Advertisement

Effect of the Activation of Phosphatidylinositol 3-Kinase by a Thiophosphotyrosine Peptide on Glucose Transport in 3T3-L1 Adipocytes (∗)

Open AccessPublished:October 27, 1995DOI:https://doi.org/10.1074/jbc.270.43.26000
      Insulin causes the activation of phosphatidylinositol 3-kinase (PI 3-kinase) through complexation of tyrosine-phosphorylated YMXM motifs on insulin receptor substrate 1 with the Src homology 2 domains of PI 3-kinase. Previous studies with inhibitors have indicated that activation of PI 3-kinase is necessary for the stimulation of glucose transport in adipocytes. Here, we investigate whether this activation is sufficient for this effect. Short peptides containing two tyrosine-phosphorylated or thiophosphorylated YMXM motifs potently activated PI 3-kinase in the cytosol from 3T3-L1 adipocytes. Introduction of the phosphatase-resistant thiophosphorylated peptide into 3T3-L1 adipocytes through permeabilization with Staphylococcus aureus α-toxin stimulated PI 3-kinase as strongly as insulin. However, under the same conditions the peptide increased glucose transport into the permeabilized cells only 20% as well as insulin. Determination of the distribution of the glucose transporter isotype GLUT4 by confocal immunofluorescence showed that GLUT4 translocation to the plasma membrane can account for the effect of the peptide. These results suggest that one or more other insulin-triggered signaling pathways, besides the PI 3-kinase one, participate in the stimulation of glucose transport.

      INTRODUCTION

      A major metabolic effect of insulin is the stimulation of glucose transport into muscle and adipose cells. The immediate basis of this effect is the insulin-induced increase in the amount of the glucose transporter isotype GLUT4 in the plasma membrane of these cells(
      • Birnbaum M.J.
      ). This increase results from an enhanced rate of trafficking of GLUT4 to the plasma membrane and possibly also some slowing of the rate of endocytosis of GLUT4 from the plasma membrane(
      • Holman G.D.
      • Leggio L.L.
      • Cushman S.W.
      ). The trafficking of GLUT4 to the cell surface may proceed by the docking and fusion of specialized secretory vesicles enriched in GLUT4(
      • Birnbaum M.J.
      ,
      • Holman G.D.
      • Leggio L.L.
      • Cushman S.W.
      ,
      • James D.E.
      • Piper R.C.
      • Slot J.W.
      ).
      An important unsolved issue is the identity of the signaling pathway(s) from the insulin receptor that triggers this translocation of GLUT4 to the plasma membrane. The insulin receptor is a tyrosine kinase. Binding of insulin to the extracellular domain activates the kinase function in the intracellular domain, and the receptor both autophosphorylates and phosphorylates substrate proteins. Among the latter is insulin receptor substrate 1 (IRS-1).
      The abbreviations used are: IRS-1
      insulin receptor substrate 1
      GTPγS
      guanosine 5′-0-(3-thiotriphosphate)
      MAP
      mitogen-activated protein
      PDGF
      platelet-derived growth factor
      PI
      phosphatidylinositol
      Z
      norleucine
      SH2
      Src homology 2
      Tyr(s)
      thiophosphotyrosine.
      IRS-1 is phosphorylated on multiple tyrosine residues, and through these, binds and so regulates at least three Src homology 2 (SH2) domain-containing proteins(
      • White M.F.
      • Kahn C.R.
      ,
      • Keller S.R.
      • Lienhard G.E.
      ). Phosphatidylinositol 3-kinase (PI 3-kinase) is one of the proteins that associates with IRS-1. It consists of an 85-kDa regulatory subunit with two SH2 domains and a 110-kDa catalytic subunit. The binding of two tyrosine-phosphorylated YMXM motifs present in IRS-1 to the two SH2 domains markedly stimulates PI 3-kinase activity(
      • Lamphere L.
      • Carpenter C.L.
      • Sheng Z.F.
      • Kallen R.G.
      • Lienhard G.E.
      ,
      • Herbst J.J.
      • Andrews G.
      • Contillo L.
      • Lamphere L.
      • Gardner J.
      • Lienhard G.E.
      • Gibbs E.M.
      ). This signal transduction pathway from the insulin receptor to PI 3-kinase thus accounts for the rapid elevation of PI 3,4-bisphosphate and 3,4,5-trisphosphate seen in vivo in response to insulin(
      • Kelly K.L.
      • Ruderman N.B.
      ,
      • Cheatham B.
      • Vlahos C.J.
      • Cheatham L.
      • Wang L.
      • Blenis J.
      • Kahn C.R.
      ). The downstream components of the pathway have not yet been elucidated. The finding that PI 3,4,5-trisphosphate stimulates some isotypes of protein kinase C in vitro suggests a role for these kinases(
      • Toker A.
      • Meyer M.
      • Reddy K.K.
      • Falck J.R.
      • Aneja R.
      • Aneja S.
      • Parra A.
      • Burns D.J.
      • Ballas L.M.
      • Cantley L.C.
      ).
      Several lines of evidence have indicated that the activation of PI 3-kinase by insulin is required for GLUT4 translocation to the plasma membrane. First, wortmannin, a potent inhibitor of the enzyme, blocks insulin stimulation of glucose transport in rat and 3T3-L1 adipocytes and in muscle tissue(
      • Okada T.
      • Kawano Y.
      • Sakakibara T.
      • Hazeki O.
      • Ui M.
      ,
      • Clarke J.F.
      • Young P.W.
      • Yonezawa K.
      • Kasuga M.
      • Holman G.D.
      ,
      • Yeh J-I.
      • Gulve E.A.
      • Rameh L.
      • Birnbaum M.J.
      ). Second, the compound LY294002, another inhibitor that is structurally unrelated to wortmannin, has also been found to prevent the stimulation of transport in 3T3-L1 adipocytes(
      • Cheatham B.
      • Vlahos C.J.
      • Cheatham L.
      • Wang L.
      • Blenis J.
      • Kahn C.R.
      ). In each study, except for the one with muscle, it was directly shown that GLUT4 translocation was blocked. Lastly, the microinjection of a dominant negative mutant of the 85-kDa subunit of PI 3-kinase into 3T3-L1 adipocytes prevented GLUT4 translocation in response to insulin (
      • Kotani K.
      • Carozzi A.J.
      • Sakaue H.
      • Hara K.
      • Robinson L.J.
      • Clark S.F.
      • Yonezawa K.
      • James D.E.
      • Kasuga M.
      ). The mutant form of this subunit is one that is unable to bind to the catalytic subunit of PI 3-kinase but can still bind to IRS-1, since it retains both SH2 domains.
      Although these studies indicate that activation of PI 3-kinase is required for insulin stimulation of glucose transport via GLUT4 translocation, they do not address the question of whether it is sufficient. The present study examines this issue. Previously, we have shown that simple peptides containing two tyrosine-phosphorylated YMXM motifs bind to and activate PI 3-kinase as well as IRS-1 itself does(
      • Herbst J.J.
      • Andrews G.
      • Contillo L.
      • Lamphere L.
      • Gardner J.
      • Lienhard G.E.
      • Gibbs E.M.
      ). Here, these peptides have been used to activate PI 3-kinase selectively in permeabilized 3T3-L1 adipocytes, and the effect on glucose transport has been determined.

