G q -coupled Receptors Transmit the Signal for GLUT4 Translocation via an Insulin-independent Pathway*

Guanosine 5 (cid:42) - O -(3-thiotriphosphate) (GTP (cid:103) S) induces the translocation of glucose transporter type 4 (GLUT4) from an intracellular pool to the cell surface and in-creases glucose uptake in adipocytes. The GTP-binding protein(s) responsible for the translocation has remained to be identified. Using a sensitive and quantita- tive method to assess the translocation of c-MYC epitope-tagged GLUT4, we obtained evidence that the activation of receptor-coupled G q (neither G i nor G s ) triggered GLUT4 translocation in cells, independently of insulin signaling pathway(s). Platelet-activating factor (PAF) induced GLUT4 translocation in the cells ex- pressing the G i - and G q -coupled PAF receptor, but the translocation was induced even after pretreatment with wortmannin, an islet-activating protein and phorbol 12,13-dibutyrate. Norepinephrine triggered GLUT4 translocation in cells expressing the G q -coupled (cid:97)

GLUT4 from an intracellular pool to the plasma membrane is thought to be a major mechanism of glucose uptake in response to insulin in these tissues (1)(2)(3)(4). 3T3-L1 adipocytes have been recognized as an ideal model to investigate the mechanism of GLUT4 translocation and glucose uptake, because these cells have a large amount of endogenous GLUT4 and insulin receptors (5).
To examine molecular mechanisms of GLUT4 translocation in 3T3-L1 adipocytes and Chinese hamster ovary (CHO) cells, we developed a highly sensitive and quantitative method to measure directly c-MYC epitope-tagged GLUT4 (GLUT4myc) on the cell surface (6). Using this system, we have found that phosphatidylinositol 3-kinase (PI 3-kinase, p85/p110 heterodimer type) is involved in signal transductions of GLUT4 translocation not only by insulin but also by platelet-derived growth factor and epidermal growth factor (7)(8)(9). Other investigators reported that PI 3-kinase plays a pivotal role in the insulin-stimulated GLUT1 and GLUT4 translocations in CHO cells and adipocytes, respectively (10 -12). Therefore, the CHO cell system as well as 3T3-L1 adipocytes are useful for studying molecular mechanisms of GLUT4 translocation (see "Discussion").
On the other hand, GLUT4 translocation is also induced in permeabilized adipocytes by treatment with guanosine 5Ј-O-(3thiotriphosphate) (GTP␥S) of nonhydrolyzable GTP analogs, as well as by insulin (13,14). The GTP␥S-induced GLUT4 translocation was also observed in 3T3-L1 adipocytes and CHO cells stably expressing the GLUT4myc, using this method of detection (6). However, the GTP-binding protein(s) responsible for the GLUT4 translocation has remained to be identified.
Some small GTP-binding proteins play important roles in intracellular vesicle trafficking or secretion (15)(16)(17)(18)(19), and these proteins are detected in GLUT4-containing vesicles of adipocytes (20,21). However, we reported that the small GTP-binding proteins, Ras, Rab3D, Rad, and Rho are unlikely to be involved in the GLUT4 translocations with insulin, phorbol 12-myristate 13-acetate (PMA), and GTP␥S treatments, and that the signaling pathway of the GTP␥S-stimulated GLUT4 translocation is different from those of insulin and PMA treatments (22). In addition, NaF plus AlCl 3 treatment, which is known to activate heterotrimeric GTP-binding proteins, stimulates GLUT4 translocation and glucose uptake in adipocytes and in CHO cells (6,23). This suggests that heterotrimeric GTP-binding protein(s) may be involved in the GTP␥S-stimulated GLUT4 translocation. Adrenergic stimulation induces GLUT4 translocation and glucose uptake in cardiac myocytes (24,25) and brown adipocytes (26,27). The GLUT4 translocation and the enhanced glucose uptake are mediated insulinindependently via adrenergic receptors coupled to heterotri-meric GTP-binding proteins. However, the molecular mechanism of insulin-independent GLUT4 translocation was not elucidated.
