Association of the Insulin Receptor with Phospholipase C-γ (PLCγ) in 3T3-L1 Adipocytes Suggests a Role for PLCγ in Metabolic Signaling by Insulin*

Phospholipase C-γ (PLCγ) is the isozyme of PLC phosphorylated by multiple tyrosine kinases including epidermal growth factor, platelet-derived growth factor, nerve growth factor receptors, and nonreceptor tyrosine kinases. In this paper, we present evidence for the association of the insulin receptor (IR) with PLCγ. Precipitation of the IR with glutathione S-transferase fusion proteins derived from PLCγ and coimmunoprecipitation of the IR and PLCγ were observed in 3T3-L1 adipocytes. To determine the functional significance of the interaction of PLCγ and the IR, we used a specific inhibitor of PLC, U73122, or microinjection of SH2 domain glutathione S-transferase fusion proteins derived from PLCγ to block insulin-stimulated GLUT4 translocation. We demonstrate inhibition of 2-deoxyglucose uptake in isolated primary rat adipocytes and 3T3-L1 adipocytes pretreated with U73122. Antilipolytic effect of insulin in 3T3-L1 adipocytes is unaffected by U73122. U73122 selectively inhibits mitogen-activated protein kinase, leaving the Akt and p70 S6 kinase pathways unperturbed. We conclude that PLCγ is an active participant in metabolic and perhaps mitogenic signaling by the insulin receptor in 3T3-L1 adipocytes.

The insulin receptor (IR) 1 is a hetero-tetramer consisting of two ␣-subunits that are entirely extracellular and two ␤-subunits that span the plasma membrane and contain intrinsic tyrosine kinase activity (1,2). One of the major metabolic effects of insulin in fat and skeletal muscle is the stimulation of glucose uptake (3). This occurs through the translocation of glucose transporters (GLUT4) from intracellular vesicles to the plasma membrane (4). Neither the molecular mechanism by which GLUT4 vesicles fuse with the plasma membrane nor the signaling proteins downstream of the IR leading to the stimulation of glucose transport have been clearly elucidated. An involvement of IRS-1 is indicated by both in vitro studies where primary rat adipocytes were transfected with an antisense ribozyme directed against rat IRS-1 (5) and in vivo studies where insulin-mediated glucose transport was attenuated in mice with targeted disruption of the IRS-1 gene (6). The ability of IRS-1 knock-out mice to transport glucose in response to insulin implies alternative mechanisms of glucose transport activation by insulin. PI 3-kinase has been demonstrated to be required for the insulin effect on glucose transport (7)(8)(9)(10).
Protein kinase C has been studied extensively as a mediator of insulin-stimulated glucose transport (11). The insulinomimetic effect of phorbol esters on glucose uptake implicates DAG as a potentiator of glucose uptake. Phorbol ester down-regulation reportedly inhibits insulin-stimulated glucose uptake in mouse soleus (12), rat heart (13), and rat adipocytes (14 -16). In 3T3-L1 adipocytes, however, insulin-stimulated glucose uptake has been reported to be refractory to down-regulation by phorbol esters (17,18). There are a number of ways that DAG can be generated in the cell in response to cell-surface receptors. An immediate release of DAG has been attributed to phosphoinositide-specific PLC activation. Sustained DAG production in many cell types, however, is the result of phosphatidylcholine (PC) hydrolysis by a PC-specific PLC and by phospholipase D (PLD) as well as de novo DAG formation (19,20). Studies have attempted to evaluate the effect of elevating intracellular DAG levels on glucose uptake by treating skeletal muscle strips (21,22) or primary rat adipocytes (23) with a bacterial PLC from Clostridium perfringens in vitro. All these studies found a stimulation of glucose uptake in the range of 30 -80% of the insulin effect. Furthermore, it has been shown that insulin can stimulate the PLC activity in Rat-1 fibroblasts expressing two different isoforms of the IR (24). These data suggest a potential role for PLC in insulin-stimulated glucose uptake.
PLC catalyzes the hydrolysis of phosphatidylinositol 4,5biphosphate to DAG and inositol 1,4,5-triphosphate, two second messengers involved in the activation of protein kinase C and the release of Ca 2ϩ from intracellular stores (25,26), respectively. In mammalian cells this hydrolytic activity is stimulated by a multitude of hormones and growth factors. One PLC isoform, PLC␥, is an excellent substrate for the epidermal growth factor receptor, its catalytic activity being stimulated by tyrosine phosphorylation (27). PLC␥ has been implicated in mitogenic signaling by the PDGF receptor. Recently, homozy-gous disruption of the PLC␥1 gene in mice has been shown to result in death on day 9 of embryonic development (28). PLC␥ has two SH2 domains (designated N-terminal and C-terminal domains), a split pleckstrin homology domain, a C2 domain (29) and an SH3 domain (30).
In our studies we have observed an interaction between the IR and PLC␥ both in vitro and in vivo. Subsequently, we used U73122 a specific PLC inhibitor and/or single cell microinjection of SH2 domain GST fusion proteins derived from PLC␥ to block several parameters of metabolic signaling, GLUT4 translocation, 2-DOG uptake, and antilipolysis.

EXPERIMENTAL PROCEDURES
Cell Culture-3T3-L1 cells were maintained in DMEM/high glucose with 50 units/ml penicillin, 50 g/ml streptomycin, and 10% FCS in a 10% CO 2 environment. Cells were differentiated 2 days post-confluency by the addition of the same media containing 500 M isobutylxanthine, 25 M dexamethasone, and 4 g/ml insulin. After 3 days, cells were grown in media containing only insulin for another 3 days. Subsequently, media were changed every 3 days until the cells were well differentiated (day 10).
