Overexpression of catalytic subunit p110alpha of phosphatidylinositol 3-kinase increases glucose transport activity with translocation of glucose transporters in 3T3-L1 adipocytes.

To elucidate the mechanisms of phosphatidylinositol (PI) 3-kinase involvement in insulin-stimulated glucose transport activity, the epitope-tagged p110α subunit of PI 3-kinase was overexpressed in 3T3-L1 adipocytes using an adenovirus-mediated gene transduction system. Overexpression of p110α was confirmed by immunoblot using anti-tagged epitope antibody. p110α overexpression induced a 2.5-fold increase in PI 3-kinase activity associated with its regulatory subunits in the basal state, an increase exceeding that of the maximally insulin-stimulated control cells, while PI 3-kinase activity associated with phosphotyrosyl protein was only modestly elevated. Overexpression of p110α induced an approximately 14-fold increase in the basal glucose transport rate, which was also greater than that observed in the stimulated control. No apparent difference was observed in the cellular expression level of either GLUT1 or GLUT4 proteins between control and p110α-overexpressing 3T3-L1 adipocytes. Subcellular fractionation revealed translocation of glucose transporters from intracellular to plasma membranes in basal p110α-overexpressing cells. The translocation of GLUT4 protein to the plasma membrane was further confirmed using a membrane sheet assay. These findings indicate that an increment in PI 3-kinase activity induced by overexpression of p110α of PI 3-kinase stimulates glucose transport activity with translocation of glucose transporters, i.e., mimics the effect of insulin.

One of the major physiological functions of insulin is to stimulate glucose uptake into insulin-sensitive cells, such as adipocytes and myocytes. This effect is primarily due to translocation of GLUT4 glucose transporters from an intracellular compartment to the plasma membrane (1,2). Binding of insulin to its receptor results in receptor autophosphorylation and activation of the receptor tyrosine kinase, followed by tyrosine phosphorylation of several intermediate proteins including insulin receptor substrate (IRS) 1 1 (3, 4). Tyrosine-phosphorylated IRSs then bind to and thereby regulate Src homology 2 (SH2) domain containing proteins.
Several lines of evidence have indicated that PI 3-kinase activation is important in insulin-stimulated glucose transport. The PI 3-kinase inhibitors, such as wortmannin and LY294002, can block the insulin-stimulated glucose transport and GLUT4 translocation in rat and 3T3-L1 adipocytes (14,15). Furthermore, inhibition of endogenous PI 3-kinase by microinjection of glutathione S-transferase (GST)-p85␣ subunit fusion protein (16) or a dominant negative mutant of the p85␣ regulatory subunit of PI 3-kinase (17) inhibits GLUT4 translocation induced by insulin in 3T3-L1 adipocytes. These findings suggest that PI 3-kinase may be required for insulin-stimulated glucose transport. However, several groups have recently reported that platelet-derived growth factor (PDGF) stimulates PI 3-kinase activity but not glucose transport activity (18,19), although conflicting results were also reported (20).
We report herein that an increase in PI 3-kinase activity induced by adenovirus-mediated overexpression of the p110␣ subunit of PI 3-kinase (p110␣ PI3K ) stimulates glucose uptake with translocation of glucose transporters in 3T3-L1 adipocytes.
Cell Culture-3T3-L1 fibroblasts were maintained in Dulbecco's modified Eagle's medium (DMEM) containing 10% donor calf serum (Life Technologies, Inc.) in an atmosphere of 10% CO 2 at 37°C. Two days after the fibroblasts had reached confluence, differentiation was induced by treating cells with DMEM containing 0.5 mM 3-isobutyl-1methylxanthine, 4 g/ml dexamethasone, and 10% fetal bovine serum for 48 h. Cells were refed with DMEM supplemented with 10% fetal bovine serum every other day for the following 4 -10 days. More than 90% of the cells expressed the adipocyte phenotype (14).
Cloning and Construct-Reverse transcription-polymerase chain reaction was performed to amplify cDNA of p110␣ PI3K using the bovine brain RNA as a substrate and oligonucleotides based on its reported sequence (7) as primers, yielding cDNA of p110␣ PI3K covering the entire coding region. A portion of human GLUT2 cDNA corresponding to residues 510 -524 was ligated to p110␣ PI3K cDNA to encode p110␣ subunit of PI 3-kinase tagged with the epitope at its C terminus.
