Subcellular membrane fractions from basal and insulin-treated 3T3-L1 adipocytes were prepared using a previously described differential centrifugation procedure (Piper et al., 1991 ; Marsh et al., 1995 ). Briefly, the plasma membrane fraction was obtained after a 20-min centrifugation at 17,200 × g followed by centrifugation through sucrose. The high-density microsomes (HDMs) were obtained by centrifuging the 17,200 × g supernatant at 38,700 × g for 20 min and the low-density microsomes (LDMs) were obtained by spinning the 38,700 × g supernatant at 150,000 × g for 75 min. These fractions have previously been characterized in detail (Piper et al., 1991 ). The HDM fraction contains large endosomal components and endoplasmic reticulum (ER), whereas the LDM fraction contains small vesicles, including those enriched in GLUT4. All fractions were resuspended in HES buffer (20 mM HEPES, 1 mM EDTA, 250 mM sucrose, pH 7.4), protein quantified using the bicinchoninic acid assay (Pierce Chemical, Rockford, IL) and stored at −80°C before use. Total membrane fractions were prepared from 3T3-L1 fibroblasts and adipocytes after homogenization in HES buffer containing protease inhibitors (10 μg/ml aprotinin, 10 μg/ml leupeptin, 250 μM phenylmethylsulfonyl fluoride). Homogenates were subjected to centrifugation at 50,000 rpm in a Beckman Coulter TLA100–3 rotor for 60 min. The membrane pellet was resuspended in HES buffer and stored at −80°C before use.

Resialylation studies were performed essentially as described by Teuchert et al. (1999) . Cells were incubated in serum-free medium overnight and insulin (20 nM) was added for 30 min at 37°C. Cells were then washed five times with ice-cold phosphate-buffered saline (PBS) containing 0.1 mM CaCl₂, 1 mM MgCl₂ (PBS⁺⁺) and biotinylated twice for 20 min in 2 ml of PBS⁺⁺ containing 0.5 mg/ml sulfo-NHS-biotin (Pierce Chemical). Cells were washed three times with ice-cold PBS⁺⁺ containing 0.1 M glycine to quench free biotin, incubated with neuraminidase Vibrio cholerae (80 mU/ml; Roche Diagnostics, Indianapolis, IN) on ice for 1 h and then washed three times with PBS⁺⁺. Cells were then incubated with prewarmed DMEM containing fetal calf serum (10%) for different times as indicated at 37°C. Cells were washed twice with PBS⁺⁺ at 4°C and incubated with Triton X-100 (1%) containing protease inhibitors (see above) for 20 min at 4°C. Cells were scraped from the dish and centrifuged at 14,000 rpm for 10 min at 4°C. The cell lysate was incubated with streptavidin agarose beads at 4°C for 16 h and washed three times in PBS⁺⁺ containing 1% Triton X-100/0.1% SDS. Samples were then heated to 60°C for 10 min and subjected to SDS-PAGE and immunoblotted with anti-IRAP.

Basal 3T3-L1 adipocytes were homogenized by passaging twice through a 25-gauge needle followed by passaging twice through a 27-gauge needle in 50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 10 mM EDTA containing protease inhibitors. Cell lysates were solubilized using 1% Triton X-100 on ice for 30 min. The solubilized lysate was cleared by centrifugation for 30 min at 4°C in a microcentrifuge. Aliquots of the soluble proteins were incubated overnight with relevant antibodies bound to protein A-Sepharose. Immunoprecipitated proteins were resolved by SDS-PAGE together with aliquots of the supernatant and starting material.

Protein G and protein A beads (Pierce Chemical) were incubated with 1% bovine serum albumin (BSA) for 30 min. Beads were then incubated with either monoclonal GLUT4 antibody 1F8, nonspecific mouse IgG, anti-Syntaxin 6, or anti-Syntaxin 16 antibodies. Antibodies were cross-linked to the beads by using 20 mM dimethyl suberimidate (Pierce Chemical) for 30 min at room temperature and cross-linked antibodies were saturated with 1% BSA for 30 min at room temperature. LDMs from noninfected, HA-GLUT4–infected or HA-TAIL–infected 3T3-L1 adipocytes were incubated separately with each of the specific and nonspecific antibody-coupled beads overnight at 4°C. The beads were washed and adsorbed material was eluted with SDS sample buffer and subjected to SDS-PAGE together with aliquots of the starting material.

