Guanosine 5′-O-(3-Thiotriphosphate) (GTPγS) Stimulation of GLUT4 Translocation is Tyrosine Kinase-dependent*

Guanosine 5′-O-(3-thiotriphosphate) (GTPγS) treatment of permeabilized adipocytes results in GLUT4 translocation similar to that elicited by insulin treatment. However, although the selective phosphatidylinositol 3-kinase inhibitor, wortmannin, completely prevented insulin-stimulated GLUT4 translocation, it was without effect on GTPγS-stimulated GLUT4 translocation. In addition, insulin was an effective stimulant, whereas GTPγS was a very weak activator of the downstream Akt serine/threonine kinase. Consistent with an Akt-independent mechanism, guanosine 5′-O-2-(thio)diphosphate inhibited insulin-stimulated GLUT4 translocation without any effect on the Akt kinase. Surprisingly, two functionally distinct tyrosine kinase inhibitors, genistein and herbimycin A, as well as microinjection of a monoclonal phosphotyrosine specific antibody, inhibited both GTPγS- and insulin-stimulated GLUT4 translocation. Phosphotyrosine immunoblotting and specific immunoprecipitation demonstrated that GTPγS did not elicit tyrosine phosphorylation of insulin receptor or insulin receptor substrate-1. In contrast to insulin, proteins in the 120–130-kDa and 55–75-kDa range were tyrosine-phosphorylated following GTPγS stimulation. Several of these proteins were identified and include protein-tyrosine kinase 2 (also known as CAKβ, RAFTK, and CADTK), pp125 focal adhesion tyrosine kinase, pp130 Crk-associated substrate, paxillin, and Cbl. These data demonstrate that the GTPγS-stimulated GLUT4 translocation utilizes a novel tyrosine kinase pathway that is independent of both the phosphatidylinositol 3-kinase and the Akt kinase.

It has been well established that the insulin stimulation of glucose uptake primarily results from the translocation of the GLUT4 1 glucose transporter isoform from intracellular storage sites to the cell surface membrane in muscle and adipose tissues. Although the molecular mechanism and signaling cascade(s) regulating the intracellular trafficking of GLUT4-containing vesicles have not been completely elucidated, several important effector molecules have recently been identified. Insulin binding to the insulin receptor results in tyrosine autophosphorylation of the ␤-subunit and activation of its intrinsic tyrosine kinase (1). Subsequently, the insulin receptor tyrosine kinase phosphorylates several intracellular proteins on tyrosine residues, most notably insulin receptor substrate 1 (IRS1) (1). The phosphorylation of these substrates creates recognition sites for additional effector proteins containing Src homology 2 (SH2) domains, thereby generating multisubunit signaling complexes (1,2). In particular, the tyrosine phosphorylation of IRS1 induces the association and activation of phosphatidylinositol (PI) 3-kinase (3). The targeting and/or activation of PI 3-kinase is well documented to be necessary for GLUT4 translocation (4,5). Recently, it has been shown that insulin stimulates the Akt kinase, also known as protein kinase B or RAC-PK (6). Insulin-stimulated Akt kinase activity is dependent on PI 3-kinase activation, and expression of a membrane-targeted and constitutively active Akt kinase results in persistent GLUT4 translocation (7,8). Consistent with a role for GLUT4 trafficking, expression of a dominant interfering Akt mutant inhibited insulin-stimulated GLUT4 translocation (9).
In addition to insulin, various other stimuli display insulinomimetic properties and can induce the translocation of GLUT4-containing vesicles to the plasma membrane. For example, introduction of guanosine 5Ј-O-(3-thiotriphosphate) (GTP␥S), a nonhydrolyzable GTP analogue, into adipocytes rapidly stimulates GLUT4 translocation to a similar extent as insulin (10,11). In addition, GTP␥S can stimulate GLUT4 translocation in the absence of ATP, suggesting that ATP is required at an early step(s) in the insulin-signaling pathway and that a GTP-binding protein(s) functions at a more distal step(s) (11). The stimulatory effect of GTP␥S can also be mimicked by treatment with AlF 4 Ϫ , which is characteristic of the involvement of a heterotrimeric GTP binding protein (12,13). Consistent with this interpretation, adrenergic stimulation can induce GLUT4 translocation and glucose uptake in both cardiac myocytes (14,15) and brown adipocytes (16,17). Furthermore, in transfected Chinese hamster ovary cells and 3T3L1 adipocytes, activation of receptors coupled to G q also stimulate GLUT4 translocation (18). In this regard, skeletal muscle appears to have two distinct pathways mediating GLUT4 translocation. Similar to adipocytes, insulin-stimulated glucose transport in muscle is PI 3-kinase dependent (19 -21). In contrast, muscle contraction/exercise or hypoxia stimulates glucose transport through a PI 3-kinase-independent pathway, which may utilize a distinct and separate pool of GLUT4 intracellular vesicles (22). It has been suggested that this alternative pathway leading to GLUT4 translocation may be mediated by an increase in cytoplasmic calcium levels (23)(24)(25), consistent with the known functional role of G proteins in stimulating increases in intracellular calcium (26 -30).
