The Trimeric GTP-binding Protein (Gq/G11) α Subunit Is Required for Insulin-stimulated GLUT4 Translocation in 3T3L1 Adipocytes*

To investigate the potential role of trimeric GTP-binding proteins regulating GLUT4 translocation in adipocytes, wild type and constitutively active Gq(Gq/Q209L), Gi (Gi/Q205L), and Gs (Gs/Q227L) α subunit mutants were expressed in 3T3L1 adipocytes. Although expression of neither the wild type nor Gi/Q205L and Gs/Q227L α subunit mutants had any effect on the basal or insulin-stimulated translocation of a co-expressed GLUT4-enhanced green fluorescent protein (EGFP) fusion protein, expression of Gq/Q209L resulted in GLUT4-EGFP translocation in the absence of insulin. In contrast, microinjection of an inhibitory Gq/G11 α subunit-specific antibody but not a Gi or Gsα subunit antibody prevented insulin-stimulated endogenous GLUT4 translocation. Consistent with a required role for GTP-bound Gq/G11, expression of the regulators of G protein signaling (RGS4 and RGS16) also attenuated insulin-stimulated GLUT4-EGFP translocation. To assess the relationship between Gq/G11 function with the phosphatidylinositol 3-kinase dependent pathway, expression of a dominant-interfering p85 regulatory subunit, as well as wortmannin treatment inhibited insulin-stimulated but not Gq/Q209L-stimulated GLUT4-EGFP translocation. Furthermore, Gq/Q209L did not induce thein vivo accumulation of phosphatidylinositol-3,4,5-trisphosphate (PIP3), whereas expression of the RGS proteins did not prevent the insulin-stimulated accumulation of PIP3. Together, these data demonstrate that insulin stimulation of GLUT4 translocation requires at least two independent signal transduction pathways, one mediated through the phosphatidylinositol 3-kinase and another through the trimeric GTP-binding proteins Gq and/or G11.

It has been well established that insulin stimulation results in increased glucose uptake in striated muscle and adipose tissue primarily through the translocation of the GLUT4 glucose transporter protein from intracellular storage sites to the plasma membrane (1)(2)(3)(4). This occurs through the activation of the intrinsic tyrosine kinase of the insulin receptor and subsequent phosphorylation of downstream effector proteins. In particular, tyrosine phosphorylation of the insulin receptor substrate (IRS) 1 proteins provides docking site for p85, the regulatory subunit of the type I phosphatidylinositol (PI) 3-kinase, resulting in activation of the catalytic p110 subunit (5)(6)(7). Multiple studies using various pharmacological inhibitors, dominant-interfering mutants, and constitutively active constructs are all consistent with the necessary role of PI 3-kinase activity in insulin-stimulated GLUT4 translocation (8 -14). However, currently it is unclear whether the association of the PI 3-kinase with an IRS protein is a required event and/or the nature of the signaling pathways functioning downstream of phosphatidylinositol (3,4,5)trisphosphate (PIP 3 ) formation. For example, several studies expressing dominant-interfering IRS PH/phosphotyrosine-binding domains have demonstrated a complete inhibition of insulin-stimulated IRS tyrosine phosphorylation and insulin-stimulated DNA synthesis, yet there was no effect on GLUT4 translocation (15,16). Thus, although PI 3-kinase function may be required for insulin-stimulated GLUT4 translocation, it is not clear whether alternative pathway(s) to IRS can couple to PI 3-kinase activation. Similarly, the activation of both protein kinase B and the atypical protein kinase C/ isoforms are both dependent on PI 3-kinase activity (17)(18)(19)(20)(21)(22)(23)(24). In this regard, several studies have directly implicated protein kinase B in insulin-stimulated GLUT4 translocation, whereas other reports have provided evidence supporting a role for protein kinase C but not protein kinase B (17,(25)(26)(27)(28)(29)(30)(31).
In contrast to insulin, various other stimuli display insulinomimetic properties and can induce the translocation of GLUT4 to the plasma membrane. For example, hyperosmolarity has potent insulin-like properties on glucose metabolism and induces the translocation of GLUT4 in a PI 3-kinase independent but tyrosine kinase-dependent manner (32,33). Similarly, introduction of guanosine 5Ј-[␥-thio]triphosphate (GTP␥S), a nonhydrolyzable GTP analog, into adipocytes rapidly stimulates GLUT4 translocation (34 -38). As with osmotic shock, the GTP␥S stimulation of GLUT4 translocation also requires tyrosine kinase activation and is independent of the PI 3-kinase. Furthermore, skeletal muscle appears to have two distinct pathways mediating GLUT4 translocation. For example, as observed in adipocytes, insulin-stimulated glucose transport in skeletal muscle is dependent on PI 3-kinase activity (39,40). However, muscle contraction/exercise or hypoxia stimulates glucose transport and GLUT4 translocation through a distinct pathway independent of the PI 3-kinase (41)(42)(43).
