The Exocytotic Trafficking of TC10 Occurs through both Classical and Nonclassical Secretory Transport Pathways in 3T3L1 Adipocytes

Abstract

To examine the structural determinants necessary for TC10 trafficking, localization, and function in adipocytes, we generated a series of point mutations in the carboxyl-terminal targeting domain of TC10. Wild-type TC10 (TC10/WT) localized to secretory membrane compartments and caveolin-positive lipid raft microdomains at the plasma membrane. Expression of a TC10/C206S point mutant resulted in a trafficking and localization pattern that was indistinguishable from that of TC10/WT. In contrast, although TC10/C209S or the double TC10/C206,209S mutant was plasma membrane localized, it was excluded from both the secretory membrane system and the lipid raft compartments. Surprisingly, inhibition of Golgi membrane transport with brefeldin A did not prevent plasma membrane localization of TC10 or H-Ras. Moreover, inhibition of trans-Golgi network exit with a 19°C temperature block did not prevent the trafficking of TC10 or H-Ras to the plasma membrane. These data demonstrate that TC10 and H-Ras can both traffic to the plasma membrane by at least two distinct transport mechanisms in adipocytes, one dependent upon intracellular membrane transport and another independent of the classical secretory membrane system. Moreover, the transport through the secretory pathway is necessary for the localization of TC10 to lipid raft microdomains at the plasma membrane.

The carboxyl-terminal CAAX sequences of small GTP-binding proteins provide important sites for posttranslational modification necessary for the appropriate trafficking, membrane localization, and signaling specificity of this class of proteins (39). In particular, recent studies comparing the carboxyl terminal sequences of H-Ras versus K-Ras4B (K-Ras) have revealed several critical aspects of this process. In the case of H-Ras, following farnesylation on cysteine 186, the last three amino acids are removed and the new carboxyl-terminal cysteine is carboxymethylated (7, 17, 20, 35). H-Ras is further processed by an endoplasmic reticulum-Golgi resident palmitoyl transferase at a second and/or third cysteine residue (amino acids 181 and 184) located two and five amino acids upstream (underlined) DESGPGCMSCKCVLS. These lipid modifications provide for the entry of H-Ras into the general membrane secretory pathway, resulting in its trafficking through the endoplasmic reticulum and Golgi complex. H-Ras ultimately localizes to the plasma membrane, where the palmitoyl groups favor association with caveolin-enriched lipid raft microdomains (3, 10, 12, 27, 38). In contrast, K-Ras has a series of positively charged amino acids (Lys) adjacent to the CAAX sequence (GKKKKKKSKTKCVIM) that function as a second targeting domain and are thought to exclude K-Ras from the secretory membrane trafficking system (3, 10, 21, 37). Thus, following farnesylation, carboxyl-terminal trimming and carboxymethylation, K-Ras inserts directly into non-lipid raft regions of the plasma membrane through a cytosolic pathway. These different trafficking and localization characteristics appear to account for the differences in signaling specificity between H-Ras and K-Ras (37, 40, 47).

TC10 is an unusual member of the Rho family of small GTP-binding proteins that has been implicated in the regulation of insulin-stimulated GLUT4 translocation (8, 43). Most Rho-family members contain a single carboxyl cysteine residue in the appropriate context for geranylgeranylation and interact with guanylnucleotide dissociation inhibitors (4, 31). However, the carboxyl-terminal region of TC10 contains sequences similar to that of H-Ras and is predicted to undergo both farnesylation and palmitoylation (underlined) (KKHTVKKRIGSRCINCCLIT). In addition, TC10 also contains two short basic regions (underlined) that could potentially function as a Golgi membrane exclusion sequence in analogy to K-Ras (KKHTVKKRIGSRCINCCLIT). However, recent sucrose gradient flotation analysis demonstrated the cofractionation of TC10 with caveolin, suggesting a trafficking pattern similar to H-Ras that directs localization to caveolin-enriched lipid raft microdomains (8). We have now used a combination of Triton X-100 solubility and fluorescence microscopy to establish the localization of TC10 to plasma membrane lipid raft microdomains. Mutational analysis identified cysteines 210 and 209 as the critical residues responsible for this targeting and indicated that the association of TC10 with lipid raft microdomains is essential for its ability to regulate insulin-dependent GLUT4 translocation in adipocytes. Surprisingly however, inhibition of Golgi membrane transport by temperature block or brefeldin A (BFA) treatment did not impair TC10 or H-Ras localization to the cell surface but instead resulted in altered targeting to non-lipid raft subdomains of the plasma membrane. These data demonstrate that both TC10 and H-Ras can traffic to the plasma membrane via an alternative transport mechanism independent of the membrane secretory pathway.

MATERIALS AND METHODS

Antibodies.

The caveolin 1 polyclonal and p115 monoclonal antibodies were purchased from Transduction Laboratories. Monoclonal antihemagglutinin (anti-HA) antibodies were obtained from Santa Cruz and Sigma, and the rabbit polyclonal GLUT4 antibody IA02 was isolated as previously described (25). The monoclonal anti-vesicular stomatitis virus G (anti-VSV-G) antibody was purchased from Accurate Chemical and Scientific Corp. The Giantin polyclonal antibody was kindly provided by Isabelle Moosbrugger (Institute of Immunology and Molecular Genetics, Karlsruhe, Germany). Fluorescent secondary antibodies were purchased from Jackson Immunoresearch Laboratories. Horseradish peroxidase-conjugated secondary antibodies were from Pierce.

