Intracellular Delivery of Phosphatidylinositol (3,4,5)-Trisphosphate Causes Incorporation of Glucose Transporter 4 into the Plasma Membrane of Muscle and Fat Cells without Increasing Glucose Uptake*

Insulin stimulates glucose uptake into muscle and fat cells by translocating glucose transporter 4 (GLUT4) to the cell surface, with input from phosphatidylinositol (PI) 3-kinase and its downstream effector Akt/protein kinase B. Whether PI 3,4,5-trisphosphate (PI(3,4,5)P 3 ) suffices to produce GLUT4 translocation is unknown. We used two strategies to deliver PI(3,4,5)P 3 intracellu- larly and two insulin-sensitive cell lines to examine Akt activation and GLUT4 translocation. In 3T3-L1 adipocytes, the acetoxymethyl ester of PI(3,4,5)P 3 caused GLUT4 migration to the cell periphery and increased the amount of plasma membrane-associated phospho-Akt and GLUT4. Intracellular delivery of PI(3,4,5)P role of study the role of inositol in the mobilization of GLUT4 to the cell periphery and its full insertion the cell membrane, a process required for insulin stimulation of glucose uptake. The following conclusions can be made from our findings: (i) Intracellular delivery of PI(3,4,5)P 3 results in mobilization of intracellular GLUT4 to the cell periphery. This effect was observed using two different PI(3,4,5)P 3 delivery methods in two different insulin-sen-sitive cell lines, 3T3-L1 adipocytes and L6 myoblasts. Increased peripheral localization of GLUT4 was evident by formation of a peripheral “rim” of the transporter in 3T3-L1 adipocytes and L6 myoblasts and by increased GLUT4 detection on plasma membrane lawns derived from 3T3-L1 adipocytes treated with PI(3,4,5)P 3 -AM. The intracellular compartment(s) of GLUT4 was concomitantly modified by the exogenous PI(3,4,5)P 3 , evinced by a morphological change in the perinuclear GLUT4 distribution from conical to crown-like in L6 myoblasts. (ii) Carrier-mediated delivery of PI(3,4,5)P 3 caused surface insertion of the mobilized GLUT4 molecules, as demonstrated by a gain in surface myc epitope detected in nonpermeabilized L6 myoblasts expressing GLUT4myc upon challenge with PI(3,4,5)P 3 . (iii) GLUT4 mobilization and surface insertion induced by PI(3,4,5)P 3 correlated with the plasma membrane migration and activation of Akt2/PKB (cid:1) in 3T3-L1 adipocytes and L6 myoblasts. This kinase is thought to be involved in the mimetic” effects induced by carrier-dependent PI(3,4,5)P 3

We have shown previously that delivery of the acetoxymethyl (AM) ester of PI (3,4,5)P 3 to 3T3-L1 adipocytes to increase intracellular levels of this lipid product of PI 3-kinase did not increase glucose uptake, although it could partially rescue the inhibitory action of wortmannin on stimulation of glucose uptake by insulin (31). However, the effect of PI(3,4,5)P 3 -AM on GLUT4 localization and Akt activation was not analyzed. Here we demonstrate that PI(3,4,5)P 3 -AM causes Akt phosphorylation and distribution of GLUT4 to the periphery of 3T3-L1 adipocytes. Moreover, using an additional strategy developed by Ozaki et al. (32) to deliver PI(3,4,5)P 3 into cells we examine the cell surface exposure of myc-tagged GLUT4 in L6 muscle cells. We find that carrier-mediated delivery of PI(3,4,5)P 3 causes Akt activation as well as translocation and incorporation of GLUT4myc into the plasma membrane of L6 myoblasts. Yet, again, this increase in the GLUT4myc content at the cell surface was not accompanied by an increase in 2-deoxyglucose uptake. We propose that PI(3,4,5)P 3 suffices to elicit translocation and insertion of GLUT4 at the plasma membrane. Because such action is insufficient to mimic the insulin-dependent gain in glucose uptake, we propose that input(s) additional to PI(3,4,5)P 3 are required for the culmination of increased transport of glucose into muscle and fat cells through the translocated transporters.
Cell Culture-3T3-L1 fibroblasts (a kind gift from Dr. G. Holman, University of Bath, UK) were grown in a monolayer culture in Dulbecco's modified Eagle's medium supplemented with 20% calf serum and 1% antibiotic solution (10,000 units/ml penicillin and 10 mg/ml streptomycin) in an atmosphere of 9% CO 2 at 37°C. 3T3-L1 fibroblasts were differentiated into adipocytes as described previously (33). L6 muscle cells stably overexpressing GLUT4 tagged with a myc epitope at the first exofacial loop (L6GLUT4myc) were grown as described previously (15). Experiments were performed in myoblast monolayers or in detached, rounded-up myoblasts as described by Foster et al. (34).
Carrier-mediated delivery of PI(3,4,5)P 3 was performed as follows: A 1 mM stock solution of phosphoinositides and 1 mM carrier (neomycin) was prepared in HEPES-buffered saline (140 mM NaCl, 20 mM Na-HEPES, 2.5 mM MgSO 4 , 1 mM CaCl 2 , and 5 mM KCl, pH 7.4). Phosphoinositides and carrier were mixed (10 M each) with serum-free ␣-minimum Eagle's medium in glass test tubes, incubated at room temperature for 10 min, followed by bath sonication (10 s at 35% maximum intensity, Sonic Dismembrator, model 300, Fisher). Phos-phoinositides and carrier-containing medium were then applied to the cells for the indicated time period. Seven different lots of dipalmitoyl PI(3,4,5)P 3 were used for the various assays described in this study using carrier-mediated phosphoinositides delivery . Lot numbers CF-1-178, CF-II-47, CF-41-138, CF-41-162, and CF-66-107 reproducibly elicited most or all of the different biological effects that they were tested  for. The biological activity of CF-41-138 was confirmed in two independent laboratories (at York University and at the Hospital for Sick Children, Toronto, ON). Two other lots, CF-II-29 and CSB-30-162, showed no biological effect in any of the cell-based assays tested and were not included in the results presented in this study. NMR analysis of these two lots confirmed that they were chemically equivalent to the active lots, and the disparity is likely the result of hydrolysis of esters or phosphoesters during shipment and storage.