      EXPERIMENTAL PROCEDURES

      Materials

      2-Deoxy-D-[2,6-3H]glucose and [U-14C]sucrose were from Amersham. [γ-32P]ATP was from ICN. Phosphatidylinositol was from Avanti Polar Lipids. GTPγS was from Sigma. Staphylococcus aureus α-toxin (catalog no. 616385) was from Calbiochem.

      Peptide Synthesis

      The non-phosphorylated and Tyr(P) peptides were synthesized as described previously(
      • Herbst J.J.
      • Andrews G.
      • Contillo L.
      • Lamphere L.
      • Gardner J.
      • Lienhard G.E.
      • Gibbs E.M.
      ). The Tyr(S) peptide was synthesized on an Applied Biosystems model 433 peptide synthesizer using Fmoc (N-(9-fluorenyl)methoxycarbonyl) procedures. Upon completion of peptide synthesis, the amino-terminal 9-fluorenylmethoxycarbonyl group was removed, and the resin was acetylated(
      • Herbst J.J.
      • Andrews G.
      • Contillo L.
      • Lamphere L.
      • Gardner J.
      • Lienhard G.E.
      • Gibbs E.M.
      ). The peptide was then thiophosphorylated by the method of Kitas et al.(
      • Kitas E.
      • King E.
      • Bannwarth W.
      ) and was then cleaved and processed according to the method of King et al.(
      • King D.S.
      • Fields C.G.
      • Fields G.B.
      ).

      Peptide Binding

      The relative affinities of peptides for binding to both SH2 domains of PI 3-kinase were determined by a Biacore assay as described in detail previously(
      • Herbst J.J.
      • Andrews G.
      • Contillo L.
      • Lamphere L.
      • Gardner J.
      • Lienhard G.E.
      • Gibbs E.M.
      ). In this assay, the Tyr(P) peptide YZPZSPK, biotinylated on the ∈-amino group of lysine, was bound to strepavidin covalently linked to the dextran surface of the flow cell of the Biacore instrument (Pharmacia Biotech Inc.). A glutathione S-transferase fusion protein containing both SH2 domains of PI 3-kinase (residues 312-722 of the 85-kDa subunit), either alone or after mixture with various concentrations of the peptide being analyzed, was injected into the flow cell. Values dependent upon the refractive index, which varies with the amount of glutathione S-transferase-SH2 domain bound to the immobilized peptide, were measured in arbitrary reflectance units.

      Cell Culture

      3T3-L1 cells were carried as fibroblasts and differentiated into adipocytes as described in (
      • Frost S.C.
      • Lane M.D.
      ), with the exception that the cells were grown in 6-well plates coated with collagen. The collagen solution (0.3 mg/ml in 3% acetic acid of rat tail collagen from Sigma, catalog no. C8897) was applied to each well, swirled, and then removed. The plates were dried with the lid off in the cell culture hood under an ultraviolet lamp and then washed with serum-free Dulbecco's modified Eagle's medium. Collagen-coated dishes were used because permeabilized cells (see below) remained attached to these, whereas they occasionally peeled off uncoated plates as sheets. Cells used for immunofluorescent staining were grown on 18-mm square, number 1 glass coverslips. The adipocytes were used between days 7 and 11 after initiation of differentiation (day 0).