TC10 cloning and site-directed mutagenesis.

The C206-to S TC10 (TC10/C206S), TC10/C209S, TC10/C210S, TC10/C206,210S, TC10/K194,195,199,200S (TC10/ΔK), and TC10/ΔK/C206,209S point mutants were generated from the human wild-type TC10 (TC10/WT) cDNA using PCR with oligonucleotide primers containing the appropriate base substitutions. The enhanced green fluorescent protein (EGFP)-TC10 and EGFP-Ras constructs were made by subcloning into the pEGFP-C1 vector (Clontech). GFP-Syn3/TM was prepared as previously described (42). The VSV-G protein cDNA was obtained from The University of Iowa Gene Transfer Vector Core Facility. GLUT1-EGFP and HA-TC10 were prepared as described previously (14, 43).

Cell culture and transient transfection of 3T3L1 adipocytes.

Murine 3T3L1 preadipocytes were purchased from the American Type Culture Collection repository. Cells were cultured in Dulbecco's modified Eagle's medium supplemented with 25 mM glucose and 10% calf serum at 37°C with 8% CO₂. Cells were differentiated into adipocytes with insulin (1 μg/ml), 1 μM dexamethasone, and 0.5 mM isobutyl-1-methylxanthine as previously described (32). Differentiated adipocytes were electroporated using the Gene Pulser II (Bio-Rad) with settings of 0.16 kV and 950 μF (29). Following electroporation, cells were plated on glass coverslips and allowed to recover in complete medium prior to stimulation with 100 nM insulin for 30 min.

Immunofluorescence and image analysis.

Transfected adipocytes were washed in phosphate-buffered saline (PBS) and fixed for 15 min in 4% paraformaldehyde containing 0.2% Triton X-100. Cells were washed briefly in PBS and then blocked in 5% donkey serum (Sigma) plus 1% bovine serum albumin (Sigma) for 1 h. Primary and secondary antibodies were used at 1:100 dilutions in blocking solution, and samples were mounted on glass slides with Vectashield (Vector Labs). Cells were imaged using confocal fluorescence microscopy (Zeiss LSM 510). Images were then imported into Adobe Photoshop (Adobe Systems, Inc.) for processing, and composite files were generated.

Fluorescent analysis of proteins insoluble in cold Triton X-100.

Differentiated 3T3L1 adipocytes were electroporated with the constructs of interest and plated on collagen-coated coverslips. After an 18-h recovery period, the cells were incubated for 15 min with 1% Triton X-100-PBS at 0°C as described previously (30).

RESULTS

The carboxyl-terminal CAAX targeting domain routes TC10 through the secretory membrane transport pathway.

TC10 has been reported to traffic through the secretory membrane system and localize to caveolin-enriched lipid raft domains in the plasma membrane of adipocytes (43). To identify the structural motifs responsible for this localization pattern, we generated a series of point mutations within the carboxyl-terminal targeting domain of TC10 (Fig. 1A). We then examined the effect of these point mutations on TC10 localization in intact 3T3L1 adipocytes (Fig. 1B). As previously observed, expression of TC10/WT for 18 h resulted in both a perinuclear and plasma membrane distribution (Fig. 1B, panel 1). Mutation of cysteine 206 to serine (TC10/C206S) had no significant effect on its subcellular distribution (Fig. 1B, panel 2). In contrast, expression of TC10/C209S resulted in a nearly exclusive plasma membrane localization with only a trace of protein present in any intracellular membrane compartment (Fig. 1B, panel 3). Moreover, mutation of both cysteines 206 and 209 (TC10/C206,209S) resulted in plasma membrane localization, but this mutant also displayed a slightly more peripheral or cytosolic distribution compared to TC10/WT, TC10/C206S, or TC10/C209S (Fig. 1B, panel 4). In any case, mutation of the terminal cysteine residue to serine (TC10/C210S) resulted in a complete lack of membrane localization, with essentially all the protein confined to the cytosol (Fig. 1B, panel 5), whereas mutation of the four upstream lysine residues to serine (TC10/ΔK) was without effect on TC10's trafficking and localization pattern (Fig. 1B, panel 6). Finally, mutation of the four upstream lysines together with cysteines 206 and 209 (TC10/ΔK/C206,209S) resulted in a pattern that was indistinguishable from the TC10/C206,209S mutant (Fig. 1B, panel 7). For comparison, the expression patterns of H-Ras/WT and K-Ras/WT are also shown (Fig. 1B, panels 8 and 9). These data demonstrate that TC10 can localize to the plasma membrane in the complete absence of palmitoylation but that the farnesylation cysteine is essential for membrane localization. In addition, the upstream lysine residues are not required for the trafficking and localization of TC10 to the plasma membrane. Together, these data are in excellent agreement with that established for the posttranslational processing and specific targeting of the homologous H-Ras carboxyl terminal targeting motif (3, 10, 12, 27, 38).