Receptor Association Assays-For the preparation of whole cell lysates from insulin-stimulated 3T3-L1 adipocytes, cells were starved overnight with 0.05% FCS and DMEM, 5 mM glucose. Upon stimulation with insulin (100 nM), 10-cm plates were scraped with 1.0 ml of cold lysis buffer (50 mM HEPES, 150 mM NaCl, 10 mM EDTA, 1% Triton X-100, 200 mM sodium fluoride, 20 mM sodium pyrophosphate, 2 mM phenylmethylsulfonyl fluoride, 10% glycerol, 4 mM sodium orthovanadate, aprotinin (800 KIU/ml), 15 mM benzamidine, and 2 mM dichloroacetic acid, pH 7.4, at 4°C and lightly homogenized using a Dounce homogenizer). The homogenates were centrifuged at 2500 rpm, and the fat was aspirated. 100-l aliquots of the supernatants were incubated for 90 min at 4°C with GST fusion proteins containing the N-terminal SH2 domains of p85 (amino acids 321-440), the SH2 domain of GAP (amino acid residues 177-278), the C-terminal SH2 domain of PLC␥, residues 1-216 of Syp including both SH2 domains, or the SH2 domain of SHC at a final concentration of 1 M and 100 l of a 50% suspension of glutathione-Sepharose beads (Amersham Pharmacia Biotech) that had been prewashed in lysis buffer. The recombinant SH2 GST fusion proteins were purified from Escherichia coli as described previously (31). The reaction mixture was centrifuged at 14,000 rpm; the pellets and 100-l aliquots of the supernatants were boiled in Laemmli's buffer, and the proteins were separated by SDS-PAGE (7.5%). The proteins were then electrotransferred to PVDF membranes (Immobilon, Millipore) and blotted with pY20 antibody (Transduction Laboratories) followed by a goat anti-mouse IgG-horseradish peroxidase conjugate and visualized by enhanced chemiluminescence (ECL, Amersham Pharmacia Biotech).
For the in vitro stimulated insulin receptors, IRs were purified by wheat germ agglutinin (WGA) affinity chromatography from 3T3-L1 adipocytes that had been starved overnight with 0.05% FCS and DMEM, 5 mM glucose, as published (31,32). After the receptors were stimulated with insulin (100 nM) in the presence of 2 mM vanadate at 4°C for 30 min, they were phosphorylated with 50 M ATP, 10 mM MnCl 2 in kinase buffer (10 mM HEPES, 0.05% Triton X-100, pH 7.4) for 30 min at 4°C. Following termination of phosphorylation with lysis buffer, the receptors were used for GST fusion protein association experiments on glutathione-Sepharose beads as described for whole cell lysates above.
Immunoprecipitation-Confluent 10-cm dishes of 3T3-L1 cells were stimulated with 100 nM insulin for 1 min, and cytoskeletal and soluble fractions of 3T3-L1 cells were prepared as described by Yang et al. (33) with minor modifications. Briefly, the soluble fraction was extracted for 5 min with cold microtubule stabilization buffer (0.1 M Pipes, pH 6.9, 2 M glycerol, 1 mM EGTA, and 1 mM magnesium acetate, 10 g/ml aprotinin, 200 M sodium orthovanadate, and 1 mM phenylmethylsulfonyl fluoride) containing 0.2% Triton X-100. The cell components remaining on the dish, termed the cytoskeletal fraction, were scraped with a high detergent buffer (50 mM HEPES, 150 mM NaCl, 1 mM EDTA, 1% deoxycholate, 1% Triton X-100, 0.1% sodium dodecyl sulfate, 10 g/ml aprotinin, and 200 mM sodium orthovanadate, pH 7.4). Following centrifugation at 2500 rpm, the fat was aspirated and the pellet discarded. The soluble and cytoskeletal fractions were incubated with a mixture of monoclonal antibodies to PLC␥ (5 g/300 l of lysate) for 1 h at 4°C followed by a 1-h incubation with anti-mouse IgG agarose (34). Antibodies were from Upstate Biotechnology Inc. (Lake Placid, NY). The immunocomplexes were pelleted by centrifugation, washed three times with the corresponding buffer, and then boiled in Laemmli's sample buffer. The proteins were then transferred to PVDF membranes and blotted with a rabbit antiphosphotyrosine antibody (35). After incubation with a secondary horseradish peroxidase-conjugated goat antirabbit antibody, the proteins were visualized by ECL.
Microinjection-3T3-L1 adipocytes were trypsinized on day 7 postdifferentiation and reseeded on acid-washed coverslips in preparation for microinjection on days 10 -12. The proteins to be microinjected, the GST fusion proteins of C-terminal and N-terminal SH2 domains derived from PLC␥, were dissolved in buffer containing 5 mM sodium phosphate, pH 7.2, 100 mM KCl. Microinjection of the fusion proteins was performed using a semiautomated Eppendorf microinjection system. Sheep IgG (10 mg/ml) was coinjected for detection of the injected cells. Cells were allowed to recover for 60 min following injection, stimulated for 20 min with 1.67 nM insulin, and fixed for immunostaining.
Immunostaining for GLUT4 -Immunostaining for GLUT4 was performed as previously published (10). Briefly, 3T3-L1 adipocytes were fixed in 3.7% formaldehyde in PBS for 5 min on ice and 5 min at room temperature. After washing, permeabilization, and blocking, cells were incubated with (1 g/ml) polyclonal anti-GLUT4 antibody (F349) in PBS with 1% FCS overnight at 4°C. Subsequently, the GLUT4 staining was visualized by fluorescein-conjugated anti-rabbit IgG (Jackson Immunoresearch Laboratories). The microinjected cells were identified by incubation with 7-amino-4-methylcoumarin-3-acetic acid-conjugated donkey anti-sheep IgG. Immunofluorescence microscopy was used to evaluate the results. Amino-4-methylcoumarin-3-acetic acid-positive, injected cells were scored as positive for GLUT4 translocation if they were observed to have a ring of fluorescence at the cell periphery. Coverslips were read blind by two independent investigators.
