the accessibility of the exofacial HAepitope of the Glut4-GFP reporter construct to antibody labeling was measured in intact cells. Infected L1 adipocytes were treated with or without insulin then labeled at 4°C. The cells were then detached from the plate and analyzed by flow cytometry (supplementary material Fig. S2). This technique allows rapid measurement of fluorescence values in a large number of individual cells (∼5000 infected cells/sample). The mean fluorescence ratios (MFR) of surface Glut4 (anti-HA) to total Glut4 (GFP) were calculated (Fig. 5A). Cells expressing HA-Glut4-GFP alone showed low surface labeling under basal conditions, which increased approximately sevenfold with insulin stimulation. Nearly identical results were observed in cells co-expressing WT-synapsin, indicating that expression of Flag-synapsin does not alter Glut4 traffic. In contrast, expression of S10A synapsin caused a three- to four-fold increase in surface Glut4 in basal cells. Insulin caused a further twofold translocation of Glut4 in S10A synapsin expressing cells (Fig. 5B), indicating that these cells are still capable of responding to insulin. Therefore, S10A synapsin specifically affects basal Glut4 retention. The effect of S10A synapsin on Glut4 localization is specific for the serine-to-alanine mutation. Expression of S10D synapsin did not affect either basal retention or insulin-stimulated translocation of Glut4. Serine-to-aspartic acid or serine-to-glutamic acid mutants can act like pseudo-phosphorylated forms of synapsin (Hosaka et al., 1999; Kao et al., 2002). This indicates that S10A synapsin is inhibiting Glut4 retention because it is unable to be phosphorylated, rather than simply because of a change in the synapsin primary sequence.

Fig. 4.

S10A synapsin redistributes Glut4 in adipocytes. L1 adipocytes prepared as described in Fig. 3 were treated with or without insulin (100 nM, 30 minutes), then fixed and analyzed. zstack confocal images of GFP fluorescence were collected from ∼60 infected adipocytes of each type and maximum projections were analyzed. (A) Representative images. (B) Percentage of cells in the maximum projections with Glut4 at or near the plasma membrane (rim + diffuse fluorescence) in basal (white) or insulin-stimulated (black) cells. (C) Average object sizes (contiguous area with green intensity greater than the threshold in maximum projections; mean ± s.e.m., standardized to Glut4 basal). (D) Histograms of intensity distributions (shown on z-axis) of GFP fluorescence in the representative images shown in A (G4, G4/S, G4/S10A, left to right).

Fig. 4.

S10A synapsin redistributes Glut4 in adipocytes. L1 adipocytes prepared as described in Fig. 3 were treated with or without insulin (100 nM, 30 minutes), then fixed and analyzed. zstack confocal images of GFP fluorescence were collected from ∼60 infected adipocytes of each type and maximum projections were analyzed. (A) Representative images. (B) Percentage of cells in the maximum projections with Glut4 at or near the plasma membrane (rim + diffuse fluorescence) in basal (white) or insulin-stimulated (black) cells. (C) Average object sizes (contiguous area with green intensity greater than the threshold in maximum projections; mean ± s.e.m., standardized to Glut4 basal). (D) Histograms of intensity distributions (shown on z-axis) of GFP fluorescence in the representative images shown in A (G4, G4/S, G4/S10A, left to right).

Fig. 5.

S10A synapsin inhibits basal intracellular retention of Glut4 in adipocytes. L1 adipocytes differentiated on culture dishes were infected as described in Fig. 3. Cells were treated ± insulin (100 nM, 30 minutes), placed on ice, then labeled with anti-HA (detected with PC7). After washing, cells were lifted from the plate with collagenase at 4°C, and analyzed by flow cytometry. MFR, mean fluorescence ratio (PC7/GFP). ∼5000 cells/sample. Mean values from uninfected cells were used to correct for autofluorescence and non-specific binding. MFRs were standardized to G4 basal samples. (A) Average MFR ± s.d. from three experiments. (B) Average foldstimulation (MFR Insulin/MFR basal) ± s.d. from three experiments. (C) Comparison of the effects of S10A synapsin and S10A dsRED, a fusion protein with the first 118-amino acids of S10A synapsin fused to dsRED. White, basal; black, insulin; grey, ratio. Statistical analysis (Bonferroni, Scheffe's and Tukey tests; Origin 7.5) was done after standardizing to either the basal or insulin stimulated controls (***P<0.01 significant difference; ns, no significant difference, P>0.05).

Fig. 5.

S10A synapsin inhibits basal intracellular retention of Glut4 in adipocytes. L1 adipocytes differentiated on culture dishes were infected as described in Fig. 3. Cells were treated ± insulin (100 nM, 30 minutes), placed on ice, then labeled with anti-HA (detected with PC7). After washing, cells were lifted from the plate with collagenase at 4°C, and analyzed by flow cytometry. MFR, mean fluorescence ratio (PC7/GFP). ∼5000 cells/sample. Mean values from uninfected cells were used to correct for autofluorescence and non-specific binding. MFRs were standardized to G4 basal samples. (A) Average MFR ± s.d. from three experiments. (B) Average foldstimulation (MFR Insulin/MFR basal) ± s.d. from three experiments. (C) Comparison of the effects of S10A synapsin and S10A dsRED, a fusion protein with the first 118-amino acids of S10A synapsin fused to dsRED. White, basal; black, insulin; grey, ratio. Statistical analysis (Bonferroni, Scheffe's and Tukey tests; Origin 7.5) was done after standardizing to either the basal or insulin stimulated controls (***P<0.01 significant difference; ns, no significant difference, P>0.05).

Interestingly, expression of a fusion protein containing only the first 118 amino acids of S10A synapsin increased cell surface Glut4 fourfold (Fig. 5C; S10A dsRED). This construct includes the hydrophilic N-terminal A and B domains and phosphorylation site 1, but lacks the large, highly conserved, hydrophobic C-domain that contains the synapsin, Ca2+- and ATP-binding sites (Brautigam et al., 2004; Hosaka and Sudhof, 1999). The A and B domains exist in an extended conformation whereas the ∼300 amino acid C-domain is independently folded.