FIG. 1.

The TC10 carboxyl-terminal cysteine residues C209 and C210 are necessary for plasma membrane localization. (A) Summary of the carboxyl-terminal targeting domains of TC10, H-Ras, and K-Ras, illustrating the point mutations (underlined) introduced into the CAAX domain of TC10. Green, predicted for nesylated residues; blue, predicted palmitoylated residues; red, basic residues. (B) Differentiated 3T3L1 adipocytes were electroporated with 50 μg of cDNAs encoding for the HA epitope-tagged human TC10/WT (panel 1), TC10/C206S (panel 2), TC10/C209S (panel 3), TC10/C206,209S (panel 4), TC10/C210S (panel 5), TC10/ΔK (panel 6), TC10/ΔK/C206,209S (panel 7), H-Ras/WT (panel 8), and K-Ras/WT (panel 9) constructs. Eighteen hours later, the cells were fixed, and the subcellular localization was determined by confocal fluorescent microscopy.

We next determined whether the presence of TC10 in intracellular membranes reflected newly synthesized protein in transit to the plasma membrane. Adipocytes were transfected with EGFP-TC10/WT, allowed to recover for 3 h, and then incubated in the absence or presence of the protein synthesis inhibitor cycloheximide (CHX) for various times (Fig. 2). In the absence of CHX, TC10 expression for 3 h resulted in localization to both the plasma membrane and the perinuclear endomembrane system (Fig. 2, panels 1 and 6). This distribution persisted for up to 6 h (Fig. 2, panels 2 to 5). Similarly, the newly synthesized EGFP-TC10 was distributed between the plasma membrane and perinuclear region following 1 h of CHX treatment (Fig. 2, panel 7). However, the amount of perinuclear localized protein began to decline by 2 h and was markedly depleted from the perinuclear membranes by 4 to 6 h (Fig. 2, panels 8 to 10). Thus, treatment with CHX resulted in a time-dependent chase of the newly synthesized TC10 protein from the perinuclear membrane compartments to the plasma membrane. In addition, the perinuclear TC10 colocalized with both Golgi (Giantin and p115) and trans-Golgi network (TGN) (syntaxin 6) markers (data not shown). Thus, together these data indicate that newly synthesized TC10 protein can transit through the secretory membrane system en route to the plasma membrane.

FIG. 2.

Expressed TC10 traffics through the exocytotic membrane system en route to the plasma membrane. Differentiated 3T3L1 adipocytes were electroporated with 50 μg of the cDNA encoding for the EGFP-TC10/WT fusion as described in Materials and Methods. Following a 3-h recovery period, the cells were incubated in the absence (panels 1 to 5) or presence (panels 6 to 10) of CHX (10 μg/ml) for the indicated times (0, 1, 2, 4, and 6 h). The cells were then fixed and visualized by confocal fluorescent microscopy. These fields are representative of cells from three independent determinations.

Functional blockade of the secretory membrane transport pathway does not inhibit the plasma membrane localization of TC10 or H-Ras.

Several studies examining H-Ras trafficking in fibroblasts have observed that collapse of Golgi membranes with BFA inhibited the transport of H-Ras to the plasma membrane (3, 10). More recently, an alternative endoplasmic reticulum-Golgi-independent transport pathway in Saccharomyces cerevisiae has been observed to target ras2p to the plasma membrane through a process that occurs prior to palmitoylation (5). Therefore, to examine the requirement of Golgi membranes for TC10 trafficking, adipocytes were transfected and then immediately plated into media supplemented with or without BFA (Fig. 3). Consistent with our previous results, at 8 h following transfection K-Ras was primarily found at the plasma membrane, with only a small amount of intracellular localization (Fig. 3A, panel 1). As a marker for exocytotic membrane processing to the plasma membrane, we also coexpressed a GFP fusion protein containing the syntaxin 3-transmembrane domain (GFP-Syn3/TM) and compared it with the endogenous Golgi marker p115 (41, 43) (Fig. 3A, panels 2 to 4). The GFP-Syn3/TM construct was chosen as a control because it is a type II integral membrane protein that is topologically similar to CAAX-containing proteins. As expected, treatment with BFA had no significant effect on the plasma membrane localization of K-Ras (Fig. 3A, panel 5). In contrast, the perinuclear localized GFP-Syn3/TM and p115 were completely dispersed and targeting of GFP-Syn3/TM to the plasma membrane was prevented (Fig. 3A, panels 6 to 8).

FIG.3.

BFA treatment collapses Golgi membranes but does not prevent TC10, H-Ras, or K-Ras trafficking to the plasma membrane. Differentiated 3T3L1 adipocytes were electroporated with 50 μg of the GFP-Syn3/TM and 50 μg of the HA-K-Ras (A), HA-H-Ras (B), or HA-TC10 (C) cDNA. The cells were either left untreated (panels 1 to 4) or immediately incubated with BFA (5 μg/ml) (panels 5 to 8) for 8 h. The cells were then fixed and colabeled with a polyclonal anti-HA antibody and a monoclonal anti-p115 antibody and processed for confocal fluorescent microscopy as described in Materials and Methods. The merged images for the HA-epitope, GFP-Syn3/TM, and p115 are presented in panels 4 and 8.