Fluorescent Detection of Fluorophore-conjugated PI (3,4,5)P 3 and Indirect Immunofluorescent Detection of PI (3,4,5)P 3 -For fluorescent detection of PI(3,4,5)P 3 , NBD-PI(3,4,5)P 3 or Bodipy-PI(3,4,5)P 3 were incubated with carrier and applied to detached rounded-up L6GLUT4myc myoblasts (see details below) or 3T3-L1 adipocytes grown on glass coverslips, respectively. Cells were then washed three times with icecold PBS and fixed for 30 min with 3% paraformaldehyde. Excess fixative was quenched with 50 mM NH 4 Cl, and coverslips were mounted using Dako mounting medium. For indirect immunofluorescent detection of PI(3,4,5)P 3 , cells were treated with carrier only or with PI(3,4,5)P 3 as described above, fixed, and quenched as detailed above. Cells were then permeabilized using 0.2% saponin in PBS for 3 min at room temperature, followed by 10-min blocking with 5% goat serum in PBS and incubation for 1 h at room temperature with anti-PI(3,4,5)P 3 mouse IgG. Coverslips were washed three times and reacted further with Alexa 488-conjugated anti-mouse IgG antibody. Images were obtained using a Zeiss Laser Scanning Confocal Microscope 510 (Thornwood, NY). Background nonspecific fluorescence was determined by the same protocol in the absence of primary antibody.
GLUT4 Localization in Permeabilized 3T3-L1 Adipocytes Pretreated with PI(3,4,5)P 3 -AM-Determination of GLUT4 distribution by immunofluorescence was performed after fixing and permeabilizing the cells. Briefly, fully differentiated 3T3-L1 adipocytes were trypsinized and replated at a lower density on poly-L-lysine-coated coverslips. The cells were allowed to settle for 24 h and then serum starved for 3 h prior to treatment for 15 min with either 100 nM insulin or 150 M PI(3,4,5)P 3 -AM. Coverslips holding the cells were then immersed in 4% paraformaldehyde in PBS for 20 min at 4°C. Excess paraformaldehyde was reacted with 50 mM NH 4 Cl in PBS for 5 min at room temperature. Cells were then permeabilized with 0.1% Triton X-100 for 30 min and blocked with 5% goat serum in PBS for 30 min. Rabbit polyclonal antibodies against the GLUT4 C terminus (35) were used at a dilution of 1:250 to label GLUT4 within the cells, and Cy3-conjugated goat anti-rabbit antibodies were used to label the primary antibodies. Fluorescence was detected using a Zeiss Laser Scanning Confocal Microscope 510.
Immunofluorescent Detection of GLUT4, Akt/PKB, and p-S473Akt in Plasma Membrane Lawns from Adipocytes Pretreated with PI(3,4,5)P 3 -AM-Differentiated 3T3-L1 adipocytes grown on glass coverslips were serum deprived for 2 h. Cells were then treated for 20 min with 100 nM insulin or 150 M PI(3,4,5)P 3 -AM. Control experiments were also performed using palmitate, BCECF-AM, BAPTA-AM, and PI (150 M, 20 min). Plasma membrane lawns (sheets) were prepared as described previously (33,36) with slight modifications. After the various treatments, cells were placed on ice and washed twice with ice-cold PBS. Hypotonic swelling buffer (23 mM KCl, 10 mM HEPES, 2 mM MgCl 2 , 1 mM EGTA, pH 7.5) was added in three quick rinses. Four ml of breaking buffer (70 mM KCl, 30 mM HEPES, 5 mM MgCl 2 , 3 mM EGTA, 1 mM dithiothreitol, 1 mM Na 3 VO 4 , 0.5 mM phenylmethylsulfonyl fluoride, 1 M pepstatin A, 1 M leupeptin, and 1 mM microcystin, pH 7.5) were added to each well, and the solution was aspirated up and down using a 1.0-ml pipetter to promote cell breakage. The coverslips were washed twice in breaking buffer and incubated with cold 3% paraformaldehyde in PBS for 10 min on ice followed by three washes in PBS. Excess fixative was quenched with 50 mM NH 4 Cl and PBS for 5 min followed by three washes with PBS at room temperature. The lawns were subsequently blocked by a 1-h incubation in 5% goat serum in PBS at room temperature, then incubated for 60 min with antibodies (1:250 dilution of anti-GLUT4, 1:150 dilution of anti-Akt, 1:250 of anti-pS473Akt) at room temperature and washed three times in PBS. Fluorescein isothiocyanate-and Cy3-conjugated species-specific secondary antibodies (1: 250) were added for 45 min then rinsed with four washes with PBS. Coverslips were mounted with Dako mounting solution. Images were obtained using a Leica DM1RB inverted fluorescent microscope with a 40ϫ objective. All images were collected under identical gain settings.
Images were analyzed using ImageJ software by comparing pixel intensities of lawns and their respective backgrounds among the various conditions tested in each experiment.