S10A synapsin does not inhibit general endocytic trafficking

To determine whether elevated Glut4 at the plasma membrane in basal cells expressing S10A synapsin was due to a nonspecific effect on endocytic cycling, the steady state distribution of the transferrin receptor in adipocytes was determined by flow cytometry. The transferrin receptor rapidly cycles between the plasma membrane and endosomes under both basal and insulin-stimulated conditions, and is commonly used as a marker for general clathrin-mediated endocytosis and constitutive exocytosis from endosomes. Insulin increased the levels of the transferrin receptor at the cell surface ∼2.5-fold in adipocytes (Fig. 6A), in agreement with previous observations (Tanner and Lienhard, 1987). S10A synapsin had no effect on transferrin receptor distribution under either basal or insulin stimulated conditions.

When Glut4 is exogenously expressed in fibroblasts, it is not efficiently retained within cells (40-50% of the Glut4 is localized to the cell surface, compared to less than 10% in adipocytes) (El-Jack et al., 1999; Lampson et al., 2001; Zeigerer et al., 2002). The intracellular Glut4 co-localizes with the transferrin receptor in endosomes in these cells. Consistent with the transferrin receptor data, S10A synapsin did not increase cell surface Glut4 content in fibroblast cells (Fig. 6B). Approximately 40% of the total cycling pool (total antibody accumulated in cells incubated at 37°C for 120 minutes) was at the cell surface in fibroblasts expressing HA-Glut4-GFP alone, whereas 30% was at the plasma membrane in cells co-expressing S10A synapsin. The total cycling pool was approximately equal in all three cell types, and was ∼90% of the total pool in all three (as measured in fixed, permeabilized cells labeled with anti-HA, data not shown).

Fig. 6.

S10A synapsin does not inhibit general endocytic trafficking. (A) Surface transferrin receptor content in basal (white) or insulinstimulated (black) cells. L1 adipocytes differentiated, infected and treated with insulin as described in Fig. 6 were labeled with PE-antitransferrin receptor antibody at 4°C and mean fluorescence intensity (MFI) determined by flow cytometry. Average MFI ± s.d. of samples from four experiments, standardized to G4 basal. (B) Infected fibroblasts were labeled for 5 minutes (surface Glut4, white) or 120 minutes (total cycling pool, grey) with anti-HA antibody at 37°C, fixed, permeabilized and anti-HA detected with Cy3-labeled secondary antibody. Samples were analyzed by wide-field microscopy. MFRs were calculated for 18 fields of cells (six fields per experiment, three experiments, 20-50 cells/field). Average MFRs ± s.d., expressed as % total cycling HA-Glut4.

Fig. 6.

S10A synapsin does not inhibit general endocytic trafficking. (A) Surface transferrin receptor content in basal (white) or insulinstimulated (black) cells. L1 adipocytes differentiated, infected and treated with insulin as described in Fig. 6 were labeled with PE-antitransferrin receptor antibody at 4°C and mean fluorescence intensity (MFI) determined by flow cytometry. Average MFI ± s.d. of samples from four experiments, standardized to G4 basal. (B) Infected fibroblasts were labeled for 5 minutes (surface Glut4, white) or 120 minutes (total cycling pool, grey) with anti-HA antibody at 37°C, fixed, permeabilized and anti-HA detected with Cy3-labeled secondary antibody. Samples were analyzed by wide-field microscopy. MFRs were calculated for 18 fields of cells (six fields per experiment, three experiments, 20-50 cells/field). Average MFRs ± s.d., expressed as % total cycling HA-Glut4.

Synapsin IIb is phosphorylated at site 1 (Ser10) in adipocytes

The redistribution of Glut4 by S10A synapsin implies that phosphorylation of synapsin is required for proper intracellular retention of Glut4 in basal adipocytes. To examine whether synapsin is phosphorylated in basal adipocytes, western blotting with a site-1-specific phospho-synapsin antibody was used to examine the phosphorylation state of synapsin (Fig. 7). Synapsin phosphorylation was detected in brain lysates (Fig. 7A), consistent with previous reports showing constitutive phosphorylation in neurons (Bonanomi et al., 2005; Hosaka et al., 1999; Menegon et al., 2000). Wild-type synapsin is also phosphorylated in basal L1 adipocytes. The percentage of the total pool of synapsin IIb that is phosphorylated at site 1 in adipocytes is higher than the percentage phosphorylated in the brain, when corrected for equal amounts of synapsin IIb loaded. The steady state levels of site 1 phosphorylation were the same in basal cell lysates and in lysates from adipocytes after 5, 10 and 30 minutes of insulin stimulation (Fig. 7B; and data not shown). However, insulin might induce a change in phosphorylation of a small, specific pool of synapsin that is undetectable in the whole-cell lysates. S10A synapsin and λ-phosphatase-treated wild-type synapsin (λ-PPase) were used as blotting controls and showed no labeling (Fig. 7B,C).

We have shown that adipocytes (both isolated primary adipocytes and cells differentiated in culture) express synapsin IIb. We used four different experimental approaches: PCR, library screening, northern analysis and western analysis (Fig. 1). While synapsin I has previously been found in non-neuronal cells (pancreatic β-cells and liver epithelial cells), this is the first report to show expression of synapsin II outside neurons and to present functional data supporting a role for synapsin in membrane traffic in non-neuronal cells.