In other experiments, BFA also completely blocked the appearance of newly synthesized VSV-G protein at the plasma membrane (data not shown). In contrast to K-Ras, H-Ras was both perinuclear and plasma membrane localized, with a distribution similar to those of GFP-Syn3/TM and the Golgi marker p115 (Fig. 3B, panels 1 to 4). Surprisingly however, BFA treatment did not prevent the localization of H-Ras to the plasma membrane but completely disrupted the appearance of intracellular membrane-localized H-Ras protein (Fig. 3B, panel 5). The plasma membrane localization of H-Ras occurred despite the inhibition of GFP-Syn3/TM plasma membrane localization and dispersion of the perinuclear GFP-Syn3/TM and p115 (Fig. 3B, panels 6 to 8). Similar to H-Ras, TC10 was localized to both plasma membrane and the perinuclear region (Fig. 3C, panels 1 to 4). Nevertheless, although BFA treatment collapsed the Golgi membranes and prevented GFP-Syn3/TM trafficking to the plasma membrane, TC10 was still found at the plasma membrane with near complete disappearance of any intracellular TC10 protein (Fig. 3C, panels 5 to 8). Quantitation of the intracellular distribution of newly synthesized TC10, H-Ras, and K-Ras demonstrated that BFA treatment had no discernible effect on the extent of TC10, H-Ras, or K-Ras plasma membrane localization (data not shown).

The surprising observation that TC10 and H-Ras can still accumulate at the plasma membrane in the presence of BFA in adipocytes suggests the presence of an alternative, membrane-independent exocytic trafficking pathway. To further investigate this possibility, we took advantage of the known property of reduced temperature to specifically block TGN membrane vesicle exit. Typically 20°C is widely used to block TGN exit in fibroblasts (15); however, 19°C is more effective at blocking TGN exit in adipocytes, while still allowing efficient vesicular transport from the endoplasmic reticulum to the Golgi (Fig. 4 and reference 36). As controls, cells transfected with the VSV-G cDNA and maintained at 19°C for 24 h resulted in a perinuclear localization of VSV-G protein, with no detectable localization to the plasma membrane (Fig. 4A, panel 2). The VSV-G protein was concentrated in the Golgi complex, as it strongly colocalized with the Golgi protein marker Giantin (Fig. 4A, panels 1 and 2). This was not due to a loss of cell viability, as warming the cells to 37°C for 4 h resulted in the trafficking of VSV-G protein to the plasma membrane (Fig. 4A, panels 3 and 4). It should be noted that two of the Giantin-positive cells in panels 1 and 3 were not transfected with the VSV-G cDNA and therefore were VSV-G protein negative (Fig. 4A, panels 1 to 4). In any case, although incubation at 19°C completely prevented the membrane transport of VSV-G protein, the trafficking of newly synthesized H-Ras and TC10 to the plasma membrane was completely unaffected (Fig. 4B, panels 1 to 4). In addition, both H-Ras and TC10 were localized to the perinuclear region identical to cells incubated at 37°C. Similarly, the localization of K-Ras in cells incubated at 19°C was also identical to cells incubated at 37°C with a predominant plasma membrane distribution and a small amount of diffuse intracellular labeling (Fig. 4B, panels 5 and 6). Essentially identical results were obtained in cells incubated at 16°C, which prevents endoplasmic reticulum membrane transport exit (data not shown).

FIG. 4.

Temperature block of TGN membrane transport does not affect the plasma membrane accumulation of TC10, H-Ras, or K-Ras. (A) Differentiated 3T3L1 adipocytes were electroporated with 50 μg of the VSV-G cDNA (panels 1 to 4) and incubated for 24 h at 19°C. The cells were either left untreated (panels 1 and 2) or warmed to 37°C (panels 3 and 4) for 4 h. The cells were fixed and processed for confocal fluorescent microscopy as described in Materials and Methods. (B) Differentiated 3T3L1 adipocytes were electroporated with 50 μg of the VSV-G cDNA (panels 1 to 6) plus 50 μg of EGFP-H-Ras (panels 1 and 2), EGFP-TC10 (panels 3 and 4), or EGFP-K-Ras (panels 5 and 6) and incubated for 24 h at 19°C. The cells were fixed and processed for confocal fluorescent microscopy as described in Materials and Methods.