Even though these alternative signaling pathways leading to GLUT4 translocation apparently bypass the initial insulin signal transduction steps, they are likely to induce GLUT4 translocation through activation of a common convergent signal * This work was supported by Grants DK33823, DK48781, and DK29525 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  transduction step downstream and/or in parallel to the PI 3-kinase. Currently, there are several lines of evidence suggesting the involvement of regulated GTP-binding protein(s). Introduction of GDP␤S into adipocytes inhibited insulin-stimulated GLUT4 translocation, whereas the stimulatory effect of GTP␥S was mimicked by treatment with NaF plus AlCl 3 , which is characteristic of the involvement of heterotrimeric GTP-binding proteins (35-37, 44, 45). Consistent with this interpretation, adrenergic stimulation can induce GLUT4 translocation and glucose uptake in cardiac and skeletal muscle as well as white and brown adipocytes (46 -54). Furthermore, in transfected Chinese hamster ovary cells and 3T3L1 adipocytes, expression and activation of several receptors that couple to G q /G 11 resulted in GLUT4 translocation (50,55). More recently, endothelin-1 (G q /G 11 -coupled receptor agonist) has been observed to stimulate glucose transport and GLUT4 translocation in 3T3L1 adipocytes (56).
Although these data provide circumstantial evidence that the trimeric GTP-binding protein G q and/or G 11 can couple to GLUT4 translocation, the relationship between insulin signaling and adrenergic activation on GLUT4 translocation remains unclear. Furthermore, whether receptor agonists that induce GLUT4 translocation directly function through G q /G 11 activation has not yet been determined. To address these issues, we have observed that expression of a constitutively active ␣ subunit mutant of G q , but not G s or G i , is sufficient to induce GLUT4 translocation in 3T3L1 adipocytes. In addition, this G q signaling pathway displays the same characteristics as osmotic shock and GTP␥S stimulation, in that it is PI 3-kinase-independent but requires tyrosine kinase activation. Furthermore, inhibition of G q /G 11 function prevented the insulin-stimulated translocation of GLUT4. Together, these data strongly suggest that insulin must utilize at least two independent pathways (PI 3-kinase-dependent and G q /G 11 -dependent) that functionally integrate to induce GLUT4 translocation in 3T3L1 adipocytes.

EXPERIMENTAL PROCEDURES
Materials-A mouse monoclonal anti-GLUT4 antibody (RDI-GLUT4abmX) was purchased from Research Diagnostics, Inc. (Flanders, NJ). The cDNA for the cation-independent mannose-6-phosphate receptor (CI/MPR) was kindly provided by Dr. Richard Roth (Stanford University, CA). Texas Red-conjugated transferrin was obtained from Molecular Probes (Eugene, OR). Rabbit polyclonal G q /G 11 ␣, G␣ i1/2 and G s ␣ subunit (carboxyl-terminal) antibodies were purchased from Calbiochem (La Jolla, CA). A sheep polyclonal anti-maltose-binding protein was generously gifted by Dr. Morris Birnbaum (University of Pennsylvania). Cy5-conjugated Donkey anti-mouse IgG and fluorescien isothiocyanate-conjugated Donkey anti-sheep IgG were purchased from Jackson Immunoresearch Laboratories (West Grove, PA). Vectashield was obtained from Vector Laboratories (Burlingame, CA). Mini-prep DNA kit was purchased from Invitrogen (Carlsbad, CA).
Plasmids-The GLUT4-enhanced green fluorescent protein (EGFP) fusion construct was prepared as described previously and cloned into the mammalian expression plasmid pcDNA3 (57). Similarly, a CI/MPR-EGFP fusion protein was generated by cloning EGFP at the carboxylterminal domain of CI/MPR. The G q /WT and constitutively active G q / Q209L ␣ subunits were obtained from Dr. John Exton (Vanderbilt University) in the mammalian expression plasmid pCMV. The G s /WT ␣ subunit in the mammalian expression plasmid pCR3.1 was provided by Dr. Mario Ascoli (Department of Pharmacology, The University of Iowa). G i2 /WT and constitutively active mutants G i2 /Q205L and G s / Q227L in the mammalian expression plasmid pcDNA1 were purchased from the American Type Tissue Culture (Manassas, VA). The cDNAs for RGS4, RGS16, and the inactive RGS16 mutant RGS16/N131A were subcloned into pcDNA3 using the BamHI/XhoI sites (for RGS4) and the EcoRI/XhoI sites (for RGS16 and RGS16/N131A). The wild type and dominant-interfering mutants of the p85 regulatory subunit of the type I PI 3-kinase p85/WT and ⌬p85 were kindly provided by Dr. Wataru Ogawa (Kobe University, Kobe, Japan) and were subcloned into pcDNA3 using the BamHI/EcoRI sites. The GRP cDNA was generously provided by Dr. Michael Czech (University of Massachusetts). The PH domain was then subcloned carboxyl-terminal to EGFP generating the EGFP-GRP/PH fusion construct.