To examine the role of synapsin in Glut4 traffic, wild-type synapsin IIb and two mutant forms of synapsin, S10A and S10D, were expressed in L1 adipocytes. Ser10 is the only site of phosphorylation reported in synapsin IIb, and the S10A and S10D mutations affect synapsin function in a dominant fashion when expressed in neurons (Bonanomi et al., 2005; Chi et al., 2001; Chi et al., 2003; Kao et al., 2002). S10A synapsin acted as a dominant inhibitor of basal Glut4 retention in adipocytes (Figs 4, 5). S10A synapsin had no effect on Glut4 distribution after insulin stimulation (Fig. 5) and did not increase cell surface transferrin receptor in adipocytes or Glut4 in fibroblasts (Fig. 6). Although many mutant proteins inhibit insulin-stimulated Glut4 exocytosis in adipocytes, S10A synapsin is one of only a small number of proteins that specifically inhibits basal Glut4 retention. Others include activated signaling molecules (PI3 kinase and Akt) that presumably modify the activity of the retention machinery proteins (Eyster et al., 2005; Frevert et al., 1998) and overexpression of `cargo' (Glut4 or IRAP), presumably through saturation of the retention mechanism (Carvalho et al., 2004; Waters et al., 1997). Glut4 retention is also inhibited by siRNA-induced knockdown of two proteins, AS160 and Golgin-160 (Eguez et al., 2005; Larance et al., 2005; Williams et al., 2006). Interestingly, expression of a dominant-interfering mutant of syntaxin 16 transiently inhibits Glut4 retention by delaying clearance of Glut4 from the rapidly cycling endosomal or plasma membrane pool back into the slowly cycling or non-cycling storage compartment after insulin withdrawal (Proctor et al., 2006).

The effect of S10A synapsin on Glut4 localization is specific for the serine to alanine mutation. Overexpression of WT or S10D synapsin did not affect Glut4 clustering or translocation. This indicates that it is the inability to phosphorylate the S10A synapsin that is inhibiting Glut4 retention, rather than overexpression of synapsin. Consistent with this, wild-type synapsin is highly phosphorylated at site 1 when expressed in basal adipocytes (Fig. 7). Insulin did not change the steady state level of synapsin phosphorylation. Although a number of treatments that stimulate synaptic vesicle exocytosis increase the steady state level of synapsin site 1 phosphorylation, a change in synapsin phosphorylation is not observed with all stimuli and is not absolutely required for regulated synaptic vesicle exocytosis (Bonanomi et al., 2005; Chi et al., 2001). The function of synapsin phosphorylation in the synaptic vesicle cycle remains unclear (Hosaka et al., 1999). It is possible that site 1 synapsin phosphorylation may be required for `turnover' of synapsin complexes during membrane trafficking. S10A synapsin may act as a dominant negative by preventing reuse and/or recycling of the components of a synapsin complex. It is also possible that insulin is changing the phosphorylation of a specific pool of synapsin that is not detected in the whole cell lysates.

Fig. 7.

Synapsin is phosphorylated at site 1 in basal adipocytes. Western blotting with site-1-specific phospho-synapsin (pSyn) and pansynapsin (Syn) antibodies. (A) Samples from mouse brain (Br) or 3T3-L1 adipocytes expressing WT-Flag-synapsin IIb (L1). (B) Samples from basal or insulin-stimulated (100 nM, 30 minutes) 3T3-L1 adipocytes expressing WT or S10A Flag-synapsin IIb. (C) Anti-Flag immunoprecipitates of WTFlag-synapsin incubated ± λ-phosphatase (λ-PPase) ± phosphatase inhibitors.

Fig. 7.

Synapsin is phosphorylated at site 1 in basal adipocytes. Western blotting with site-1-specific phospho-synapsin (pSyn) and pansynapsin (Syn) antibodies. (A) Samples from mouse brain (Br) or 3T3-L1 adipocytes expressing WT-Flag-synapsin IIb (L1). (B) Samples from basal or insulin-stimulated (100 nM, 30 minutes) 3T3-L1 adipocytes expressing WT or S10A Flag-synapsin IIb. (C) Anti-Flag immunoprecipitates of WTFlag-synapsin incubated ± λ-phosphatase (λ-PPase) ± phosphatase inhibitors.

Dominant inhibitory proteins act by binding to and affecting the activity of endogenous wild-type proteins. Therefore, S10A synapsin inhibits Glut4 retention by affecting the activity of a component of the retention machinery that binds to synapsin. Since Glut4 retention is inhibited by expression of S10A synapsin, but not wild-type or S10D synapsin, phosphorylation at site 1 is expected to decrease the physical or functional interaction of synapsin with the retention machinery. S10A synapsin might inhibit proteins that prevent Glut4 storage compartments from moving to and fusing with the plasma membrane, such as AS160 (Eguez et al., 2005; Larance et al., 2005). Alternatively, S10A synapsin might inhibit proteins involved in Glut4 recycling from endosomes to storage compartments or delivery from the biosynthetic pathway, such as Rab 11, GGA, sortilin, syntaxin 16 and Golgin-160 (Li and Kandror, 2005; Lin et al., 1997; Morris et al., 1998; Proctor et al., 2006; Shi and Kandror, 2005; Watson et al., 2004; Williams et al., 2006; Zeigerer et al., 2002; Zeigerer et al., 2004). In either case, Glut4 would be redistributed from the storage compartments into the rapidly cycling endosomal or plasma membrane pool. This redistribution would account for both the increased plasma membrane Glut4 content and the increased perinuclear clustering observed in the S10A synapsin expressing cells, since the early and sorting endosomes that contain the transferrin receptor are normally clustered near the nucleus in adipocytes (Shewan et al., 2003; Zeigerer et al., 2002).

Synapsins are expressed in a number of non-neuronal cell types including cells that are not specialized for regulated secretion (i.e. fibroblasts and epithelial cells). This indicates that like many proteins originally identified as important in synaptic vesicle exocytosis, the synapsins also play a role in general membrane trafficking. However, the specific function of synapsin in cells remains unclear. In liver cells, synapsin I is found in juxtanuclear punctate compartments closely associated with Golgi membranes (Bustos et al., 2001). Its distribution largely overlaps with TGN markers and with the transferrin receptor. Synapsin IIb has a similar distribution in adipocytes, co-localizing with Glut4 in the perinuclear region of the cell (Fig. 3), where both TGN markers and the transferrin receptor are localized (Shewan et al., 2003; Zeigerer et al., 2002). Like Glut4, synaptic vesicle proteins are recycled from the cell surface in part through clathrin-mediated endocytosis and recycling back into regulated secretory vesicles from endosomes (Linstedt and Kelly, 1991; Matteoli et al., 1992; Provoda et al., 2000). Synapsins are not found in clathrincoated vesicles with the other synaptic vesicle proteins. However, they are found on tubular extensions of the recycling endosomes, as well as on the reserve synaptic vesicles (Bloom et al., 2003). These data suggest that the synapsins may be involved in the recycling of proteins from endosomes back to synaptic vesicles. Consistent with this idea, synapsin knockdown leads to a significant decrease in the number of synaptic vesicles and the amount of synaptic vesicle protein in neurons (Bloom et al., 2003; Gitler et al., 2004a; Hilfiker et al., 1998; Lonart and Simsek-Duran, 2006; Pieribone et al., 1995; Rosahl et al., 1995; Ryan et al., 1996; Sun et al., 2006; Takei et al., 1995).

