INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

₂-Adrenergic receptors (₂ARs) are members of the superfamily of G protein-coupled receptors (GPCRs) and display desensitization in response to ₂-adrenergic agonists as well as counterregulation by several growth factor receptors with intrinsic tyrosine kinase activity (Morris and Malbon, 1999). Sequestration of the ₂AR plays a central role in agonist-induced and insulin-induced regulation of -adrenergic signaling (Lefkowitz, 1998; Morris and Malbon, 1999). ₂ARs are sequestered in response to insulin and this loss of the surface complement of receptors plays a critical role in the counterregulatory physiological effects of insulin on catecholamine action (Karoor et al., 1998). Although information has been gained on agonist-induced trafficking of GPCRs (Carman and Benovic, 1998; Gagnon et al., 1998), much less is known about how counterregulation by tyrosine kinases receptors influences GPCR trafficking.

₂ARs are substrates for insulin-stimulated tyrosine phosphorylation (Karoor and Malbon, 1998). In vivo, insulin stimulates the phosphorylation on two major tyrosine residues, Y350 and Y364, both residues located in the C-terminal cytoplasmic domain of the ₂AR (Karoor et al., 1995). Phosphorylation of the Y350 residue creates an SH2-binding site to which several molecules can dock, including Grb2, the p85 catalytic domain of phosphatidylinositol 3-kinase, and the GTPase dynamin (Shih and Malbon, 1998). In vitro, insulin stimulates the purified insulin receptor to phosphorylate recombinant ₂-adrenergic receptor on these same residues (Baltensperger et al., 1996; Doronin et al., 2000). The phosphorylation of the ₂-adrenergic receptor impairs its function, a blockade that requires Grb2 with an intact SH2 domain (Shih and Malbon, 1998). Although inducing a rapid, profound sequestration of the ₂AR, -adrenergic agonists and insulin display differences in the pathways by which the sequestration occurs (Karoor et al., 1998). The sequestration in response to insulin, but not -adrenergic agonist, can be blocked with inhibitors of phosphatidylinositol 3-kinase (PI3K), such as wortmannin or LY294002 [2-(4-morpholynyl)-8-phenyl-4H-1-benzopyran-4-one) (Wang et al., 2000).

Recently, the nonreceptor tyrosine kinase Src has been shown to be involved in -adrenergic agonist-induced desensitization (Luttrell et al., 1996), associating with the ₂AR and the scaffold protein gravin, also known as AKAP250 (Fan et al., 2001a,b), leading to its phosphorylation and activation of G protein receptor kinases (Ruiz-Gomez and Mayor, 1997; Sarnago et al., 1999). Whether Src functions in insulin-stimulated counterregulation of the ₂AR is not known. Investigation of the role of Src in this insulin-stimulated response is the focus of the current studies. Herein, we report that insulin activates Src, mediating signaling from the insulin receptor to the level of PI3-kinase activation and the trafficking of the ₂AR.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Materials

The plasmid encoding the enhanced green fluorescent protein (eGFP)-tagged human ₂AR (in pCDNA3) was obtained from Dr. Jeffrey Benovic (Kimmel Cancer Center, Thomas Jefferson University, Philadelphia, PA). The pCDNA3GFP-GLUT4 expression vector was a gift of Dr. Jeffrey Pessin (Department of Physiology and Biophysics, University of Iowa, Iowa City, IA), and the expression vector harboring the constitutively active Src Y527F mutant (CA-Src) was generously provided by Dr. Joan Brugge (Department of Cell Biology, Harvard Medical School, Boston, MA). To analyze and characterize an expression of green fluorescent protein (GFP)-tagged ₂AR, the following antibodies were used: anti-₂AR (CM02, antipeptide antibody to the exofacial domain of the ₂AR; Wang et al., 1989a) and anti-GFP rabbit polyclonal antibodies (Quantum Biotechnologies, Montreal, Quebec, Canada); goat anti-rabbit antibody conjugated with horseradish peroxidase (Kirkegarrd and Perry Laboratories, Gaithersburg, MD); and mouse anti-insulin receptor  subunit (p95; Transduction Laboratories, Lexington, KY).

Cell Culture

Human epidermoid carcinoma A431 wild-type cells and A431 clones stably expressing either eGFP-tagged ₂AR or eGFP-tagged GLUT4 glucose transporters were cultured in DMEM containing 10% fetal bovine serum, 100 U/ml penicillin, and 100 µg/ml streptomycin (George et al., 1988; Wang et al., 1991; Shih and Malbon, 1994, 1996; Shih et al., 1999). The mouse 3T3-L1 and 3T3-F442A cells (American Type Culture Collection, Manassas, VA) were cultured in the same medium with an addition of 1× nonessential amino acids (Invitrogen, Carlsbad, CA) and were induced to the adipocyte phenotype by the treatment with dexamethasone and 1-methylisobutylxanthine, as described previously (Green and Kehinde, 1975; Wang et al., 1992). Before analysis for insulin action, A431, 3T3-L1, and 3T3-F442 cells were serum starved overnight (Karoor et al., 1995).