In the present study, we have examined the mechanism by which GTP␥S stimulates GLUT4 translocation by comparing its signaling properties with those of insulin. Our data demonstrate that GTP␥S-stimulated GLUT4 translocation is independent of the PI 3-kinase, the Akt kinase, and changes in intracellular calcium ion concentration. However, GTP␥S-stimulated GLUT4 translocation occurred through the activation of a novel tyrosine kinase pathway, which does not involve the insulin receptor or IRS1 but may require the tyrosine phosphorylation of protein-tyrosine kinase 2 (PYK2, also known as CAK␤, RAFTK, and CADTK), pp125 focal adhesion tyrosine kinase (pp125 FAK ), pp130 Crk-associated substrate (pp130 Cas ), paxillin, and Cbl.

EXPERIMENTAL PROCEDURES
Cell Culture and Differentiation of 3T3L1 Adipocytes-Murine 3T3L1 preadipocytes were obtained from the American Type Tissue Culture repository and were cultured in Dulbecco's modified Eagle's medium (DMEM) containing 25 mM glucose and 10% calf serum at 37°C. Confluent cultures were induced to differentiate into adipocytes 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 containing 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, more than 95% of the cell population morphologically differentiated into adipocytes. All studies were performed on adipocytes between 8 and 12 days after initiation of differentiation (day 0). Prior to all experimental treatments, the differentiated adipocytes were serumstarved in DMEM containing 25 mM glucose and 0.1% bovine serum albumin for 2 h at 37°C.
Streptolysin-O Permeabilization of 3T3L1 Adipocytes-3T3L1 adipocytes were permeabilized with streptolysin-O (SL-O) as described by Robinson et al. (11) with minor modifications. Briefly, 3T3L1 adipocytes were washed three times with intracellular (IC) buffer (140 mM potassium glutamate, 20 mM Hepes, pH 7.15, 7.5 mM MgCl 2 , 5 mM EGTA, 5 mM NaCl, 2 mM CaCl 2 ) and incubated in IC buffer containing 0.8 IU/ml of SL-O (Murex Diagnostics Inc., Atlanta, GA) for 5 min at 37°C. Under these conditions, more than 95% of the cells were permeabilized based upon incorporation of propidium iodide. Following SL-O permeabilization, the cells were washed two times with ICR buffer (IC buffer containing 1 mg/ml bovine serum albumin, 1 mM dithiothreitol and enriched with either 10 mM MgATP or an ATP-regenerating system: 40 IU/ml creatine phosphokinase, 5 mM creatine phosphate, and 1 mM ATP). Unless otherwise indicated, all experimental treatments were performed by incubating the cells in ICR buffer containing various additions for 15 min at 37°C.
Single Cell Microinjection-3T3L1 adipocytes used for microinjection were grown on 60-mm tissue culture dishes. The cells were incubated in Krebs-Ringer bicarbonate Hepes buffer (pH 7.4), containing 2 mM pyruvate, 0.5% bovine serum albumin, and 2.5 mM glucose for 45 min prior to microinjection. Adipocytes were microinjected with antibodies over a 45-min period using an Eppendorf model 5171 micromanipulator and given injections of approximately 0.1 pl directly into the cell cytoplasm with an Eppendorf model 5246 transjector. Following microinjection of approximately 100 -300 cells/dish, the buffer was changed to fresh DMEM containing 0.1% bovine serum albumin, and the cells were allowed to recover for 90 min at 37°C, prior to permeabilization and treatment.
Plasma Membrane Sheet Assay-Preparation of plasma membrane sheets from the adipocytes was performed essentially by the method of Robinson et al. (11). Briefly, cells cultured on 35-or 60-mm dishes, following the appropriate treatment as described in each figure legend, were rinsed once in ice-cold phosphate-buffered saline (PBS) and incubated with 0.5 mg/ml of poly-L-lysine in PBS for 30 s. The cells were then swollen in a hypotonic buffer (23 mM KCl, 10 mM Hepes, pH 7.5, 2 mM MgCl 2 , 1 mM EGTA) by three successive rinses. The swollen cells were sonicated for 3 s at power setting 4.5 with a model 550 Fisher sonic dismembrator fitted with a 5-mm microtip set 1 cm above the surface of the cell monolayer in 10 ml of sonication buffer (70 mM KCl, 30 mM Hepes, pH 7.5, 5 mM MgCl 2 , 3 mM EGTA, 1 mM dithiothreitol, 0.1 mM phenylmethylsulfonyl fluoride). The bound plasma membrane sheets were washed two times with sonication buffer and used for either indirect immunofluorescence or immunoblot analysis as described below.