Gene Transduction-The recombinant adenoviruses Adex1CAlacZ (24) and Adex1CAp110␣, which encode Escherichia coli lacZ and the epitope-tagged bovine p110␣ PI3K , respectively, were constructed by homologous recombination between the expression cosmid cassette and the parental virus genome as described previously (25,26). 3T3-L1 adipocytes were incubated with DMEM containing the adenoviruses for 1 h at 37°C, and the growth medium was then added. Experiments were performed 3 to 4 days after the infection. When the adenovirus Adex1CAlacZ was applied at a multiplicity of infection of 200 -300 pfu/cell, lacZ gene expression was observed in more than 90% of 3T3-L1 adipocytes on postinfection day 3 (data not shown). Infection with Adex1CAlacZ or Adex1CAp110␣ resulted in no apparent differences in extent of differentiation into adipocytes, numbers of differentiated adipocytes, or morphological features in 3T3-L1 adipocytes, as compared with untreated cells on postinfection day 3. In addition, 3T3-L1 adipocytes infected with Adex1CAlacZ exhibited no significant differences in glucose transport activity as compared with noninfected cells on postinfection day 3 (data not shown). Therefore, in the present study, recombinant adenoviruses were applied at a multiplicity of infection of approximately 200 -300 pfu/cell, and 3T3-L1 adipocytes infected with Adex1CAlacZ virus were used as a control.
Immunoblotting-3T3-L1 adipocytes were lysed and boiled in Laemmli buffer containing 10 mM dithiothreitol and subjected to SDSpolyacrylamide (7.5% for p110␣ PI3K and 10% for glucose transporters) gel electrophoresis and transferred onto nitrocellulose filters. The filters were incubated with antisera, subsequently with 125 I-protein A (Amersham), and subjected to autoradiography.
Glucose Transport Assay-3T3-L1 adipocytes in a 12-well culture dish were serum-starved for 3 h in DMEM containing 0.2% bovine serum albumin, and glucose-free incubation was performed for 45 min in Krebs-Ringer phosphate buffer (14). Cells were then incubated with or without various concentrations of insulin for 15 min, and 2-deoxy-D-[ 3 H]glucose uptake was measured as described previously (23).
PI 3-Kinase Assay-The cells were serum-starved, as described above, and solubilized in ice-cold lysis buffer containing 20 mM Tris, pH 7.5, 137 mM NaCl, 1 mM CaCl 2 , 10 g/ml aprotinin, 1 mM phenylmethylsulfonyl fluoride, and 100 M sodium orthovanadate. Lysates were extracted by centrifugation at 15,000 ϫ g for 10 min and incubated with anti-phosphotyrosine antibody or anti-p85␣ antibody. Immunocomplexes were precipitated with protein A-Sepharose (Pharmacia Biotech Inc.). PI 3-kinase activity was assayed in the immunoprecipitates as reported previously (12). The results were quantitated using an image analyzer BAS2000 (Fujix).
Subcellular Fractionation-Fractionation of subcellular membranes from 3T3-L1 adipocytes was done essentially as described (27). Aliquots of subcellular membrane fractions containing equal amounts of protein were subjected to SDS-polyacrylamide gel electrophoresis and immunoblotted as described above.
Preparation of Plasma Membrane Sheets and Immunofluorescence-3T3-L1 fibroblasts plated on sterile glass coverslips were induced to differentiate into adipocytes. Plasma membrane sheets were prepared by sonication as described (28). Adherent plasma membranes were fixed in 2% paraformaldehyde and processed for indirect immunofluorescence using anti-GLUT4 antiserum (1:100 dilution) followed by rhodamine-conjugated secondary antibodies.

RESULTS
Overexpression of p110␣ PI3K was achieved utilizing an adenovirus-mediated gene transduction system in 3T3-L1 adipocytes, as demonstrated by immunoblotting with the antibody against the tagged epitope (Fig. 1A). Control (LacZ-expressing) and p110␣ PI3K -overexpressing 3T3-L1 (p110␣-L1) adipocytes were incubated with or without insulin for 15 min, and the PI 3-kinase activity was measured. In control cells, insulin stimulated the coimmunoprecipitation of PI 3-kinase activity with tyrosine-phosphorylated protein (maximum approximately 35fold) (Fig. 1, B and C) and that with the regulatory subunits of PI 3-kinase (maximum approximately 2-fold) (Fig. 1D) in a dose-dependent manner. Similar results were obtained in pa-rental 3T3-L1 adipocytes. Overexpression of p110␣ PI3K induced only a very modest elevation of PI 3-kinase activity in precipitates with the anti-phosphotyrosine antibody in the basal condition (Fig. 1C). Thus, a very small portion of exogenously expressed p110␣ PI3K bound to the tyrosine-phosphorylated protein in the basal state. Insulin addition to p110␣-L1 adipocytes markedly and dose-dependently increased the PI 3-kinase activity in the precipitates with the anti-phosphotyrosine antibody, and activity was significantly greater than that observed in insulin-stimulated control adipocytes (Fig. 1C).