Akt Activity in 3T3-L1 Adipocyte Lysates Treated with PI(3,4,5)P 3 -AM-3T3-L1 adipocytes were stimulated for 30 min with 100 nM insulin or 150 M PI(3,4,5)P 3 -AM, lysed, and the in vitro kinase activities of Akt1, Akt2, and Akt3 were determined as described previously (33). Briefly, cells grown on 6-well plates were lysed in buffer containing 50 mM HEPES, pH 7.6, 150 mM NaCl, 10% glycerol (v/v), 1% Triton X-100 (v/v), 30 mM Na 4 P 2 O 7 , 10 mM NaF, 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, 1 mM benzamidine, 1 mM Na 3 VO 4 , 1 mM dithiothreitol, and 100 nM okadaic acid. Antibody (2 g/condition) precoupled to protein A-/protein G-Sepharose beads (20 l of 100 mg/ml each/condition) was added to 250 g of total protein from cell lysates. Antibody-coupled beads were washed twice with ice-cold PBS and once with ice-cold lysis buffer before use. The indicated enzymes were immunoprecipitated by incubating with the antibody-bead complex for 2-3 h under constant rotation at 4°C. Immunocomplexes were isolated and washed four times with 1 ml of wash buffer (25 mM HEPES, pH 7.8, 10% glycerol (v/v), 1% Triton X-100 (v/v), 0.1% bovine serum albumin, 1 M NaCl, 1 mM dithiothreitol, 1 mM phenylmethylsulfonyl fluoride, 1 M microcystin, and 100 nM okadaic acid) and twice with 1 ml of kinase buffer (50 mM Tris/HCl, pH 7.5, 10 mM MgCl 2 , 10 nM okadaic acid, and 1 mM dithiothreitol). The complexes were then incubated under constant agitation for 30 min at 30°C with 30 l of reaction mixture (kinase buffer containing 5 M ATP, 2 Ci of [␥ 32 P]ATP plus 100 M Crosstide as substrate in Akt assays). After the reaction, 30 l of the supernatant was transferred onto Whatman p81 filter paper and washed four times for 10 min with 3 ml of 175 mM phosphoric acid and once with distilled water for 5 min. Filters were air dried and subjected to liquid scintillation counting.

Immunofluorescent Detection of GLUT4myc in Intact, Rounded-up Myoblasts upon Carrier-mediated Delivery of PI(3,4,5)P 3 -Rounded-up
myoblasts were used to detect surface GLUT4myc as described previously (34). Briefly, quiescent L6 myoblasts were detached from their substratum by incubation in Ca 2ϩ -and Mg 2ϩ -free PBS for 10 min. Dislodged cells were suspended in HEPES-buffered RPMI and allowed to begin reattaching onto poly-L-lysine-coated glass coverslips for 10 min. These rounded-up cells were then treated for 10 min with 100 nM insulin or PI(3,4,5)P 3 plus carrier, each 10 M unless otherwise stated, for 10 min, as described above. After rinsing with PBS, coverslips with rounded-up myoblasts were fixed with 4% paraformaldehyde for 3 min, and then excess paraformaldehyde was reacted with 0.1 M glycine for 10 min. Cells were then incubated at 4°C with 5% goat serum in PBS for 15 min. Monoclonal antibody against the myc epitope (1:150) was then added for 1 h to label surface GLUT4myc followed by Alexa 488conjugated goat anti-mouse IgG (1:200) for 50 min. Cells were postfixed with 4% paraformaldehyde for 10 min, and excess paraformaldehyde was reacted with 0.1 M glycine for 10 min. Coverslips were mounted on slides using Dako antifade solution, and fluorescence was detected using a Zeiss Laser Scanning Confocal Microscope 510.
Akt Phosphorylation in L6 Myoblasts upon Carrier-mediated Delivery of PI(3,4,5)P 3 -L6 myoblasts grown on coverslips were treated with insulin or PI(3,4,5)P 3 plus carrier as described above, washed three times with cold PBS supplemented by 100 nM okadaic acid and 1 mM Na 3 VO 4 , fixed in 3% paraformaldehyde in PBS supplemented with the phosphatase inhibitors for 30 min, after which excess paraformaldehyde was quenched with 50 mM NH 4 Cl. Cells were permeabilized with 0.1% Triton X-100 in PBS for 20 min and blocked with 1% bovine serum albumin in PBS. Immunohistochemistry-compatible anti-pS473Akt antibody (Cell Signaling) (1:250) was prepared in blocking solution and applied to the coverslips for 1 h at room temperature. After three washes in PBS, fluorophore-conjugated anti-rabbit IgG was used to detect pAkt immunoreactivity. Z-stack images were acquired using laser confocal images, collapsed into a single image, and analyzed as described below for intracellular GLUT4myc.
Quantification of Cell Surface GLUT4myc in L6 Myoblast Monolayers upon Carrier-mediated Delivery of PI (3,4,5)P 3 -Quantification of GLUT4myc at the surface of adhered myoblasts was performed as described previously with slight modifications (37). Briefly, after treatment of quiescent L6GLUT4myc myoblasts with or without 100 nM insulin or PI(3,4,5)P 3 plus carrier (10 M each) at 37°C as indicated, cells in 24-well tissue culture plates were placed on ice and rinsed twice with ice-cold PBS, blocked with 5% goat serum in PBS for 15 min, and then reacted at 4°C for 1 h with either a monoclonal anti-myc antibody (kind gift from Dr. M. Moran, University of Toronto, Canada, 8 g/ml in HEPES-modified RPMI containing 3% goat serum), or a commercial anti-myc antibody (9E10, Santa Cruz, 1:100 in 5% goat serum in PBS).
After four washes with ice-cold PBS, cells were fixed with 3% (v/v) formaldehyde in PBS at 4°C for 10 min followed by quenching with 0.1 M glycine in PBS for 10 min and incubation with secondary antibody (horseradish peroxidase-conjugated donkey anti-mouse IgG, 1:1,000 in PBS containing 3% goat serum) at 4°C for 1 h. The cell monolayers were then washed six times with PBS, then 1 ml/well of 0.4 mg/ml o-phenylenediamine dihydrochloride reagent (horseradish peroxidase substrate) was added at room temperature for 20 -30 min. The reaction was stopped by the addition of 0.25 ml of 3 N HCl, and the A 492 nm was measured. Control wells remained untreated with either primary antibody or both primary and secondary antibodies. Measurements of surface GLUT4myc in control wells were subtracted from values in all other treatments.