Preparation of Total Cell Extracts and Immunoprecipitation-Total cell extracts were prepared from 60-or 150-mm plates of 3T3L1 adipocytes following the appropriate treatment as described in each figure legend. Cells from each plate were washed two times with ice-cold PBS and scraped into 1 ml of lysis buffer (50 mM HEPES, pH 7.8, 1% Triton X-100, 100 mM NaF, 10 mM Na 3 P 2 O 7 , 2.5 mM EDTA) containing 1.0 mM phenylmethylsulfonyl fluoride, 2 mM Na 3 VO 4 , 1 mg/ml aprotinin, 10 mM leupeptin, and 1 mM pepstatin A by rotation for 15 min at 4°C. Insoluble material was separated from the soluble extract by microcentrifugation for 15 min at 4°C. Protein concentration was determined, and samples were either subjected directly to SDS-polyacrylamide electrophoresis (as described below) or immunoprecipitated for IRS1 or PYK2. Briefly, 3-5 mg of cellular protein were immunoprecipitated with 5 g of IRS1 polyclonal antibody (␣IRS1; Santa Cruz Biotechnology, Inc., Santa Cruz, CA) or PYK2 polyclonal antibody (31) for 2 h at 4°C. Immune complexes were recovered by the addition of protein A-Sepharose (Amersham Pharmacia Biotech) and subjected to SDS-polyacrylamide electrophoresis (as described below).
Precipitation with GST Fusion Proteins-GST and GST-CrkII were prepared as described previously (32). Soluble GST fusion proteins were covalently linked to agarose beads with Amino-Link (Pierce). 100 g of fusion proteins bound to the beads were incubated for 2 h at 4°C with 4 mg of cell extracts isolated from control, insulin-stimulated, or GTP␥S-stimulated cells. The beads were subsequently pelleted, washed three times with washing buffer (10 mM Tris, pH 7.4, 2 mM EDTA, 150 mM NaCl, 0.2% Triton X-100, 0.1% Nonidet P-40, 0.2 mM phenylmethylsulfonyl fluoride, and 0.2 mM sodium vanadate), and boiled in Laemmli sample buffer. The precipitated proteins were analyzed by SDS-polyacrylamide gel electrophoresis and Western blotting (as described below).
Electrophoresis and Immunoblotting-Plasma membrane sheets were scraped into a buffer containing 250 mM sucrose, 20 mM Tris, pH 7.40, 1 mM EDTA and pelleted by centrifugation at 200,000 ϫ g for 1 h at 4°C. The pelleted fraction was solubilized in 100 l of Laemmli sample buffer, and 5 g of the total solubilized fraction was separated by SDS-polyacrylamide gel electrophoresis (10% polyacrylamide). The resolved proteins were then transferred to nitrocellulose membranes and immunoblotted with polyclonal rabbit GLUT4 antibody (IRGT; Charles River).
Whole cell lysates, GST precipitates, and immunoprecipitates were separated by SDS-polyacrylamide gel electrophoresis (7.5% polyacrylamide). The resolved proteins were then transferred to Immobilon P membrane (Millipore Corp.) and immunoblotted with a monoclonal phosphotyrosine antibody (PY20:HRPO; Transduction Laboratories), a monoclonal paxillin antibody (Transduction Laboratories), a monoclonal p130 Cas antibody (Transduction Laboratories), and a polyclonal Cbl antibody (Santa Cruz Biotechnology). All immunoblots were subjected to enhanced chemiluminescence detection (Amersham Pharmacia Biotech).
Immunofluorescence and Confocal Microscopy-Sonicated plasma membrane sheets were fixed for 20 min at 4°C in a solution containing 2% paraformaldehyde, 70 mM KCl, 30 mM HEPES, pH 7.5, 5 mM MgCl 2 , and 3 mM EGTA. The fixed plasma membrane sheets were quenched for 15 min at 25°C in 100 mM glycine-PBS (pH 7.5). After three rinses in PBS, the sheets were blocked overnight at 4°C in PBS containing 5% donkey serum (Sigma). The blocked sheets were incubated at 4°C overnight with a 1:100 dilution of polyclonal rabbit GLUT4 antibody (IRGT; Charles River). For the microinjection studies, this incubation was done in combination with a 1:5000 dilution of polyclonal sheep anti-maltose binding protein antiserum (generously provided by Dr. Morris Birnbaum, University of Pennsylvania). The plasma membrane sheets were then washed for 30 min with PBS (six changes of PBS) and incubated overnight at 4°C with a 1:50 dilution of lissamine-rhodamine-conjugated donkey anti-rabbit IgG (Jackson Immunoresearch, Inc.). For the microinjection studies, this incubation was done in combination with a 1:100 dilution of fluorescein isothiocyanate-conjugate donkey anti-sheep IgG (Jackson Immunoresearch Inc.). Following incubation with the secondary antibodies, the membrane sheets were washed for 30 min with PBS (six changes of PBS) and mounted for microscopic analysis with Vectashield mounting medium (Vector Laboratories Inc.). Confocal images were obtained on a Bio-Rad MRC 600 laser confocal microscope (University of Iowa Central Microscopy Research Facility).