The data from knockout animals indicate that in neurons synapsin plays a key role in regulating the exocytosis of the reserve or resting pool of synaptic vesicles. The literature also indicates that site 1 phosphorylation is important for synapsin function. However, the exact role of this phosphorylation remains unclear. Future work will be aimed at determining the mechanism through which S10A synapsin affects Glut4 trafficking to help elucidate the role that synapsin IIb plays in membrane traffic, in particular the specialized Glut4 traffic in adipocytes.

Cell culture

3T3-L1 fibroblasts (CCL 92.1, American Type Culture Collection) were maintained in DMEM (high glucose, with 2 mM L-glutamine, 50 U/ml penicillin and 50 μg/ml streptomycin) supplemented with 10% calf serum (Hyclone). 1 day post confluence (d 0) the cells were switched to DMEM supplemented with 10% FBS (Hyclone, characterized) and induced to differentiate into adipocytes essentially as described (Frost and Lane, 1985) with minor modifications. Cells were incubated for 3 days in differentiation medium (DMEM supplemented with FBS, dexamethasone, IBMX, and insulin), 3 days in medium supplemented with FBS and insulin, then 4-5 days in medium supplemented with only FBS. Cells were used 8-11 days after initiation of differentiation. 3T3-L1 fibroblasts were used 3 days post confluence. Primary adipocytes were prepared from the epididymal fat pads of Sprague-Dawley rats (175-225 g) by collagenase digestion and fractionated by differential centrifugation (Simpson et al., 1983). Homogenates from 3T3-L1 cells, brain, and liver were also fractionated by differential centrifugation (Shewan et al., 2003). Normal primary human foreskin fibroblasts (from Greg Pari, University of Nevada, Reno, NV) were cultured as described (Sanguinetti et al., 2003).

Northern blotting

Total RNA was isolated by acid guanidium thiocyanate-phenol-chloroform extraction (RNAsol, Biotecx Laboratories). RNA samples (10 μg) were separated by electrophoresis in 1% agarose, 2.2 M formaldehyde, 1× MOPS buffer and transferred by capillary diffusion to nylon membranes. Membranes were prehybridized for 4 hours in FBY hybridization buffer (10% polyethylene glycol, 1.5× SSC, 7% SDS) then hybridized overnight with synapsin probes labeled by random priming. They were then washed at 60°C for 30 minutes in 2× SSC/0.1% SDS, followed by 30 minutes in 0.5× SSC/0.1% SDS, and two times for 30 minutes each in 0.1× SSC/0.1% SDS at 65°C. Hybridization was determined by autoradiography.

Antibodies

Polyclonal antisera were prepared by immunizing rabbits with GST-fusion proteins: Glut4 (final 30 amino acids) or synapsin II C-domain (amino acids 181-397) (Sudhof et al., 1989). Synapsin II antibodies were affinity purified using a Histagged synapsin II C-domain fusion protein. Other antibodies used: monoclonal anti-Flag (M2, SIGMA); monoclonal anti-HA (HA.11, Covance); monoclonal antiactin (Chemicon); monoclonal anti-Human transferrin receptor (HTR.D65, from I. Trowbridge, Salk Institute La Jolla, CA); PE-conjugated anti-transferrin receptor antibody (anti-CD71, eBiosciences); polyclonal phospho-synapsin (Ser9) antibody (Cell Signaling Technology). Immunoblotting was performed as described (Corley Mastick et al., 2001).

Adenovirus

cDNA encoding HA-Glut4 (from Samuel Cushman, NIH, Bethesda, MD) (Dawson et al., 2001) was amplified by PCR and inserted at the AgeI site in-frame upstream of the GFP reporter in the pAdTrack-CMV shuttle vector (He et al., 1998). This vector is engineered to express both a GFP-tagged reporter construct and a gene of interest from a single virus using tandem CMV promoters. cDNA encoding rat Synapsin IIb (from Andrew Czernik, The Rockefeller University, New York, NY) was subcloned into the multiple cloning site in the second CMV promotor in this vector. PCR was used to amplify the C-terminal domain of synapsin, adding the Flag-epitope (NcoI site to stop, ∼760 bp). Serine-to-alanine and serine-to-asparticacid mutations at amino acid 10 (S10A or S10D) in synapsin IIb were also generated by PCR (start site to KpnI site, ∼120 bp). The modifications were then subcloned into the wild type synapsin by restriction digestion and recombination. All constructs were fully sequenced to ensure that only the desired mutations were present in the expressed proteins. Recombinant adenovirus was prepared essentially as described (He et al., 1998). Adenovirus was purified from packaging cell lysates using Optiprep gradients following the manufacturer's recommended protocol for retroviral purification (Axis-Shield UK, Dundee, UK). Adipocytes and fibroblasts were infected with adenovirus as previously described with modifications as noted (Greenberg et al., 2003; Sanguinetti et al., 2003). Adipocytes were infected on day 7 of differentiation (1 day after removal of insulin). Cells were re-fed in DMEM with 10% FBS 4 hours after infection, and used for experiments 2-3 days post infection. To avoid an adverse response of the cells to viral infection, and to facilitate microscopic analysis, less than 30% of the cells on the plate were infected. All viruses expressed similar amounts of HA-Glut4-GFP (supplementary material Fig. S2A,B). The total level of Glut4 in the infected adipocytes was not higher than the level in uninfected adipocytes, when measured on a cell-by-cell basis using flow cytometry (supplementary material Fig. S2C; the Glut4 antibody used recognizes both endogenous Glut4 and HA-Glut4-GFP, supplementary material Fig. S2D). This result was verified by western blotting and by 2-deoxyglucose uptake (data not shown; cells were infected at higher MOI in these experiments so that ∼80% of the cells were infected). At this low level of expression the basal retention mechanism was not saturated (supplementary material Table S1). Equal amounts of the Flagsynapsins were expressed (supplementary material Fig. S2E) and every cell that expresses Glut4-GFP also expresses Flag-synapsin (supplementary material Fig. S2F) when cells were infected with viruses co-expressing HA-Glut4-GFP and either WT or S10A synapsin.