Cell Culture and Transient Transfection-3T3L1 preadipocytes were cultured in DMEM containing 25 mM glucose, 10% calf serum at 37°C in a 8% CO 2 atmosphere and induced to differentiate into an adipocyte phenotype as described previously (58). Confluent cultures were induced to differentiate by incubation of the cells with DMEM containing 25 mM glucose, 10% fetal bovine serum, 1 g/ml insulin, 1 mM dexamethasone, and 0.5 mM isobutyl-1-methylxanthine. After 4 days, the medium was changed to DMEM, 25 mM glucose, 10% fetal bovine serum, and 1 g/ml insulin for an additional 4 days. The medium was then changed to DMEM containing 25 mM glucose and 10% fetal bovine serum. Under these conditions, greater than 95% of the cell population morphologically differentiated into adipocytes. The adipocytes were maintained for an additional 4 -8 days prior to use. The differentiated 3T3L1 adipocytes were transfected by electroporation as described previously (59). Briefly, the adipocytes were put into suspension by mild trypsinization and electroporated with plasmids under low voltage conditions (160 V, 950 microfarad). The cells were then allowed to adhere to tissue culture dishes for 36 h, and the adipocytes were then serumstarved for 2 h prior to experiments. In some experiments, the electroporated adipocytes were seeded on cover slips.
Single Cell Microinjection and Plasma Membrane Sheet Assay-3T3L1 adipocytes were grown on 35-mm tissue culture dishes, and prior to microinjection, the medium was changed to Lebovitz's L-15 medium containing 0.1% bovine serum albumin. Differentiated 3T3L1 adipocytes were impaled using an Eppendorf model 5171 micromanipulator and injected with 2-4 mg/ml antibodies plus 4 mg/ml MBP-Ras in 100 mM KCl, 5 mM Na 2 PO 4 , pH 7.2, directly into the cell cytoplasm with an Eppendorf model 5246 transjector. Preparation of plasma membrane sheets from the microinjected 3T3L1 adipocytes was performed as described previously (59). Briefly, the cells were rinsed once in ice-cold phosphate-buffered saline (PBS) and incubated with 0.5 mg/ml poly-Dlysine (Sigma) for 1 min. Cells were then swollen in a hypotonic buffer (23 mM KCl, 10 mM HEPES, 2 mM MgCl 2 , 1 mM EGTA, pH 7.5) by three successive rinses. The swollen cells were sonicated for 3 s at power setting 5 with a Fisher Sonic Dismembrator model 550 fitted with a 5-mm microtip set 0.7 cm above the surface of the cell monolayer in 10 ml of sonication buffer (70 mM KCl, 30 mM HEPES, 6 mM MgCl 2 , 3 mM EGTA, 1 mM dithiothreitol, 0.1 mM phenylmethylsulfonyl fluoride, pH 7.5). The bound plasma membranes sheets were washed twice in sonication buffer and subjected to immunofluorescence confocal microscopy as described below.
Immunofluorescence and Confocal Microscope-The isolated plasma membrane sheets were fixed with sonication buffer containing 2% paraformaldehyde for 30 min, and the reaction was quenched by incubation with PBS containing 100 mM glycine for 20 min. After three rinses in PBS, the sheets were blocked overnight at 4°C in 5% donkey serum in PBS. The blocked sheets were incubated at room temperature for 1 h with a 1:100 dilution of the monoclonal GLUT4 antibody in combination with a 1:1000 dilution of polyclonal sheep MBP antiserum. The plasma membrane sheets were then washed three times with PBS and incubated for 1 h with a 1:100 dilution of fluorescien isothiocyanate-conjugated donkey anti-sheep IgG in combination with a 1:100 dilution of Cy5-conjugated donkey anti-mouse IgG. Following incubation with the secondary antibodies, the membrane sheets were rinsed three times in PBS and mounted for microscopic analysis using Vectashield mounting medium. Confocal images were obtained on a Bio-Rad MRC 600 laser confocal microscope (The University of Iowa Central Microscopy Research Facility).
Transferrin Receptor Internalization-Differentiated 3T3L1 adipocytes were electroporated and allowed to recover as described above. The cells were then placed in serum-free DMEM for 2 h and incubated with 10 g/ml of Texas Red-conjugated transferrin at 4°C for 30 min. The cells were then extensively washed in serum-free DMEM to remove all the unbound Texas Red-cojugated transferrin, and the cells were warmed to 37°C for 30 min. The cells were then fixed with 4% paraformaldehyde in PBS for 20 min and visualized by confocal fluorescent microscopy.
Intact Cell PIP 3 Determination-The in vivo production of PIP 3 was determined by a modification of the ARNO/PH domain GFP fusion protein assay (60). Briefly, 3T3L1 adipocytes were transfected with a cDNA encoding for the GRP/PH domain fused to EGFP (EGFP-GRP/ PH). The GRP/PH domain has high affinity for PIP 3 with substantially lower affinity for other phosphatidylinositol phosphates (60, 61). Because the insulin stimulation of PIP 3 formation occurs at the plasma membrane, the plasma membrane localization of the EGFP-GRP/PH fusion protein directly reflects the accumulation of PIP 3 .