Immunofluorescent Detection of Perinuclear GLUT4myc in Adhered L6 Myoblasts upon Carrier-mediated Delivery of PI (3,4,5)P 3 -L6 myoblasts grown on glass coverslips were exposed to the different treatments (insulin, PI(3,4,5)P 3 plus carrier, or carrier alone) and then fixed and permeabilized to detect intracellular GLUT4myc by immunofluorescence (38). Images at a magnification of ϫ100 using a Zeiss Laser Scanning Confocal Microscope 510 were acquired from several fields of view from at least three separate experiments. To obtain an integrated image of total intracellular protein with minimal contribution of the surface signal, the whole cell was scanned along the z axis by optical slicing 0.4 m apart. A single composite image (collapsed XY projection) of the optical cuts in which the outline of the nucleus was discernible was generated using LSM 5 Image software (Carl Zeiss, Inc.) and used for subsequent analyses. For this study, the "cone" morphology was defined as a characteristic polarized clustering of GLUT4myc around the nucleus, which spans less than two-thirds of the nuclear perimeter and exhibits a peak-like distribution. In contrast, the "crown" morphology was defined by a distinctly uniform, tight ring of GLUT4myc about the nuclear perimeter. The number of cells exhibiting these morphologies is expressed as percent of the total number of cells analyzed.
Determination of 2-Deoxyglucose upon Carrier-mediated Delivery of PI(3,4,5)P 3 -2-Deoxyglucose uptake was measured as described previously (39). Briefly, L6 myoblast monolayers or 3T3-L1 adipocytes were serum deprived for 4 h and exposed to insulin, carrier alone or PI(3,4,5)P 3 plus carrier, as indicated. The monolayers were washed twice with HEPES-buffered saline, and any remaining liquid was aspirated. Cells were then incubated for 5 min in HEPES-buffered saline containing 10 M unlabeled 2-deoxyglucose and 2-deoxy-D-[ 3 H]glucose (0.5 Ci/ml). The radioactivity was then aspirated and the cells washed three times in ice-cold saline solution. In some experiments, the wash solution contained 2 mM HgCl 2 to ensure blockade of any possible efflux of free 2-deoxyglucose during the wash. This addition did not change the results, suggesting that there was no free hexose in the cell under the conditions of the assay. Background radioactivity counts not inhibitable by 10 M cytochalasin B added to the transport solution were subtracted from all values. The intracellular content of 2-deoxy-D-[ 3 H]glucose was determined and expressed as pmol of 2-deoxyglucose/mg of protein/min.
Statistical Analysis-Statistical analyses were performed with Prism 3.0 software (San Diego). Experiments with more than two groups of cells were analyzed using analysis of variance with Tukey's post hoc analysis. When indicated, two groups were compared using Student's t test. A p value of Ͻ0.05 was considered statistically significant.

PI(3,4,5)P 3 -AM Elicits GLUT4 Translocation in 3T3-L1
Adipocytes-PI(3,4,5)P 3 -AM is a derivative of PI(3,4,5)P 3 in which all of the negatively charged phosphates are masked by AM esters to allow the compound to cross cell membranes readily (31). Within the cytoplasm, PI(3,4,5)P 3 -AM is hydrolyzed to PI(3,4,5)P 3 by endogenous intracellular esterases (31). 3T3-L1 adipocytes were treated for 15 min with 150 M PI(3,4,5)P 3 -AM or 100 nM insulin, and then cellular localization of GLUT4 was assessed in permeabilized cells by indirect immunofluorescence using antibodies directed against the C terminus of GLUT4 (Fig. 1A). Under basal conditions GLUT4 exhibited a perinuclear distribution with very low amounts at the cell periphery. Insulin and PI(3,4,5)P 3 -AM each caused a redistribution of GLUT4 to the cell periphery and a concomitant decrease in the level of this transporter in the perinuclear region. The redistribution of GLUT4 was visualized further by detection of the transporters on plasma membrane lawns (sheets) of 3T3-L1 adipocytes. Fig. 1B shows representative immunofluorescence images of lawns derived from cells pretreated with insulin or PI(3,4,5)P 3 -AM. This method allows a semiquantitative analysis of fluorescence intensity (40), and a summary of multiple image analyses is shown in the lower panel of Fig. 1B. The amount of GLUT4 associated with the plasma membrane lawn was low under basal conditions, and treatment of cells with insulin caused a 3.7-fold increase in the plasma membrane lawn content of GLUT4 (p Ͻ 0.05). Cell pretreatment with PI(3,4,5)P 3 -AM increased the amount of GLUT4 associated with plasma membrane lawns to 2.5-fold (p Ͻ 0.05) that of nontreated basal values.
To determine whether the mobilization of glucose transporters seen in response to PI(3,4,5)P 3 -AM was caused by nonspecific effects of acetate or formaldehyde (liberated upon hydrolysis of PIP 3 -AM) we measured GLUT4 levels in plasma membrane lawns derived from 3T3-L1 adipocytes pretreated with other AM-containing compounds (BAPTA-AM and BCECF-AM). Neither of these agents had any effect on GLUT4 levels associated with plasma membrane lawns (Table I). This result is in agreement with a study showing no effect of BAPTA-AM on GLUT4 abundance in membrane lawns of 3T3-L1 adipocytes (41). Similarly, pretreatment of cells with phosphatidylinositol or the fatty acid palmitate did not increase the amount of GLUT4 associated with membrane lawns (Table I). Taken together, these results suggest that elevation of cellular PI(3,4,5)P 3 can specifically mobilize GLUT4 toward the membrane of 3T3-L1 adipocytes.