Primary Rat Adipocyte Isolation and 2-Deoxyglucose Uptake Determination-Primary rat adipocytes were isolated from epididymal fat pads of 180 -220 g male Sprague-Dawley rats, and 2-DOG uptake assays were performed as published with minor modifications (36). Following a 10-min pretreatment with Me 2 SO, U73122, or U73343, the cells were stimulated with 0 or 8.4 nM of insulin for 15 min. 2-DOG uptake assay was initiated at 37°C with the addition of 2-deoxy-D-[1,2- 2-Deoxyglucose Uptake in 3T3-L1 Adipocytes-3T3-L1 adipocytes were reseeded into 12-well plates 7 days after differentiation. Transport assays were performed between days 12 and 14. Cells were maintained in 2% calf serum DMEM, 5 mM glucose with no antibiotics for 12-16 h before the assay. Cells were incubated in assay buffer (DMEM, 0.5% BSA, 5 mM glucose, 25 mM HEPES, pH 7.4) in a 37°C 10% CO 2 incubator for 60 min then pretreated with Me 2 SO or U73122 or U73343 for 10 min. After 30 min of insulin stimulation, 1 Ci of 2-deoxy-D- [1,[2][3] H]glucose (NEN Life Science Products) was added and the uptake assay terminated after 15 min by adding 0.1 mM phloretin. After 4 washes in ice-cold PBS, the cells were solubilized in 0.1 N NaOH, and following neutralization radioactivity was measured by scintillation counting. Uptake of tracer-labeled 2-DOG was corrected for cellular protein.
Antilipolysis-The antilipolytic effect of insulin was measured in 3T3-L1 adipocytes 12-14 days post-differentiation. Cells were washed with a Krebs-Ringer phosphate-HEPES buffer (4% BSA) and incubated at 37°C with varying concentrations of insulin for 30 min followed by isoproterenol (100 nM) for 90 min. Subsequently, glycerol released into the buffer was determined by a colorimetric method (37). Glycerol release was corrected for cellular protein.
PI 3-Kinase Activity-In vitro phosphorylation of phosphatidylinositol was carried out in IRS-1 immunoprecipitates as described previously (38). 3T3-L1 adipocytes, which had been serum-starved overnight and pretreated for 10 min with 10 M U73122 or U73343, were incubated in the absence or the presence of 100 nM insulin for 5 min. After lysis in 50 mM HEPES, 150 mM NaCl, 10 mM EDTA, 1% Triton X-100, 200 mM sodium fluoride, 10% glycerol, 4 mM orthovanadate, aprotinin (800 KIU/ml), and 15 mM benzamidine, lysates were subjected to immunoprecipitation with anti-IRS-1 antibody (Upstate Biotechnology) overnight at 4°C, followed by protein A-Sepharose (Sigma) pelleting. Washed immunocomplexes were incubated with phosphatidylinositol (Avanti) and [␥-32 P]ATP (NEN Life Science Products) (3000 Ci/mmol) for 10 min at room temperature. Reactions were stopped with 20 l of 8 N HCl and 160 l of CHCl 3 :methanol (1:1) and centrifuged. The lower organic phase was removed and applied to potassium oxalate (1%)-coated silica gel thin layer chromatography (TLC) plates (Merck). After the lipid products were resolved on the TLC plates (38), they were visualized by autoradiography and quantitated by densitometry.
Akt Activation-3T3-L1 adipocytes were serum-starved 20 h in DMEM, 5 mM glucose. Cells were pretreated with Me 2 SO, wortmannin (100 nM, 30 min), or U73122 (3 M, 10 min) and then stimulated with insulin (16.7 nM) for 30 min. Cells were then lysed, boiled in Laemmli's sample buffer, and proteins were resolved by SDS-PAGE. Following transfer to PVDF membranes, proteins were immunoblotted with an Akt antibody (Santa Cruz).
MAP Kinase Activity-3T3-L1 adipocytes were maintained in 2% calf serum in DMEM, 5 mM glucose overnight and were further serumstarved in 0.5% BSA in the same medium for 1 h prior to pretreatment with the MEK inhibitor PD098059 (30 M, 30 min), U73122 (3 or 10 M, 10 min), or U73343 (3 or 10 M, 10 min). Cells were insulin-stimulated (16.7 nM) for 15 min, lysed directly in Laemmli's sample buffer containing 2 mM sodium orthovanadate, 100 mM sodium fluoride, 10 mM sodium pyrophosphate, 1 mM phenylomethylsulfonyl fluoride, 400 KIU/ml of aprotinin, 750 M benzamidine, and 1 mM dichloroacetic acid. The samples were boiled and sonicated, and the proteins were resolved on a 10% SDS-PAGE gel. After being transferred to PVDF paper the proteins were immunoblotted with an active MAP kinase antibody (Promega).
Statistical Analysis-Results were analyzed by analysis of variance, and groups were compared using the Bonferroni post-analysis test.