We report here that G q -coupled receptors, but not G i -nor G s -coupled receptors, trigger GLUT4 translocation in an insulin-independent manner in 3T3-L1 adipocytes and CHO cells.

MATERIALS AND METHODS
Cells and Materials-The parent cell lines used in this study were CHO-GLUT4myc, a CHO cell line expressing GLUT4myc, constructed by inserting a human c-MYC epitope (14 amino acids) into the first ectodomain of GLUT4, and 3T3-L1-GLUT4myc, a 3T3-L1 fibroblast line expressing GLUT4myc. The 3T3-L1-GLUT4myc fibroblasts were induced to differentiate into adipocytes as described previously (6). All other reagents were of analytical grade.
Cell Surface Anti-c-MYC Antibody Binding Assay (GLUT4myc Translocation Assay)-The CHO-GLUT4myc cells expressing various receptors in 24-well plates were incubated in 500 l of KRH buffer (6) for 30 min at 37°C and then with indicated concentrations of ligands for indicated periods at 37°C. GLUT4myc translocation was measured as described previously (6).
The 3T3-L1-GLUT4myc-PAFR and 3T3-L1-GLUT4myc-␣ 1a AR adipocytes in 24-well plates were incubated in 500 l of KRH buffer for 20 min at 37°C and then treated with indicated concentrations of ligands for indicated periods at 37°C. GLUT4myc translocation was measured after fixation with 2% paraformaldehyde, as described previously (6).
2-Deoxyglucose Uptake Measurement-Cells in 24-well plates were treated with indicated concentrations of ligands for indicated periods at 37°C. 2-Deoxyglucose uptake was measured as described previously (6).
Down-regulation of Protein Kinase C with Phorbol 12, and Pretreatment with Islet-activating Protein (IAP) or Wortmannin-Cells were pretreated with or without PDBu (100 ng/ml) for 24 h at 37°C in medium for down-regulating protein kinase C and with or without IAP (100 ng/ml) for 24 h at 37°C in medium that abolishes G i -coupled pathway(s). To inactivate PI 3-kinases, the cells were pretreated with or without indicated concentrations of wortmannin for 20 min at 37°C.
cAMP Assay-The cAMP levels were measured using Yamasa radioimmunoassay kits as described (36).
[Ca 2ϩ ] i was calculated based on the formula by Grynkiewicz et al. (38).
To examine which heterotrimeric GTP-binding protein(s) is responsible for the GTP␥S-stimulated GLUT4 translocation, we used both 3T3-L1-GLUT4myc adipocytes and CHO-GLUT4myc cells. Platelet-activating factor (PAF) receptor (PAFR) is thought to transmit the signal via IAP-sensitive (G i ) and -insensitive (G q ) heterotrimeric GTP-binding proteins (28,37,39). As shown in Fig. 1, 3T3-L1-GLUT4myc adipocytes stably expressing PAF receptors showed that PAF-stimulated GLUT4myc translocation in a dose-and time-dependent manner, whereas the parent 3T3-L1-GLUT4myc adipocytes did not respond to any concentrations of PAF (Fig. 1, A and B). PAF treatment also increased the rate of glucose uptake in 3T3-L1-GLUT4myc-PAFR adipocytes, in proportion to GLUT4myc translocation (Fig. 1, A and C). Almost the same dose-and time-dependent GLUT4myc translocation and glucose uptake in response to PAF were observed in CHO-GLUT4myc-PAFR cells but not in the parent CHO-GLUT4myc cells (Fig. 1, D--F). The CHO-GLUT4myc cells have relatively large amounts of endogenous GLUT1 compared with GLUT4myc, and the GLUT1 might affect the rate of glucose uptake. The translocated-GLUT4myc in response to PAF increased the rate of glucose uptake in CHO-PAFR-GLUT4myc cells, compared with CHO-PAFR cells ( Table I). The enhanced glucose uptake in CHO-PAFR cells was attributed to GLUT1 translocation, because PAF stimulated GLUT1myc translocation in CHO- PAFR-GLUT1myc cells. 