Akt isoforms undergo PI 3-kinase-dependent recruitment to membranes to achieve full activation (42,43). Therefore, we determined the effect of cell pretreatment with PI(3,4,5)P 3 -AM on the levels of Akt/PKB associated with plasma membrane lawns. 3T3-L1 adipocytes were treated with either 150 M PI(3,4,5)P 3 -AM or 100 nM insulin, and the presence of specific Akt/PKB isoforms was detected by immunofluorescence on plasma membrane lawns (Fig. 2B). PI(3,4,5)P 3 -AM pretreatment augmented the association of Akt2/PKB␤ with the adhered plasma membranes, whereas insulin pretreatment increased the association of Akt2/PKB␤ and Akt3/PKB␥. Despite quantitative differences between the effects of PI(3,4,5)P 3 -AM and insulin, it is clear that both increase Akt2PKB␤ activity and its association with the plasma membrane. No increase in Akt1/PKB␣ associated with plasma membrane lawns was observed in response to either insulin or PI(3,4,5)P 3 -AM. It is possible that the insulin stimulation of Akt1/PKB␣ activity detected in cell lysates ( Fig. 2A) is associated with the cytosol or with intracellular membranes that do not copurify with membrane lawns. This is consistent with a recent study demonstrating that Akt2/PKB␤, but not Akt1/PKB␣, translocates to plasma membrane fractions of 3T3-L1 adipocytes in response to insulin stimulation (44). In any case, the above results indicate that PI(3,4,5)P 3 -AM emulates insulin in mobilizing and activating Akt2/PKB␤, the Akt/PKB isoform believed to participate in GLUT4 translocation (25).

)P 3 Causes Phosphorylation and Increased Membrane Association of Akt in 3T3-L1
Adipocytes and L6 Myoblasts-To corroborate the effects of PI(3,4,5)P 3 -AM, we used an alternative approach developed by Ozaki et al. (32) to deliver phosphoinositides into cells. The method is based on the ability of cells to take up negatively charged lipids when presented to cells in equimolar complexes with cationic carriers via a nonendocytic mechanism. The effectiveness of this carrier-mediated delivery of PI(3,4,5)P 3 into several cell types was described recently (32) but had not been verified in 3T3-L1 adipocytes or L6 myoblasts. To this end, we  assessed carrier-mediated delivery of PI(3,4,5)P 3 into these cells by fluorescence microscopy using fluorophore-conjugated PI(3,4,5)P 3 and indirect immunofluorescence utilizing anti-PI(3,4,5)P 3 IgG. Bodipy-conjugated PI(3,4,5)P 3 preincubated with carrier (neomycin) could be readily detected within 3T3-L1 adipocytes (Fig. 3A). Similarly, the intracellular delivery of phosphoinositides into detached, rounded-up L6GLUT4myc myoblasts was confirmed with 10 M NBD-labeled PI(3,4,5)P 3 along with 10 M neomycin (Fig. 3B). Finally, cellular PI(3,4,5)P 3 content was detected in adhered L6GLUT4myc myoblasts using anti-PI(3,4,5)P 3 antibody (Fig. 3C). Cells treated with PI(3,4,5)P 3 and carrier showed a notably higher signal than control or carrier-only exposed cells. Collectively these experiments demonstrate the ability to deliver exogenous PI(3,4,5)P 3 using a carrier into 3T3-L1 adipocytes and L6 muscle cells. 3T3-L1 adipocytes were treated with carrier without or with PI(3,4,5)P 3 , membrane lawns were generated, and Akt phosphorylation on Ser-473 was determined by immunofluorescence. Fig. 4A illustrates the results and indicates that Akt was phosphorylated in lawns from 3T3-L1 adipocytes pretreated with either insulin or PI(3,4,5)P 3 plus carrier, but not with carrier alone. Similarly, the ability of carrier-dependent delivery of PI(3,4,5)P 3 to activate Akt was also tested in L6 myoblasts, cells in which Akt is phosphorylated robustly by insulin (15,34,38,45). Myoblasts were pretreated with either insulin or carrier with or without PI(3,4,5)P 3 , and pS473Akt was detected by immunofluorescence in permeabilized myoblasts. Fig.  4B shows that insulin and PI(3,4,5)P 3 plus carrier, but not carrier alone, increased the level of pAkt in L6 myoblasts.
The clonal line of L6 muscle cells used in this study stably expresses myc-tagged GLUT4. The exofacially facing myc epitope allows detection of membrane-inserted GLUT4 (15,34,38,45). Hence, this system was used to ascertain whether PI(3,4,5)P 3 causes bona fide insertion of GLUT4 into the plasma membrane. Carrier-dependent delivery of various phosphoinositides was tested in this system. GLUT4myc appearance at the cell surface was assessed first by immunofluorescent detection of myc epitope exposure in intact (nonpermeabilized) rounded-up myoblasts. Fig. 5C shows that in untreated control cells very little GLUT4myc was available at the surface of myoblasts examined by confocal fluorescence microscopy. Insulin treatment caused a clear increase in cell surface GLUT4myc, visualized as a rim of fluorescence in intact cells. PI(3,4,5)P 3 plus carrier (10 M each) also increased the amount of GLUT4myc at the cell surface, whereas the carrier alone (10 M) or with PI(4,5)P 2 (10 M, not shown) had no effect on surface GLUT4myc. To complement this single-cell assay, the extent of GLUT4myc exposure at the surface of cell populations was determined on adhered L6 myoblast monolayers and quantified by a colorimetric assay based on horseradish peroxidase-linked secondary antibody. As shown in Fig. 5D, insulin elicited a 2.8 Ϯ 0.3-fold rise in surface GLUT4myc (p Ͻ 0.001), and PI(3,4,5)P 3 plus carrier (10 M each) caused a 1.7 Ϯ 0.1-fold gain in surface transporters (p Ͻ 0.05). Higher concentrations of PI(3,4,5)P 3 plus carrier (up to 50 M each) did not elicit any further increase in surface GLUT4myc compared with the 10 M treatment. The ability of PI(3,4,5)P 3 plus carrier to increase GLUT4myc exposure was not prevented by the PI 3-kinase inhibitor wortmannin (results not shown). These results demonstrate that PI(3,4,5)P 3 causes bona fide insertion of GLUT4 at the surface of L6 myoblasts.