      Permeabilization of Adipocytes and Glucose Transport

      Fully differentiated adipocytes were serum starved in Dulbecco's modified Eagle's medium for 3 h at 37°C prior to the start of the experiment. The cells in each well were then washed three times with 2 ml of IC buffer (10 mM NaCl, 20 mM HEPES, 50 mM KCl, 2 mM K2HPO4, 90 mM potassium glutamate, 1 mM MgCl2, 4 mM EGTA, 2 mM CaCl2, pH 7.4) at 37°C and incubated at 37°C for 5 min in 0.5 ml of ICR buffer (IC buffer plus 4 mM MgATP, 3 mM sodium pyruvate, 100 μg/ml bovine serum albumin, pH 7.4) containing α-toxin at 250 hemolytic units/ml (Calbiochem units, approximately 8 μg/ml), to permeabilize the plasma membranes. The medium containing α-toxin was removed, the cells were treated at 37°C with insulin, GTPγS, or peptide in 0.4 ml of ICR buffer for 15 min, or put in ICR buffer alone. Following this treatment, glucose transport into the permeabilized cells was measured by the uptake of 2-deoxyglucose, according to a method previously described for 3T3-L1 adipocytes permeabilized with streptolysin O(
      • Clarke J.F.
      • Young P.W.
      • Yonezawa K.
      • Kasuga M.
      • Holman G.D.
      ). A small volume (20 μl) of radiolabeled 2-deoxy-D-glucose and sucrose was added to the 0.4 ml on the plate, such that the final concentration was 50 μM 2-deoxy-D-glucose (0.3 μCi of 2-deoxy-D-[2,6-3H]glucose) and 50 μM sucrose (0.06 μCi of [U-14C]sucrose), and the mixture incubated for 5 min at 37°C. The cells were then rapidly washed twice with 4 ml of ice-cold IC buffer and solubilized in 1 ml of 1% Triton X-100, and the amount of 3H and 14C in each sample was then measured. The amount of sucrose remaining with the cells provided a measure of nonspecific trapping of and uptake from the medium, and the raw values for 2-deoxyglucose uptake were corrected by subtraction of the corresponding amount of this compound. This correction amounted to approximately 30% of the 2-deoxyglucose uptake by adipocytes in the basal state. Each experiment consisted of transport assays on duplicate or triplicate wells of cells in the basal state and ones treated with 100 nM insulin, 200 μM GTPγS, and a peptide or other agent at one or more concentrations. By this design, each experiment contains positive controls for insulin responsiveness and permeabilization (insulin and GTPγS, respectively); occasional experiments in which the response to either agent was less than a 2-fold stimulation of transport have not been included.

      Preparation of Lysates and Assay of PI 3-Kinase

      The method here is the one that we have described previously(
      • Lamphere L.
      • Carpenter C.L.
      • Sheng Z.F.
      • Kallen R.G.
      • Lienhard G.E.
      ,
      • Herbst J.J.
      • Andrews G.
      • Contillo L.
      • Lamphere L.
      • Gardner J.
      • Lienhard G.E.
      • Gibbs E.M.
      ). Briefly, 3T3-L1 adipocytes were homogenized, and the organelles were sedimented at 140,000 × g for 1 h. The cytosol was assayed for activity by the phosphorylation of PI with [γP32]ATP, followed by separation of the PI 3-phosphate by thin layer chromatography. Previously, we have shown that this cytosol, from both basal and insulin-treated cells, contains approximately 50% of the PI 3-kinase protein and that as the result of insulin treatment, the enzyme in the cytosol is activated due to complexation with the Tyr(P) form of IRS-1(
      • Lamphere L.
      • Carpenter C.L.
      • Sheng Z.F.
      • Kallen R.G.
      • Lienhard G.E.
      ). This method was applied to both intact cells and to cells that had been permeabilized with α-toxin and exposed to insulin, GTPγS, peptide, or ICR buffer alone for 15 min. To ensure reduction of the peptide in the medium below a concentration that activates PI 3-kinase, the cells in each well were washed three times with 2 ml of IC buffer and once with 20 mM Tris-Cl, 140 mM NaCl, pH 7.6, over a 0.5-min period at 20°C before being scrapped and homogenized in 0.33 ml of the homogenization buffer.

      GLUT4 Immunofluorescence

      3T3-L1 adipocytes on coverslips were serum starved, permeabilized, and incubated under the conditions described for the assay of hexose transport. Following these treatments, the cells were rinsed quickly with IC buffer, fixed in 3.5% paraformaldehyde for 5 min, and then permeabilized in 5 ml of TBSB (20 mM Tris, 150 mM NaCl, 1% bovine serum albumin, 0.025% NaN3, pH 7.4) containing 0.2% saponin for 5 min. The cells were then incubated with 5 μg/ml affinity-purified rabbit antibodies against the carboxyl-terminal peptide of GLUT4 (
      • Calderhead D.M.
      • Kitagawa K.
      • Tanner L.I.
      • Holman G.D.
      • Lienhard G.E.
      ) in TBSB for 30 min, washed twice in 5 ml of TBSB for 5 min, and treated with goat antibodies against rabbit immunoglobulin at 6 μg/ml conjugated to fluorescein (Vector Laboratories) in TBSB for 30 min. The coverslips were washed twice with 5 ml of TBSB for 5 min, mounted in fluorescein anti-fade reagent (Testog, Inc.), and sealed at the edges with nail varnish. The slides were then viewed on a Bio-Rad MAC 1000 confocal microscope, and the immunofluorescent images were captured and printed.

      RESULTS

      Peptides Used

      In our previous study, we found that peptides containing two tyrosine-phosphorylated YMXM motifs potently activate PI 3-kinase by binding simultaneously to the two SH2 domains on the 85-kDa regulatory subunit of this enzyme(
      • Herbst J.J.
      • Andrews G.
      • Contillo L.
      • Lamphere L.
      • Gardner J.
      • Lienhard G.E.
      • Gibbs E.M.
      ). For ease of synthesis and stability, norleucine (designated Z), an isomorph of methionine, was used instead of methionine. In this study, we have employed one of these peptides, YZPZSGSYZPZS, in which both tyrosines were phosphorylated, thiophosphorylated, or nonphosphorylated (Table 1). The thiophosphorylated peptide was examined because thiophosphoryltyrosine peptides have been found to be resistant to hydrolysis by several tyrosine phosphatases(
      • Tonks N.K.
      • Diltz C.D.
      • Fischer E.H.
      ,
      • Kuhné M.R.
      • Zhizhuang Z.
      • Rowles J.
      • Lavan B.E.
      • Shen S-H.
      • Fischer E.H.
      • Lienhard G.E.
      ). The nonphosphorylated peptide served as a control.