Assay of Phosphorylation of Src

A431 cells were stimulated with 100 nM insulin for the indicated times, collected, and then lysed in a lysis buffer (150 mM NaCl, 5 mM EDTA, 50 mM NaF, 40 mM sodium pyrophosphate, 50 mM KH₂PO₄, 10 mM Na-molybdate, 2 mM Na-orthovanadate, 20 mM Tris-HCl, pH 7.4, 1% Triton X-100, 0.5% NP-40, 6 mM dithiothreitol, 10 µg/ml aprotinin, 10 µg/ml leupeptin, and 0.2 mM phenylmethylsulfonyl fluoride). Samples (50 µg of protein/lane) were subjected to electrophoresis on 10% polyacrylamide gels, the separated proteins transferred to nitrocellulose, and the blots stained with antibodies, as described previously (Jho et al., 1997). For these experiments, the blots were stained with rabbit anti-phospho-Src (Y416) antibodies purchased from Upstate Biotechnology (Lake Placid, NY). These antibodies specifically recognize only the Y416 phosphorylated, activated form of Src family members. Other antibodies used for immunoblotting of samples of cell lysates were anti-Src, anti-Fyn, anti-Lck, anti-Lyn, and anti-Yes, each obtained from Santa Cruz Biotechnology (Santa Cruz, CA).

Inhibitor and Morpholino Antisense Oligomer Studies

A431 cells stably expressing eGFP-tagged ₂-adrenergic receptors were pretreated for 30 min with one of the following inhibitors: the PI3K inhibitor LY294002 (20 µM), the mitogen-activated protein kinase kinase (MEK) inhibitor PD98059 (2'-amino-3'-methoxyflavone) (<µM), or the inhibitor of nonreceptor Src family kinases PP2 (50 nM). After pretreatment with an inhibitor, cells were incubated with or without 100 nM insulin for 30 min followed by live imaging of receptor localization. The morpholino antisense oligomers were designed and synthesized by Gene Tools (Philomath, OR). The antisense sequence selected for Src was 5'-ATGGGTAGCAACAAGAGCAAGCC-3', with the corresponding sense sequence used as a control. Cells were treated according to the protocol provided by Gene Tools. Treatment with the morpholinos was limited to 48 h, because extending the period of treatment to >72 h resulted in significant cell death.

Src Activity Assay

Whole-cell lysates were prepared from either A431 or 3T3-L1 clones. The phosphotransferase activity of Src kinase was determined in immunoprecipitates from cell lysates by using Src assay kit (Upstate Biotechnology), following the manufacturer's instructions. The preferred Src substrate peptide KVEKIGEGTYGVVYK corresponding to amino acids 6-20 of p34^cdc2 was used in these studies. The assay is linear for incubation times to 30 min. Samples were assayed in quadruplicate and the amount of incorporated label determined by scintillation counting.

PI3K Activity Assay

Crude cell membrane fractions were resuspended in a reaction buffer containing 10 mM Tris-HCl, pH 7.5, 100 mM NaCl, 1 mM EDTA, and 1 mM EGTA. An aliquot of membrane (100 µg/50 µl) was added to a reaction mixture containing 13 mM MnCl₂, 65 µM phosphatidylinositol diphosphate, 0.2 µg/µl phosphatidylserine, and 1 mM ATP (containing 1 µCi of [³²P]ATP). The reaction was initiated by addition of the protein sample and maintained at 30°C for 20 min. The reaction was stopped with 2 N HCl and then an aliquot (160 µl) of chloroform:methanol (1:1) mixture was added. The samples were subjected to centrifugation at 10,000 × g for 1 min and the lower phase was transferred to a new tube and washed by chloroform:methanol:0.1 N HCl (1:1:1) twice. The lower phase was collected and dried under vacuum. The samples were resuspended in 10 µl of chloroform:methanol (95:5) and then spotted onto the origin of a thin layer chromatography (TLC) plate (Song et al., 2001). The plate was developed in chloroform:methanol:25% NH₄OH:H₂O (100:70:25:15). The resolved plate was dried and exposed to films Eastman Kodak, Rochester, NY) or PhosphorImager (Molecular Dynamics, Sunnyvale, CA) cassette.

Radioligand Binding Studies

The number of 2ARs was determined by radioligand binding by using the high-affinity, -adrenergic antagonist [¹²⁵I]iodocyanopindolol (ICYP). Intact A431 cells were incubated with 0.5 nM ICYP (PerkinElmer Life Sciences, Boston, MA) in the presence or absence of 10 µM propranolol at 23°C for 90 min. The incubation buffer contained 50 mM Tris-HCl, pH 7.5, 10 mM MgCl₂, and 150 mM NaCl. The cells were collected under vacuum on GF/C membranes (Whatman, Maidstone, United Kingdom) and washed rapidly three times. The amount of ICYP bound to the washed cell mass retained by the filter was quantified by use of gamma counting (George et al., 1988).