The fact that newly synthesized TC10 and H-Ras can become plasma membrane localized under conditions that inhibit exocytotic membrane transport is consistent with the function of an alternative parallel transport pathway. To compare the relative rate of these two transport processes, we next transfected cells with TC10 and immediately (within 1 min) incubated the cells with CHX for 3 h to block protein synthesis but not TC10 transcription. At the end of the 3-h protein synthesis block, CHX was removed, and within 15 min TC10 was rapidly translated and displayed a diffuse intracellular labeling pattern (Fig. 5, panel 1). However, in approximately 30% of the cells there was also a clear plasma membrane localization without any evidence for accumulation in post-endoplasmic reticulum compartments. Thirty minutes following release of the CHX block, all of the cells displayed a plasma membrane TC10 localization, with a concomitant decrease in cytosolic labeling. In addition, some cells showed significant levels of TC10 in the perinuclear region (Fig. 5A, panel 2). By 60 to 180 min after CHX removal, the intracellular localization pattern of TC10 was essentially identical to that observed for control cells that were not exposed to CHX (Fig. 5A, panels 3 to 5). This rapid exocytotic membrane-independent transport pathway was specific for TC10, as GLUT1 was also synthesized 15 min following CHX washout but at this time was perinuclear localized and was not detected in the plasma membrane (Fig. 5, panel 6). Similarly, at 30 min all the detectable GLUT1 remained in the perinuclear region, whereas TC10 was already plasma membrane localized (Fig. 5, panel 7). GLUT1 was only weakly present at the plasma membrane 60 min following translation and was readily apparent at 120 and 180 min (Fig. 5, panels 9 and 10). Thus, unlike integral membrane proteins that must traffic to the plasma membrane via the secretory membrane transport system, newly synthesized TC10 protein utilizes both a more rapid nonsecretory membrane transport system as well as the slower prototypical membrane transport pathway.

FIG. 5.

Newly synthesized TC10 rapidly accumulates at the plasma membrane prior to entry into the exocytotic membrane system. Differentiated 3T3L1 adipocytes were coelectroporated with a cDNAs (100 μg) encoding HA-TC10/WT (panels 1 to 5) and GLUT1-EGFP (panels 6 to 10) as described in Materials and Methods. Immediately following electroporation, the cells were incubated in the presence of CHX (10 μg/ml) for 3 h. The cells were then rapidly washed to remove the CHX, incubated for the times indicated, and processed for confocal fluorescent microscopy. These images are representative of three independent determinations.

Transport through the secretory membrane pathway is necessary for compartmentalization of TC10 in plasma membrane lipid raft microdomains.

Due to the unique biophysical properties of lipid raft microdomains, proteins such as caveolin that localize to these structures are relatively resistant to cold Triton X-100 extraction (2). Recently, we have also observed that TC10 and caveolin colocalize in approximately 1-μm-diameter ring-like structures in the plasma membrane of adipocytes (43). Therefore, to determine whether these TC10-containing morphologically distinct regions represent lipid raft microdomains, we assessed the ability of these domains to be extracted with cold Triton X-100 (Fig. 6). As a control for expression, TC10/WT was predominantly found at the plasma membrane and in the perinuclear region similar to that for the VSV-G protein (Fig. 6A, panels 1 and 13). As expected, the endogenous caveolin protein was predominantly localized to the plasma membrane, although it was also detected in internal membrane structures (Fig. 6A, panel 7). Consistent with previous results, the expressed TC10/C209S was exclusively localized to the plasma membrane, whereas TC10/C210S was cytosolic (Fig. 6A, panels 2 and 3). In addition, the localization pattern of TC10/ΔK was indistinguishable from TC10/WT (Fig. 6A, panel 4).

FIG. 6.

Localization of TC10 to plasma membrane lipid raft microdomains. Differentiated 3T3L1 adipocytes were coelectroporated with a cDNA (100 μg) encoding for VSV-G protein and cDNAs (100 μg) encoding for EGFP-TC10/WT (panels 1, 7, and 13), EGFP-TC10/C209S (panels 2, 8, and 14), EGFP-TC10/C210S (panels 3, 9, and 15), EGFP-TC10/ΔK (panels 4, 10, and 16), EGFP-HRas (panels 5, 11, and 17) or EGFP-KRas (panels 6, 12 and 18) as described in Materials and Methods. (A) The cells were cells were allowed to recover for 24 h, fixed, and visualized by confocal fluorescent microscopy for EGFP (panels 1 to 6), endogenous caveolin 1 (panels 7 to 12), and VSV-G protein (panels 13 to 18). These fields are representative of cells from four independent determinations. (B) The cells were incubated with ice-cold Triton X-100 (1%) for 15 min and subsequently washed to remove the extract material. The protein remaining on the coverslips was then visualized by confocal fluorescent microscopy for the presence of EGFP (panels 1 to 6), endogenous caveolin 1 (panels 7 to 12), and VSV-G protein (panels 13 to 18) as described in Materials and Methods. These images are representative of four independent determinations.

Previous studies have reported that H-Ras can also traffic through the secretory membrane system en route to the plasma membrane, whereas K-Ras is apparently excluded from intracellular membranes and traffics through a soluble pathway (3, 10). Thus, for comparison we also observed that H-Ras displayed an identical subcellular distribution as TC10/WT and that K-Ras had the same localization as TC10/C209S (Fig. 6A, panels 5 and 6). Expression of TC10/WT, TC10/C206S, TC10/C209S, TC10/ΔK, H-Ras or K-Ras had no significant effect on the localization of caveolin or VSV-G protein (Fig. 6A, panels 7 to 18).