Carrier-mediated Delivery of PI (3,4,5)P 3 Alters the Perinuclear Distribution of GLUT4myc in L6 Myoblasts-GLUT4 is mobilized to the plasma membrane from intracellular GLUT4 compartment(s), and intracellularly GLUT4 concentrates in the perinuclear region (46,47). We further assessed the effect of carrier-delivered PI(3,4,5)P 3 on perinuclear GLUT4 and quantified these effects in L6 myoblasts. To this end, we examined the morphology of the GLUT4 pool around the nucleus of L6 myoblasts in the basal state and in response to insulin or PI(3,4,5)P 3 plus carrier. The staining pattern of intracellular GLUT4myc in L6GLUT4myc myoblasts in three independent experiments is depicted in Fig. 6. In nonstimulated conditions, GLUT4myc assumed a polarized conical distribution at one pole of the nucleus in the majority (62.3 Ϯ 14.7%) of the cells. In response to insulin, this distribution changed to cover a wider area of the nuclear perimeter and appeared more even and less conical. The percent of cells that exhibited GLUT4myc cones was decreased to 9.7 Ϯ 3.5 (p Ͻ 0.01 compared with control cells), and the percent of cells with GLUT4myc signal coverage exceeding half of the nuclear perimeter rose from 35.5 Ϯ 14.1 to 63.5 Ϯ 15.4. Strikingly, PI(3,4,5)P 3 plus carrier, but not carrier alone, caused a similar change to insulin in GLUT4myc distribution. Further analysis of the mechanisms underlying these changes is beyond the scope of this study, but the results illustrate an insulin-like effect of PI(3,4,5)P 3 plus carrier on the perinuclear compartment(s) of GLUT4myc, presumably as part of the mechanism that mobilizes transporters toward the cell surface.

Carrier-mediated Delivery of PI(3,4,5)P 3 Does Not Increase Hexose Uptake into 3T3-L1 Adipocytes and L6 Myoblasts-The
above results show that carrier-mediated delivery of PI(3,4,5)P 3 suffices to induce phosphorylation of Akt, changes in the perinuclear distribution of GLUT4myc, mobilization of GLUT4 to the cell surface, and insertion into the plasma membrane. It was therefore important to examine whether glucose uptake was stimulated under the same conditions. In L6 GLUT4myc myoblasts, GLUT4myc is highly overexpressed and is the transporter determining glucose uptake (39). Strikingly, carrier-introduced PI(3,4,5)P 3 (10 M, 20 min) had no effect on 2-deoxyglucose uptake in either 3T3-L1 adipocytes or L6GLUT4myc myoblasts (Fig. 7, A and B, respectively). Higher concentrations of the lipid-carrier complex (up to 50 M) were also ineffective in stimulating glucose uptake. Pretreatments with carrier alone, with PI 3-phosphate, with a combination of PI 3-phosphate and PI(3,4,5)P 3 , or along with PI or PI(4,5)P 2 (not shown) were also without consequence on 2-deoxyglucose uptake. In contrast, 100 nM insulin caused an increase in hexose uptake of more than 7-fold and 2-fold in 3T3-L1 adipocytes and L6GLUT4myc myoblasts, respectively, the latter characteristic of the response of muscle cells in culture and isolated muscle preparations. To compare better the effect of insulin and exogenous PI(3,4,5)P 3 on glucose uptake, cells were also stimulated with a submaximal concentration of insulin (1 nM). This concentration was shown previously to cause ϳ50% the increase in GLUT4 associated with plasma membrane lawns in 3T3-L1 adipocytes (39), similar to the effect induced by PI(3,4,5)P 3 (Fig. 5B). Likewise, 1 nM insulin induced an ϳ1.5fold increase in surface GLUT4myc in L6GLUT4myc myoblasts (48), resembling the magnitude of the effect provoked by exogenous PI(3,4,5)P 3 in these cells (Fig. 5D). Submaximal (1 nM) insulin stimulated glucose uptake nearly 4.5-and 1.5-fold in 3T3-L1 adipocytes (Fig. 7A) and L6GLUT4myc myoblasts (Fig.  7B), respectively. Yet, despite the similar levels of GLUT4 translocation achieved with insulin, exogenous PI(3,4,5)P 3 had no effect on glucose uptake. These results suggest that insertion of GLUT4 into the plasma membrane in response to PI(3,4,5)P 3 does not suffice to elicit stimulation of glucose uptake. DISCUSSION It is generally accepted that type IA PI 3-kinase(s) participates in the insulin-dependent translocation of GLUT4 to the

FIG. 4. Carrier-dependent delivery of PI(3,4,5)P 3 causes phosphorylation of Akt in 3T3-L1 adipocytes and L6 myoblasts.
A, 3T3-L1 adipocytes were incubated for 20 min with 100 nM insulin, PI(3,4,5)P 3 plus carrier (10 M each), or 10 M carrier alone. Membrane lawns were generated, and Akt/PKB was detected by immunofluorescence using an anti-pS473Akt antibody. The lawn-associated fluorescent signal was quantified as described under "Experimental Procedures." Results are expressed as the mean Ϯ S.E. relative to the untreated control. Lot CF-41-138 of PI(3,4,5)P 3 was used in these studies, yielding p Ͻ 0.012 for PI(3,4,5)P 3 versus control (Student's t test). B, L6 myoblasts were incubated for 20 min with 100 nM insulin, PI(3,4,5)P 3 plus carrier (10 M each), or 10 M carrier alone. Detection of anti-pS473Akt antibody via immunofluorescence and quantification is described under "Experimental Procedures." Results are expressed as the mean Ϯ S.E. relative to the untreated control. Lot CF-41-138 of PI(3,4,5)P 3 was used in this study yielding p Ͻ 0.08 for PI(3,4,5)P 3 versus control (Student's t test). plasma membrane, based on the inhibitory action of pharmacological inhibitors and a dominant-negative mutant of the p85 subunit of the enzyme, in both adipose and muscle cells. Conversely, when the p110 catalytic subunit of class I PI 3-kinase is targeted to membranes in cultured adipocytes, surface GLUT4 increases. In vitro, class IA PI 3-kinases can phosphorylate PI, PI 4-phosphate, and PI(4,5)P 2 , resulting in the generation of PI 3-phosphate, PI(3,4)P 2 , and PI(3,4,5)P 3 , respectively. Thus, a fundamental question that arises is which of the specific PI 3-kinase product(s) mediates insulin action and in particular whether any one of these lipids suffices to elicit GLUT4 translocation to the plasma membrane and glucose uptake. Most studies have indistinctly measured glucose uptake or GLUT4 gain at the cell surface, yet these phenomena may require input from different products of the enzyme.