Epifluorescence of Receptor/GLUT4 Translocation (Live Cell Microscopy)

Clones expressing either the eGFP-tagged ₂AR or the eGFP-tagged GLUT4 were seeded onto coverslips of either four- or eight-chamber slides. Confluent cells were serum deprived overnight, treated with the drugs, and carefully washed three times with HBB buffer (1× Hanks' balanced salt solution, 0.1% bovine serum albumin, 10 mM HEPES, pH 7.5; Invitrogen). Fresh DMEM medium without phenol red and serum was added and cells were imaged rapidly in an inverted fluorescent microscope (Nikon, Tokyo, Japan), fitted with a 40× objective (oil immersion). An acquisition of digital images was performed using a cooled charge-coupled device Princeton camera and WinView software (Roper Scientific, Inc., Trenton, NJ).

Confocal Microscopy

Clones were grown on glass slides, treated with drugs as indicated, fixed with 2% paraformaldehyde in phosphate-buffer saline, washed, and then treated with SlowFade Antifade kit (Molecular Probes, Eugene, OR) to prevent photobleaching. The confocal microscopy was performed on Eclipse E600 microscope (Nikon) using argon laser (488 nm). Images acquired with C-imaging system (Compix, Cranberry Township, PA) operating SimplePCI software.

Data Presentation and Analysis

Unless otherwise noted, the values presented are mean values ± SEM and are representative of multiple (at least three), independent experiments. Adobe Photoshop5.5 and Illustrator 8.0 (Adobe Systems, Mountain View, CA) were used to prepare final images and figures.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

To facilitate study of ₂AR sequestration in response to insulin, we stably transfected human epidermoid carcinoma A431 cells, a well-characterized model of ₂-adrenergic receptor biology (Delavier-Klutchko et al., 1984; Kassis et al., 1987; Lohse et al., 1990), with an expression vector harboring eGFP-tagged human ₂AR (Fan et al., 2001a). Stable clones were selected that expressed sufficient amounts of a GFP-tagged version of the human ₂-adrenergic receptor (₂AR-GFP) to enable high-quality epifluorescence microscopy, but contributed <10% to the endogenous level of the receptor (Figure 1). Immunoblotting of wild-type (WT) and A431 clones (₂AR-GFP) stained with anti-₂AR antibodies revealed the endogenous (M_r of 65 kDa) receptor in the WT cells as well as the endogenous and the GFP-tagged versions (M_r of ~95 kDa) in the A431 stable transfectants (Figure 1A). Staining of the blots with anti-GFP antibodies, in contrast, identified only the 95 kDa-M_r, GFP-tagged species in the A431 clones and identified no immunoreactivity in blots from the WT cells. Radioligand binding with the radiolabeled, high-affinity -adrenergic antagonist [¹²⁵I]iodocyanopindolol, revealed ~35,000 receptors/WT A431 cell and ~38,000 receptors/A431-transfected cell (our unpublished data). The GFP-tagged human ₂AR has been characterized fully and displays activation, desensitization, and sequestration in response to agonist, as does the native ₂AR (Kallal et al., 1998; Fan et al., 2001b).

View larger version (44K): [in this window] [in a new window]   Figure 1.   Insulin counterregulates catecholamine action by sequestration of 2ARs. A431 cells stably transfected with an expression vector harboring the cDNA for the human 2AR tagged with enhance GFP were used. (A) Immunoblots of WT cells and clones expressing 2AR-GFP were stained with antibodies to either 2AR or GFP. (B) Fluorescent images of live A431 cells that express 2AR-GFP were collected by using epifluorescence (inverted) microscopy after treatment of the cells with either the -adrenergic agonist isoproterenol (10 µM, +Iso) or insulin (100 nM, +Ins) for 30 min (bottom row). White arrows highlight receptors localized to the cell membrane; yellow arrowheads indicate the receptors in perinuclear zone. The same experiments were performed, except that in this case the cells were fixed and the images were acquired by confocal microscopy (top row). (C) Time course of 2AR-GFP translocation after A431 cell stimulation with insulin. A431 cells were stimulated with 100 nM insulin and imaged live on the charge-coupled device camera as repetitive frames were collected from the same spot over the time period of 35 min.

We made use of both epifluorescence and confocal microscopy to examine the effects of insulin on the localization of the ₂AR (Figure 1B). Treatment with insulin (100 nM, +Ins) provoked a substantial sequestration of ₂AR to perinuclear regions of the cell (yellow arrowheads) away from the cell membrane (white arrows), within 15 min of the addition of insulin. Challenge with the ₂-adrenergic agonist isoproterenol (10 µM, +Iso) was examined as a control. ₂ARs were observed to sequester in the same manner in response to stimulation by either insulin or by isoproterenol. The results of confocal microscopy were in agreement with the epifluorescence data, revealing a significant translocation of ₂AR from the cell membrane to a perinuclear locale in response to homologous (agonist) and heterologous (insulin) stimulation. A time course of ₂AR localization in response to 100 nM insulin extended to 35 min revealed the profound redistribution of ₂AR stimulated by insulin (Figure 1C). In the absence of insulin, the majority of the ₂AR is localized to the cell membrane. Within 30-35 min of insulin treatment the bulk the receptor is now found densely packed within a perinuclear region of the cell. Washout of the insulin was followed by a full recovery of the ₂AR to the cell membrane within 60 min, trafficking back from the perinuclear region of the cell (our unpublished data). ₂ARs display a high-level of recycling after internalization, with little discernable degradation over the period of a few hours (Wang et al., 1989b). Significant degradation (20-30%) of ₂AR is only detected after 24 h of chronic agonist treatment (Shenoy et al., 2001).