      Peptide Binding to PI 3-Kinase SH2 Domains

      To determine the effect of substituting the thiophosphoryl group for the phosphoryl group on binding to the PI 3-kinase SH2 domains, the Tyr(S) and Tyr(P) peptides were compared in a Biacore assay in which the effectiveness of the peptide to compete with immobilized Tyr(P) YZPZSPK peptide for binding a glutathione S-transferase fusion protein containing both SH2 domains was measured. The Tyr(S) peptide was almost as effective as the Tyr(P) peptide in blocking binding of the SH2 domains to the immobilized peptide (Fig. 1). Thus, substitution of the thiophosphoryl groups had no major effect on the affinity for the SH2 domains. In contrast, as expected, the nonphosphorylated peptide did not bind to the SH2 domains.
      Figure thumbnail gr1
      Figure 1Binding of peptides to the SH2 domains of PI 3-kinase. The relative affinities of the peptides for binding to the glutathione S-transferase fusion protein containing both SH2 domains of PI 3-kinase were measured in the Biacore assay described under “Experimental Procedures.” Results are expressed as percent of the binding of the PI 3-kinase SH2 domain protein to the immobilized peptide in the absence of competing peptide. ○, PYZPZSGSPYZPZS; •, SYZPZSGSSYZPZS; □, YZPZSGSYZPZS.

      Stimulation of PI 3-Kinase by Peptides

      To determine whether the strong binding of the Tyr(S) peptide caused a stimulation of PI 3-kinase activity, we measured its effect on the activity of PI 3-kinase in the cytosol of 3T3-L1 adipocytes. The Tyr(S) peptide stimulated activity as potently as the Tyr(P) peptide (Fig. 2). Maximal stimulation was 4.5-fold, with half-maximal effect at 5 nM. This degree of stimulation was approximately the same as that found upon insulin treatment of the cells, which is due to the complexation of Tyr(P) form of IRS-1 with PI 3-kinase(
      • Lamphere L.
      • Carpenter C.L.
      • Sheng Z.F.
      • Kallen R.G.
      • Lienhard G.E.
      ). On the other hand, GTPγS, a compound that stimulates glucose transport in permeabilized cells (see below), did not enhance PI 3-kinase activity.
      Figure thumbnail gr2
      Figure 2Stimulation of PI 3-kinase activity by Tyr(P) and Tyr(S) peptides. The cytosolic fraction from basal 3T3-L1 adipocytes was incubated with peptide in 100 μl at room temperature for 15 min and then assayed for PI 3-kinase by the addition of a mixture of [γ32P]ATP and PI (60 μl). ○, PYZPZSGSPYZPZS; •, SYZPZSGSSYZPZS. The concentrations given are the ones in the initial 100-μl incubation mixture. The values are the means ± S.D. for triplicate measurements in a representative experiment of three that gave similar results. The activity in the cytosol prepared at the same time from 3T3-L1 adipocytes treated with 100 nM insulin for 3 min and in cytosol to which 200 μM GTPγS was added are also shown.