Several lines of evidence suggest that PI(3,4,5)P 3 is the ma-jor PI 3-kinase product required for the propagation of the insulin signal toward GLUT4. Overexpression of phosphatidylinositol phosphatases that decrease cellular PI(3,4,5)P 3 content reduce insulin-mediated GLUT4 translocation. This is the case for the 3Ј-phosphatase PTEN in one report (49), but not another (50). Conditional deletion of the PTEN gene from skeletal muscle increases the insulin response of glucose uptake in this tissue (51). However, because PTEN acts on several PIPs one cannot specifically assign a role for PI(3,4,5)P 3 versus PI(3,4)P 2 based on these observations. Like PTEN, the 5Ј-phosphatases SHIP and SKIP, which reduce PI(3,4,5)P 3 levels to elevate those of PI(3,4)P 2 , impaired insulin-mediated GLUT4 translocation and glucose metabolism (52)(53)(54). Furthermore, SHIP2 knock-out mice exhibit enhanced insulin action (55). All of these studies support the notion that PI(3,4,5)P 3 is required for insulin action, but none shows that it is sufficient.

FIG. 5. Carrier-dependent delivery of PI(3,4,5)P 3 causes GLUT4 translocation and insertion at the plasma membrane.
A and B, GLUT4 association with plasma membrane lawns of 3T3-L1 adipocytes. 3T3-L1 adipocytes grown on glass coverslips were serum starved and treated for 20 min with 100 nM insulin, or carrier with or without PI(3,4,5)P 3 (10 M each). Plasma membrane lawns were then prepared and reacted with anti-GLUT4 antibody as described under "Experimental Procedures." Shown are images representing three independent experiments (A) and the quantitation of the fluorescent signal intensity in these experiments, as detailed under "Experimental Procedures" (B). C, immunofluorescent detection of GLUT4myc in intact, rounded-up L6 myoblasts. L6 GLUT4myc cells stably expressing GLUT4myc were detached from the substratum using Ca 2ϩ -and Mg 2ϩ -free PBS for 15 min. Dislodged cells were suspended in HEPES-buffered RPMI and seeded on glass coverslips. These rounded-up L6 myoblasts were allowed to attach for 10 -20 min prior to 20 min treatment with 100 nM insulin, PI(3,4,5)P 3 plus carrier (10 M each), or carrier alone. Surface GLUT4myc was then detected by indirect immunofluorescence using an anti-myc antibody in nonpermeabilized cells as described under "Experimental Procedures." Results show representative cells of at least five experiments. Scale bar, 10 M. D, quantification of surface GLUT4myc in L6 myoblast monolayers. L6 myoblast monolayers were exposed for 20 min to 100 nM insulin or PI(3,4,5)P 3 plus carrier (10 M each), or carrier alone. The cells were then rinsed, and surface GLUT4myc was detected on intact cells using anti-myc antibody followed by secondary antibody coupled to horseradish peroxidase, as described under "Experimental Procedures." GLUT4myc levels at the surface of untreated cells are given a value of 1.0, and all other values are expressed relative to this value. Nonspecific antibody binding was measured by anti-mouse IgG alone and was subtracted from all experimental values. Results are the mean of 21 independent experiments, each performed in triplicate, using PI(3,4,5)P 3 lots CF-I-178, CF-II-47, CF-41-138 and CF-66-107 (along with carrier). **, p Ͻ 0.001, *, p Ͻ 0.05 compared with control, by analysis of variance.
Here we undertook a direct approach to analyze the role of PI(3,4,5)P 3 in insulin action. Two methods were used to deliver PI(3,4,5)P 3 across the negatively charged lipid bilayer. The first one was the use of an AM derivative of PI(3,4,5)P 3 (31), and the second one was the introduction of unmodified PI(3,4,5)P 3 along with a polycationic carrier (32). We have reported previously that intracellular delivery of PI(3,4,5)P 3 -AM alone did not elevate glucose uptake in 3T3-L1 adipocytes, but it could partly restore insulin-stimulated glucose transport inhibited by the PI 3-kinase inhibitor wortmannin (31). We concluded that PI(3,4,5)P 3 generation is a required, but not sufficient input for the stimulation of glucose uptake by insulin. The present study was designed to assess the role of inositol phospholipids in the mobilization of GLUT4 to the cell periphery and its full insertion into the cell membrane, a process required for insulin stimulation of glucose uptake. The following conclusions can be made from our findings: (i) Intracellular delivery of PI(3,4,5)P 3 results in mobilization of intracellular GLUT4 to the cell periphery. This effect was observed using two different PI(3,4,5)P 3 delivery methods in two different insulin-sensitive cell lines, 3T3-L1 adipocytes and L6 myoblasts. Increased peripheral localization of GLUT4 was evident by formation of a peripheral "rim" of the transporter in 3T3-L1 adipocytes and L6 myoblasts and by increased GLUT4 detection on plasma membrane lawns derived from 3T3-L1 adipocytes treated with PI(3,4,5)P 3 -AM. The intracellular compartment(s) of GLUT4 was concomitantly modified by the exogenous PI(3,4,5)P 3 , evinced by a morphological change in the perinuclear GLUT4 distribution from conical to crown-like in L6 myoblasts. (ii) Carrier-mediated delivery of PI(3,4,5)P 3 caused surface insertion of the mobilized GLUT4 molecules, as demonstrated by a gain in surface myc epitope detected in nonpermeabilized L6 myoblasts expressing GLUT4myc upon challenge with PI(3,4,5)P 3 . (iii) GLUT4 mobilization and surface insertion induced by PI(3,4,5)P 3 correlated with the plasma membrane migration and activation of Akt2/PKB␤ in 3T3-L1 adipocytes and L6 myoblasts. This kinase is thought to be involved in GLUT4 traffic in response to insulin (15,25,34). (iv) Strikingly, and despite the "insulin mimetic" effects induced by carrier-dependent PI(3,4,5)P 3 delivery into L6 myoblasts, glucose uptake was not stimulated.