RESULTS
In preliminary studies aimed at the identification of proteins that interact with the IR, we have observed that IR, overexpressed in Chinese hamster ovary cells, is able to associate with GST fusion proteins of the signaling protein PLC␥. 2 To assess whether we could demonstrate this interaction in a cell model where metabolic actions of insulin can be evaluated, we investigated the IR-PLC␥ interaction in differentiated 3T3-L1 adipocytes. In this study, 3T3-L1 adipocytes were stimulated with insulin, and whole cell lysates were precipitated with a panel of SH2 domain GST fusion proteins derived from the p85 domain of PI 3-kinase, GAP, Syp, and finally SHC, a protein known to interact with the epidermal growth factor receptor but not the IR. Fig. 1, panel A, shows the in vitro precipitation of the IR from 3T3-L1 adipocytes with SH2 domain proteins derived from these signaling molecules. The IR and IRS-1 were precipitated by SH2 domains derived from p85, GAP, Syp, and PLC␥. As expected, the SH2 domain from SHC did not precipitate either the IR or IRS-1.
After having observed an interaction of the IR with the PLC␥ C-terminal SH2 domain, we proceeded to investigate whether other domains in the PLC␥ molecule could also interact with the IR. We incubated insulin-stimulated WGA-purified receptors from 3T3-L1 adipocytes with various GST fusion proteins derived from PLC␥. Our results show that fusion proteins containing either the N-or C-terminal SH2 domains of PLC␥ bind IR in vitro, but the SH3 domain, as expected, does not ( Fig. 1, panel B). Identical results were obtained using whole cell lysates from 3T3-L1 adipocytes (data not shown).
We have performed immunoprecipitation experiments to determine if the observed in vitro association indeed reflects an in vivo interaction. The EGF and PDGF receptors stimulate the translocation of PLC␥ from the cytosol to the cytoskeletal component in rat hepatocytes (33), and rat embryo fibroblasts (39), respectively. Therefore, we performed the immunoprecipita-tions in both soluble and cytoskeletal fractions of whole cell lysates by using a similar fractionation procedure. 3T3-L1 adipocytes in 10-cm dishes were stimulated with insulin (100 nM) for 1 min at 37°C, extracted for 5 min at room temperature with 500 l of microtubule stabilizing buffer (0.1 M Pipes, pH 6.9, 2 M glycerol, 1 mM EGTA, 1 mM magnesium acetate, 10 g/ml aprotinin, 200 M sodium orthovanadate, and 1 mM phenylmethylsulfonyl fluoride containing 0.2% Triton X-100 to generate the soluble fraction) (33). This extraction procedure leaves the cytoskeletal architecture of the cells intact on the dish. The cells remaining on the dishes were then solubilized in a high detergent buffer (50 mM HEPES, 150 mM NaCl, 1 mM EDTA, 1% deoxycholate, 1% Triton X-100, 0.1% sodium dodecyl sulfate, containing protease and phosphatase inhibitors). This second fraction was designated the cytoskeletal fraction. Both fractions were immunoprecipitated with a mixture of monoclonal antibodies to PLC␥1 for 1 h at 4°C. PLC␥2 is not expressed in 3T3-L1 adipocytes (data not shown). Precipitated proteins were separated by SDS-PAGE (sodium dodecyl sulfate-polyacrylamide gel electrophoresis), electrotransferred to PVDF membranes, and immunoblotted with PLC␥ or antiphosphotyrosine antibodies, followed by chemiluminescent detection. Fig. 1, panel C, illustrates the results of this study. In the soluble fraction (upper panel) antibodies to PLC␥ immunoprecipitated one major and two minor phosphoproteins. The phosphoprotein at 95 kDa corresponds to the ␤-subunit of the IR. A weakly insulin-stimulated phosphorylated protein of 145 kDa was observed. This protein was identified as PLC␥ upon stripping and reblotting with anti-PLC␥ antibodies (Fig. 1, panel C,  lower panel). Immunoprecipitation of the PLC␥ appears to be very efficient as little PLC␥ remains in the supernatant and PLC␥ appears to be equally distributed in the soluble and cytoskeletal fractions. In the cytoskeletal fraction, the 95-kDa IR ␤-subunit is immunoprecipitated by the PLC␥ antibody both in the basal and insulin-stimulated state. While the degree of basal phosphorylation of the IR varies between experiments, we have consistently observed that the majority of the basally phosphorylated IR appears to be associated with the PLC␥ in the cytoskeletal fraction. Similarly, PLC␥ was also detected in IR immunoprecipitates, confirming the in vivo association, and we have also observed an insulin-stimulated transient increase in PLC␥ activity and diacylglycerol levels. 3 The lack of a dramatic effect on PLC␥ phosphorylation in this study confirms prior findings by Nishibe et al. (27) and Wahl et al. (40) who showed a weak tyrosine phosphorylation of PLC␥ in response to insulin as compared with EGF or PDGF stimulation. Our results also suggest that extensive phosphorylation of the enzyme may not be necessary in insulin signaling in contrast to signaling by other growth factors. This finding is in agreement with reports indicating that even though PLC␥ is phosphorylated extensively by the EGF and PDGF receptors, activation of the enzyme may occur in the absence of phosphorylation. Both the activation and translocation or PLC␥ to the membrane fraction in response to PDGF stimulation was independent of phosphorylation in a recent study (41). PLC␥ was also shown to be activated independently of tyrosine phosphorylation in vitro in the presence of microtubule-associated protein tau and unsaturated fatty acids (42).