Proteins were subjected to electrophoresis on 7.5 or 12% SDS-polyacrylamide gels and transblotted onto polyvinylidene difluoride. Immunolabeled proteins were visualized using horseradish peroxidase-conjugated secondary antibody and either the enhanced chemiluminescence system (Amersham Biosciences, Aylesbury, United Kingdom) or Supersignal (Pierce Chemical). Bands were quantitated by densitometry or by using a Lumi-Imager (Roche Diagnostics, Castle Hill, New South Wales, Australia).

The preparation of plasma membrane (PM) lawns was performed as described in Robinson et al. (1992) . Briefly, after incubating cells on coverslips with the appropriate treatment, adipocytes were sonicated yielding a lawn of PM fragments attached to the coverslip. Coverslips were then incubated with the relevant antibodies directed against C-terminal domains, followed by fluorescein isothiocyanate-conjugated secondary antibody (Molecular Probes, Eugene, OR). Cells were viewed using either a 63×/1.4 oil immersion objective on an Axiovert fluorescence microscope (Carl Zeiss, Thornwood, NY), equipped with an MRC-600 laser confocal imaging system (Bio-Rad, Hercules, CA), or a 100×/1.4 Plan Apo oil immersion objective on an Eclipse E600 fluorescence microscope (Nikon, Tokyo, Japan), equipped with a Radiance 2000 laser confocal imaging system (Bio-Rad).

Syntaxin 6 is highly concentrated in the TGN area in PC12 cells (Bock et al., 1997 ) and together with the t-SNARE Syntaxin 16 (Mallard et al., 2002 ) plays an important role in a vesicle transport pathway from endosomes to the TGN. The overlap between GLUT4 and Syntaxins 6 and 16 in adipocytes raised the possibility that the perinuclear GLUT4 compartment may constitute either the TGN or vesicles associated with the TGN. To further explore this possibility we examined the kinetics of IRAP resialylation as an index of the trafficking of GLUT4 via the TGN because this protein has been shown to follow similar trafficking kinetics to GLUT4 (Ross et al., 1996 ). Cells were treated with insulin to introduce a cohort of IRAP molecules into the surface membrane. Cells were then biotinylated on ice and treated with neuraminidase to remove sialic acid before reincubation for various times at 37°C. The biotinylated molecules were recovered and immunoblotted. IRAP underwent a slight but significant increase in electrophoretic mobility after neuraminidase treatment consistent with a loss of carbohydrate (Figure ​(Figure5).5). After 60 min at 37°C, there was a significant reduction in the mobility of IRAP back toward the level observed in cells not incubated with neuraminidase. These data suggest that the cohort of IRAP that was desialylated at the cell surface was resialylated indicative of retrieval back to the TGN. We did not observe significant IRAP resialylation at shorter times (0–30 min), indicating that the kinetics of this process is longer than trafficking of GLUT4 from the cell surface to the Syntaxin 16-positive compartment.

Syntaxin 6 and Syntaxin 16 have recently been shown to form a SNARE complex in HeLa cells and synaptosomes (Kreykenbohm et al., 2002 ; Mallard et al., 2002 ). This complex seems to play an important role in trafficking between endosomes and the TGN. As shown in Figure ​Figure7,7, Syntaxin 6 and Syntaxin 16 also form a stable complex in 3T3-L1 adipocytes. This did not seem to be a nonspecific association, because we could not detect any Syntaxin 4 (Figure ​(Figure7A)7A) or SNAP-23 (our unpublished data) in the Syntaxin 6-containing complexes. One possibility is that Syntaxins 6 and 16 play an integral role in the intracellular sequestration of GLUT4 in adipocytes. Indeed, Syntaxin 6 and Syntaxin 16 may be involved in the biogenesis of insulin-responsive GLUT4-containing vesicles. The characteristic ability of 3T3-L1 cells to form an insulin responsive GLUT4 compartment is markedly up-regulated soon after adipocyte differentiation (El-Jack et al., 1999 ), suggesting that the machinery required for the biogenesis of this compartment might be specifically up-regulated in these cells. Strikingly, Syntaxin 6 and Syntaxin 16 levels were increased by 2.7 ± 0.6- and 4.6 ± 2.1-fold (n = 4 ± SEM), respectively, upon differentiation of fibroblasts into adipocytes (Figure ​(Figure7B).7B). We also noted a slight increase in the levels of Syntaxin 13 after adipocyte differentiation (2.1 ± 1.1, n = 3). In contrast, the levels of a number of other t-SNAREs was either unchanged (Syntaxin 4, 0.9 ± 0.1; Syntaxin 5, 0.8 ± 0.2) or decreased (Syntaxin 7, 0.6 ± 0.2).