Microscopy

3T3-L1 cells were differentiated on glass coverslips, then infected with virus. Cells were fixed (10 minutes, 25°C) in PBS, 4% formaldehyde (methanol-free, EM grade; Polysciences). To detect Flag-synapsin, fixed cells were permeabilized in PBS, 0.3% Triton X-100 (10 minutes), blocked in PBS, 0.1% Tween, 3% BSA (30 minutes, 25°C) and incubated with anti-Flag (5 μg/ml). To measure Glut4 distribution in fibroblasts, infected cells were incubated at 37°C in the continuous presence of anti-HA antibody (5 μg/ml) for 5 minutes (equal to surface labeling measured in fixed non-permeabilized cells, data not shown) and 2 hours (to measure the total recycling pool) (Govers et al., 2004). The cells were rapidly washed, then fixed and permeabilized. Anti-HA was detected with Cy3-labeled secondary antibody, and the ratio of red fluorescence (anti-HA) to green fluorescence (GFP) was determined. Both red and green fluorescence was very low in non-infected cells, although there was a small amount of anti-HA antibody taken up non-specifically by these cells (less than 10% of that in infected cells; data not shown). These values were subtracted from the values for infected cells. Anti-HA and anti-Flag were detected with Cy3-conjugated donkey anti-mouse IgG (H+L; 0.75 μg/ml; Jackson ImmunoResearch Laboratories). In all experiments, coverslips were mounted on glass slides with Prolong Gold (Molecular Probes) and imaged by either wide-field epifluorescence or laser-scanning confocal microscopy.

Image acquisition and analysis

Wide-field images were captured using a Leica DMR fluorescence microscope (lenses: 20×/0.40 and 50×/0.75, water immersion) with an Optronics cooled CCD camera (LEI-750) operated by Image Acquire software. Images were analyzed using Adobe Photoshop CS v8.0. Confocal images were captured using a Nikon Eclipse E800 microscope (lens: 60×/0.85) equipped with a Nikon C1 confocal system operated by Nikon EZ-C1 v1.7 software. Maximum projections (∼10 0.5 μ z-slices per cell) were generated and analyzed with SimplePCI v5.3.1 (Compix). Measurement of Glut4-GFP clustering was performed using pixel intensity threshold protocols. In this analysis, all pixels are assigned a binary value, either positive for labeling (above the threshold) or negative for labeling (below the threshold). The threshold was set for intensity greater than 60% of maximum intensity (channel 150 of 255). This threshold was set to gate out diffuse cell surface staining, leaving only the brighter punctate objects and perinuclear clusters. An object is defined as a group of adjacent positive pixels. The size of the object is measured as the area of contiguous positive pixels (Fig. 4B, area of the base of the peaks). Multiple objects were identified in each cell and the mean size of all objects in montages of ∼60 cells calculated. This analysis gives information about how the fluorescence is distributed in the cell, but not the relative amounts of Glut4 in each cluster. All cells have approximately the same total amount of GFP fluorescence (supplementary material Fig. S2B).

Flow cytometry

Infected 3T3-L1 adipocytes were incubated with or without insulin (100 nM) for 30 minutes. Cells were labeled at 4°C with anti-HA (5 μg/ml) in PBS + 1% BSA, followed by biotinylated donkey anti-mouse IgG (1.3 μg/ml; Jackson Immunoresearch Laboratories), then PC7-streptavidin (0.05 μg/ml; eBioscience). Surface transferrin receptor (TfR) content was assessed by labeling cells with PE-conjugated anti-TfR antibody (CD71; 0.25 μg/ml; eBioscience). Compensation was applied to the 595 nm detector to account for GFP spill over into the 595 nm signal. After labeling, cells were detached using collagenase (100 μl, 1 μg/μl in PBS, 20 minutes, 4°C) and suspended in 500 μl ice-cold PBS, 1% BSA by gentle pipetting. Adherent cells are labeled for two reasons: ease of washing the cells (fat cells float, so washing cannot be done by simply pelleting and resuspending the cells) and to minimize damage to the very fragile adipocytes. There is no difference in signal between cells detached from the plate with collagenase, and cells detached from the plate by gentle pipetting, so collagenase does not affect the antibody labeling (supplementary material Fig. S2G). Collagenase treatment before pipetting significantly improves cell recovery.

Samples were analyzed by flow cytometry using a Coulter Epics XL equipped with an Argon laser (488 nm peak). Forward scatter, side scatter, 525 nm, 595 nm, 670 nm and 740 nm fluorescence signals were collected. The data were processed using FlowJo software (Tree Star Inc). Differentiated adipocytes were identified based on forward scatter and 670 nm auto-fluorescence. Infected cells were selected based on a gate in 525 nm fluorescence (GFP). Non-infected cells (NV) in the same sample were used to determine background auto-fluorescence and non-specific binding (supplementary material Fig. S2B,G). These values were subtracted from the calculated mean fluorescence intensities (MFI) of total Glut4 (GFP; 525 nm) and anti-HA labeled Glut4 (PC7; 740 nm) before calculating the mean fluorescence ratios (MFR; MFI anti-HA/MFI GFP) for each sample. The signal-to-noise ratio was greater than 10 for GFP and at least two- to threefold for PC7 in basal cells (insulin values were ∼sevenfold higher than basal values).