Expression of the Constitutively Active G q ␣ Subunit Stimulates GLUT4 Translocation-
Recently several studies have shown that receptor agonists coupled to G i and G q /G 11 can result in GLUT4 translocation in both transfected fibroblasts and 3T3L1 adipocytes (50,55,56). To determine the potential involvement of the G q /G 11 ␣ subunit as an intermediate in a potential signal transduction cascade leading to GLUT4 translocation, we examined the effect of G q /WT and constitutively active G q /Q209L ␣ subunit expression (Fig. 1). In these experiments we also co-transfected the cells with the GLUT4-EGFP fusion protein to easily assess only the transfected cell population. As typically observed for the endogenous GLUT4 protein, in the basal state GLUT4-EGFP was localized to the perinuclear region and small vesicles scattered throughout the cytoplasm (data not shown). This distribution was identical to cells co-expressing GLUT4-EGFP and the G q /WT ␣ subunit (Fig. 1A,  panel a). As expected, insulin stimulation resulted in the translocation of the GLUT4-EGFP protein to the plasma membrane detected as a continuous rim of cell surface fluorescence in both GLUT4-EGFP transfected and G q /WT co-transfected cells (Fig.  1A, panel b). In contrast, cells expressing G q /Q209L resulted in the constitutive association of the GLUT4-EGFP protein with the plasma membrane in the absence of insulin (Fig. 1A, panel  c). The subsequent insulin stimulation had no further effect on the extent but did slightly increase the number of cells displaying GLUT4-EGFP cell surface rim fluorescence (Fig. 1A, panel  d). In control experiments, expression of wild type G s and G i2 (G s /WT, G i2 /WT) or constitutively active G s and G i2 (G s /Q227L, G i /Q205L) had no effect on either the basal or insulin-stimulated translocation of GLUT4-EGFP (Fig. 1A, panels e-l).
Quantitation of these data by scoring the transfected cells for GLUT4-EGFP plasma membrane rim fluorescence is presented in Fig. 1B. Together, these data suggest that expression of constitutively active of G␣ q is sufficient to induce GLUT4 translocation, which is not mimicked by the expression of constitutively active G s or G i2 .
Microinjection of the G q /G 11 ␣ Subunit Antibody Inhibits Insulin-stimulated GLUT4 Translocation-Although the stimulation of GLUT4-EGFP translocation by G q /Q209L appeared to be specific for this family of trimeric GTP-binding proteins, the overexpressed proteins can potentially impinge upon signaling pathways not necessarily utilized by the endogenous protein. Thus, to examine the role of the endogenous G q /G 11 ␣ subunit, we microinjected 3T3L1 adipocytes with several specific antibodies and examined the translocation of the endogenous GLUT4 proteins by single cell immunofluorescence in isolated plasma membrane sheets (Fig. 2). In these experiments, all the antibodies were prepared against the carboxylterminal region of the individual ␣ subunits, which inhibits the coupling between the trimeric GTP-binding protein with their respective receptors (62)(63)(64). To identify the specific 3T3L1 adipocytes that were microinjected following preparation of the plasma membrane sheets, the antibodies were also co-injected with the carboxyl-terminal domain of Ras fused to the maltosebinding protein (MBP-Ras) as a plasma membrane marker (58). Following co-microinjection with antibody and MBP-Ras, the isolated plasma membrane sheets demonstrated a specific MBP immunofluorescence signal from the microinjected cells compared with the surrounding nonmicroinjected cells. Microinjection of the G q /G 11 antibody inhibited the insulin-stimulated translocation of the endogenous GLUT4 protein com- pared with the surrounding nonmicroinjected cells (Fig. 2,  panels a and b). In this particular field, the G q /G 11 antibody prevented the translocation of GLUT4 in all three of the microinjected cells. In contrast, the G i1/2 or G s ␣ subunit-specific antibodies were essentially without effect (Fig. 2, panels c-f).
Scoring of approximately 60 plasma membrane sheets/condition from three independent experiments demonstrated that the G q /G 11 , G i1/2 and G s -specific ␣ subunit antibodies inhibited the insulin-stimulated GLUT4 translocation in 53.3 Ϯ 10.1, 7.8 Ϯ 1.7, and 7.3 Ϯ 2.5% of the microinjected 3T3L1 adipocytes, respectively.
Expression of RGS Proteins Inhibit Insulin-stimulated GLUT4 Translocation-In an alternative approach to inhibit G q function, we attenuated G q activation by expression of two RGS proteins and evaluated their effects upon insulin-stimulated GLUT4 translocation. RGS proteins are a recently identified family of intracellular GTPase-activating proteins that accelerate GTP hydrolysis, thus limiting the duration of trimeric G protein ␣ subunit activation (65)(66)(67)(68). Expression of the empty vector had no effect on the basal or insulin-stimulated distribution of GLUT4-EGFP (Fig. 3, panels a and b). Similarly, expression of RGS4 and RGS16, which are both relatively specific for G q and G i family members, had no effect on the basal distribution of the GLUT4-EGFP fusion protein (Fig. 3,  panels c and e). In contrast, these RGS proteins inhibited insulin-stimulated GLUT4-EGFP translocation (Fig. 3, panels d and f). However, RGS16 N131A , which does not have GTPaseactivating protein activity for the ␣ subunits, had no detectable effect on either the basal or insulin-stimulated distribution of GLUT4-EGFP (Fig. 3, panels g and h). Together, the effects of RGS proteins, specific GTP-binding protein ␣ subunit antibodies, and expression of constitutively active ␣ subunits strongly support a critical role for G q function in insulin-stimulated GLUT4 translocation.