The maximal increase in surface GLUT4myc caused by exogenously delivered PI(3,4,5)P 3 is smaller than the response elicited by insulin. It is possible that the efficiency of the PI(3,4,5)P 3 delivery systems is insufficient to mimic quantitatively the rise in PI(3,4,5)P 3 caused by the hormone, although it suffices to mobilize GLUT4 to a significant extent. Indeed, for most of the end points measured (GLUT4 translocation, Akt translocation and/or phosphorylation) the response to PI(3,4,5)P 3 was smaller than the insulin effect. It is also very likely that the duration of the PI(3,4,5)P 3 produced in response to the hormone is critical for accurate signal transduction to metabolic end points. Indeed, longer duration or higher magnitude of plasma membrane localization of the PH domain of Akt (presumably detecting PI(3,4,5)P 3 and PI(3,4,)P 2 ) determines GLUT4 translocation in response to insulin or PDGF (56). Furthermore, the subcellular localization of PI(3,4,5)P 3 may be critical. Accordingly, we have shown recently that insulindependent actin remodeling clusters PI(3,4,5)P 3 in specific submembrane and membrane loci through segregation of the p85 and p110␣ subunits of PI 3-kinase (57). Finally, it is possible that signals additional to PI(3,4,5)P 3 are required for the full extent of translocation. These could involve carefully balanced levels of other phosphoinositides such as PI(3,4)P 2 , PI 3-phos-  (20 M), or with carrier alone, as indicated. Cells were washed, and 2-deoxyglucose uptake was measured as described under "Experimental Procedures." Results are the mean Ϯ S.E. of three independent experiments, each performed in triplicate, using PI(3,4,5)P 3 lots positive for Akt activation and GLUT4 translocation. In each experiment, basal (nonstimulated cells) 2-deoxyglucose uptake was assigned a value of 1.0. **, p Ͻ 0.001, *, p Ͻ 0.01 compared with control. phate or PI 5-phosphate (58,59) or signals complementary to the PI 3-kinase pathway. It must be stressed that the exogenously delivered PI(3,4,5)P 3 may be metabolized to some of those forms during the course of the loading period, and hence the biological responses observed here cannot be uniquely ascribed to PI(3,4,5)P 3 . Alternatively, a PI 3-kinase independent pathway has been proposed in adipocytes to contribute to the full mobilization of GLUT4, involving the sequential engagement of Cbl, CAP, and the small GTPase TC10 (60,61). However, this signaling cascade does not appear to operate in muscle cells (62), and the exact role of TC10 in adipocytes is still being defined (63).
The significant though partial translocation of GLUT4 to the plasma membrane caused by PI(3,4,5)P 3 was not accompanied by a parallel increase in 2-deoxyglucose uptake. Based on insulin dose-response curves of GLUT4 translocation and hexose uptake, we calculate that a partial increase in uptake should have been detectable for the extent of GLUT4 translocation by PI(3,4,5)P 3 in both cell types, if translocation alone sufficed to stimulate glucose uptake. The lack of stimulation of glucose uptake is not the result of lack of sensitivity of the assay because increases in 2-deoxyglucose uptake as small as 10% can be detected with statistical accuracy. Rather, it is possible that a threshold of PI(3,4,5)P 3 is required to stimulate glucose uptake (64), similar to a threshold proposed for eliciting GLUT4 translocation (56). Alternatively, again, other phosphoinositides might be required to impart glucose uptake properties to the translocated transporters in response to insulin. Interestingly, and coincident with the results reported here, Maffucci et al. (58) recently showed GLUT4 translocation in response to individual, different phosphoinositides without concomitant stimulation of glucose uptake. Here we show that adding PI 3-phosphate along with PI(3,4,5)P 3 also failed to provoke an increase in glucose uptake (Fig. 7). Hence, it is possible that in parallel to the PI 3-kinase lipid products, insulin generates additional inputs required for the stimulation of glucose uptake through the translocated GLUT4. These could include so far unidentified downstream targets of the serine kinase activity of proteins phosphorylated by PI 3-kinase or PI 3-kinase-independent signals (17,65). Regardless of the identity of the PI(3,4,5)P 3 -independent signal, it is clear that exogenously delivered PI(3,4,5)P 3 can cause a significant GLUT4 mobilization to the surface and insertion into the plasma membrane. Future work should focus on identifying the nature of the PI(3,4,5)P 3 complementary signal that is required to stimulate glucose uptake.
In conclusion, exogenous delivery of PI(3,4,5)P 3 to muscle and fat cells in culture causes a mobilization and membrane insertion of GLUT4, partly mimicking the response to insulin. The exogenous lipid also remodels the perinuclear GLUT4 pool and causes modest activation and redistribution of Akt. Despite measurable GLUT4 translocation, there is no concomitant stimulation of glucose uptake in PI(3,4,5)P 3 -stimulated cells. Signals additional to PI(3,4,5)P 3 or a specific amount and/or localization of PI(3,4,5)P 3 may be required to produce stimulation of glucose uptake through the translocated transporters.