Akt Kinase Activity Assay-Akt protein kinase activity was determined as described by Moule et al. (33). 3T3L1 adipocyte detergent cell extracts were prepared as described above, and 600 g of cellular protein were immunoprecipitated with 5 g of Akt antiserum (kindly provided by Dr. Kelly Moule, University of Bristol) for 2 h at 4°C. Immune complexes were recovered by the addition of protein A-Sepharose (Amersham Pharmacia Biotech). The protein A-Sepharose beads were washed and resuspended in 40 l of assay buffer (20 mM MOPS, pH 7.0, 1 mM EDTA, 1 mM EGTA, 0.01% Brij 35, 5% glycerol) containing 0.1% 2-mercaptoethanol, 2.5 M cAMP-dependent protein kinase inhibitor peptide. The activity of Akt in the immunoprecipitate was measured using histone H2B (0.5 mg/ml) as an in vitro substrate. The reaction was initiated by the addition of 100 M [␥-32 P]ATP (10 Ci) for 20 min at 30°C. The reaction was stopped by the addition of 2ϫ Laemmli sample buffer (25 l), after which the samples were boiled for 5 min at 100°C and electrophoresed on a 16% SDS-polyacrylamide gel. The gel was stained with Coomassie Blue, dried, and subjected to autoradiography.

GTP␥-S, but Not Insulin, Stimulation of GLUT4 Translocation in 3T3L1
Adipocytes Is Wortmannin-insensitive-As previously observed (10,11), treatment of SL-O permeabilized 3T3L1 adipocytes with insulin resulted in the translocation of GLUT4 to the plasma membrane as detected by GLUT4 immunoblotting of isolated plasma membrane sheets (Fig. 1A, lanes  1 and 2). Similarly, the addition of GTP␥S to the permeabilized cells also induced the translocation of GLUT4 (Fig. 1A, lanes 3  and 4). Since it is well established that activation and/or ap-propriate intracellular targeting of the PI 3-kinase is necessary for insulin-stimulated GLUT4 translocation, we next examined the role of PI 3-kinase in GTP␥S-stimulated GLUT4 translocation. In the control, unstimulated cells, there was a low level of GLUT4 immunofluorescence detected from the isolated plasma membrane sheets (Fig. 1B, panel 1). Pretreatment with the selective PI 3-kinase inhibitor wortmannin (100 nM) slightly reduced the amount of GLUT4 present in the isolated plasma membrane sheets from unstimulated cells (Fig. 1B, panel 4). As expected, insulin stimulated a large increase in the translocation of GLUT4 to the plasma membrane, which was completely inhibited by pretreatment with wortmannin (Fig. 1B, panels 2 and 5). In contrast, although GTP␥S stimulated a similar extent of GLUT4 translocation compared with insulin, pretreatment with wortmannin was without effect (Fig. 1B, panels 3  and 6). In addition, pretreatment of the 3T3L1 adipocytes with higher concentrations of wortmannin (1 M) also inhibited insulin-stimulated GLUT4 translocation but did not reduce the translocation induced by GTP␥S (data not shown).
GTP␥S Is an Ineffective Activator of the Akt Protein Kinase-Recently, it has been reported that insulin stimulation increases Akt protein kinase activity in a PI 3-kinase-dependent manner, and stable overexpression of a constitutively active membrane-bound form of Akt kinase resulted in the persistent translocation of GLUT4 (7,8). We therefore examined the ability of GTP␥S to stimulate Akt kinase as a potential common point of convergence between the insulin and GTP␥S signaling pathways leading to GLUT4 translocation. Activation of Akt protein kinase activity was assessed by immunoprecipitation and analysis of in vitro protein kinase activity (Fig. 2). Isolation of Akt from unstimulated cells demonstrated a basal level of Akt protein kinase activity that was reduced by pretreatment with wortmannin (Fig. 2, lanes 1 and 2). Insulin stimulation resulted in a marked activation of Akt protein kinase activity, which was also attenuated by wortmannin (Fig. 2, lanes 5 and  6). Treatment with GTP␥S resulted in an intermediate stimulation of Akt protein kinase activity (Fig. 2, lane 3). Nevertheless, pretreatment with wortmannin completely inhibited the GTP␥S-stimulated activation of Akt protein kinase activity (Fig. 2, lane 4). Since Akt kinase activation is accompanied by dual phosphorylation on serine and threonine residues (34), we also examined the effect of insulin and GTP␥S on the SDSpolyacrylamide gel electrophoretic mobility of Akt. Consistent with Akt kinase activity, insulin stimulation resulted in a marked reduction of Akt electrophoretic mobility, whereas GTP␥S stimulation had little effect (data not shown). Furthermore, pretreatment of the cells with wortmannin completely prevented the reduction in Akt electrophoretic mobility. Thus, together these data demonstrate that GTP␥S stimulation of GLUT4 translocation occurs in a pathway that is independent of both PI 3-kinase and Akt kinase activation.