Having established the expression patterns of the various proteins, the cells were then extracted with ice-cold Triton X-100, and the resulting membrane fragments were visualized by fluorescent microscopy as described by Nichols et al. (30). Since caveolin is localized to Triton X-100-resistant lipid raft microdomains, the caveolin protein remained attached to the coverslips in a punctate pattern consistent with morphologically distinct ring-shaped structures (Fig. 6B, panels 7 to 12). In contrast, VSV-G protein, being a non-lipid-raft-localized protein, was almost completely extracted by cold Triton X-100 (Fig. 6B, panels 13 to 18). Consistent with plasma membrane lipid raft microdomain compartmentalization, TC10/WT was resistant to cold Triton X-100 and remained colocalized in the caveolin-positive structures (Fig. 6B, panel 1), as was the TC10/ΔK mutant (Fig. 6B, panel 4). In contrast, although TC10/C209S was plasma membrane localized (Fig. 6A), it was completely extracted with cold Triton X-100 (Fig. 6B, panel 2). Since TC10/C210S was predominantly cytosolic, it was also completely extracted by cold Triton X-100 (Fig. 6B, panel 3). Similarly, the TC10/C206,209S and TC10/ΔK/C206,209S mutants were also completely extracted with cold Triton X-100 (data not shown). As controls, H-Ras was resistant to cold Triton X-100 extraction, whereas K-Ras was completely sensitive (Fig. 6B, panels 5 and 6).

Since TC10/C206S only contains a single potential palmitoylation site, we also examined its lipid raft association by cold Triton X-100 extraction. Similar to that of TC10/WT, expression of TC10/C206S also resulted in both plasma membrane and perinuclear localization (Fig. 7A, panels 1 to 3). As previously observed, caveolin was resistant to cold Triton X-100, whereas the coexpressed VSV-G protein was almost completely extracted (Fig. 7B, panels 5 to 12). Although TC10/C206S was resistant to cold Triton X-100, the degree of extraction was variable, ranging from essentially none to partial extraction (Fig. 7B, panels 1 to 4). Nevertheless, in all cases the residual TC10/C206 protein was colocalized with caveolin. Together, these data demonstrate that cysteine 209 and cysteine 210 are both necessary for TC10 to associate with the plasma membrane lipid raft microdomains in adipocytes. In contrast, cysteine 206 appears to help stabilize the association of TC10 to lipid rafts, although this residue is not essential for TC10 localization and trafficking. The fact that BFA treatment prevented the endomembrane trafficking of H-Ras and TC10 without affecting plasma membrane localization suggests that, in the absence of Golgi membrane processing, H-Ras and TC10 traffic to the plasma membrane in a palmitoylation-independent manner, perhaps through a mechanism similar to K-Ras.

FIG. 7.

The TC10/C206S mutant displays partial resistance to Triton X-100 extraction. Differentiated 3T3L1 adipocytes were coelectroporated with cDNAs (100 μg) encoding VSV-G protein and EGFP-TC10/C206S as described in Materials and Methods. (A) The cells were allowed torecover for 24 h, fixed, and visualized by confocal fluorescent microscopy for endogenous caveolin 1 (panel 1), EGFP-TC10/C206S (panel 2), and VSV-G protein (panel 3). (B) The cells were incubated with ice cold Triton X-100 (1%) for 15 min and subsequently washed to remove the extract material. The protein remaining on the coverslips was then visualized by confocal fluorescent microscopy for the presence of EGFP-TC10/C206S (panels 1 to 4), endogenous caveolin 1 (panels 5 to 8), and VSV-G protein (panels 9 to 12) as described in Materials and Methods. These images are representative of the range of TC10/C206S extraction from five independent determinations.

Lipid raft compartmentalization is required for TC10 to modulate the insulin-dependent translocation of GLUT4.

To examine the potential functional consequences of TC10 targeting, we next assessed the effect of these TC10 mutants on insulin-stimulated GLUT4 translocation (Fig. 8). As previously reported (8, 43), expression of TC10/WT had no significant effect on the basal distribution of the coexpressed GLUT4-EGFP reporter construct (Fig. 8, panels 1 and 3). In empty vector transfected cells, insulin induced a marked redistribution of the GLUT4-EGFP fusion protein to the plasma membrane that was nearly completely inhibited by the coexpression of TC10/WT (Fig. 8, panels 1 to 4). Similarly, expression of TC10/C206S also resulted in an inhibition of insulin-stimulated GLUT4 translocation (Fig. 8, panels 5 and 6). In contrast, expression of TC10/C209S, TC10/C206,209S, or TC10/C210S had no significant effect on the insulin-stimulated translocation of the GLUT4-EGFP reporter to the plasma membrane (Fig. 8, panels 7 to 12). However, expression of TC10/ΔK strongly inhibited insulin-dependent translocation of GLUT4 (Fig. 8, panels 13 and 14), whereas the TC10/ΔK/C206,209S mutant had no significant inhibitory effect (Fig. 8, panels 15 and 16). These data demonstrate that the TC10 constructs that are either cytosolic or localized to non-lipid raft membrane microdomains have no significant influence on insulin-stimulated GLUT4 translocation. In contrast, TC10 constructs that are targeted to the plasma membrane lipid raft microdomains have a dramatic inhibitory effect on GLUT4 translocation. Thus, the appropriate compartmentalization of TC10 is required for its ability to functionally modulate GLUT4 translocation.