To unravel the details of the downstream signaling from insulin to the sequestration of ₂ARs, we first investigated roles of three likely signaling enzymes: PI3-kinase, the Src family of nonreceptor tyrosine kinases, and the mitogen-activated protein kinase cascade. Using selective inhibitors of each enzyme, we investigated whether inhibition would promote blockade of insulin-stimulated sequestration of ₂ARs (Figure 2A). A431 clones were treated with insulin and the sequestration of GFP-tagged ₂ARs examined by epifluorescence microscopy at 30 min postchallenge. The PI3-kinase inhibitor LY294002 effectively blocked the sequestration of ₂AR induced by insulin. The ability of isoproterenol to induce the agonist-specific sequestration of ₂AR, unlike insulin action, was unaffected by inhibition of PI3-kinase. Treatment with the PD98059 inhibitor of MEK in the mitogen-activated protein kinase pathway, in contrast to the effects of LY294002, did not alter the ability of insulin to sequester ₂AR. Previously, it was shown that the nonreceptor tyrosine kinase Src plays a role in agonist-specific desensitization and internalization of ₂ARs, so the effects of the PP2 inhibitor of Src family kinases were tested on insulin-induced sequestration of ₂ARs. We found that PP2 was a potent inhibitor of insulin-stimulated sequestration of ₂ARs (Figure 2A). These results piqued our interest in further exploring the seemingly obligate role of Src in this action of insulin.

View larger version (64K): [in this window] [in a new window]   Figure 2.   Inhibitors of PI3-kinase, inhibitors of Src phosphotyrosine kinases, and depletion of Src block insulin-induced sequestration of 2AR. (A) A431 clones were untreated (control) or treated with the PI3-kinase inhibitor LY294002, Src kinase inhibitor PP2, or the MEK inhibitor PD98059 and then stimulated with either isoproterenol (10 µM, +Iso) or insulin (100 nM, +Ins). After 30 min with the hormones, the cells were examined by epifluorescence microscopy. White arrows highlight cell membrane-localized receptors; yellow arrowheads highlight perinuclear-localized receptors, sequestered from the surface. (B) A431 cells were treated with morpholinos antisense to Src for 48 h to deplete cellular levels of Src as well as sense morpholinos as a control. Cells were then stimulated either with 100 nM insulin (+Ins) or without insulin (control). After 30 min with the hormones, the cells were examined by epifluorescence microscopy. White arrows highlight cell membrane-localized receptors; yellow arrowheads highlight perinuclear-localized receptors, sequestered from the surface. (C) Immunoblot of total cell lysates of A431 cells that were untreated (control) or treated with morpholinos either antisense (antisense) or sense (sense) to Src. The lysates were subjected to SDS-PAGE, the resolved proteins transferred to nitrocellulose blots, and stained with an anti-Src antibody. The blots were stained with antibody against the insulin receptor -subunit (p95 IR) to establish equivalence of sample loading.

Based upon the effects of the Src family kinase inhibitor, we explored the effects of the depletion of Src on the ability of insulin to provoke the sequestration of ₂ARs in A431 cells (Figure 2B). Depletion of Src was accomplished using morpholino oligomers (morpholinos) antisense to Src, and sense morpholinos were used as a control. Treating the cells with antisense morpholinos for 48 h resulted in a loss of insulin-stimulated sequestration of ₂AR, whereas treatment with sense morpholinos did not (Figure 2B). Immunoblots of whole-cell lysates stained with anti-Src antibodies demonstrate a depletion of Src (Figure 2C). Attempts to prolong the exposure to the antisense morpholinos resulted in cell death, suggesting that some level of Src expression is required for A431 cell viability. Parallel experiments performed with thioate(S)-modified oligodeoxynucleotides antisense to Src provided comparable data to the morpholinos (our unpublished data).

The activation of PI3-kinase is an early and essential response in the insulin-signaling cascade, so we explored whether Src kinase acts upstream or downstream of PI3-kinase (Figure 3). A431 cells were treated with insulin and the activation of PI3-kinase examined by assay of its product PI(3,4,5)P3 by using TLC. Insulin stimulated a rapid and robust increase in PI3-kinase activity, as shown in a representative chromatogram (Figure 3A). Analysis from replicate experiments revealed a two- to threefold increase in PI3-kinase activity in response to stimulation with 100 nM insulin in these cells (Figure 3B). As a control, we examined the effects of the addition of the PI3-kinase inhibitor LY294002, which abolished the ability of insulin to activate PI3-kinase. Remarkably, treating the cells with the Src family inhibitor PP2 proved equally effective as LY294002 in inhibiting insulin activation of PI3-kinase. These data suggest a critical role of Src kinase activation either upstream or collateral with the activation of PI3-kinase.