In marked contrast, even without insulin stimulation, PI 3-kinase activity was greatly increased in precipitates with the regulatory subunits of PI 3-kinase, an increase exceeding that of the maximally insulin-stimulated control cells (Fig. 1D). Thus, exogenously expressed p110␣ PI3K bound to its regulatory subunits and exhibited its activity in 3T3-L1 adipocytes. Incubation of p110␣-L1 adipocytes with insulin further stimulated the PI 3-kinase activity measured in precipitates with the regulatory subunits of PI 3-kinase, and the increment was greater than that observed with insulin in control adipocytes (Fig. 1D). Thus, exogenously expressed p110␣ PI3K contributed to the insulin-induced increment in PI 3-kinase activity, indicating that insulin stimulated the PI 3-kinase activity in the complex of the endogenous regulatory subunits with not only endogenous catalytic subunits but also exogenously expressed p110␣ PI3K .
2-Deoxy-[ 3 H]glucose uptake in response to a 15-min incubation with 10 Ϫ10 -10 Ϫ6 M insulin was measured in control and p110␣-L1 adipocytes. Insulin dose-dependently stimulated 2-deoxyglucose uptake in control 3T3-L1 adipocytes (Fig. 2), and the maximally stimulated values were similar (approximately 12-fold) in control and parental 3T3-L1 adipocytes. Overexpression of p110␣ PI3K induced a marked increase (approximately 14-fold) in the basal glucose transport rate. Insulin further stimulated the glucose uptake in p110␣-L1 in a dosedependent manner (maximum approximately 1.4-fold) (Fig. 2). To begin addressing the mechanism whereby overexpression of p110␣ PI3K stimulated hexose transport activity, cellular expression levels of the two glucose transporter isoforms (GLUT1 and GLUT4), known to be expressed in 3T3-L1 adipocytes, were determined. As shown in Fig. 3, A and B, no apparent differences in cellular expression levels of GLUT1 and GLUT4 were observed between control and p110␣-L1 adipocytes.
We next tested whether overexpression of p110␣ PI3K affected the subcellular distribution of glucose transporters by immunoblot analysis of the plasma membrane (PM) fraction and the intracellular low density microsome (LDM) fraction (Fig. 3, C  and D). Insulin caused 4-and 6-fold increases in the amounts of plasma membrane GLUT1 and GLUT4, respectively, in control adipocytes. Corresponding decreases in GLUT1 and GLUT4 were observed in the LDM fraction of these cells. Strikingly, p110␣ PI3K overexpression induced translocation of GLUT1 and GLUT4 from the LDM fraction to the PM fraction, in a fashion similar to the insulin effect in control cells. Insulin addition to p110␣-L1 adipocytes had no apparent effects on the subcellular distributions of either GLUT1 or GLUT4.
We further confirmed the translocation of GLUT4 protein to the plasma membrane using the membrane sheet assay method. Fig. 4 demonstrates cell surface GLUT4 expression in control and p110␣-L1 adipocytes treated in the absence or presence of insulin (100 nM) for 15 min. In control cells, very little GLUT4 staining was observed on the plasma membrane, whereas insulin treatment increased surface GLUT4 staining substantially (Fig. 4). On the other hand, cell surface GLUT4 expression in p110␣-L1 adipocytes was already intense in the basal state, and insulin appeared to have virtually no further effects, consistent with the results obtained by the subcellular fractionation method. These findings demonstrate that overexpression of p110␣ PI3K exerts stimulatory effects on the translocation of glucose transporters from intracellular to plasma membranes.

DISCUSSION
In this study, overexpression of p110␣ PI3K , using an adenovirus-mediated gene transduction system, induced translocation of glucose transporters from intracellular low density microsomes to the plasma membrane and, thus, increased the glucose transport rate without changing the total amount of glucose transporter protein. Recently, several findings on the relationship between PI 3-kinase activity and glucose transport in 3T3-L1 adipocytes have been reported. PDGF caused no significant stimulation of glucose transport activity in 3T3-L1 aidipocytes, despite increasing PI 3-kinase activity to a level approaching that elicited by insulin (18,19). We obtained similar results in PDGF-treated 3T3-L1 adipocytes (data not shown). In addition, introduction of a thiophosphotyrosine peptide into permeabilized 3T3-L1 adipocytes stimulated PI 3-kinase to the same extent as insulin, while having little stimulatory effect on glucose transport activity (29). These authors concluded that another signaling pathway, in addition to the activation of PI 3-kinase, might be required for the stimulation of GLUT4 translocation and glucose transport activity.