      Effect of Peptides on Glucose Transport

      The effect of the Tyr(P) and Tyr(S) peptides on glucose transport was examined in α-toxin permeabilized 3T3-L1 adipocytes. α-toxin is a 34-kDa protein that inserts into the plasma membrane and then oligomerizes to form a 3-nm-diameter aqueous pore that allows passage of molecules of about 5 kDa and less(
      • Fussle R.
      • Bhakdi S.
      • Sziegoleit A.
      • Tranumjensen J.
      • Kranz T.
      • Wellensiek H.
      ). It has been previously used to introduce membrane-impermeant molecules into rat and 3T3-L1 adipocytes(
      • Baldini G.
      • Hohman R.
      • Charron M.J.
      • Lodish H.F.
      ,
      • Klarlund J.K.
      • Khalaf N.
      • Kozma L.
      • Czech M.P.
      ). An appropriate concentration of α-toxin for permeabilization was determined by allowing intact cells to take up 2-deoxy-[3H]glucose, which is trapped by phosphorylation as 2-deoxyglucose 6-phosphate(
      • Rosen O.M.
      • Smith C.J.
      • Fung C.
      • Rubin C.S.
      ), and then incubating cells with various concentrations of α-toxin in the ICR buffer (see “Experimental Procedures”) and measuring the appearance of radiolabel in the medium. We chose a concentration of α-toxin that caused a release of 50% of the 2-deoxyglucose 6-phosphate in 5 min.
      Previous studies have found that glucose transport can be measured in adipocytes permeabilized by electric discharge (
      • Shibata H.
      • Robinson F.W.
      • Benzing C.F.
      • Kono T.
      ) or with the detergent streptolysin O(
      • Clarke J.F.
      • Young P.W.
      • Yonezawa K.
      • Kasuga M.
      • Holman G.D.
      ), because the rate of hexose uptake catalyzed by the glucose transporter is much greater than the rate of either its uptake or of the loss of the 2-deoxyglucose 6-phosphate product through the pores. This also proved to be the case in the 3T3-L1 adipocytes permeabilized with α-toxin. Cytochalasin B, a specific inhibitor of glucose transport(
      • Calderhead D.M.
      • Kitagawa K.
      • Tanner L.I.
      • Holman G.D.
      • Lienhard G.E.
      ), at 25 μM, reduced the 2-deoxyglucose uptake into permeabilized cells in the basal and insulin-treated states to 24 and 9% of the uninhibited rates, respectively, such that residual 2-deoxyglucose uptake was the same for the cells in the two states.
      The effects of insulin, GTPγS, and the Tyr(S) peptide on the uptake of 2-deoxyglucose by the α-toxin permeabilized 3T3-L1 adipocytes are shown in Fig. 3. Both insulin and GTPγS stimulated transport approximately 3-fold. These results are consistent with the previous finding that insulin stimulation of glucose transport is preserved in streptolysin O-permeabilized 3T3-L1 adipocytes and that GTPγS also enhances transport in these permeabilized cells(
      • Clarke J.F.
      • Young P.W.
      • Yonezawa K.
      • Kasuga M.
      • Holman G.D.
      ). Earlier studies have also shown that both insulin and GTPγS elicit the translocation of GLUT4 to the plasma membrane in α-toxin permeabilized rat adipocytes (
      • Baldini G.
      • Hohman R.
      • Charron M.J.
      • Lodish H.F.
      ) and streptolysin O-permeabilized 3T3-L1 adipocytes(
      • Robinson L.J.
      • Pang S.
      • Harris D.S.
      • Heuser J.
      • James D.E.
      ). The stimulation of transport by GTPγS that we observed provides further evidence that the 3T3-L1 adipocytes were permeabilized by α-toxin, since GTPγS did not enhance transport in the absence of α-toxin treatment.
      Figure thumbnail gr3
      Figure 3Glucose transport in permeabilized adipocytes. 3T3-L1 adipocytes were permeabilized with α-toxin and then left untreated (basal) or exposed to 100 nM insulin, 200 μM GTPγS, or peptide for 15 min, followed by the assay of 2-deoxyglucose (dGlc) uptake for 5 min as described under “Experimental Procedures.” In the case of treatment with the peptide plus insulin, the peptide was added 5 min before insulin, and exposure to insulin was for 10 min before the transport assay. Panel A, 10 μMSYZPZSGSSYZPZS; panel B, 10 and 100 μM YZPZSGSYZPZS. The values are the means ± the range for measurements on uptake by two 35-mm wells of cells in a representative experiment. Similar results were obtained with the thiophosphorylated peptide in four other separate experiments of this type. ∗, p ≤ 0.05.
      The Tyr(S) peptide, at 10 μM, consistently caused a small (approximately 1.5-fold) but statistically significant increase in glucose transport, whereas the nonphosphorylated peptide had no effect (Fig. 3, panels A and B, respectively). In other experiments where the Tyr(S) peptide was tested at higher concentrations (30 and 100 μM), the stimulation of transport was no greater; also, lower concentrations of the peptide (1 and 0.1 μM) did not enhance the transport significantly. Moreover, the combination of the Tyr(S) peptide (10 μM) and insulin (100 nM) gave the same rate of transport as insulin alone (panel A). Thus, the peptide did not inhibit the effect of insulin.
      In one experiment of an analog of the product of the PI 3-kinase, reaction was examined for its effect on glucose transport. Permeabilized cells were treated with a water-soluble PI 3,4,5-trisphosphate (the dioctanoyl compound(
      • Toker A.
      • Meyer M.
      • Reddy K.K.
      • Falck J.R.
      • Aneja R.
      • Aneja S.
      • Parra A.
      • Burns D.J.
      • Ballas L.M.
      • Cantley L.C.
      ); kindly provided by Drs. J. R. Falck and L. C. Cantley) at 20 μM. No stimulation of glucose transport occurred (data not shown).

      PI 3-Kinase Activity in Permeabilized Cells

      To obtain direct evidence that the Tyr(S) peptide entered the permeabilized 3T3-L1 adipocytes and bound to and activated PI 3-kinase, we measured the activity of the enzyme in the cytosol prepared from permeabilized cells (Fig. 4). As was the case for intact cells (Fig. 2), insulin treatment of permeabilized cells enhanced the activity. Treatment with the Tyr(S) peptide caused an increase in activity of the same magnitude as insulin. The effect of Tyr(S) peptide cannot be explained by residual peptide from the medium that was not removed in washing the cells before homogenization, since peptide added to intact cells did not stimulate PI 3-kinase activity (Fig. 4). Thus, these results indicate that the Tyr(S) peptide activated PI 3-kinase in the permeabilized cells. In contrast, GTPγS treatment of the permeabilized cells did not lead to enhanced PI 3-kinase activity in the cytosol. Although this finding suggests that GTPγS does not activate the enzyme in the permeabilized cells, it is also possible that an activation was reversed upon preparation of the cytosol.
      Figure thumbnail gr4
      Figure 4PI 3-kinase activity in the cytosol from permeabilized 3T3-L1 adipocytes. Cells were permeabilized with α-toxin and treated with no addition (basal), 100 μM insulin, 10 μMSYZPZSGSSYZPZS, or 200 μM GTPγS for 15 min, and then the cytosolic fraction was prepared and assayed for PI 3-kinase activity as described under “Experimental Procedures.” Cells exposed to Tyr(S) peptide in the absence of toxin (intact cells) were also assayed. The values are the means ± S.D. for triplicate measurements in a representative experiment. Two other separate experiments gave similar results.