G q Stimulates the Translocation of Other Insulin-regulated Trafficking Proteins without Affecting Plasma Membrane Endocytosis-In addition to GLUT4, insulin stimulation in adipocytes also results in the translocation of several constitutive recycling proteins including the GLUT1, cation-independent mannose-6-phosphate receptor (CI/MPR), and the transferrin receptor (TfR), albeit to lesser degree than GLUT4 (2-4). To examine whether G q can modulate general vesicle trafficking, we examined the plasma membrane translocation of the EGFP epitope-tagged CI/MPR (CI/MPR-EGFP). In the absence of in-sulin, CI/MPR-EGFP was primarily localized to the perinuclear region and in small vesicles scattered throughout the cytoplasm (Fig. 4, panel a). Insulin stimulation increased the amount of CI/MPR-EGFP at the cell surface, although to a lesser extent than that observed for GLUT4 (Fig. 4, panel b). Nevertheless, co-expression of RGS16 completely prevented the insulin-stimulated translocation of CI/MPR-EGFP (Fig. 4,  panel c). In addition, co-expression of constitutively active G q / Q209L induced a marked increase in the plasma membrane localized CI/MPR-EGFP protein (Fig. 4, panel d). Essentially identical results were obtained when we examined the subcellular localization of GLUT1-EGFP (data not shown).
Increased protein levels at the plasma membrane can arise either because of enhanced exocytosis and/or because of decreased endocytosis. To distinguish between these possibilities, we next determined the effect of G q /Q209L expression on TfR internalization (Fig. 5). Empty vector and G q /Q209L electroporated 3T3L1 adipocytes were cell surfaced labeled with Texas Red-conjugated transferrin at 4°C (Fig. 5A, panels a and c). To identify the cells that were transfected, the cells were co-transfected with GLUT4-EGFP. As is apparent, at 4°C the labeled TfR was predominantly confined to the cell surface membrane. However, following warming to 37°C for 30 min, the labeled TfR decreased at the cell surface and redistributed to the cell interior (Fig. 5A, panels b and d). However, in the presence of G q /Q209L, the GLUT4-EGFP protein was persistently plasma membrane localized (Fig. 5A, panel d). Quantitation of these data by counting the number of cells displaying either cell surface or internalized TfR is presented in Fig. 5B. Together, these data demonstrate that G q functions in the insulin stimulation of general vesicle trafficking by enhancing the rate of vesicle exocytosis without any apparent effect on plasma membrane endocytosis.
G q /G 11 Does Not Function Upstream of the PI 3-Kinase-It has been well established that PI 3-kinase plays a necessary role in insulin-stimulated GLUT4 translocation to the plasma membrane (8 -14). Therefore, to examine whether G q /G 11 could be functioning upstream and thereby activating the PI 3-kinase, we took advantage of the recently established property of the GRP PH domain to bind PIP 3 with high affinity (60,61). In these experiments, we co-expressed the constitutively active G q /Q209L mutant protein with the PH domain of the GRP protein tagged with EGFP (EGFP-GRP/PH). As observed in Fig. 6, The EGFP-GRP/PH fusion protein was strongly local- FIG. 2. Microinjection of a G q /G 11 -specific antibody inhibits insulin-stimulated GLUT4 translocation. Differentiated 3T3L1 adipocytes were microinjected with approximately 0.1 pl of MBP-Ras (6 mg/ml) mixed 1:1 with 4 mg/ml of a G q /G 11 antibody (panels a and b), 4 mg/ml of a G i1/2 antibody (panels c and d), or 4 mg/ml of a G s antibody (panels e and f). Following microinjection, the cells were incubated with 100 nM insulin for 30 min at 37°C. A, plasma membrane sheets were prepared and subjected to confocal immunofluorescence microscopy with an MBP-specific antibody (panels a, c, and e) or a GLUT4-specific antibody (panels b, d, and f). These are representative fields from three to four independent experiments in which a total of 155 plasma membrane sheets were analyzed for GLUT4. The plasma membrane sheets derived from the microinjected cells are indicated by the arrows.
ized to the nucleus as well as distributed throughout the cell cytoplasm (Fig. 6, panel a). Because expressed EGFP was also concentrated in the nucleus of 3T3L1 adipocytes, this phenomenon is a property of EGFP itself and was not due to the presence of the GRP/PH domain. In any case, insulin stimulation induced in a rapid translocation of the cytosolic GRP/PH-EGFP to the plasma membrane because of the formation of PIP 3 (Fig. 6, panel b). Co-expression of G q /Q209L had no significant effect on either the basal state distribution of the GRP/PH-EGFP fusion protein or on the insulin-stimulated plasma membrane rim fluorescence (Fig. 6, panels c and d). To extend these findings, we next determined whether RGS protein expression affected the insulin-stimulated formation PIP 3 in intact 3T3L1 adipocytes. Similar to G q /Q209L, expression of RGS16 had no effect on either the basal state distribution of the GRP/PH-EGFP fusion protein or the insulin-induced plasma membrane rim fluorescence (Fig. 6, panels e and f). Together, these data indicate that G q /G 11 does not function in an upstream pathway, resulting in the activation of the PI 3-kinase and formation of PIP 3 in the plasma membrane.