GDP␤S Inhibits Insulin Stimulation of GLUT4 Translocation-The data presented above suggest that the GTP␥S stimulation of GLUT4 translocation occurs in a pathway(s) downstream to and/or in parallel with the PI 3-kinase and Akt kinase. To further characterize the relationship between insulin and GTP␥S stimulation, we next incubated the SL-O-permeabilized cells with 100 M GDP␤S prior to insulin treatment (Fig. 3). In the absence of insulin, GDP␤S had no significant effect on GLUT4 translocation as detected by GLUT4 immunofluorescence of isolated plasma membrane sheets (Fig. 3A,  panels 1 and 5). Insulin stimulation resulted in a dose-dependent increase in the amount of plasma membrane-associated GLUT4 protein (Fig. 3A, panels 1-4). Pretreatment with GDP␤S resulted in the inhibition of GLUT4 translocation at low (1 and 10 nM) insulin concentrations but not at high (100 nM) insulin concentrations (Fig. 3A, panels 5-8). Insulin stimulation also resulted in a dose-dependent decrease in Akt SDSpolyacrylamide gel electrophoretic mobility (Fig. 3B, lanes 1, 2,  5, and 8). However, pretreatment with GDP␤S had no effect on the insulin stimulation of the Akt gel shift and, hence, phosphorylation and presumably protein kinase activation (Fig. 3B,  lanes 3, 6, and 9). These data further support a model in which GTP␥S stimulation of GLUT4 translocation is mediated by a mechanism downstream of the Akt kinase.
GTP␥S Stimulation of GLUT4 Translocation Is Tyrosine Kinase-dependent-Recently, several studies have suggested that G protein-coupled receptors can modulate tyrosine ki-nase signaling pathways. Having ruled out any potential role for PI 3-kinase and Akt protein kinase, we next assessed whether GTP␥S-stimulated GLUT4 translocation was mediated by a tyrosine kinase-dependent mechanism. This possibility was initially examined using the relatively specific tyrosine kinase inhibitors genistein and herbimycin A (Fig.  4). In unstimulated cells, there was little GLUT4 detectable in the isolated plasma membrane sheets, and pretreatment with either genistein or herbimycin A had no effect (Fig. 4,  panels 1, 4, and 7). As expected, both genistein and herbimycin A were effective inhibitors of insulin-stimulated GLUT4 translocation (Fig. 4, panels 2, 5, and 8), as both of these agents inhibit the insulin receptor tyrosine kinase and IRS1 tyrosine phosphorylation (data not shown). However, to our surprise, pretreatment with these tyrosine kinase inhibitors also prevented the GTP␥S-stimulated translocation of GLUT4 (Fig. 4, panels 3, 6, and 9).
Although genistein and herbimycin A are relatively specific tyrosine kinase inhibitors, it is possible that the prevention of GLUT4 translocation was due to some other nonspecific effect. To further confirm the requirement for tyrosine kinase activity, we next utilized single cell microinjection of a monoclonal phosphotyrosine antibody (Fig. 5). To identify the plasma membrane sheets derived from the small fraction of microinjected cells, the carboxyl-terminal domain of Ras fused to the maltosebinding protein (MBP-Ras) was co-injected (35). Plasma membrane sheets isolated from cells that were microinjected with MBP-Ras and a nonspecific mouse IgG were detected by MBP-Ras immunofluorescence (Fig. 5, panel 1). Insulin stimulation increased GLUT4 immunofluorescence in the plasma membrane sheets derived from both mouse IgG microinjected and noninjected cells (Fig. 5, panel 4). As expected, microinjection of the monoclonal PY20 phosphotyrosine antibody inhibited the insulin-stimulated translocation of GLUT4 compared with the GLUT4 translocation in the neighboring noninjected cells (Fig.   FIG. 3. GDP␤S inhibits insulin-stimulated GLUT4 translocation but not Akt phosphorylation. A, differentiated 3T3L1 adipocytes were serum-starved for 2 h, permeabilized in the absence (control; panels [1][2][3][4] or presence of 100 M GDP␤S (panels 5-8) with 0.8 IU/ml SL-O for 5 min at 37°C. The cells were then incubated with 0 (panels 1 and 5), 1 (panels 2 and 6), 10 (panels 3 and 7), or 100 nM (panels 4 and 8) insulin for 15 min. Plasma membrane sheets were prepared and subjected to immunofluorescence microscopy as described under "Experimental Procedures." B, differentiated 3T3L1 adipocytes were serum-starved for 2 h and then permeabilized in the absence (Ϫ; lanes 1, 2, 4, 5, 7, 8, and 10) or presence (ϩ; lanes 3, 6, and 9) of 100 M GDP␤S with 0.8 IU/ml SL-O for 5 min at 37°C. The cells were then incubated for an additional 15 min with 0 ( lanes  1, 4, 7, and 10), 1 (lanes 2 and 3), 10 (lanes 5 and 6), or 100 nM (lanes 8 and 9) insulin. The cells were detergent-solubilized and subjected to immunoblotting with an Akt antibody as described under "Experimental Procedures." These are representative observations from two independent experiments.