The association of PLC␥ and the IR in vivo and in vitro suggested a role for PLC␥ in IR signal transduction. The following experiments were designed to determine if PLC␥ might be involved in GLUT4 translocation and glucose uptake in 3T3-L1 adipocytes. Single cell microinjection of 3T3-L1 adipo-cytes coupled with immunofluorescent microscopy of GLUT4 proteins has become an established technique for studying IR signal transduction (10). Consequently, we examined the ability of SH2 domain proteins of PLC␥ to inhibit GLUT4 translocation following microinjection of single 3T3-L1 adipocytes. The translocation of GLUT4 to the cell surface is visualized using a GLUT4 antibody and a fluorescently labeled secondary antibody. A visual presentation of the inhibition of GLUT4 translocation by the microinjection of the GST fusion protein derived from the N-terminal SH2 domain of PLC␥ is shown in Fig. 2 4 The amino steroid U73122 has been used as a specific inhibitor of phosphatidylinositol-PLC in many cell systems. U73122 is postulated to inhibit PLC by a reversible interaction with the Ca 2ϩ -binding site on the enzyme (43). In opposum kidney cells and FTRL-5 thyroid cells, U73122 abolished parathyroid hormone-stimulated diglyceride accumulation (44) and inositol triphosphate production (45), respectively. In a study of the involvement of phosphatidylinositol and PC-specific PLCs in transforming growth factor-␤ signaling in A549 human lung carcinoma cells, Halstead et al. (46) were able to demonstrate that U73122 did not inhibit PC-specific PLC. One of the advantages of using this PLC inhibitor U73122 is the availability of an inactive dihydro analog, U73343, which can be used as a negative control. In our experiments, pretreatment of 3T3-L1 adipocytes with 10 M U73122 for 10 min greatly diminished insulin-induced translocation of GLUT4 compared with Me 2 SO controls. In Fig. 3 the top panel provides a visual depiction of the inhibition of insulin-stimulated GLUT4 translocation by pretreatment with U73122. The left panel shows GLUT4 staining in 3T3-L1 adipocytes in the unstimulated state. In the center panel the characteristic ring of GLUT4 staining in the plasma membrane in response to insulin stimulation is observed. Pretreatment with U73122 results in a pattern of GLUT4 staining similar to the unstimulated state (right panel). Results of three experiments are quantitated in the bottom panel of Fig. 3. U73122 caused a 50% decrease in cells positive for GLUT4 translocation. In contrast, U73343 did not have a significant effect.
To measure glucose uptake directly in a cell that is a genuine target of insulin action, we examined the effect of the U73122 on primary rat adipocytes. Pretreatment of primary rat adipocytes with U73122 completely blocked insulin-stimulated 2-DOG uptake compared with Me 2 SO vehicle controls. The weak analogue U73343 caused a modest but statistically significant reduction in insulin-stimulated uptake (Fig. 4, panel  A). A dose-response curve for inhibition by U73122 showed that the EC 50 for the effect was 2 M, a value similar to published values for PLC inhibition in other cell lines (Fig. 4, panel B). Considering that our immunoprecipitation and GLUT4 translocation experiments were performed in 3T3-L1 adipocytes, we proceeded to examine the effect of the inhibitor U73122 on glucose transport in 3T3-L1 adipocytes. In these cells insulin stimulation led to 5-8-fold increase in glucose uptake (Fig. 5). U73122 did not completely block glucose transport in 3T3-L1 adipocytes as it did in primary rat adipocytes. However, it did cause an approximately 50% reduction in insulin-stimulated glucose transport.
We were concerned that the inhibitor might be toxic to the cells. Consequently, we examined other insulin effects to see if they too were blocked by the inhibitor. Initially, we focused on a second metabolic end point of insulin signaling, namely antilipolysis. 3T3-L1 adipocytes were pretreated with Me 2 SO, U73122 (10 M), or U73343 (10 M) for 10 min. After being stimulated with varying concentrations of insulin (30 min), cells were treated with 100 nM isoproterenol. The stimulatory effect of isoproterenol on lipolysis was inhibited by insulin with an EC 50 of 0.2 nM (Fig. 6). First, the inhibitor U73122 had no effect on the ability of isoproterenol to stimulate lipolysis. Second, U73122 had no effect on the ability of insulin to suppress lipolysis. Therefore, the lipolytic pathway seems to be intact in 3T3-L1 adipocytes treated with U73122.
The autophosphorylation of IR and IRS-1 phosphorylation is the initial step in insulin signaling which lead to the pleiotropic effects of this hormone, including GLUT4 translocation and glucose uptake. It is possible for the observed inhibitory effect of U73122 on glucose uptake to be due an effect on receptor or IRS-1 phosphorylation. To address this issue we proceeded to test the effect of the PLC inhibitor on IR autophosphorylation and IRS-1 phosphorylation. After pretreating serum-starved 3T3-L1 adipocytes with Me 2 SO vehicle, U73122 (10 M), or U73343 (10 M) for 10 min, we stimulated 3T3-L1 adipocytes with increasing concentrations of insulin for 5 min. Cells were lysed in boiling sample buffer and proteins resolved by SDS-PAGE. Tyrosine-phosphorylated proteins were visualized by electrotransfer to PVDF membranes followed by antiphosphotyrosine immunoblotting. Fig. 7, panel A, shows that neither IR nor IRS-1 phosphorylation were affected by U73122 or the inactive analog U73343.
Many studies have shown that insulin-stimulated glucose transport requires PI 3-kinase activity. IRS-1 associated PI 3-kinase activity was measured on 3T3-L1 cells pretreated with the inhibitor U73122 and its inactive analog, U73343. Cells were lysed in the presence of phosphatase and protease inhibitors, and IRS-1 was immunoprecipitated overnight at 4°C. A PI 3-kinase assay was performed on the immunoprecipitates by using phosphatidylinositol as substrate. Lipid products were separated by thin layer chromatography and detected by autoradiography. Activity was quantified by densitometry. Insulin stimulated PI 3-kinase 13-fold; however, the magnitude of the stimulation was 50% in the presence of U73122 (Fig. 7, panel   B). The inactive analog U73343 had no significant effect.