The authors would like to thank Parul Matani and Sarah Carpenter for mRNA analysis and protein isolation for synapsin antibody purification. Research supported by American Diabetes Association grant 7-03-IN-04 and by National Institutes of Health grant DK 56197.

Bloom, O., Evergren, E., Tomilin, N., Kjaerulff, O., Low, P., Brodin, L., Pieribone, V. A., Greengard, P. and Shupliakov, O. (
2003
). Colocalization of synapsin and actin during synaptic vesicle recycling.
J. Cell Biol.
161
,
737
-747.
Bonanomi, D., Menegon, A., Miccio, A., Ferrari, G., Corradi, A., Kao, H. T., Benfenati, F. and Valtorta, F. (
2005
). Phosphorylation of synapsin I by cAMPdependent protein kinase controls synaptic vesicle dynamics in developing neurons.
J. Neurosci.
25
,
7299
-7308.
Brautigam, C. A., Chelliah, Y. and Deisenhofer, J. (
2004
). Tetramerization and ATP binding by a protein comprising the A, B, and C domains of rat synapsin I.
J. Biol. Chem.
279
,
11948
-11956.
Bryant, N. J., Govers, R. and James, D. E. (
2002
). Regulated transport of the glucose transporter GLUT4.
Nat. Rev. Mol. Cell Biol.
3
,
267
-277.
Bustos, R., Kolen, E. R., Braiterman, L., Baines, A. J., Gorelick, F. S. and Hubbard, A. L. (
2001
). Synapsin I is expressed in epithelial cells: localization to a unique trans-Golgi compartment.
J. Cell Sci.
114
,
3695
-3704.
Cain, C. C., Trimble, W. S. and Lienhard, G. E. (
1992
). Members of the VAMP family of synaptic vesicle proteins are components of glucose transporter-containing vesicles from rat adipocytes.
J. Biol. Chem.
267
,
11681
-11684.
Carvalho, E., Schellhorn, S. E., Zabolotny, J. M., Martin, S., Tozzo, E., Peroni, O. D., Houseknecht, K. L., Mundt, A., James, D. E. and Kahn, B. B. (
2004
). GLUT4 overexpression or deficiency in adipocytes of transgenic mice alters the composition of GLUT4 vesicles and the subcellular localization of GLUT4 and insulin-responsive aminopeptidase.
J. Biol. Chem.
279
,
21598
-21605.
Chi, P., Greengard, P. and Ryan, T. A. (
2001
). Synapsin dispersion and reclustering during synaptic activity.
Nat. Neurosci
4
,
1187
-1193.
Chi, P., Greengard, P. and Ryan, T. A. (
2003
). Synaptic vesicle mobilization is regulated by distinct synapsin I phosphorylation pathways at different frequencies.
Neuron
38
,
69
-78.
Corley Mastick, C., Sanguinetti, A. R., Knesek, J. H., Mastick, G. S. and Newcomb, L. F. (
2001
). Caveolin-1 and a 29-kDa caveolin-associated protein are phosphorylated on tyrosine in cells expressing a temperature-sensitive v-Abl kinase.
Exp. Cell Res.
266
,
142
-154.
Dawson, K., Aviles-Hernandez, A., Cushman, S. W. and Malide, D. (
2001
). Insulinregulated trafficking of dual-labeled glucose transporter 4 in primary rat adipose cells.
Biochem. Biophys. Res. Commun.
287
,
445
-454.
De Camilli, P., Harris, S. M., Jr, Huttner, W. B. and Greengard, P. (
1983
). Synapsin I (Protein I), a nerve terminal-specific phosphoprotein. II. Its specific association with synaptic vesicles demonstrated by immunocytochemistry in agarose-embedded synaptosomes.
J. Cell Biol.
96
,
1355
-1373.
Eguez, L., Lee, A., Chavez, J. A., Miinea, C. P., Kane, S., Lienhard, G. E. and McGraw, T. E. (
2005
). Full intracellular retention of GLUT4 requires AS160 Rab GTPase activating protein.
Cell Metab.
2
,
263
-272.
El-Jack, A. K., Kandror, K. V. and Pilch, P. F. (
1999
). The formation of an insulinresponsive vesicular cargo compartment is an early event in 3T3-L1 adipocyte differentiation.
Mol. Biol. Cell
10
,
1581
-1594.
Eyster, C. A., Duggins, Q. S. and Olson, A. L. (
2005
). Expression of constitutively active Akt/protein kinase B signals GLUT4 translocation in the absence of an intact actin cytoskeleton.
J. Biol. Chem.
280
,
17978
-17985.
Foster, L. J. and Klip, A. (
2000
). Mechanism and regulation of GLUT-4 vesicle fusion in muscle and fat cells.
Am. J. Physiol. Cell Physiol.
279
,
C877
-C890.
Frevert, E. U., Bjorbaek, C., Venable, C. L., Keller, S. R. and Kahn, B. B. (
1998
). Targeting of constitutively active phosphoinositide 3-kinase to GLUT4-containing vesicles in 3T3-L1 adipocytes.
J. Biol. Chem.
273
,
25480
-25487.
Frost, S. C. and Lane, M. D. (
1985
). Evidence for the involvement of vicinal sulfhydryl groups in insulin-activated hexose transport by 3T3-L1 adipocytes.
J. Biol. Chem.
260
,
2646
-2652.
Gitler, D., Takagishi, Y., Feng, J., Ren, Y., Rodriguiz, R. M., Wetsel, W. C., Greengard, P. and Augustine, G. J. (
2004a
). Different presynaptic roles of synapsins at excitatory and inhibitory synapses.
J. Neurosci.
24
,
11368
-11380.
Gitler, D., Xu, Y., Kao, H. T., Lin, D., Lim, S., Feng, J., Greengard, P. and Augustine, G. J. (
2004b
). Molecular determinants of synapsin targeting to presynaptic terminals.
J. Neurosci.
24
,
3711
-3720.
Govers, R., Coster, A. C. and James, D. E. (