      GLUT4 Distribution by Immunofluorescence

      The subcellular distribution of GLUT4 in α-toxin permeabilized 3T3-L1 adipocytes treated with the various agents was assessed by confocal immunofluorescence. In cells in the basal state, there was strong staining in the perinuclear region and punctate staining through the cell (Fig. 5A). Previously, in brown fat cells GLUT4 has been found by immunoelectron microscopy to be in tubules and vesicles concentrated in the juxtanuclear trans Golgi region and also scattered throughout the cytoplasm(
      • Slot J.W.
      • Geuze H.J.
      • Gigengack S.
      • Lienhard G.E.
      • James D.E.
      ), and the staining pattern in the 3T3-L1 adipocytes is thus consistent with the same locations in this cell type. Upon treatment with insulin (Fig. 5B) or GTPγS (Fig. 5C), the 3T3-L1 adipocytes developed a distinct ring of staining at the cell border. This finding thus provides direct evidence for translocation of GLUT4 to the cell surface in the α-toxin permeabilized cells. Interestingly, the Tyr(S) peptide also caused a noticeable increase in staining at the cell border (Fig. 5D), even though, as described above, it caused an increase in glucose transport only 20% as large as did insulin and GTPγS. Several possible interpretations of this result are considered under “Discussion.”
      Figure thumbnail gr5
      Figure 5GLUT4 distribution in permeabilized 3T3-L1 adipocytes by immunofluorescence. 3T3-L1 adipocytes on coverslips were permeabilized with α-toxin, treated with various agents, and fixed with formaldehyde; GLUT4 was then detected by confocal immunofluorescence, as described under “Experimental Procedures.” Shown are representative images of basal cells (A) and cells treated with 100 nM insulin (B), 200 μM GTPγS (C), and 10 μMSYZPZSGSSYZPZS (D). The large clear circles within the cells are fat droplets. Similar results were obtained in four repetitions of this entire experiment.

      DISCUSSION

      The main finding of this study is that activation of PI 3-kinase is not sufficient to stimulate glucose transport in 3T3-L1 adipocytes to the same extent as insulin. This finding suggests that one or more additional signaling pathways are involved. While this study was in progress, Wiese et al.(
      • Wiese R.J.
      • Mastick C.C.
      • Lazar D.F.
      • Saltiel A.R.
      ) reached the same conclusion by a different approach. These authors found that treatment of 3T3-L1 adipocytes with PDGF caused no significant stimulation of glucose transport, even though it led to an increase in PI 3-kinase activity similar to that elicited by insulin. Despite the fact that these two lines of investigation point to the same conclusion, it should be noted that alternative explanations for the results remain possible. For example, the stimulation of glucose transport may require that PI 3-kinase be activated at a specific subcellular location, and insulin may, to some extent, activate PI 3-kinase in different subcellular locations than do the Tyr(S) peptide and PDGF. Also, PDGF activates some signaling pathways not activated by insulin(
      • White M.F.
      • Kahn C.R.
      ,
      • Keller S.R.
      • Lienhard G.E.
      ,
      • Claesson-Welsh L.
      ), and one of these may have inhibited the stimulation of glucose transport.
      Our observations that the PI 3-kinase-activating peptides did increase glucose transport to a modest extent and did cause redistribution of GLUT4 to the cell surface, as assessed by immunofluorescence, are consistent with the conclusion, derived from the use of inhibitors (see the Introduction), that activation of PI 3-kinase is involved in GLUT4 translocation and transport stimulation. There are, however, complicating factors that should be considered in the interpretation of these observations. With regard to the stimulation of transport, it is possible that translocation of the transporter isotype GLUT1 contributes to part or all of this effect. 3T3-L1 adipocytes contain considerable intracellular GLUT1, part of which is in vesicles with GLUT4 and part of which is in separate vesicles, and a portion of this isotype also translocates to the cell surface in response to insulin (
      • Calderhead D.M.
      • Kitagawa K.
      • Tanner L.I.
      • Holman G.D.
      • Lienhard G.E.
      ,
      • Piper R.C.
      • Hess L.J.
      • James D.E.
      ). Studies with inhibitors indicate that the activation of PI 3-kinase is also necessary for the smaller increases in cell surface GLUT1 and its associated transport activity caused by insulin(
      • Clarke J.F.
      • Young P.W.
      • Yonezawa K.
      • Kasuga M.
      • Holman G.D.
      ,
      • Hara K.
      • Yonezawa K.
      • Sakaue H.
      • Ando A.