Although the majority of GLUT4 is found in small tubulovesicular elements in muscle and adipocytes, a finite pool (10–15%) is also located in the TGN (Bryant et al., 2002 ). This definition is based upon immunoelectron microscopy localization studies, which find a significant pool of GLUT4 in a tubulovesicular compartment adjacent to the Golgi. In the present study, we provide evidence that GLUT4 recycles via a perinuclear compartment that has a number of characteristics consistent with it being a subdomain of the TGN. Although this compartment is morphologically distinct from the highly tubular TGN38 compartment, it is often immediately adjacent to the latter compartment (Figure ​(Figure3).3). This is in agreement with previous studies showing that there is little colocalization between TGN38 and GLUT4 in adipocytes (Martin et al., 1994 ). In addition, both Syntaxins 6 and 16 are thought to function in the TGN acting as t-SNAREs for transport vesicles arriving from the endosomal system (Mallard et al., 1998 ). Thus, the significant overlap between GLUT4 and these t-SNAREs in the perinuclear region is consistent with this representing the TGN. Although the TGN was originally defined as the sorting and exit site of the Golgi, its structure of cisternae and tubulovesicular elements has been poorly defined. Three-dimensional electron microscopic analysis of the Golgi has recently led to the formulation of a model in which molecules are sorted in the secretory pathway in multiple cisternae of the TGN, each of which served as an exit site (Ladinsky et al., 2002 ). This suggests that trans-cisternae and tubulovesicular elements may be separate but interconnected. Likewise, entry sites into different subdomains of the TGN are feasible and may explain the discrepancy we find in kinetics between entry of HA-GLUT4 in the Syntaxin 16-positive compartment and resialylation of IRAP. The majority of TGN38 and sialyl transferase is located in the trans-cisternae of the Golgi in other cell types (Bennett and O'Shaughnessy, 1981 ; Roth et al., 1985 ; Ladinsky and Howell, 1992 ; Ladinsky et al., 2002 ), whereas a significant pool of GLUT4 in the TGN area colocalizes with other TGN recycling proteins, such as the cation-dependent mannose 6-phosphate receptor (Martin et al., 2000a ). It may thus be that the Syntaxin 6/16-positive compartment is a specialized compartment that has arisen from the TGN. Consistent with such a model are our previous studies in atrial cardiomyocytes (Slot et al., 1997 ). This is an unusual cell that possesses both a regulated secretory pathway and an insulin-responsive glucose transport system. A considerable proportion (~60%) of GLUT4 is localized to secretory granules in these cells (Slot et al., 1997 ). The GLUT4 trafficking and the regulated secretory pathways seem to merge in the TGN, again suggesting that the flux of GLUT4 through this pathway is considerable. Intriguingly, Syntaxin 6 is also found in regulated secretory granules where it plays a role in granule maturation (Wendler et al., 2001 ). Conversely, it has been reported that during granule maturation Syntaxin 6 is removed from immature granules by AP-1/clathrin-coated vesicles and then delivered to endosomes (Klumperman et al., 1998 ). Interestingly, GLUT4 is also found in clathrin-coated AP-1–positive vesicles within the TGN area (Gillingham et al., 1999 ; Martin et al., 2000b ), suggesting that GLUT4 is rapidly cycling between the TGN and endosomes in basal adipocytes, which may call into question the idea of a stable intracellular storage compartment. Based on this evidence, we speculate that GLUT4 may be retained in a compartment that is generated from the TGN and that is possibly analogous to immature secretory granules in secretory cells. The t-SNARES Syntaxin 6 and 16 may play an integral role in the maturation of this compartment in adipocytes and we are currently in the process of testing this hypothesis.