FIG. 8.

Targeting to lipid rafts is required for TC10 to modulate GLUT4 translocation. Differentiated 3T3L1 adipocytes were coelectroporated with 50 μg of GLUT4-EGFP plus 200 μg of the empty vector (panels 1 and 2), TC10/WT (panels 3 and 4), TC10/C206S (panels 5 and 6), TC10/C209S (panels 7 and 8), TC10/C206,209S (panels 9 and 10), TC10/C210S (panels 11 and 12), TC10/ΔK (panels 13 and 14), or TC10/ΔK/C206,209S (panes 15 and 16) mutant cDNAs. Eighteen hours later, the cells were then serum starved and incubated for 30 min in the absence (panels 1, 3, 5, 7, 9, 11, 13, and 15) or presence (panels 2, 4, 6, 8, 10, 12, 14, and 16) of 100 nM insulin. The cells were then fixed, and the subcellular localization was determined by confocal fluorescent microscopy. These fields are representative of cells from three or four independent determinations.

DISCUSSION

Insulin stimulates glucose uptake in adipocytes and muscle by recruiting the GLUT4 facilitative glucose transporter from intracellular compartments to the cell surface. This process is initiated by the activation of the insulin receptor tyrosine kinase, followed by the tyrosine phosphorylation of IRS proteins that subsequently interact with the type IA phosphatidylinositol (PI) 3-kinase (for reviews, see references 11 and 44). Although these signaling events are necessary for the insulin-stimulated translocation of GLUT4, several studies have suggested that activation of the PI 3-kinase is not sufficient (16, 22, 23, 26, 45). Recently, a second insulin receptor-mediated signaling pathway has been proposed to function in concert with the PI 3-kinase pathway. This pathway involves the tyrosine phosphorylation of the CAP/Cbl/APS complex that associates with and recruits the CrkII/C3G proteins to plasma membrane lipid raft microdomains (6). In vitro and in vivo studies have demonstrated that C3G can function as a guanylnucleotide exchange factor TC10 (8). Thus, these data suggest that the plasma membrane lipid raft microdomain targeting of the CAP/Cbl/APS/CrkII/C3G signaling complex functions to direct C3G to the appropriate location for the efficient insulin-dependent activation of TC10.

One prediction of this model is that TC10 must be a resident component of plasma membrane lipid raft microdomains. The data presented here demonstrate that following its initial biosynthesis, TC10 can enter the secretory membrane trafficking system while en route to the plasma membrane. This was based upon the ability of CHX to efficiently chase newly synthesized TC10 out of the perinuclear region. Consistent with these data in adipocytes, it has recently been reported that TC10 can also traffic through the secretory system to the plasma membrane in fibroblasts (28). These localization properties of TC10 are similar to that reported for the trafficking and targeting of H-Ras (3, 10, 12, 27, 38). However, although the majority of TC10 localizes to the plasma membrane under steady-state conditions, it remains to be determined whether TC10 may also play a functional role in intracellular membranes, as has recently been proposed for H-Ras (9).

The carboxyl-terminal domain of TC10 contains three consensus cysteine residues in the appropriate context for palmitoylation and farnesylation. Thus, to determine whether the carboxyl-terminal residues of TC10 also provide information to direct its membrane trafficking, localization, and functional properties, we prepared site-directed mutants of cysteines 206, 209, and 210. As expected, expressed TC10/C210S was completely confined to the cytoplasm, with no evidence for membrane localization. This is consistent with prenylation being the first requisite step in CAAX box posttranslational modification (7, 18-20, 34). Since this protein was not membrane bound, it did not affect insulin-dependent GLUT4 translocation. On the other hand, the expressed TC10/C206S protein was indistinguishable from TC10/WT in terms of its plasma membrane localization and inhibition of GLUT4 translocation. However, the interaction of TC10/C206S with lipid raft microdomains was functionally weaker than that for TC10/WT or H-Ras based upon relative Triton X-100 extractability. This suggests that cysteine 206 is also palmitoylated but that a single palmitoylation on cysteine 209 is sufficient for lipid raft association. Consistent with the latter possibility, a single palmitoylation of H-Ras at cysteine 184 was also found to be sufficient to confer appropriate trafficking, plasma membrane microdomain localization, and function (20, 21, 46).