2004
). Insulin increases cell surface GLUT4 levels by dose dependently discharging GLUT4 into a cell surface recycling pathway.
Mol. Cell. Biol.
24
,
6456
-6466.
Greenberg, C. C., Meredith, K. N., Yan, L. and Brady, M. J. (
2003
). Protein targeting to glycogen overexpression results in the specific enhancement of glycogen storage in 3T3-L1 adipocytes.
J. Biol. Chem.
278
,
30835
-30842.
He, T. C., Zhou, S., da Costa, L. T., Yu, J., Kinzler, K. W. and Vogelstein, B. (
1998
). A simplified system for generating recombinant adenoviruses.
Proc. Natl. Acad. Sci. USA
95
,
2509
-2514.
Hilfiker, S., Schweizer, F. E., Kao, H. T., Czernik, A. J., Greengard, P. and Augustine, G. J. (
1998
). Two sites of action for synapsin domain E in regulating neurotransmitter release.
Nat. Neurosci.
1
,
29
-35.
Hosaka, M. and Sudhof, T. C. (
1999
). Homo- and heterodimerization of synapsins.
J. Biol. Chem.
274
,
16747
-16753.
Hosaka, M., Hammer, R. E. and Sudhof, T. C. (
1999
). A phospho-switch controls the dynamic association of synapsins with synaptic vesicles.
Neuron
24
,
377
-387.
Huttner, W. B., Schiebler, W., Greengard, P. and De Camilli, P. (
1983
). Synapsin I (protein I), a nerve terminal-specific phosphoprotein. III. Its association with synaptic vesicles studied in a highly purified synaptic vesicle preparation.
J. Cell Biol.
96
,
1374
-1388.
Kao, H. T., Song, H. J., Porton, B., Ming, G. L., Hoh, J., Abraham, M., Czernik, A. J., Pieribone, V. A., Poo, M. M. and Greengard, P. (
2002
). A protein kinase Adependent molecular switch in synapsins regulates neurite outgrowth.
Nat. Neurosci.
5
,
431
-437.
Krueger, K. A., Ings, E. I., Brun, A. M., Landt, M. and Easom, R. A. (
1999
). Sitespecific phosphorylation of synapsin I by Ca2+/calmodulin-dependent protein kinase II in pancreatic betaTC3 cells: synapsin I is not associated with insulin secretory granules.
Diabetes
48
,
499
-506.
Lampson, M. A., Schmoranzer, J., Zeigerer, A., Simon, S. M. and McGraw, T. E. (
2001
). Insulin-regulated release from the endosomal recycling compartment is regulated by budding of specialized vesicles.
Mol. Biol. Cell
12
,
3489
-3501.
Larance, M., Ramm, G., Stockli, J., van Dam, E. M., Winata, S., Wasinger, V., Simpson, F., Graham, M., Junutula, J. R., Guilhaus, M. et al. (
2005
). Characterization of the role of the Rab GTPase-activating protein AS160 in insulinregulated GLUT4 trafficking.
J. Biol. Chem.
280
,
37803
-37813.
Laurie, S. M., Cain, C. C., Lienhard, G. E. and Castle, J. D. (
1993
). The glucose transporter GluT4 and secretory carrier membrane proteins (SCAMPs) colocalize in rat adipocytes and partially segregate during insulin stimulation.
J. Biol. Chem.
268
,
19110
-19117.
Li, L. V. and Kandror, K. V. (
2005
). Golgi-localized, gamma-ear-containing, Arf-binding protein adaptors mediate insulin-responsive trafficking of glucose transporter 4 in 3T3-L1 adipocytes.
Mol. Endocrinol.
19
,
2145
-2153.
Lin, B. Z., Pilch, P. F. and Kandror, K. V. (
1997
). Sortilin is a major protein component of Glut4-containing vesicles.
J. Biol. Chem.
272
,
24145
-24147.
Linstedt, A. D. and Kelly, R. B. (
1991
). Synaptophysin is sorted from endocytotic markers in neuroendocrine PC12 cells but not transfected fibroblasts.
Neuron
7
,
309
-317.
Lonart, G. and Simsek-Duran, F. (
2006
). Deletion of synapsins I and II genes alters the size of vesicular pools and rabphilin phosphorylation.
Brain Res.
1107
,
42
-51.
Longuet, C., Broca, C., Costes, S., Hani, E. H., Bataille, D. and Dalle, S. (
2005
). Extracellularly regulated kinases 1/2 (p44/42 mitogen-activated protein kinases) phosphorylate synapsin I and regulate insulin secretion in the MIN6 beta-cell line and islets of Langerhans.
Endocrinology
146
,
643
-654.
Mastick, C. C. and Falick, A. L. (
1997
). Association of N-ethylmaleimide sensitive fusion (NSF) protein and soluble NSF attachment proteins-alpha and -gamma with glucose transporter-4-containing vesicles in primary rat adipocytes.
Endocrinology
138
,
2391
-2397.
Mastick, C. C., Aebersold, R. and Lienhard, G. E. (
1994
). Characterization of a major protein in GLUT4 vesicles. Concentration in the vesicles and insulin-stimulated translocation to the plasma membrane.
J. Biol. Chem.
269
,
6089
-6092.
Matsumoto, K., Ebihara, K., Yamamoto, H., Tabuchi, H., Fukunaga, K., Yasunami, M., Ohkubo, H., Shichiri, M. and Miyamoto, E. (
1999
). Cloning from insulinoma cells of synapsin I associated with insulin secretory granules.
J. Biol. Chem.
274
,
2053
-2059.
Matteoli, M., Takei, K., Perin, M. S., Sudhof, T. C. and De Camilli, P. (
1992
). Exoendocytotic recycling of synaptic vesicles in developing processes of cultured hippocampal neurons.
J. Cell Biol.
117
,
849
-861.