5, panels 2 and 5).
Similarly, microinjection of the monoclonal PY20 antibody also prevented the GTP␥S-stimulated translocation of GLUT4 (Fig. 5, panels 3 and 6). The fact that two independent methods to inhibit tyrosine kinase activity (genistein or herbimycin A treatment and PY20 microinjection) prevented both insulin-and GTP␥S-stimulated GLUT4 translocation provides compelling evidence that GTP␥S utilizes a tyrosine kinase pathway to mediate GLUT4 translocation.
Insulin and GTP␥S Stimulate the Tyrosine Phosphorylation of Discrete Proteins-Based upon the inhibition of insulin-and GTP␥S-stimulated GLUT4 translocation by genistein, herbimycin A, and the PY20 phosphotyrosine antibody, we next compared the tyrosine phosphorylation of proteins induced by insulin and GTP␥S treatment (Fig. 6). Insulin stimulation resulted in increased tyrosine phosphorylation of the insulin receptor ␤-subunit and IRS1 (Fig. 6A, lanes 3 and 4). In contrast, GTP␥S stimulation had no effect on either insulin receptor ␤ subunit or IRS1 tyrosine phosphorylation (Fig. 6A, lanes  1 and 2). Instead, GTP␥S appeared to induce the tyrosine phosphorylation of several proteins in the molecular mass range of 120 -130 kDa and 55-75 kDa (Fig. 6A, compare lanes  1 and 2 with lanes 3 and 4). To more carefully examine the potential for IRS1 tyrosine phosphorylation, we also specifically immunoprecipitated IRS1 from both insulin-and GTP␥Sstimulated adipocytes. Insulin stimulated the tyrosine phosphorylation of IRS1 in the IRS1 immunoprecipitates (Fig. 6B,  lanes 3 and 4), whereas there was no detectable IRS1 tyrosine phosphorylation following GTP␥S stimulation (Fig. 6B, lanes 1  and 2). IRS1 immunoblotting of that same membrane demonstrated the presence of equal amounts of IRS1 protein (Fig. 6B,  lanes 5-8). The apparent increase in molecular weight following both insulin and GTP␥S stimulation most likely reflects the serine/threonine phosphorylation of IRS1.
GTP␥S Stimulates the Tyrosine Phosphorylation of PYK2, pp130 Cas , Paxillin, and Cbl-One interesting tyrosine kinase that has recently been implicated in the intracellular signaling of several trimeric G protein-coupled receptors and by osmotic shock is PYK2 (31, 36 -39). Since osmotic shock, like GTP␥S, stimulated GLUT4 translocation through a PI 3-kinase-and Akt-independent pathway (31), we next examined the tyrosine phosphorylation of PYK2 by immunoprecipitation. As previously reported, in unstimulated cells there was no detectable PYK2 tyrosine phosphorylation, whereas osmotic shock induced a marked tyrosine phosphorylation of PYK2 (Fig. 7A,  lanes 1 and 2). In contrast, there was no detectable tyrosine phosphorylation of PYK2 in response to insulin (Fig. 7A, lane  3). Although permeabilization alone resulted in a slight increase in the phosphotyrosine content of PYK2, GTP␥S was an effective activator of PYK2 tyrosine phosphorylation (Fig. 7A,  lanes 4 and 5). Immunoblotting of the same Western blot with a PYK2 antibody indicated the presence of equal amounts of PYK2 in each lane (Fig. 7B, lanes 1-5). DISCUSSION Over the past several years, significant progress has been made in our understanding of the insulin signal transduction pathway leading to GLUT4 vesicle trafficking. It is generally accepted that insulin stimulation results in the activation of the insulin receptor tyrosine kinase, leading to the tyrosine phosphorylation of the IRS family of intracellular docking proteins (1,2). The tyrosine phosphorylation of IRS1 induces the association/targeting and/or activation of PI 3-kinase (3). Multiple studies have demonstrated that PI 3-kinase function is necessary for insulin-stimulated GLUT4 translocation (4,5). Although the downstream target(s) of PI 3-kinase involved in GLUT4 translocation has not been identified, recent studies indicate that activation of PI 3-kinase results in the activation of the Akt serine/threonine-protein kinase (7,8).