However, the results of the present study indicate that PI 3-kinase activation alone stimulates translocation of GLUTs and, thus, activates glucose transport activity. One interpretation of these findings is that activation of PI 3-kinase by insulin is qualitatively different from that by either PDGF or the thiophosphotyrosine peptide, including the possibility that insulin activates PI 3-kinase in different subcellular locations than do the peptide and PDGF. A previous study (30) showed that activated PI 3-kinase has the same intracellular location as tyrosine-phosphorylated IRS1 in insulin-stimulated adipocytes, whereas the thiophosphotyrosine peptide appears to inhibit the association of PI 3-kinase activity with IRS1. The tyrosyl-phosphorylated IRS1-p85 complex formed in response to insulin was demonstrated to be localized in a very low density vesicle subpopulation. These vesicles could be distinguished from vesicles containing the insulin receptor which was endocytosed from the plasma membrane (30). On the other hand, tyrosyl-phosphorylated PDGF receptors, the p85 subunit of PI 3-kinase, and activated PI 3-kinase are all found in isolated clathrin-coated vesicles after PDGF stimulation in 3T3-L1 cells, indicating that both receptor and activated PI 3-kinase enter the endocytic pathway (31). These data suggest that the subcellular redistribution of PI 3-kinase activity in response to PDGF is different from that induced by insulin. Increased PI-3 kinase activity was observed in every LDM subpopulation, including a very low density vesicle subpopulation, when p110␣-L1 adipocytes were homogenized and frac- tionated on a sucrose gradient (data not shown). Subcellular redistribution of PI 3-kinase activity may be the key determinant in signal generation.
The increase in PI-3 kinase activity induced by overexpressing p110␣ PI3K is due to an increased amount of the protein, whereas insulin stimulates enzyme activity. In this respect, the increased activity resulting from overexpression is quantitative, rather than qualitative and, thus, might be nonphysiological. Overexpression of Ras (32) or Raf-1 (33) also increased the glucose transport activity. However, levels of total GLUT4 and GLUT1 expression were changed in stable cell lines of 3T3-L1 adipocytes overexpressing Ras and Raf-1, respectively. In contrast, in the present study, overexpression of p110␣ PI3K in 3T3-L1 adipocytes exerted essentially no effect on the total expression level while having a marked effect on the subcellular distribution of glucose transporters, which is qualitatively similar to the physiological effect of insulin.
Overexpression of p110␣ PI3K induced greater glucose transport activity than that observed in maximally insulin-stimulated control cells. A slightly larger amount (1.3-fold) of GLUT1 glucose transporter protein was observed in the PM fraction of p110␣-L1 cells as compared with that in insulin-stimulated control cells. However, it may be difficult that the small difference in the amount of GLUT1 protein in the PM fraction can entirely explain the difference in glucose transport activity between basal p110␣-L1 cells and insulin-stimulated control cells. Furthermore, insulin produced no apparent increase in the amount of glucose transporter protein in the PM fraction of p110␣-L1 cells, despite further stimulating glucose transport activity in these cells. These findings suggest that insulin and/or an increment in PI 3-kinase activity may also stimulate the intrinsic activity of glucose transporters.
One of the advantages of transient expression using an adenovirus-mediated gene transduction system is that long-term effects of gene transduction are much smaller than those in cells stably expressing the gene product. Furthermore, transient expression is not hampered by the problem of cell selection. Long-term selection of stable cell lines can lead to the establishment of specific cell lines which express many other genes necessary for survival during the selection period. We achieved expression of the intended protein in almost all 3T3-L1 adipocytes, while avoiding the selection process, using the adenovirus-mediated gene transduction system.
The downstream components of the pathway from PI 3-kinase through glucose transport have yet to be elucidated. Recently, several downstream targets of PI 3-kinase have been identified, such as Akt (34) and the small G-protein Rac (35). However, Rac does not couple PI 3-kinase to insulin-stimulated glucose transport in 3T3-L1 adipocytes (36). Other as yet unknown pathway(s) may be involved in insulin-stimulated glucose transport. p110-L1 adipocytes may serve as a model for studying the steps downstream from PI 3-kinase.