G q /G 11 Is Not a Downstream Target for the PI 3-Kinase-Because G q /G 11 does not appear to function as an upstream activator of the PI 3-kinase, it was also possible that G q /G 11 may be a downstream target in this pathway. To address this issue, we initially examined the effect of a dominant-interfering mutant of the PI 3-kinase p85 regulatory subunit (⌬p85) on GLUT4 translocation (Fig. 7). As previously observed, insulin stimulation and expression of G q /Q209L resulted in the translocation of GLUT4-EGFP from intracellular storage sites to the plasma membrane (Fig. 7, panels a, d, and g). Co-expression of the wild type p85 PI 3-kinase regulatory subunit had no effect on the basal subcellular distribution of GLUT4-EGFP (Fig. 7,  panel b). As expected, both insulin-and G q /Q209L-stimulated GLUT4-EGFP translocation was also not altered by expression of the p85 wild type protein (Fig. 7, panels e and h). Even though expression of the dominant-interfering ⌬p85 mutant  (panels a and b) or G q /Q209L (panels c and d). The cells were allowed to recover for 36 h and subsequently incubated with Texas Red-conjugated transferrin for 30 min at 4°C. The cells were immediately fixed (panels a and c) or incubated for 30 min at 37°C prior to fixing (panels b and d). A, the cells were then visualized by confocal fluorescence microscopy. B, the number of cells displaying cell surface TfR were scored and quantified from three independent experiments. appeared to somewhat concentrate the GLUT4-EGFP protein in the perinuclear region of the cell, there was a complete inhibition of insulin-stimulated GLUT4-EGFP translocation (Fig. 7, panels c and f). In contrast, expression of ⌬p85 had no significant effect on GLUT4-EGFP translocation induced by expression of G q /Q209L (Fig. 7, panel i).
To further assess the relationship between G q /G 11 function and PI 3-kinase activity, we next utilized the selective PI 3-kinase inhibitor wortmannin (Fig. 8A). In this experiment, the GLUT4-EGFP transfected cells were stimulated with insulin for 30 min followed by addition of wortmannin for various times in the continuous presence of insulin. Because previous studies have established that the initial events in GLUT4 internalization are independent of the PI 3-kinase (69 -71), the loss of plasma membrane localized GLUT4-EGFP in the continuous presence of insulin demonstrates that wortmannin effectively inhibits insulin-stimulated exocytosis. Consistent with the expression of the dominant-interfering ⌬p85 mutant, wortmannin treatment did not result in a loss of plasma membrane localized GLUT4-EGFP in cells persistently activated by expression of G q /Q209L. Together, these data provide compelling evidence that the G q /G 11 plays an important function in the insulin-stimulated translocation of GLUT4 by a pathway independent of the PI 3-kinase.
G q Stimulation of GLUT4 Translocation Is Independent of Phospholipase C-In multiple cell types and in in vitro reconstitution experiments, the activated G q ␣ subunit can directly stimulate phospholipase C␤ (PLC␤) which generates inositol 1,4,5-triphosphate and diacylglycerol as important downstream second messengers (72)(73)(74)(75)(76). Therefore, to assess the possible involvement of PLC on G q /Q209L-stimulated GLUT4 translocation, we examined the effect of PLC inhibitor U73221 on GLUT4-EGFP exocytosis. Because the G q /Q209L-induced GLUT4-EGFP translocation required several hours following transfection, we examined the time course of GLUT4-EGFP internalization in a manner analogous to that of wortmannin (Fig. 8B). In this case, treatment with U73221 for up to 3 h had no significant effect on the amount of plasma membrane-associated GLUT4-EGFP in cells continuously stimulated with either insulin or G q /Q209L. These data indicate that the G q / Q209L signaling pathway leading to GLUT4 translocation in 3T3L1 adipocytes is independent of PLC activation.
G q Stimulation of GLUT4 Translocation Is Tyrosine Kinasedependent-Recently, we and others have reported that GLUT4 translocation induced by several insulinomimetic agents, osmotic shock, ceramide, exercise/contraction, and GTP␥S are independent of PI 3-kinase activation (32,(37)(38)(39)(40)(41)(42)(43)77). In particular, both the osmotic shock and GTP␥S-stimulated GLUT4 translocation require the activation of a tyrosine kinase pathway distinct from the insulin receptor and IRS protein tyrosine phosphorylation (32,37). Therefore to determine whether G q /Q209L-stimulated GLUT4 translocation was tyrosine kinase-dependent, we examined the effects of the tyrosine kinase inhibitors, herbimycin A, and genistein (Fig. 9). As expected, following 30 min of insulin stimulation the addition of genistein resulted in a time-dependent decrease in the amount of the GLUT4-EGFP localized to the plasma membrane (Fig. 9A). In parallel, genistein treatment also resulted in a parallel decrease in plasma membrane associated GLUT4-EGFP from G q /Q209L-stimulated cells (Fig. 9A). Essentially identical results were also obtained when the tyrosine kinase inhibitor herbimycin A was used (Fig. 9B). Thus, the downstream pathway utilized by G q /Q209L requires a tyrosine kinase signal in a manner similar to that reported for both osmotic shock and GTP␥S-stimulated GLUT4 translocation.