2 Next, we examined the effect of wortmannin on the PAFstimulated GLUT4 translocation in 3T3-L1-GLUT4myc-PAFR adipocytes and CHO-GLUT4myc-PAFR cells; wortmannin completely inhibited the insulin-stimulated GLUT4 translocation and glucose uptake in 3T3-L1 adipocytes and CHO-GLUT4myc cells, by abolishing phosphatidylinositol (PI) 3-kinase activity (7,12). As shown in Figs. 2A and 3A, wortmannin inhibited the insulin-stimulated GLUT4myc translocation dose dependently and abolished the translocation at 10 Ϫ7 M in both lines of cells. This inhibitory effect on the insulin-stimulated translocation was closely related to that on PI 3-kinase (p85/ p110 heterodimer type) activity. 3 However, the PAF-induced GLUT4myc translocations were 3.8-fold in 3T3-L1-GLUT4myc-PAFR adipocytes and 9.4-fold in CHO-GLUT4myc-PAFR cells, even in the presence of 10 Ϫ6 M wortmannin. Therefore, the PAF-stimulated GLUT4myc translocation pathway was independent of PI 3-kinases including p110␥ (40).
Some PAF-stimulated physiological responses are mediated through G i that is sensitive to IAP (39). The PAF-stimulated GLUT4myc translocation of 3T3-L1-GLUT4myc-PAFR adipocytes was partially inhibited but remained to a certain extent (from 4.4-fold increase to 2.4-fold increase) ( Fig. 2B) (see "Discussion"). On the other hand, the PAF-stimulated GLUT4myc translocation of CHO-GLUT4myc-PAFR cells was not affected significantly by treatment of 100 ng/ml IAP (Fig. 3B), a treatment that abolished the G i coupling (39). The IAP-insensitive pathway observed in the both cell lines is thought to be mediated by G q , which activates phosphoinositide-specific phospholipase C ␤ (PLC␤) to hydrolyze phosphatidylinositol 4,5bisphosphate (PIP 2 ) (41,42). The breakdown products, inositol 1,4,5-trisphosphate (IP 3 ) and 1,2-diacylglycerol, activate a Ca 2ϩ channel and protein kinase C, respectively (43,44). Therefore, we asked whether the IAP-insensitive GLUT4myc translocation by PAF is due to protein kinase C activation or to [Ca 2ϩ ] i increase.
To examine the effects of [Ca 2ϩ ] i on the PAF-stimulated GLUT4myc translocation, we used two different Ca 2ϩ ionophores. Ionomycin (1 M) and PAF (2 ϫ 10 Ϫ8 M) treatments elevated [Ca 2ϩ ] i in CHO-GLUT4myc-PAFR cells, from basal levels (lower than 100 nM) up to about 900 and 600 nM, respectively, as in Fig
As shown in Fig. 6, A and B, norepinephrine stimulated GLUT4myc translocation in a dose-and time-dependent manner in CHO-GLUT4myc-␣ 1b AR cells but not in the parent CHO-GLUT4myc cells. Norepinephrine stimulated glucose uptake in CHO-GLUT4myc-␣ 1b AR in almost the same dose-dependent manner as GLUT4myc translocation (Fig. 6C). The fold increase of glucose uptake, however, is not in proportion to that of GLUT4myc translocation, because the norepinephrinestimulated glucose uptake resulted from the translocations of both the endogenous GLUT1 and exogenous GLUT4myc in CHO cells. The GLUT4myc translocated to the cell surface in response to norepinephrine took up more glucose in CHO-␣ 1b AR-GLUT4myc cells, compared with CHO-␣ 1b AR cells (Table II). The enhanced glucose uptake in the CHO-␣ 1b AR cells was probably due to GLUT1 translocation, because norepinephrine stimulated GLUT1myc translocation in CHO- GLUT1myc-␣ 1b AR cells. 4 Like the PAF-stimulated GLUT4myc translocation in CHO-GLUT4myc-PAFR, the norepinephrinestimulated GLUT4myc translocation in CHO-GLUT4myc-␣ 1b AR cells was not inhibited by wortmannin (Fig. 6D), PDBu (Fig. 6E), and by simultaneous wortmannin and PDBu (Fig.  6F) treatments. In addition, Ca 2ϩ ionophores did not trigger GLUT4myc translocation, and the norepinephrine-stimulated translocation was reduced by guanosine 5Ј-O-(2-thiodiphosphate) (GDP␤S) pretreatment, 4 while norepinephrine-activated G q accelerates PIP 2 breakdown via PLC␤. Therefore, the norepinephrine-stimulated GLUT4myc translocation was thought to be mediated by G q directly or through unknown pathway(s) after G q activation but probably not to be a secondary phenomenon by PIP 2 breakdown.