Insulin activates a whole spectrum of serine threonine kinases including MAP kinase, p70 S6 kinase, and Akt. Two of these kinases, p70 S6 kinase and Akt, have been demonstrated to be downstream of PI 3-kinase (47)(48)(49)(50). Our purpose in investigating the effect of the PLC inhibitor U73122 on these downstream effectors of insulin signaling was 2-fold. We wanted to determine whether any of these kinases might be downstream of PLC and to preclude a nonspecific inhibition of signaling pathways by the inhibitor U73122.
By having observed a 50% reduction in insulin-stimulated PI 3-kinase activity, we proceeded to investigate if the two serine threonine kinases downstream of PI 3-kinase, p70 S6 kinase and Akt, could be activated normally in response to insulin in the presence of the PLC inhibitor. For the p70 S6 kinase studies, 3T3-L1 adipocytes were starved 20 h in serum-free me- dium, pretreated with increasing concentrations of U73122 for 10 min, or rapamycin, an inhibitor of p70 S6 kinase activation (20 nM, 30 min), and then stimulated with insulin (100 nM) for 30 min. Cells were scraped and solubilized in lysis buffer with protease and phosphatase inhibitors and boiled in Laemmli's sample buffer. Equal amounts of protein were resolved by SDS-PAGE and electrotransferred to PVDF membranes. Proteins were immunoblotted with an antibody against p70 S6 kinase (Upstate Biotechnology). Activation of p70 S6 kinase is associated with a shift in electrophoretic mobility on a reducing gel. Insulin caused a change in migration of p70 S6 kinase consistent with its activation (Fig. 8, panel A). Rapamycin completely blocked this shift. PLC inhibitor U73122 had no effect on the ability of insulin to activate p70 S6 kinase at any concentration. These results were confirmed using a phosphospecific p70 S6 kinase antibody (New England Biolabs) (data not shown).
Akt, the serine threonine kinase downstream of PI 3-kinase, has been demonstrated to be regulated by the lipid products of PI 3-kinase. The expression of a constitutively active Akt in 3T3-L1 cells has been shown to stimulate both glucose transport and GLUT4 translocation (51). Therefore, it was impor-tant to determine whether insulin could activate Akt in the presence of U73122. 3T3-L1 adipocytes were serum-starved 20 h and were pretreated with either wortmannin, an inhibitor of PI 3-kinase (100 nM, 30 min), or U73122 (3 M, 10 min). Cells were then stimulated with insulin (16.7 nM) for 30 min and lysed and solubilized in lysis buffer and boiled in Laemmli's buffer. Equal amounts of protein (30 g) were resolved by SDS-PAGE (7.5%). The proteins were transferred onto PVDF membranes and blotted with an Akt antibody (Santa Cruz Biotechnology). A shift in electrophoretic mobility of Akt corresponds to its activation (Fig. 8, panel B). In response to insulin, Akt was clearly shifted. Wortmannin abolished the shift. However, the Akt shift persisted in the presence of U73122. The ability of insulin to cause a shift in the mobility of Akt implies that Akt is being activated normally in response to insulin. Thus, the activation of p70 S6 kinase and Akt in the presence of U73122 suggests that the insulin-stimulated PI 3-kinase activity is sufficient to activate downstream effectors despite being reduced by 50%.
The mitogenic effects of insulin in fibroblasts have been shown to be mediated by the activation of the Ras/Raf/MAP kinase cascade (38,52). Activation or inhibition of this pathway in 3T3-L1 adipocytes has no effect GLUT4 translocation or glucose uptake (10,53,54). These cells are terminally differentiated and do not undergo insulin-stimulated DNA synthesis. However, stimulation of this pathway has been shown to lead to activation of immediate early gene transcription such as c-fos (55). PLC␥ has been shown to be involved in the activation of Ras in response to EGF, possibly through phosphorylation of GAP by PKC. Another possible point of entry of PLC into this pathway is through the phosphorylation of Raf by PKC (56). Consequently, we determined the effect of PLC inhibition on MAP kinase activation. 3T3-L1 adipocytes were pretreated with PD098059, an inhibitor of MAP kinase activating enzyme, MEK (30 M, 30 min), or U73122 (3 or 10 M, 10 min), or U73343 (10 M, 10 min). Cells were stimulated with insulin (16.7 nM) for 15 min. A rabbit polyclonal antibody to the dually phosphorylated MAP kinase (Promega) was used for immunoblotting. Results are shown in Fig. 8, panel C. Insulin-stimulated MAP kinase activity was appreciably blocked by MEK inhibitor PD098059 as expected. U73122 was also able to inhibit MAP kinase activity in 3T3-L1 adipocytes while the inactive analog U73343 was without effect. DISCUSSION Our understanding of vesicular trafficking of GLUT4 in response to insulin has been growing rapidly and has primarily been based on the groundwork laid by the study of the controlled release of neurotransmitters from small synaptic vesi- cles in neurons in response to depolarizing stimuli. In neuronal cells, proteins in transport vesicles named v-SNAREs (soluble N-ethylmaleimide-sensitive factor attachment protein receptor), such as vesicle-associated membrane protein 2 (VAMP2), cellubrevin, and synaptobrevin, dock with corresponding receptor proteins in the plasma membrane named t-SNAREs such as syntaxin and synaptosome associated protein 25 (SNAP-25) (57)(58)(59). This docking is then followed by recruitment of proteins ␣-SNAP and N-ethylmaleimide-sensitive factor to the docking complex. N-Ethylmaleimide-sensitive factor is an ATPase which is activated by binding to the N terminus of syntaxin. Hydrolysis of ATP causes disruption of the docking complex exposing the two lipid membranes in close proximity. The final step is fusion of the membranes which is regulated by Ca 2ϩ in neuronal cells and, in many cases, is the result of activation of PLC. In pituitary cells, GnRH stimulation of LH secretion is regulated by Ca 2ϩ which is released in response to PLC activation (60). The actual fusion reaction has been studied extensively in the trafficking of coated vesicles. Fusion is thought to be mediated by recruitment of small G-proteins ARF1 (ADP-ribosylation factor) and Rho leading to activation of PLD generating phosphatidic acid from phosphatidylcholine. The phosphatidic acid activates phosphatidyl-4-phosphate 5-kinase generating more phosphatidylinositol 4,5-biphosphate which in turn activates both PLD and the GTPase activity of ARF1, hence deactivating the complex and causing disassembly of the coat proteins in the Golgi (61). This feedback loop thus generates high local concentrations of phosphatidic acid which was thought to be fusogenic. More recently, however, overexpression of DAG kinase in yeast resulting in increased phosphatidic acid and decreased DAG levels was shown to cause an inhibition of vesicle assembly in the Golgi (62). Therefore, it appears that DAG may be the major fusogen required for the production of secretory vesicles in the Golgi rather than phosphatidic acid.