In contrast to the TC10/C206S construct, the TC10/C209S point mutant localized to the plasma membrane without any evidence of endomembrane accumulation and was indistinguishable from K-Ras localization. This suggests that elimination of cysteine 209 also prevented the efficient palmitoylation of cysteine 206. Consistent with this interpretation, RhoB is known to undergo farnesylation at cysteine 193, as well as palmitoylation at two upstream cysteines, C189 and C192 (1). In addition, palmitoylation of RhoB at C192 appears to increase the efficiency C189 palmitoylation. Therefore, assuming that TC10 can undergo sequential palmitoylation then, although TC10/C209S can be farnesylated at C210, this mutant may not undergo efficient palmitoylation at C206 following initial biosynthesis and endomembrane trafficking. However, it is possible that cysteine 206 may undergo palmitoylation once the TC10/C209S mutant is present at the cell surface, which in turn may help anchor it at the plasma membrane. Regardless, even though TC10/C209S was plasma membrane localized, it had no significant effect on insulin-stimulated GLUT4 translocation and was readily extracted with cold Triton X-100, indicating it was not present in lipid rafts at the cell surface. These data are also consistent with the trafficking of TC10/C209S through a pathway independent of the exocytotic membrane transport system and its inability to localize to plasma membrane lipid raft microdomains. Thus, the localization and trafficking patterns of TC10/C209S is analogous to that of K-Ras and demonstrates that lipid raft microdomain localization requires both prenylation and palmitoylation.

To confirm that palmitoylation is not necessary for plasma membrane localization and can occur independent of secretory membrane transport, we generated a double TC10/C206,209S mutant. This construct was plasma membrane localized but was not detected in any intracellular membrane compartments similar to that of TC10/C209S and K-Ras/WT. As expected, this mutant was excluded from lipid raft microdomains and was not functional in terms of insulin-stimulated GLUT4 translocation. Interestingly however, TC10/C206,209S was more peripherally plasma membrane localized, with detectable amounts in the cytosol, suggesting a weaker plasma membrane association. Since TC10, unlike K-Ras, does not contain a strong polybasic region, the lack of both a palmitoylation and a polybasic signal in the TC10/C206,209S appears to account for its weak plasma membrane interaction. Consistent with this interpretation, we have also observed that the related Rho-family member Cdc42 displays a localization pattern very similar to that of TC10/C206,209S when expressed in adipocytes (24). As predicted, Cdc42 has only a single prenylation site (underlined) and lacks both an extended polybasic motif and palmitoylation sites (AAEAIEPKKSRRCVLL).

It is well established that both H-Ras and TC10 can traffic to the cell surface by an exocytotic membrane pathway in mammalian cells. However, recent studies have suggested the presence of an alternative transport pathway independent of palmitoylation for ras2p in yeast (5). Potentially, this may represent the pathway utilized by K-Ras in mammalian cells and is consistent with the similar membrane transport properties observed when palmitoylation is blocked in the TC10 mutants. Furthermore, BFA-induced collapse of Golgi membranes back into the endoplasmic reticulum through the inhibition of Arf-dependent vesicle formation did not inhibit the transport of either H-Ras or TC10 to the plasma membrane. Thus, despite association of TC10 and H-Ras with the endoplasmic reticulum, they are able to traffic to the plasma membrane without transiting through the Golgi complex. Consistent with this interpretation, inhibition of TGN vesicle budding by reduced temperature also did not impair the plasma membrane localization of TC10 or H-Ras. In fact, following translation TC10 initially localizes to the plasma membrane and subsequently becomes associated with the exocytotic membrane system. Thus, these data support a model in which adipocytes contain two transport pathways for the trafficking of these small GTP binding proteins (Fig. 9). Under normal conditions the majority of TC10 and H-Ras are lipid raft microdomain associated, suggesting that palmitoylation and exocytotic membrane transport are the primary mechanism for plasma membrane targeting. However, a second potentially soluble, parallel pathway is present that operates in the absence of palmitoylation and directs these CAAX targeting sequences to the non-lipid raft plasma membrane regions. Alternatively, TC10 and H-Ras may be continuously exchanging between these two pathways through a regulatory balance of palmitoylation-depalmitoylation reactions. This possibility would be analogous to the role of palmitate cycling on PSD-95, which regulates synaptic clustering and glutamate receptor function in hippocampal neurons (13). Furthermore, recent studies have suggested that GTP loading can regulate the distribution of H-Ras between lipid raft and non-lipid raft microdomains (33). Thus, it remains possible that the spatial compartmentalization of TC10 may be defined by the relative rates of palmitoylation and depalmitoylation reactions controlled by the GTP-GDP cycle. Future work is needed to resolve these issues and to define the nature of the parallel cytosolic transport pathway.

FIG. 9.

Hypothetical model depicting the trafficking of newly synthesized TC10 through both membrane and soluble transport pathways to the plasma membrane. Following translation on free ribosomes, the newly synthesized TC10 protein undergoes farnesylation and subsequently associates with the endoplasmic reticulum (ER) for proteolytic trimming and carboxymethylation. The majority of the modified TC10 protein enters the Golgi system, where palmitoylation probably occurs, providing entry into the membrane secretory transport pathway en route to the lipid raft regions (caveolae) of the plasma membrane. This pathway is in equilibrium with a soluble trafficking pathway for TC10 molecules that have either escaped palmitoylation or have undergone depalmitoylation along the membrane transport pathway. This second pathway results in the localization of TC10 to the disordered, nonraft regions of the plasma membrane. Under normal conditions, the membrane transport pathway predominates; however, when blocked (e.g., by BFA or reduced temperature), a significant proportion of TC10 is redirected to the soluble transport pathway and becomes readily apparent.

FIG. 1—Continued.