Heterotrimeric GTP-binding Proteins and [Ca 2ϩ
] i -The heterotrimeric GTP-binding proteins are associated with signal transduction from cell surface receptors (48). ␣-Subunits of the GTP-binding proteins have been classified into four, G s , G i , G q , and G 12 , based on amino acid sequence homology (47). We find that G q class can mediate GLUT4 translocation and the resultant glucose uptake, because three independent G q -coupled receptors triggered GLUT4 translocation in 3T3-L1 adipocytes and CHO cells. The use of CHO cells to study the molecular mechanisms of GLUT4 translocation has been controversial (49). In CHO cells, a relatively smaller amount of intracellular GLUT4myc than in 3T3-L1 adipocytes was translocated to the cell surface, and most was retained intracellularly, even after insulin treatment (50). The high sensitivity of our method made feasible the detection of GLUT4 on the cell surface, following insulin-induced translocation in CHO cells. The machinery for insulin-stimulated GLUT4 translocation in CHO cells is not completely identical with that in 3T3-L1 adipocytes, but CHO cells seem to possess a basic machinery for insulinstimulated translocation of exogenously expressed GLUT4, which mimics that of adipocytes (6,13,14,23,45,46,50,51). CHO cells are clonally stable, even after long term culture or transfection with various expression plasmids. On the other hand, 3T3-L1 adipocytes are relatively unstable and difficult to obtain stable clones expressing exogenous cDNAs. Actually, we couldn't establish 3T3-L1-GLUT4myc adipocytes stably expressing ␣ 1b -AR. The 3T3-L1 adipocytes are an ideal cell culture model so far, but CHO cells also seem to be useful to study molecular mechanisms involved in GLUT4 translocation, in addition to 3T3-L1 adipocytes.
The physiological effects of norepinephrine and PAF after IAP pretreatment are thought to be mediated mainly by activating G q class (Figs. 2B, 3B, and 7D), although other unknown pathways need to be considered (39). The PAF-and norepinephrine-stimulated GLUT4 translocations were not affected by wortmannin (Figs. 2A, 3A, and 6D) but were attenuated with GDP␤S pretreatment. 4 Therefore, activation of G q class seems to trigger GLUT4 translocation directly or via unknown

cells) in CHO-␣ 1b AR cells and CHO-␣ 1b AR-GLUT4myc cells
The CHO-␣ 1b AR cells and CHO-␣ 1b -AR-GLUT4myc cells were estimated to have almost the same amount of ␣ 1b AR and were treated with 10 Ϫ5 M norepinephrine or buffer alone (Ϫ) for 10 min. 2-Deoxyglucose uptake was measured, as described under "Materials and Methods." Values are means Ϯ S.E. of three determinations. Asterisk shows significant difference (p Ͻ 0.005, Student's t test) from the norepinephrine-stimulated increment of CHO-␣ 1b AR cells.  pathway(s), but not by an insulin-dependent pathway(s). Among G q class, G q , G 14 , and G 15 (or G 16 ) isotypes have different activities to stimulate PLC␤ subtypes (47,52). We have not identified which isotype of G q class is important for the GLUT4 translocation. G i2 ␣-deficient transgenic mice have been generated, and insulin-stimulated glucose uptake was impaired in these mice (53). An IAP-sensitive pathway involved in PAFstimulated GLUT4 translocation in 3T3-L1-GLUT4myc adipocytes (Fig. 2B) may be related to G i2 ␣. Of course, the involvement of G 12 class or ␤␥ subunits in the GLUT4 translocation cannot be excluded.