View larger version (26K): [in this window] [in a new window]   Figure 3.   Either LY294002 or Src inhibitor PP2 blocks activation of PI3-kinase by insulin. A431 cells were untreated (), treated with LY294002 (LY), or treated with the Src inhibitor PP2 (PP2), and then stimulated with 100 nM insulin for 30 min. (A) PI3-kinase activity was determined in the cells by measurement of its product phosphatidylinositol(3,4,5)P3 by using TLC. A representative experiment is shown. (B) Summation of the effects of LY294002 and PP2 on insulin stimulated PI3-kinase activity, displayed as arbitrary units with the control set as "1." The results are mean values from three separate experiments.

In the absence of reports in the literature regarding Src activation in response to insulin, we examined the obvious tenet that insulin stimulates Src activation (Figure 4). Src activation first was measured using activation-specific antibodies that recognize only the Y416 phosphorylated form of pp60Src. A431 cells treated with 100 nM insulin displayed a rapid activation of Src, detectable within 10 min and peaking within 30 min (Figure 4A). This time course for Src activation is not unlike that observed for the activation of insulin-sensitive pathways, such as GLUT4 translocation and glycogen synthase. Equivalent loading in these experiments was established by immunoblotting with antibodies to p60Src itself (Figure 4, A and B). The dose response for Src activation by insulin was investigated. Activation of Src was detected at the lowest concentration of insulin tested, 25 nM (Figure 4B). Maximal activation of Src occurred at concentrations of insulin from 50 to 100 nM, the same range of insulin concentration typically maximal for insulin-stimulated GLUT4 translocation and activation of glycogen synthase. Direct measurement of Src activity was performed using a preferred substrate, which likewise demonstrated activation of Src in response to insulin (Figure 4C). We next measured the Src activity in cells that were treated with either LY294002 or PP2. As predicted, treatment with PP2 inhibitor blocked insulin-stimulated activation of Src (Figure 4D). PP2 completely inhibits insulin-stimulated Src activity, but not the ambient levels of Src. The basis for this lack of inhibition of basal Src activity by PP2 is not clear. Treatment with LY294002, in sharp contrast, suppresses the ambient level of Src activity, but does not block Src activation. These data do demonstrate that activation of PI3-kinase is not obligate for insulin to activate Src, but that PI3-kinase seems necessary to sustain the ambient level of Src activity. Perhaps association with some element of the ₂AR signaling complex in the "basal", insulin-unstimulated state (Fan et al., 2001a,b) may alter Src susceptibility of PP2 inhibition.

View larger version (20K): [in this window] [in a new window]   Figure 4.   Insulin stimulates activation of Src. A431 cells were stimulated with insulin and the activation of Src determined by use of activation-specific antibodies for Src (pp60Src; A and B) or by analysis of Src activity by using a preferred substrate peptide (KVEKIGEGTYGVVYK; C). (A) Cells were stimulated with 100 nM insulin and the time course of Src activation was measured using activation-specific Src antibodies to stain immunblots of samples from the whole-cell lysates. Loading equivalence was checked using an antibody that recognizes Src (p60Src). (B) Cells were stimulated with 25-100 nM insulin for 30 min and Src activation measured using activation-specific Src antibodies to stain immunblots from the whole-cell lysates. Loading equivalence was checking using an antibody that recognizes Src (p60Src). (C) Cells were stimulated with 100 nM insulin and the time course of Src activation measured by assay of the phosphorylation of a Src preferred substrate by using Src immunoprecipitated with anti-Src protein A/G-agarose from whole-cell lysates. The results displayed are mean values of three separate experiments. (D) Cells were treated with either LY294002 or the Src kinase inhibitor PP2 and then stimulated with 100 nM insulin for 30 min. Src activation was measured by assay of the phosphorylation of Src preferred substrate by using Src immunoprecipitated from whole-cell lysates. The results displayed are mean values ±SEM of five separate experiments.

The GLUT4 hexose transporter is insulin sensitive; insulin stimulates its translocation from intracellular vesicles to the plasma membrane (Olefsky, 1999). In many respects, the sequestration of the ₂AR from the cell membrane to perinuclear vesicles in response to insulin seems functionally to be the inverse of insulin-stimulated trafficking of GLUT4 transporters, a process that moves GLUT4 from the perinuclear vesicles to the cell membrane. To examine this intriguing relationship, A431 cells were transfected with a expression vector harboring an eGFP-tagged GLUT4 transporter and then the response to a challenge with insulin was examined by epifluorescence microscopy (Figure 5). In sharp contrast to the situation with the ₂AR, the GLUT4 transporter is shown to reside in a perinuclear locale (see yellow arrowheads) in the absence of insulin and to be readily translocated to the cell membrane (white arrows) after a 30-min challenge with 100 nM insulin (Figure 5, a and d). Treating the cells with the PI3-kinase inhibitor LY294002 blocks the ability of insulin to translocate GLUT4 to the cell membrane (Figure 5, b and e), much as it blocks the ability of insulin to translocate the ₂AR to the perinuclear space (Figure 2A). These data are consistent with the observations that inhibition of PI3-kinase in many cells blocks insulin action. Treatment with the Src inhibitor PP2 also blocks the ability of insulin to translocate GLUT4 to the cell membrane (Figure 5, c and f). Thus, in the context of the A431 cells both PI3-kinase and Src activation seem obligate for insulin action with respect to ₂AR sequestration and GLUT4 translocation.