Changes in G␣ q versus G␣ 11 Protein Levels during 3T3L1 Adipocyte Differentiation-Because both G␣ q and G␣ 11 appear to have similar signaling specificity and function, we also examined the relative expression of the ␣ subunits during 3T3L1 adipocyte differentiation (Fig. 10). Immunoblotting of whole cell extracts indicated that the expression of G␣ q slightly decreased following 4 days of adipocyte differentiation (Fig. 10,  top panel, lanes 1-4). In contrast, the expression of G␣ 11 appeared to slightly increase following 12 days of adipocyte differentiation compared with the undifferentiated cells (Fig. 10,  middle panel, lanes 1-4). In contrast, GLUT4 protein expression was not apparent until 8 days post adipocyte differentiation and was markedly increased following 12 days (Fig. 10,   FIG. 6. G q is not involved in the insulin stimulation of PIP 3 formation in intact cells. Differentiated 3T3L1 adipocytes were electroporated with 40 g of the EGFP-GRP/PH cDNA and 200 g of the pcDNA3 empty vector (panels a and b), G q /Q209L (panels c and d), or RGS16 (panels e and f). The cells were allowed to recover for 36 h and subsequently incubated in the absence (panels a, c, and e) or presence (panels b, d, and f) of 100 nM insulin for 30 min at 37°C. The cells were then fixed and subjected to confocal fluorescence microscopy, and the represented images were obtained. These are representative fields from three to four independent experiments. FIG. 7. G q stimulated GLUT4 translocation is independent of PI 3-kinase function. Differentiated 3T3L1 adipocytes were electroporated with 25 g of the GLUT4-EGFP cDNA, 100 g of G q /Q209L (panels g, h, and i), and 200 g of the pcDNA3 empty vector (panels a, d, and g), p85/WT (panels b, e, and h), or the dominant-interfering ⌬p85 mutant (panels c, f, and i). The cells were allowed to recover for 36 h and subsequently incubated in the absence (panels a-c) or presence (panels d-f) of 100 nM insulin for 30 min at 37°C. The cells were then fixed and subjected to confocal fluorescence microscopy, and the represented images were obtained. These are representative fields from three to four independent experiments.
bottom panel, lanes 1-4). In any case, it should be noted that the affinities of the G␣ q and G␣ 11 antibodies are not known, and therefore the absolute amounts of G␣ q and G␣ 11 cannot be compared. Thus, these data only indicate relative changes in expression within a given isoform. DISCUSSION Insulin stimulation of the PI 3-kinase is necessary for the plasma membrane translocation of GLUT4 in both skeletal muscle and adipocytes (8 -14). However, recent studies have also demonstrated that insulin must generate additional signals that function in conjunction with a PI 3-kinase-dependent pathway to stimulate GLUT4 translocation. For example, activation of PI 3-kinase by interleukin-4 or through engagement of the integrin receptors is not sufficient to induce GLUT4 translocation (78,79). Furthermore, addition of a cell permeable analog of PIP 3 was unable to induce GLUT4 translocation alone, whereas treatment of cells with wortmannin, insulin, and the PIP 3 analog resulted in enhanced glucose uptake and GLUT4 translocation (80). These data provide compelling evidence that although the PI 3-kinase pathway is necessary, there is at least one additional signaling pathway that is independent of PI 3-kinase activation.
In this regard, several lines of evidence have suggested the involvement of GTP-binding protein(s) in regulating GLUT4 trafficking. As previously indicated, adrenergic stimulation, GTP␥S and AlF 4 Ϫ treatments can induce GLUT4 translocation (35-37, 44, 45). More importantly, expression and activation of G protein-coupled receptors that signal through both G q and G i FIG. 8. G q stimulated GLUT4 translocation is independent of PI 3-kinase and PLC activation. Differentiated 3T3L1 adipocytes were electroporated with 40 g of the GLUT4-EGFP cDNA and 200 g of either the pcDNA3 empty vector (squares) or G q /Q209L (circles). The cells were allowed to recover for 36 h and subsequently incubated in the absence (circles) or in the presence (squares) of 100 nM insulin at 37°C for 30 min. A, following this 30-min preinsulin stimulation, both the continuous G q /Q209L and insulin-stimulated cells were incubated with 100 nM wortmannin for the times indicated. B, following the 30-min preinsulin stimulation, both the continuous G q /Q209L and insulin-stimulated cells were incubated with 50 M U73221 for the times indicated. The number of cells displaying plasma membrane GLUT4-EGFP translocation were quantitated by confocal fluorescence microscopy. These values represent the quantitation of greater than 200 cells from three to four independent experiments.