The GLUT4 translocations triggered by G q -coupled receptors were not inhibited by down-regulating protein kinase C (Figs. 2, 3, and 6), and Ca 2ϩ mobilization with ionophores did not cause GLUT4 translocation (Fig. 4, A, C, and E). Therefore, GLUT4 translocation does not seem to be a secondary phenomenon following PIP 2 breakdown, as evoked by G q activation. However, the GLUT4 translocation required a certain amount of [Ca 2ϩ ] i , because 20 -40 M BAPTA-AM inhibited the GLUT4 translocation and Ca 2ϩ mobilization (Fig. 4, B, D, and E). It is considered that [Ca 2ϩ ] i and some Ca 2ϩ -binding proteins play important roles in membrane traffic (54). Considering that the GLUT4 translocation is one of the regulated membrane traffic systems, it seems reasonable that GLUT4 translocation requires a certain amount of [Ca 2ϩ ] i . However, the possibility that BAPTA-AM affects the GLUT4 translocation by effects other than chelating Ca 2ϩ would need to be ruled out.
Physiological Aspects of PAF-stimulated GLUT4 Translocation-PAF receptors are expressed mainly in hematopoietic cells, and GLUT1 (but not GLUT4) is expressed in hematopoietic cells. PAF induces GLUT1 translocation and glucose uptake by about 1.5-2-fold in CHO-GLUT1myc-PAFR cells. 2 PAF exerts reactions such as chemotaxis, phagocytosis, and smooth muscle contraction (37) that require fuel for ATP. To supply the fuel, GLUT1 may translocate to the cell surface and take up glucose via G q coupling in response to PAF. Numerous receptors couple with G q , and most ligand-stimulated reactions require an energy supply. Therefore, our evidence provides new insights into G q functions.
Physiological Aspects of Norepinephrine-stimulated GLUT4 Translocation-The finding that norepinephrine stimulates glucose uptake via G q might seem inconsistent with data that catecholamines usually antagonize many insulin actions, especially in glucose and lipid metabolism in adipocytes and hepatocytes. However, norepinephrine did not antagonize the insulin-stimulated GLUT4 translocation in CHO-GLUT4myc-␣ 1b AR cells, and the simultaneous stimulation of norepinephrine and insulin worked additively as for GLUT4 translocation. 4 CHO cells are not physiologically adrenergic target cells. However, norepinephrine treatment or physical exercise stimulates glucose uptake by translocating GLUT4 in cardiomyocytes that have a large amount of endogenous GLUT4, independently of insulin (24,25); they reported that norepinephrine-stimulated glucose uptake was mediated by ␣ 1 -adrenergic receptor but not by ␣ 2 -or ␤-adrenergic receptors. The ␣ 1 -adrenergic receptors are classified into at least three types, ␣ 1a (␣ 1C ), ␣ 1b (␣ 1B ), and ␣ 1d (␣ 1A/D ) (31)(32)(33)(34), and all the three types of receptors couple to G q class (52). This is consistent with our results (Figs. 5-7). In cardiomyocytes not only fatty acid but also endogenous glycogen and glucose supplied from extracellular fluid are consumed to generate ATP during physical exercises (55). Interestingly, dysfunction and hypertrophy of cardiomyocytes were observed in GLUT4-knockout mice (56). They considered that GLUT4 played an important role in cardiomyocytes and that a low energy supply to cardiomyocytes owing to deficient GLUT4 might lead to dysfunction and hy-pertrophy. Therefore, it seems reasonable that norepinephrine translocates GLUT4 via G q coupling and takes up glucose to supply the fuel for ATP during physical exercises, independently of insulin signaling pathway(s).