View larger version (66K): [in this window] [in a new window]   Figure 5.   Inhibition of PI3-kinase and Src phosphotyrosine kinases blocks insulin-induced translocation of GLUT4 to the cell membrane. A431 clones stably transfected with an expression vector harboring the mouse eGFP-tagged GLUT4 were either untreated (control) or treated with the PI3-kinase inhibitor LY294002 or the Src kinase inhibitor PP2, and then either stimulated with insulin (100 nM, +Ins) or untreated (basal). After 30 min of insulin stimulation, the cells were washed, fixed, and analyzed by epifluorescence confocal microscopy. White arrows highlight cell membrane-localized GLUT4; yellow arrowheads highlight GLUT4 residing in the cytoplasm.

We extended these results obtained in A431 cells to a well-studied model of insulin action, the mouse 3T3-L1 adipocyte. The 3T3-L1 cells were induced to differentiate into adipocytes by treatment with dexamethasone and methylisobutylxanthine (Green and Kehinde, 1975). The ability of insulin to stimulated activation of Src was measured using antibodies that recognize the activation-specific, Y416-phosphorylated form (pp60Src). The adipocyte cultures were stimulated with insulin and the whole-cell extracts were subjected to immunoblotting and staining with the anti-pp60Src antibody (Figure 6A). Sample loading was compared using immunoblotting and staining with antibodies specific for the p95 -subunit of the insulin receptor. Insulin stimulated a dose-dependent activation of Src, with activation observed at the lowest concentration of insulin tested, 25 nM. The dose response with respect to insulin and Src activation in these adipocytes was comparable with that observed in the A431 cells (compare Figures 4B and 6A). Direct measurement of activity by using a Src-specific substrate was performed in 3T3-L1 cultures treated with 100 nM insulin (Figure 6B). The time course was similar to that observed in the A431 cells (compare Figures 4A and 6B). To further extend these studies, we used the mouse 3T3-L1 derivative F442A cell line. We were unable to obtain suitable levels of expression of GFP-tagged ₂AR in the 3T3-L1 cells, but were successful using the F442A derivative for these same purposes. Mouse 3T3-F442A cells can be induced by various hormones to differentiate into adipocytes and have been used as a model for study of insulin action (Chen et al., 1989). Cells were transfected with GFP-tagged ₂AR and examined for the effects of insulin (Figure 6C). ₂AR sequestration in response to insulin was similar to that observed in the A431 cells. Similarly, treatment with the Src inhibitor PP2 blocked the ability of insulin to sequester ₂ARs in this model cell line of insulin action (compare Figures 2A and 6C). These data confirm the results obtained in the A431 cells, demonstrating not only insulin-stimulated Src activation but also insulin-stimulated sequestration of ₂AR in well-studied models of insulin action.

View larger version (23K): [in this window] [in a new window]   Figure 6.   Insulin stimulates activation of Src and sequestration of 2AR in mouse adipocytes in culture. Mouse 3T3-L1 cells were induced to differentiate into adipocytes. (A) Adipocyte cultures were stimulated with 25-200 nM insulin for 30 min and Src activation measured using activation-specific Src antibodies to stain immunblots of samples from the whole-cell lysates. Loading equivalence was determined by using an antibody that recognizes insulin receptor -subunit (p95 IR) to stain the blots also. (B) Adipocyte cultures were stimulated with 100 nM insulin and the time course of Src activation measured by assay of the phosphorylation of a Src preferred substrate by using Src immunoprecipitated from whole-cell lysates. The results displayed are mean values of three separate experiments. (C) Mouse 3T3-F442 cells, a cell line derived from the 3T3-L1 cell line and also used as a model of insulin signaling, were transiently transfected with GFP-tagged 2ARs and used to study the effects of PP2 inhibitor on the ability of insulin (30 min, 100 nM, +Ins) to stimulate 2AR sequestration. The images shown are representative of at least three separate experiments.