Recent studies have indicated that the trafficking of the GLUT4 containing vesicles utilizes proteins analogous to those involved in regulated neuronal vesicle transport (63). In particular, the t-SNARE protein syntaxin 4 and the v-SNARE proteins VAMP2, and cellubrevin (VAMP3) are important for the docking of the vesicle with the plasma membrane (64 -68). Furthermore, ARF and Rho have been shown to be recruited to the plasma membrane in response to insulin (69), and insulin is reported to activate PLD in adipocytes (70). Since the PI 3-kinase inhibitor, wortmannin, blocks insulin-induced PLD activation, it has been postulated that PI 3-kinase might act through Rho/ARF to activate PLD and DAG/PKC signaling in the plasma membrane (71). The more recent finding that the lipid products of PI 3-kinase activate atypical PKC-, which also appears to play an important role in the activation of glucose transport, has strengthened the PI 3-kinase/PKC interrelationship in insulin signaling in rat adipocytes (72).
The requirement for PI 3-kinase activation for glucose transport has been mentioned in the Introduction. However, whether PI 3-kinase activation is sufficient for the stimulation of glucose transport is still a matter of controversy. Both insulin and growth factors like PDGF stimulate PI 3-kinase. However, PDGF stimulates glucose uptake only 2-3-fold, whereas insulin stimulates it 8 -10-fold. This disparity has been resolved by the demonstration that both PDGF and insulin stimulate PI 3-kinase activity in the plasma membrane, but only insulin stimulates PI 3-kinase activity in the low density microsomal compartment (73)(74)(75). Furthermore, the minimal activation in glucose transport in response to PDGF stimulation has been attributed to an increase in GLUT1 concentration in plasma membranes of 3T3-L1 adipocytes (76).
While growth factor stimulation of cells expressing endogenous levels of receptors resulted in minimal stimulation of glucose transport, overexpression, for instance, of EGF receptors in 3T3-L1 adipocytes resulted in the stimulation of both GLUT4 translocation and glucose transport equivalent to levels achieved by insulin (77). Kamohara et al. (78) overexpressed both PDGF receptors and GLUT4 with an insert of a c-myc epitope in Chinese hamster ovary cells and were able to observe a 3.5-fold stimulation of GLUT4 translocation by PDGF. Interestingly, mutations in the PI 3-kinase and PLC␥-binding sites of the PDGF receptor in the study by Kamohara et al. (78) indicated a PLC␥-dependent component of glucose uptake in Chinese hamster ovary cells coexpressing GLUT4myc and PDGF receptors. Since overexpression appears to alter the contribution of growth factors to the stimulation of glucose transport in these studies, caution needs to be exercised in interpreting data based on overexpression.
Adenovirus-mediated expression of a constitutively active PI 3-kinase in 3T3-L1 adipocytes has yielded varied results. Overexpressing the bovine catalytic p110␣ sununit of PI 3-kinase fused to an amino acid tag derived from the C terminus of GLUT 2, Katagiri et al. (79) demonstrated a stimulation of glucose transport above that induced by insulin in control cells. Despite the 14-fold increase in the basal glucose transport rates in the overexpressers, a dose-dependent stimulation of glucose transport by insulin could still be observed. In a more recent study of adenovirus-mediated expression of constitutively active PI 3-kinase in 3T3-L1 adipocytes, coexpression of the p110␣ catalytic subunit with the inter-SH2 domain of the p85 regulatory subunit resulted in PI 3-kinase activities that exceeded insulin-stimulated activity levels but was only partially able to stimulate glucose transport (80). A recent study by the same group using a constitutively active PI 3-kinase targeted to GLUT4 vesicles demonstrated that targeting PI 3-kinase reduced its ability to activate glucose transport (81). Glucose uptake increased only 2-fold with the targeted PI 3-kinase compared with 5-fold with the untargeted constitutively active PI 3-kinase. The ability of insulin to stimulate glucose transport above levels achievable by PI 3-kinase overexpression in all these studies may indicate that alternative signaling pathways may be required to achieve maximal glucose transport.