In addition to the identification of this signaling pathway of insulin, there are several insulinomimetic agents that can also stimulate glucose transport in adipocytes and muscle. In particular, nonhydrolyzable GTP analogues have been reported to increase glucose transport activity by inducing the translocation of GLUT4 (10,11). We therefore reasoned that examining the GTP␥S stimulation of glucose transport might provide new insight by identifying common and/or distinct pathways leading to GLUT4 translocation. In this regard, two previous studies have reported contradictory findings for the role of PI 3-kinase in GTP␥S-stimulated glucose transport regulation. A study by Clarke et al. (40) suggested that the ability of GTP␥S to stimulate 2-deoxy-D-glucose uptake by permeabilized cells was partially blocked by wortmannin (1 M). In contrast, it was later reported by Herbst et al. (41) that incubation of permeabilized 3T3L1 adipocytes with GTP␥S (200 M) did not increase PI 3-kinase activity over basal level. Although the basis for this difference is not apparent, our data demonstrate that GTP␥S-stimulated GLUT4 translocation is completely wortmannin-insensitive at concentrations sufficient to inhibit both the receptor tyrosine kinase-activated PI 3-kinase and the G ␤␥responsive PI 3-kinase isoform (42).
One recently established function of PI 3-kinase is to couple receptor tyrosine kinase activation to the Akt protein kinase, also known as RAC-PK or protein kinase B (6,43,44). The Akt kinase is directly activated by dual phosphorylation on serine and threonine residues (34). In addition, the pleckstrin homology domain of Akt can bind phosphatidylinositol 3,4-bisphosphate in vitro, which also has been shown to stimulate its protein kinase activity (45)(46)(47). In either case, insulin activates the Akt kinase in a PI 3-kinase dependent manner. Furthermore, overexpression of a membrane-targeted constitutively active Akt results in the persistent translocation of GLUT4 (7,8). Since it has been suggested that there are PI 3-kinaseindependent pathways leading to Akt activation (48), it remained formally possible that GTP␥S activated Akt, hence GLUT4 translocation, by a wortmannin-insensitive mecha- nism. However, our data demonstrated that GTP␥S is a very weak activator of Akt phosphorylation and protein kinase activity. In addition, wortmannin completely inhibited this small extent of Akt activation, yet had no effect on GTP␥S-stimulated GLUT4 translocation. Thus, these data clearly establish the presence of a PI 3-kinase-and Akt-independent pathway leading to GLUT4 translocation in 3T3L1 adipocytes.
In an alternative approach to identify the pathway(s) utilized by GTP␥S, we observed that GDP␤S was an effective inhibitor of insulin-stimulated GLUT4 translocation at physiological but not at supraphysiological (saturating) insulin concentrations. The inability of GDP␤S to prevent insulin-stimulated GLUT4 translocation at supraphysiological insulin concentrations has been previously reported, but the effect of GDP␤S was not examined at physiological insulin levels (49). There are several plausible reasons accounting for the ineffectiveness of GDP␤S at high insulin concentrations, including the possibility that insulin can mediate GLUT4 translocation by more than one signaling pathway, i.e. one pathway that is dependent on a GTP binding protein(s) and another one that is not. The former pathway would be the predominant signaling pathway under normal physiological stimulation, whereas the later pathway would be activated only by high insulin concentrations. Alternatively, at high insulin concentrations, the downstream effectors are maximally activated and would require substantially greater concentrations of GDP␤S to be inhibited. In any case, the fact that GDP␤S inhibited the insulinstimulated translocation of GLUT4 without any effect on Akt kinase is consistent with the presence of at least one GTPbinding protein functioning downstream of Akt. Together, these data provide compelling evidence that neither PI 3-kinase nor Akt kinase are involved in the GTP␥S stimulation of GLUT4 translocation and that GTP␥S functions at a site distal to the Akt kinase.
Although it is generally thought that calcium does not play a significant role in the insulin stimulation of GLUT4 translocation, it has been hypothesized to be necessary for the exercise/ contraction-stimulated GLUT4 translocation in skeletal mus- FIG. 5. Insulin-and GTP␥S-stimulated GLUT4 translocation is inhibited by microinjection of the PY20 monoclonal phosphotyrosine antibody. Differentiated 3T3L1 adipocytes were microinjected with 2 mg/ml MBP-Ras mixed with 5 mg/ml mouse IgG (panels 1 and 4) or with a 5 mg/ml concentration of the PY20 phosphotyrosine antibody (panels 2, 3, 5, and 6) as described under "Experimental Procedures." Following microinjection, the cells were either treated with 10 nM insulin (panels 1, 2, 4, and 5) or permeabilized and incubated with 100 M GTP␥S (panels 3 and 6) for 15 min at 37°C. Plasma membrane sheets were then prepared and subjected to immunofluorescence microscopy with an antibody specific for MBP-Ras (panels 1-3) or with a GLUT4-specific antibody (panels 4 -6). The arrows depict the locations of the plasma membrane sheets that were MBP-positive. These are representative immunofluoresence observations from three independent experiments.