GPCRs signal via complexes that include scaffold proteins (e.g., AKAP79 and AKAP250), protein kinase A, protein kinase C, protein phosphatases, and other molecules (Shih et al., 1999; Dodge and Scott, 2000; Edwards and Scott, 2000; Fraser et al., 2000; Lin et al., 2000; Miller et al., 2000; Cong et al., 2001; Fan et al., 2001a,b). Phosphorylation of the ₂ARs in response to insulin at Y350 creates a docking site to which Grb2, the p85 regulatory subunit of PI3-kinase, dynamin, and Src may bind via SH2 domains (Shih and Malbon, 1996, 1998; Shih et al., 1999). We sought to investigate the association of Src with the ₂AR-signaling complex and ascertain its activation state, by using anti-pp60Src antibodies. A431 cells were either untreated (), or treated with insulin (+Ins, 100 nM) for 30 min, lysed, and subjected to pull-down assays performed with antibodies to the ₂ARs (Figure 7A). The ₂AR complexes were then subjected to immunoblotting and staining for content of Src or pp60Src. The results demonstrate a sharp increase in the amount of Src associated with the ₂AR complex (Figure 7A). Staining with activation-specific antibodies for pp60Src revealed activated Src accumulation with the ₂AR complex in response to insulin. Staining with antibodies against ₂AR established the equivalence of sample loading. Similar pull-down assays were performed with antibodies against Src (Figure 7B). Staining of the Src-associated complexes with anti-₂AR antibodies revealed the presence of ₂ARs when the complexes were pulled down with antibodies against Src. The results obtained with Src pull-downs from insulin-treated cells confirmed those obtained in the ₂AR-targeted pull-downs (Figure 7, A and B). We next examined whether the insulin-stimulated increase in Src association with ₂AR complexes was specific for Src or would be manifest for other members of the Src family. The results of these ₂AR-targeted pull-down assays reveal that Src, but not Lyn (both 53- and 56-kDa M_r splice variants), Fyn, and Yes, shows insulin-induced increase in association of Src with ₂AR signaling complexes (Figure 7C). Association of inactive Src with signaling complexes, including those containing the ₂AR (Fan et al., 2001a,b) and IIb3 integrins (Obergfell et al., 2002), provides an explanation for the association of Src with ₂AR, although not activated (Figure 7C). Insulin is shown to stimulate both a sharp increase in Src association with ₂AR (Figure 7C) and more importantly a sharp increase in Src activation (Figure 7A).

View larger version (13K): [in this window] [in a new window]   Figure 7.   Insulin stimulates Src, but not Lyn, Fyn, or Yes, association with 2AR signaling complexes. A431 cells were stimulated without () or with 100 nM insulin (+Ins) for 30 min and whole-cells lysates were prepared and used for pull-down immunoprecipitations (IPs) with antibodies specific for either 2ARs (A and C) or Src (B). The immunoprecipitates were washed and subjected to SDS-PAGE and immunoblotting. The blots were stained with antibodies to one of the following proteins: activated Src (pp60Src); Scr (p60Src); 2AR (2AR); Lyn, Fyn, or Yes. A representative blot is shown for each.

These studies demonstrate an obligate role of Src in the counterregulatory effects of insulin on ₂AR action, exerted at the level of ₂AR sequestration. In both the A431 cells and the 3T3-L1 adipocytes, we observed the ability of insulin to translocate Src to ₂AR signaling complexes and to activate Src. To further probe the linkage between Src activation and ₂AR sequestration, we first transfected A431 cells with an expression vector that harbors a constitutively active Y527F mutant form of chicken Src (CA-Src). Clones stably expressing CA-Src were then transiently transfected with the plasmid harboring the eGFP-tagged ₂AR. The influence of expression of CA-Src on the localization of GFP-tagged ₂AR was explored (Figure 8). The epifluorescence images reveal that in A431 cells, in the absence of insulin stimulation, the ₂ARs are localized largely to the cell membrane (Figure 8A). Expression of the CA-Src, in contrast, provoked redistribution of the ₂ARs to the perinuclear space (Figure 8B). In cells expressing the CA-Src, the pattern of ₂AR-GFP intracellular localization resembled the one induced by cell stimulation with insulin (Figures 2 and 8). Expression of CA-Src led to the activation of PI3-kinase activity (Figure 8, C and D). Thus, either activation of Src by insulin or the introduction of a constitutively active version of Src leads to sequestration of the ₂ARs from the cell membrane, a hallmark of the counterregulatory effect of insulin on catecholamine action.

View larger version (22K): [in this window] [in a new window]   Figure 8.   Expression of constitutively active Src activates PI3-kinase and provokes 2AR sequestration in the absence of insulin stimulation. A431 clones stably transfected with the constitutively active mutant of Src (CA-Src) or with empty vector (control) were transiently transfected with GFP-tagged 2ARs. At 30 h posttransfection, the clones were examined by epifluorescence microscopy (A and B). The images displayed are representative of numerous fields in replicate cultures. The PI3-kinase activity was assayed in the wild-type A431 cells and the clones expressing the constitutively active Src mutant under basal, unstimulated conditions (C and D). The image shown is representative of at least three separate experiments (C), data from which are summarized in D.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Src plays a pivotal role of many signaling pathways (Thomas and Brugge, 1997; Martin, 2001). Receptor-based pathways known to couple with Src kinases include integrins, cytokine receptors, immune recognition receptors, various ion channels, and GPCRs (Abram and Courtneidge, 2000). Src family kinases phosphorylate substrates that regulate cellular events such as transcription, differentiation, adhesion/migration, cell cycle progression, and apoptosis. Src itself, Fyn, Lyn, Lck, Yes, Hck, and other Src family members have been implicated in a variety of these events, depending on the cellular context of the signaling (Thomas and Brugge, 1997). Fyn has been shown to be involved in insulin growth factor-1 signaling events (Sun et al., 1996), but links between Src and insulin signaling are only first revealed in the present work.