FIG. 9. G q stimulated GLUT4 translocation is tyrosine kinase-dependent. Differentiated 3T3L1 adipocytes were electroporated with 40 g of the GLUT4-EGFP cDNA and 200 g of either the pcDNA3 empty vector (squares) or G q /Q209L (circles). The cells were allowed to recover for 36 h and subsequently incubated in the absence (circles) or in the presence (squares) of 100 nM insulin at 37°C for 30 min. A, following this 30-min preinsulin stimulation, both the continuous G q /Q209L and insulin-stimulated cells were incubated with 200 M genistein for the times indicated. B, following the 30 min preinsulin stimulation, both the continuous G q /Q209L and insulin-stimulated cells were incubated with 1 g/ml herbimycin A for the times indicated. The number of cells displaying plasma membrane GLUT4-EGFP translocation were quantitated by confocal fluorescence microscopy. These values represent the quantitation of greater than 200 cells from three to four independent experiments. are also effective stimulators of GLUT4 translocation (50,55,56). Furthermore, the adrenergic stimulation of GLUT4 translocation is not inhibited by pertussis toxin, strongly suggesting the direct involvement of G q . However, it is important to recognize that G q and G 11 have overlapping activities and that adipocytes express both these trimeric GTP-binding proteins. Because the relative amount of G␣ 11 was found to increase during adipocyte differentiation, G 11 may play a more significant role than G q in mediating the insulin stimulation of GLUT4 translocation. However, the relative amounts of G␣ q and G␣ 11 were not determined, and therefore it is not currently possible to functionally distinguish between G q and/or G 11 in mediating GLUT4 translocation in 3T3L1 adipocytes.
In any case, we have directly examined the role of G q /G 11 ␣ subunits in relationship with insulin-stimulated GLUT4 translocation in 3T3L1 adipocytes. Our data demonstrate that expression of constitutively active G q but not G i or G s ␣ subunits results in GLUT4 translocation. In addition, inhibition of G q / G 11 activation by expression of RGS proteins and by microinjection of G q /G 11 antibodies inhibited insulin-stimulated GLUT4 translocation. The ability of G q to induce GLUT4 translocation apparently resulted from an increase in the rate of exocytosis because there was no discernable change in the plasma membrane endocytosis of the constitutively recycling transferrin receptor. Furthermore, we have observed that this G q /G 11 signaling pathway leading to GLUT4 translocation is distinct from the PI 3-kinase dependent pathway. This is based upon the ability of constitutively active G q (G q /Q209L) to induce GLUT4 translocation in the presence of wortmannin and when co-expressed with a dominant-interfering p85 (⌬p85) mutant. In addition, expression of G q /Q209L did not stimulate PIP 3 formation. Similarly, inhibition of G q /G 11 function had no effect on insulin-stimulated PIP 3 generation. Inhibition of G q function through the expression of RGS proteins also did not affect the insulin stimulation of plasma membrane PIP 3 formation in intact cells. Thus, the simplest interpretation of these data is that G q functions downstream of the PI 3-kinase in mediating the insulin stimulation of GLUT4 translocation.
Alternatively, these characteristics are also reminiscent of the insulinomimetic actions of osmotic shock, exercise/contraction, ceramide, and, importantly, GTP␥S-stimulated GLUT4 translocation, all of which occur independent of PI 3-kinase activation. Although the induction of GLUT4 translocation by these alternative pathways does not involve the insulin receptor or IRS protein tyrosine phosphorylation, these activation events are tyrosine kinase-dependent. Similarly, the G q -dependent translocation of GLUT4 was inhibited by two structurally distinct tyrosine kinase inhibitors. Thus, it remains formally possible that this alternative pathway is the site of G q action.
In this regard, a recent study by Imamura et al. (81) has also observed that insulin stimulation of GLUT4 translocation occurs through a G q /G 11 -specific pathway. Although these results are in general agreement with our findings, these data also indicate that G q functions as an upstream activator of PI 3-kinase signaling, whereas our findings are more consistent with either a downstream or parallel signaling pathway. At present the basis for this difference is not apparent, and further studies will be necessary to distinguish between these signaling events.
In any case, it is also important to recognize that G q /G 11 ␣ subunits have been well documented to activate PLC␤ and thereby induce the formation of inositol 1,4,5-triphosphate and subsequent release of intracellular calcium (82). However, previous studies have documented that insulin does not stimulate any changes in intracellular calcium, which we have also con-firmed by fura2 measurements (Refs. 83 and 84 and data not shown). Moreover, inhibition of PLC activation did not prevent the G q /G 11 -dependent translocation of GLUT4. Thus, to reconcile this apparent discordance, insulin must activate either a subset of G q /G 11 ␣ subunits that are specifically coupled to GLUT4 translocation and not PLC␤ or alternatively that G q / G 11 provides a required permissive function for GLUT4 translocation. In any case, the data presented in this manuscript strongly implicate a necessary role for G q /G 11 ␣ subunit function in mediating the insulin-stimulated translocation of GLUT4 in 3T3L1 adipocytes. Future studies will be necessary to determine the downstream signal events mediated by G q necessary for the insulin-stimulated translocation of GLUT4 to the plasma membrane.