As mentioned under "Result," Akt has been demonstrated to be downstream of PI 3-kinase. Both serine and threonine phosphorylation of Akt by PDK1 (82) and its activity is stimulated by the lipid products of PI 3-kinase (47). Recent data have linked Akt activation to the stimulation of glucose transport in 3T3-L1 adipocytes (51) and rat adipose cells (83) using constitutively active Akt constructs. In 3T3-L1 cells, retroviral expression of the myristoylated Akt resulted in a stimulation of glucose uptake that was 70% of the maximum insulin stimulation of controls. This effect on glucose transport was accompanied by about a 55% increase in GLUT1 expression in total membrane fractions and maximal stimulation of GLUT4 trans-location to the plasma membrane in the absence of insulin stimulation. In the latter study, the overexpression of wild type Akt cotransfected with epitope-tagged GLUT4 resulted in GLUT4 translocation to the membrane that was 80% that observed in control cells maximally stimulated with insulin. Furthermore, overexpression of the myristoylated Akt resulted in dramatic translocation of GLUT4 beyond levels achievable by insulin stimulation in control cells (150%). However, a kinase-defective Akt mutant only inhibited insulin-stimulated GLUT4 translocation 20% compared with controls. As mentioned earlier, these overexpression studies need to be interpreted with caution. Further studies are required to demonstrate to what extent Akt activation can account for the physiological stimulation of glucose transport by insulin in adipose tissue, and the mechanism by which Akt activates GLUT4 trafficking (84).
In the present study we have found that PLC inhibitor U73122 can partially inhibit endogenous PI 3-kinase activity in 3T3-L1 adipocytes, yet Akt can be activated by insulin in the presence of the inhibitor. Our results support the idea of an absolute requirement of PI 3-kinase and a modulatory effect of PLC on glucose transport. Our finding of a difference in sensitivity of glucose transport to U73122 in primary rat adipocytes FIG. 7. Effect of PLC inhibition on IR and IRS-1 phosphorylation and IRS-1-associated PI 3-kinase activity. Panel A, insulinstimulated phosphorylation of IR and IRS-1. 3T3-L1 adipocytes serumstarved overnight were pretreated with 10 M U73122, U73343, or Me 2 SO (DMSO) vehicle for 10 min and stimulated with increasing concentrations of insulin. Cells were lysed, and the proteins were resolved by SDS-PAGE. Following electrophoretic transfer to PVDF filters, phosphorylated proteins were immunoblotted with antiphosphotyrosine antibodies and visualized by ECL. IR and IRS1 phosphorylation are unaffected by the PLC inhibitor U73122. Panel B, IRS-1 associated PI3-kinase activity. Serum-starved 3T3-L1 adipocytes were pretreated for 10 min with 10 M U73122, U73343, or Me 2 SO (DMSO) vehicle and stimulated with 100 nM insulin for 5 min. After cell lysis, PI 3-kinase activity was determined in IRS-1 immunoprecipitates using phosphatidylinositol as substrate. Lipid products were resolved by TLC and visualized by autoradiography. Activity was quantitated by densitometry. Insulin stimulation increased PI 3-kinase activity about 13-fold. U73122 pretreatment caused about a 50% reduction in the insulin stimulation (*, p Ͻ 0.05 versus Me 2 SO). Mean Ϯ S.E. of three experiments is shown. Upon lysis the proteins were resolved by SDS-PAGE (7.5%) as in the preceding experiment. Proteins were immunoblotted with an Akt antibody. Panel C, MAP kinase. Cells were starved 20 h with 2% calf serum and 1 h with 0.5% BSA in DMEM, 5 mM glucose. Following pretreatment with 30 M PD098059 for 30 min, or 3 or 10 M U73122 or 10 M U73343 for 10 min, 3T3-L1 adipocytes were stimulated with 16.7 nM insulin for 15 min. Cells were then lysed with Laemmli's sample buffer, boiled, and sonicated. The proteins were subsequently resolved by SDS-PAGE (10%), transferred to PVDF membranes, and immunoblotted with an antibody against the dually phosphorylated active MAP kinase antibody. and 3T3-L1 adipocytes is intriguing. In this context, we would like to postulate that a differential expression of PKC isozymes in primary rat adipocytes and 3T3-L1 adipocytes may account for the relative importance of PLC in the regulation of glucose uptake in these cell types. Both primary rat adipocytes and 3T3-L1 adipocytes express ␣ and ␤ isozymes of PKC (18). In primary rat adipocytes insulin stimulates the translocation of ␣, ␤, and ⑀ isozymes of PKC from the cytosol to the plasma membrane (85). Although 3T3-L1 adipocytes express ␣ and ␤ isozymes of PKC and these isozymes appear to translocate to the plasma membrane fraction in response to insulin, overexpression of these isozymes did not affect basal or insulin-stimulated glucose transport. Overexpression of PKC-, however, resulted in increases in basal and insulin-stimulated glucose transport (86). PKC-⑀ has been shown to be expressed in 3T3-L1 fibroblasts and to be increased in expression upon differentiation into 3T3-L1 adipocytes (87). We suggest that a predominance of signaling by DAG-regulated conventional and novel PKC isozymes in primary adipocytes may reflect a greater dependence of glucose transport on PLC activity.
In a study published by Van Epps-Fung et al. (88) during the final stages of the preparation of this manuscript, EGF was demonstrated to stimulate glucose transport and GLUT4 translocation in 3T3-L1 adipocytes overexpressing the EGF receptor. EGF-stimulated glucose transport was shown to be inhibited 56% by 50 M U73122, a concentration that is five times the level used in the present study. Although insulin stimulation of PLC activity was low, insulin-stimulated glucose transport was also inhibited 56% by U73122. These results lend support to our findings of an interaction of the IR with PLC␥ in 3T3-L1 adipocytes. We have found that inhibition of PLC␥ by a specific inhibitor or microinjection of PLC␥ SH2 domains blocks insulin-stimulated GLUT4 translocation and glucose uptake. Our results suggest that PLC␥ is a signaling molecule involved in modulating glucose transport in insulinsensitive tissues.