Menegon, A., Dunlap, D. D., Castano, F., Benfenati, F., Czernik, A. J., Greengard, P. and Valtorta, F. (
2000
). Use of phosphosynapsin I-specific antibodies for image analysis of signal transduction in single nerve terminals.
J. Cell Sci.
113
,
3573
-3582.
Morris, N. J., Ross, S. A., Lane, W. S., Moestrup, S. K., Petersen, C. M., Keller, S. R. and Lienhard, G. E. (
1998
). Sortilin is the major 110-kDa protein in GLUT4 vesicles from adipocytes.
J. Biol. Chem.
273
,
3582
-3587.
Pieribone, V. A., Shupliakov, O., Brodin, L., Hilfiker-Rothenfluh, S., Czernik, A. J. and Greengard, P. (
1995
). Distinct pools of synaptic vesicles in neurotransmitter release.
Nature
375
,
493
-497.
Porton, B., Kao, H. T. and Greengard, P. (
1999
). Characterization of transcripts from the synapsin III gene locus.
J. Neurochem.
73
,
2266
-2271.
Proctor, K. M., Miller, S. C., Bryant, N. J. and Gould, G. W. (
2006
). Syntaxin 16 controls the intracellular sequestration of GLUT4 in 3T3-L1 adipocytes.
Biochem. Biophys. Res. Commun.
347
,
433
-438.
Provoda, C. J., Waring, M. T. and Buckley, K. M. (
2000
). Evidence for a primary endocytic vesicle involved in synaptic vesicle biogenesis.
J. Biol. Chem.
275
,
7004
-7012.
Rosahl, T. W., Spillane, D., Missler, M., Herz, J., Selig, D. K., Wolff, J. R., Hammer, R. E., Malenka, R. C. and Sudhof, T. C. (
1995
). Essential functions of synapsins I and II in synaptic vesicle regulation.
Nature
375
,
488
-493.
Ryan, T. A., Li, L., Chin, L. S., Greengard, P. and Smith, S. J. (
1996
). Synaptic vesicle recycling in synapsin I knock-out mice.
J. Cell Biol.
134
,
1219
-1227.
Sanguinetti, A. R., Cao, H. and Corley Mastick, C. (
2003
). Fyn is required for oxidative- and hyperosmotic-stress-induced tyrosine phosphorylation of caveolin-1.
Biochem. J.
376
,
159
-168.
Shewan, A. M., van Dam, E. M., Martin, S., Luen, T. B., Hong, W., Bryant, N. J. and James, D. E. (
2003
). GLUT4 recycles via a trans-Golgi network (TGN) subdomain enriched in Syntaxins 6 and 16 but not TGN38: involvement of an acidic targeting motif.
Mol. Biol. Cell
14
,
973
-986.
Shi, J. and Kandror, K. V. (
2005
). Sortilin is essential and sufficient for the formation of Glut4 storage vesicles in 3T3-L1 adipocytes.
Dev. Cell
9
,
99
-108.
Simpson, I. A., Yver, D. R., Hissin, P. J., Wardzala, L. J., Karnieli, E., Salans, L. B. and Cushman, S. W. (
1983
). Insulin-stimulated translocation of glucose transporters in the isolated rat adipose cells: characterization of subcellular fractions.
Biochim. Biophys. Acta
763
,
393
-407.
Sudhof, T. C. (
2004
). The synaptic vesicle cycle.
Annu. Rev. Neurosci.
27
,
509
-547.
Sudhof, T. C., Czernik, A. J., Kao, H. T., Takei, K., Johnston, P. A., Horiuchi, A., Kanazir, S. D., Wagner, M. A., Perin, M. S., De Camilli, P. et al. (
1989
). Synapsins: mosaics of shared and individual domains in a family of synaptic vesicle phosphoproteins.
Science
245
,
1474
-1480.
Sun, J., Bronk, P., Liu, X., Han, W. and Sudhof, T. C. (
2006
). Synapsins regulate usedependent synaptic plasticity in the calyx of Held by a Ca2+/calmodulin-dependent pathway.
Proc. Natl. Acad. Sci. USA
103
,
2880
-2885.
Takei, Y., Harada, A., Takeda, S., Kobayashi, K., Terada, S., Noda, T., Takahashi, T. and Hirokawa, N. (
1995
). Synapsin I deficiency results in the structural change in the presynaptic terminals in the murine nervous system.
J. Cell Biol.
131
,
1789
-1800.
Tanner, L. I. and Lienhard, G. E. (
1987
). Insulin elicits a redistribution of transferrin receptors in 3T3-L1 adipocytes through an increase in the rate constant for receptor externalization.
J. Biol. Chem.
262
,
8975
-8980.
Waters, S. B., D'Auria, M., Martin, S. S., Nguyen, C., Kozma, L. M. and Luskey, K. L. (
1997
). The amino terminus of insulin-responsive aminopeptidase causes Glut4 translocation in 3T3-L1 adipocytes.
J. Biol. Chem.
272
,
23323
-23327.
Watson, R. T., Khan, A. H., Furukawa, M., Hou, J. C., Li, L., Kanzaki, M., Okada, S., Kandror, K. V. and Pessin, J. E. (
2004
). Entry of newly synthesized GLUT4 into the insulin-responsive storage compartment is GGA dependent.
EMBO J.
23
,
2059
-2070.
Williams, D., Hicks, S. W., Machamer, C. E. and Pessin, J. E. (
2006
). Golgin-160 is required for the Golgi membrane sorting of the insulin-responsive glucose transporter GLUT4 in adipocytes.
Mol. Biol. Cell
17
,
5346
-5355.
Zeigerer, A., Lampson, M. A., Karylowski, O., Sabatini, D. D., Adesnik, M., Ren, M. and McGraw, T. E. (
2002
). GLUT4 retention in adipocytes requires two intracellular insulin-regulated transport steps.
Mol. Biol. Cell
13
,
2421
-2435.
Zeigerer, A., McBrayer, M. K. and McGraw, T. E. (
2004
). Insulin stimulation of GLUT4 exocytosis, but not its inhibition of endocytosis, is dependent on RabGAP AS160.
Mol. Biol. Cell
15
,
4406
-4415.

Supplementary information