      • Kotani K.
      • Kitamura T.
      • Kitamura Y.
      • Ueda H.
      • Stephens L.
      • Jackson T.R.
      • Hawkins P.T.
      • Dhand R.
      • Clark A.E.
      • Holman G.D.
      • Waterfield M.D.
      • Kasuga M.
      ). Moreover, we have previously shown that microinjection of the Tyr(P) peptide into Xenopus oocytes stimulates glucose transport to the same extent as does insulin-like growth factor I(
      • Gould G.W.
      • Jess T.J.
      • Andrews G.C.
      • Herbst J.J.
      • Plevin R.J.
      • Gibbs E.M.
      ). The immediate basis of this effect on transport in oocytes has not been determined but may be GLUT1 translocation to the plasma membrane.
      In regard to the redistribution of GLUT4 elicited by the Tyr(S) peptide, the qualitative nature of the immunofluorescence methodology precludes a definitive interpretation. It is possible that modest translocation of GLUT4 to the plasma membrane, consistent with the observed increase in transport (about 20% of the insulin effect), appears as a distinct increase in staining at the cell border. Alternatively, the peptide may cause GLUT4 vesicles to migrate to the plasma membrane as effectively as insulin does but not trigger their fusion with the membrane. This situation would probably not be distinguishable from completed translocation by immunofluorescence. Another possibility is that the peptide causes an increase in GLUT4 in the plasma membrane as large as insulin does but that an additional process is required for stimulation of transport. The determination of GLUT4 subcellular distribution by other methods, such as quantitative immunoelectron microscopy(
      • Slot J.W.
      • Geuze H.J.
      • Gigengack S.
      • Lienhard G.E.
      • James D.E.
      ), will be required to decide which of these possibilities is in fact occurring.
      Although our data suggest that a signaling pathway, in addition to the activation of PI 3-kinase, is necessary for the stimulation of GLUT4 translocation and glucose transport, there is at present evidence against any of the other known signaling pathways from the insulin receptor being required for this effect (reviewed in (
      • Lavan B.E.
      • Lienhard G.E.
      )). Most relevant here are studies in which specific inhibitors of signaling pathways were used, since if two signaling mechanisms are necessary, specific inhibition of only one will block the insulin stimulation of transport, whereas specific activation of only one will not elicit it. Insulin treatment of 3T3-L1 adipocytes also causes the stimulation of the 70-kDa ribosomal S6 kinase, the activation of the SH2 domain-containing tyrosine phosphatase PTP2C, the elevation of the GTP form of Ras, and the activation of the mitogen-activated protein (MAP) kinase cascade. Specific inhibition of the activation of the 70-kDa S6 kinase by rapamycin does not block insulin-stimulated GLUT4 translocation or glucose transport(
      • Fingar D.C.
      • Hausdorff S.F.
      • Blenis J.
      • Birnbaum M.J.
      ). Similarly, inhibition of the activation of PTP2C through the microinjection of antibodies or isolated SH2 domains does not prevent insulin-stimulated GLUT4 translocation(
      • Hausdorff S.F.
      • Bennett A.M.
      • Neel B.G.
      • Birnbaum M.J.
      ). In the case of Ras, inhibition of GTP loading by microinjection of a dominant negative mutant or a neutralizing antibody or by prevention of Ras isoprenilation does not impair GLUT4 translocation or the increase in glucose transport(
      • Hausdorff S.F.
      • Franioni J.V.
      • Birnbaum M.J.
      ,
      • Beusch J.E-B.
      • Bhuripanyo P.
      • Carel K.
      • Leitner J.W.
      • Hsieh P.
      • DePaolo D.
      • Draznin B.
      ). Although no complete reports have yet appeared regarding specific inhibitors of the MAP kinase cascade, a recent abstract presents data showing that a specific inhibitor of MAP kinase kinase does not inhibit insulin stimulation of glucose transport(
      • Lazar D.F.
      • Wiese R.J.
      • Mastick C.C.
      • Cautrecasas P.
      • Saltiel A.R.
      ). Moreover, neither PDGF nor epidermal growth factor, both of which activate the MAP kinase cascade in 3T3-L1 adipocytes, increases glucose transport significantly(
      • Wiese R.J.
      • Mastick C.C.
      • Lazar D.F.
      • Saltiel A.R.
      ,
      • Fingar D.C.
      • Birnbaum M.J.
      ,
      • van den Berghe N.
      • Ouwens D.M.
      • Maassen J.A.
      • van Mackelenbergh M.G.H.
      • Sips H.C.M.
      • Krans H.M.J.
      ). Taken together, these results thus imply that there may be an as yet undiscovered signaling pathway specific to insulin that is necessary for GLUT4 translocation and transport stimulation.