Glucose uptake into brown adipocytes is enhanced directly by norepinephrine released from sympathetic nerves (26,27). As the enhanced glucose uptake occurs without increase in plasma insulin levels and is not inhibited by wortmannin, a PI 3-kinase inhibitor, the phenomenon is probably independent of insulin actions. This is consistent with our findings with 3T3-L1 adipocytes (Fig. 7). However, there is controversy regarding GLUT4 translocation by subcellular fractionation, despite the enhanced glucose uptake (26,27). The discrepancy may relate to different sensitivities of tests to assess GLUT4 translocation or by different cell origins. We examined norepinephrine-stimulated GLUT4myc translocation (Fig. 7, E and F) and glucose uptake (Fig. 7G) in 3T3-L1-GLUT4myc adipocytes expressing ␣ 1a -AR, because ␣ 1a -AR also activated G q and translocated GLUT4myc in CHO-GLUT4myc-␣ 1a -AR cells, 2 like ␣ 1b -AR. As shown in Fig. 7E, the norepinephrine treatment increased the anti-c-MYC antibody binding on the cell surface in 3T3-L1 adipocytes that do not express exogenous GLUT4myc. The GLUT4myc-unrelated increase of the antibody binding in 3T3-L1 adipocytes was only observed by the norepinephrine treatment but not by the treatments with insulin, plateletderived growth factor, and epidermal growth factor (6,8,9). In CHO cells, the GLUT4myc-unrelated increase of the antibody binding was not observed by the treatment of norepinephrine. We have found an approximate 120-kDa protein in 3T3-L1 and 3T3-L1-GLUT4myc-␣ 1a -AR adipocytes, which is different from GLUT4myc protein and cross-reacts with the anti-c-MYC antibody. 2 The adipocyte-specific 120-kDa protein may translocate to the cell surface by the norepinephrine treatment and increase the surface antibody binding, in addition to GLUT4myc. Therefore, we assessed the norepinephrine-stimulated GLUT4myc translocation in 3T3-L1-GLUT4myc-␣ 1a -AR adipocytes by subtracting each surface antibody binding of 3T3-L1 adipocytes in the absence and presence of norepinephrine (Fig. 7F). Our results were consistent with those in brown adipocytes reported by Omatsu-Kanbe and Kitasato (26). Brown adipocytes play a key role in thermogenesis especially in newborn animals. Norepinephrine-induced glucose uptake may thus contribute to the thermogenesis in brown adipocytes.
Another possible target of norepinephrine is capillary endothelial cells. Endothelial cells have diverse characteristics, depending on organs or tissues. Endothelial cells of muscle and fat have abundant GLUT4, and the GLUT4 translocated to the luminal plasma membrane was thought to help take more glucose into the muscle and fat, in response to insulin (57). Considering that endothelial cells are dominated by the autonomic nervous system, norepinephrine may stimulate the GLUT4 translocation.
Contractile stimuli were seen to enhance the glucose uptake of skeletal myocytes, independently of insulin (58 -61), but the exact mechanism has not been identified. The released bradykinin by muscle contraction (62) may via G q activate bradykinin receptors (63,64) and the translocated GLUT4 may help to take up glucose to generate ATP for repeated muscle contractions.
As we find that three independent G q -coupled receptors trigger GLUT4 translocation to take up glucose, new insights into G q function have been forthcoming.