FIG. 6. Insulin and GTP␥S stimulate the tyrosine phosphorylation of discrete proteins. Differentiated 3T3L1 adipocytes were serum-starved for 2 h and either left nonpermeabilized (lanes 3, 4, 7, and 8) or permeabilized with 0.8 IU/ml SL-O for 5 min at 37°C (SLO; lanes 1, 2, 5, and 6). The cells were then untreated (C; lanes 1, 3, 5, and 7) or incubated with 100 nM insulin (I; lanes 4 and 8) or 100 M GTP␥S (G; lanes 2 and 6) for 5 min at 37°C. A, whole cell detergent extracts were prepared and subjected to Western blotting with the PY20 monoclonal phosphotyrosine antibody. B, whole cell detergent extracts were immunoprecipitated with an IRS1 antibody, and the immunoprecipitates were subjected to immunoblotting with the PY20 phosphotyrosine antibody (lanes 1-4) and the IRS1 antibody (lanes [5][6][7][8]. This is a representative immunoblot from 3-5 independent experiments. IP, immunoprecipitation; IB, immunoblot. cle (23)(24)(25). In analogy with GTP␥S, the exercise/contractionstimulated GLUT4 translocation occurs by a wortmannininsensitive pathway (22). In addition, agonists of G q -coupled receptors, receptors that elicit increases in intracellular calcium, have also been reported to induce GLUT4 translocation in both transfected Chinese hamster ovary cells and 3T3L1 adipocytes (18). Thus, it seemed possible that GTP␥S-stimulated GLUT4 translocation occurred via mediation of an increase in intracellular calcium. However, both an inhibitor of intracellular calcium release and calcium chelators failed to prevent the GTP␥S-stimulated translocation of GLUT4 (data not shown).
Having ruled out PI 3-kinase, Akt kinase, and changes in intracellular calcium, we next assessed whether GTP␥S stimulated GLUT4 translocation in a tyrosine kinase-dependent pathway. To our surprise, we observed that two tyrosine kinase inhibitors (genistein and herbimycin A) as well as microinjection of the PY20 monoclonal phosphotyrosine antibody were effective blockers of GTP␥S-stimulated GLUT4 translocation. Since GTP␥S did not induce the tyrosine phosphorylation of either the insulin receptor or IRS1 (which would activate the PI 3-kinase and Akt kinase), the ability to block the GTP␥S signal with genistein, herbimycin A, and PY20 microinjection suggests the presence of a novel tyrosine kinase pathway. Consistent with this interpretation, GTP␥S stimulation resulted in the specific tyrosine phosphorylation of proteins in the 120 -130-kDa and 55-75-kDa ranges.
To search for the identity of these proteins, we took advantage of the ability of SH2 domains to bind with high affinity to tyrosine-phosphorylated proteins. Using a GST-CrkII fusion, we identified several proteins (pp125 FAK , pp130 Cas , paxillin, and Cbl) that displayed a slight increase in tyrosine phosphorylation following GTP␥S treatment (data not shown). On the other hand, immunoprecipitation of PYK2, a newly identified soluble tyrosine kinase that has been implicated in the intracellular signaling of several trimeric G protein-coupled receptors and by osmotic shock (31, 36 -39), demonstrated that PYK2 was markedly tyrosine-phosphorylated in response to GTP␥S but not to insulin. The fact that both GTP␥S and osmotic shock stimulation result in GLUT4 translocation by both a PI 3-kinase-and Akt kinase-independent pathway provides suggestive evidence that PYK2 may be an important kinase regulating these trafficking events.
In summary, we have observed that GTP␥S-stimulated GLUT4 translocation occurs by a novel tyrosine kinase pathway in 3T3L1 adipocytes. This alternative signaling pathway does not require tyrosine phosphorylation of the insulin receptor, IRS1, or the targeting and/or activation of either PI 3-kinase or Akt kinase. However, GTP␥S stimulation results in the tyrosine phosphorylation of several distinct proteins, including PYK2, pp125 FAK , pp130 Cas , paxillin, and Cbl. Whether or not any of these tyrosine-phosphorylated proteins are necessary and/or sufficient for the GTP␥S stimulation of GLUT4 translocation is an important future issue for investigation.