      Acknowledgments

      We thank Dr. Gwyn Gould for the initial development of the permeabilization procedure, Drs. A. J. Milici and Stuart Ross for assistance with the immunofluorescence, and Drs. J. R. Falck and Lewis Cantley for the dioctanoyl PI 3,4,5-trisphosphate.

      REFERENCES

        • Birnbaum M.J.
        Int. Rev. Cytol. 1992; 137: 239-297
        • Holman G.D.
        • Leggio L.L.
        • Cushman S.W.
        J. Biol. Chem. 1994; 269: 17516-17524
        • James D.E.
        • Piper R.C.
        • Slot J.W.
        Trends Cell Biol. 1994; 4: 120-126
        • White M.F.
        • Kahn C.R.
        J. Biol. Chem. 1994; 269: 1-4
        • Keller S.R.
        • Lienhard G.E.
        Trends Cell Biol. 1994; 4: 115-119
        • Lamphere L.
        • Carpenter C.L.
        • Sheng Z.F.
        • Kallen R.G.
        • Lienhard G.E.
        Am. J. Physiol. 1994; 266: E486-E494
        • Herbst J.J.
        • Andrews G.
        • Contillo L.
        • Lamphere L.
        • Gardner J.
        • Lienhard G.E.
        • Gibbs E.M.
        Biochemistry. 1994; 33: 9376-9381
        • Kelly K.L.
        • Ruderman N.B.
        J. Biol. Chem. 1993; 268: 4391-4398
        • Cheatham B.
        • Vlahos C.J.
        • Cheatham L.
        • Wang L.
        • Blenis J.
        • Kahn C.R.
        Mol. Cell. Biol. 1994; 14: 4902-4911
        • Toker A.
        • Meyer M.
        • Reddy K.K.
        • Falck J.R.
        • Aneja R.
        • Aneja S.
        • Parra A.
        • Burns D.J.
        • Ballas L.M.
        • Cantley L.C.
        J. Biol. Chem. 1994; 269: 32358-32367
        • Okada T.
        • Kawano Y.
        • Sakakibara T.
        • Hazeki O.
        • Ui M.
        J. Biol. Chem. 1994; 269: 3568-3573
        • Clarke J.F.
        • Young P.W.
        • Yonezawa K.
        • Kasuga M.
        • Holman G.D.
        Biochem. J. 1994; 300: 631-635
        • Yeh J-I.
        • Gulve E.A.
        • Rameh L.
        • Birnbaum M.J.
        J. Biol. Chem. 1995; 270: 2107-2111
        • Kotani K.
        • Carozzi A.J.
        • Sakaue H.
        • Hara K.
        • Robinson L.J.
        • Clark S.F.
        • Yonezawa K.
        • James D.E.
        • Kasuga M.
        Biochem. Biophys. Res. Commun. 1995; 209: 343-348
        • Kitas E.
        • King E.
        • Bannwarth W.
        Int. J. Pept. Protein Res. 1994; 43: 146-153
        • King D.S.
        • Fields C.G.
        • Fields G.B.
        Int. J. Pept. Protein Res. 1990; 36: 255-266
        • Frost S.C.
        • Lane M.D.
        J. Biol. Chem. 1985; 260: 2646-2652
        • Calderhead D.M.
        • Kitagawa K.
        • Tanner L.I.
        • Holman G.D.
        • Lienhard G.E.
        J. Biol. Chem. 1990; 265: 13800-13808
        • Tonks N.K.
        • Diltz C.D.
        • Fischer E.H.
        J. Biol. Chem. 1988; 263: 6722-6730
        • Kuhné M.R.
        • Zhizhuang Z.
        • Rowles J.
        • Lavan B.E.
        • Shen S-H.
        • Fischer E.H.
        • Lienhard G.E.
        J. Biol. Chem. 1994; 269: 15833-15837
        • Fussle R.
        • Bhakdi S.
        • Sziegoleit A.
        • Tranumjensen J.
        • Kranz T.
        • Wellensiek H.
        J. Cell Biol. 1981; 91: 83-94
        • Baldini G.
        • Hohman R.
        • Charron M.J.
        • Lodish H.F.
        J. Biol. Chem. 1991; 266: 4037-4040
        • Klarlund J.K.
        • Khalaf N.
        • Kozma L.
        • Czech M.P.
        J. Biol. Chem. 1993; 268: 7646-7649
        • Rosen O.M.
        • Smith C.J.
        • Fung C.
        • Rubin C.S.
        J. Biol. Chem. 1978; 253: 7579-7583
        • Shibata H.
        • Robinson F.W.
        • Benzing C.F.
        • Kono T.
        Arch. Biochem. Biophys. 1991; 285: 97-104
        • Robinson L.J.
        • Pang S.
        • Harris D.S.
        • Heuser J.
        • James D.E.
        J. Cell Biol. 1992; 117: 1181-1196
        • Slot J.W.
        • Geuze H.J.
        • Gigengack S.
        • Lienhard G.E.
        • James D.E.
        J. Cell Biol. 1991; 113: 123-135
        • Wiese R.J.
        • Mastick C.C.
        • Lazar D.F.
        • Saltiel A.R.
        J. Biol. Chem. 1995; 270: 3442-3446
        • Claesson-Welsh L.
        J. Biol. Chem. 1994; 269: 32023-32026
        • Piper R.C.
        • Hess L.J.
        • James D.E.
        Am. J. Physiol. 1991; 260: C570-C580
        • Hara K.
        • Yonezawa K.
        • Sakaue H.
        • Ando A.
        • Kotani K.
        • Kitamura T.
        • Kitamura Y.
        • Ueda H.
        • Stephens L.
        • Jackson T.R.
        • Hawkins P.T.
        • Dhand R.
        • Clark A.E.
        • Holman G.D.
        • Waterfield M.D.
        • Kasuga M.
        Proc. Natl. Acad. Sci. U. S. A. 1994; 91: 7415-7419
        • Gould G.W.
        • Jess T.J.
        • Andrews G.C.
        • Herbst J.J.
        • Plevin R.J.
        • Gibbs E.M.
        J. Biol. Chem. 1994; 269: 26622-26625
        • Lavan B.E.
        • Lienhard G.E.
        Biochem. Soc. Trans. 1994; 22: 676-680
        • Fingar D.C.
        • Hausdorff S.F.
        • Blenis J.
        • Birnbaum M.J.
        J. Biol. Chem. 1993; 268: 3005-3008
        • Hausdorff S.F.
        • Bennett A.M.
        • Neel B.G.
        • Birnbaum M.J.
        J. Biol. Chem. 1995; 270: 12965-12968
        • Hausdorff S.F.
        • Franioni J.V.
        • Birnbaum M.J.
        J. Biol. Chem. 1994; 269: 21391-21394
        • Beusch J.E-B.
        • Bhuripanyo P.
        • Carel K.
        • Leitner J.W.
        • Hsieh P.
        • DePaolo D.
        • Draznin B.
        J. Biol. Chem. 1995; 270: 2036-2040
        • Lazar D.F.
        • Wiese R.J.
        • Mastick C.C.
        • Cautrecasas P.
        • Saltiel A.R.
        Diabetes. 1995; 44: 206
        • Fingar D.C.
        • Birnbaum M.J.
        Endocrinology. 1994; 134: 728-735
        • van den Berghe N.
        • Ouwens D.M.
        • Maassen J.A.
        • van Mackelenbergh M.G.H.
        • Sips H.C.M.
        • Krans H.M.J.
        Mol. Cell. Biol. 1994; 14: 2372-2377