Insulin signaling dominates two cellular events, mitogenesis, via the mitogen-activated protein kinase cascade, and metabolic regulation, highlighted by insulin action at skeletal muscle, liver, and adipose tissue (Czech and Corvera, 1999; Olefsky, 1999; Saltiel and Kahn, 2001). Catecholamines generally act in opposition to insulin. Catecholamines stimulate glycogen breakdown, protein degradation, gluconeogenesis, and lipolysis, whereas insulin alone acts to counteract each of these major metabolic pathways. The ability of insulin to counterregulate the ₂ARs is an essential element of insulin action (Morris and Malbon, 1999). Insulin has been shown to provoke the phosphorylation of ₂AR on specific tyrosyl residues confined to the cytoplasmic, C-terminal tail of the receptor (Karoor et al., 1995; Doronin et al., 2000). Phosphorylation of Y350 creates a docking site for proteins with SH2 domains (Baltensperger et al., 1996). The integrity of this residue and its phosphorylation in response to insulin are essential for the counterregulation of ₂AR function and ₂AR sequestration by insulin (Shih and Malbon, 1998).

Sequestration of ₂ARs in response to insulin is equal to or greater than that stimulated by -adrenergic agonist, translocating up to 85% of the cellular complement of ₂ARs away from the cell membrane (Karoor et al., 1998). Src has been implicated in agonist-induced sequestration of the ₂AR, through its ability to phosphorylate and activate the G protein-coupled receptor kinase (GRK2) (Ruiz-Gomez and Mayor, 1997; Sarnago et al., 1999; Fan et al., 2001a). In the current study, we demonstrate that counterregulation of ₂ARs by insulin requires Src activation. We speculate that the mobilization and activation of Src by the ₂AR signaling complex in response to insulin may lead subsequently to activation of GRK2 (Sarnago et al., 1999), which would also act to terminate ₂AR signaling in counterregulation by insulin.

Based upon the advanced knowledge of Src structure and biology, we were surprised to find no literature implicating Src signaling in insulin action. Our results clearly demonstrate that Src is activated and associates with ₂AR signaling complexes in response to insulin. Src, but not Lyn, Fyn, or Yes, displays increased insulin-induced association with ₂AR signaling complexes. Inhibition of Src blocks insulin-stimulated ₂AR sequestration. This insulin-induced activation of Src required for ₂AR sequestration was observed not only in studies performed in human epidermoid carcinoma A431 cells but also in the mouse 3T3-L1 adipocytes, a well-accepted model for study of insulin action (Green and Kehinde, 1975). In A431 cells, inhibition of Src with PP2 blocks the sequestration of ₂ARs and the well-known translocation of GLUT4 to the cell membrane in response to insulin.

The interrelationships between insulin and catecholamine action are many in metabolic regulation, epitomized by the counterregulation of -catecholamine action exerted through insulin-stimulated phosphorylation of the ₂AR at two dominant sites, Y350 and Y364, both located in the cytoplasmic, C-terminal tail of the receptor (Figure 9). These sites of phosphorylation have been demonstrated in vivo through metabolic labeling and peptide chemistry as well as in vitro by using purified insulin receptor and r₂ARs (Karoor et al., 1995; Baltensperger et al., 1996). Phosphorylation of Y364 generates a form of the phospho-₂AR that is a potent, feedback inhibitor of the insulin receptor tyrosine kinase (Doronin et al., 2002). Phosphorylation of the Y350 residue impairs the ability of the ₂AR to signal to Gs and facilitates the sequestration of the ₂AR in response to insulin. We found that the Y350F mutation blocks the ability of insulin to counterregulate the ₂AR- and insulin-induced receptor sequestration (Karoor et al., 1995). Thus, the ability of this site to be phosphorylated and therefore to create a docking site for proteins with SH2 domains is essential for many regulatory events designed to uncouple the ₂AR and sequester it away from the cell membrane.

View larger version (28K): [in this window] [in a new window]   Figure 9.   Schematic model of insulin counterregulation of 2ARs. The phosphorylation of Y350 of the 2-adrenergic receptor creates an SH2-binding site to which Src is recruited upon insulin stimulation. The 2AR is organized in a complex with protein kinase A, protein kinase C, GRK2, and other signaling elements by the AKAP 250 scaffold protein gravin. The effects of insulin can be mimicked by the expression of constitutively activated Src. Inhibition of Src kinase with PP2, by expression of a DN-Src mutant, or the Y350F mutation of the 2-adrenergic receptor abolishes the ability of Src to be recruited/activated in response to -adrenergic agonist and thereby precludes both desensitization and eventually full internalization of the receptor-Src complex. In this manner, the internalization of 2AR in response to insulin has many similarities to the internalization stimulated by -adrenergic agonists after activation. GLUT4 glucose transporters, in contrast to the situation for 2ARs, are localized within the cell in the absence of insulin. Stimulation by insulin leads to a translocation of GLUT4-laden vesicles to the cell membrane for fusion. Thus, GLUT4 and the 2-adrenergic receptor seem to traffic in opposite manners in response to insulin. Whether these two signaling molecules can be localized to the same vesicles, or different vesicles, remains to be established.

Still unresolved is the question how Src is activated in response to insulin. Based upon several lines of data demonstrating Src involvement in agonist-induced sequestration of ₂ARs and insulin-stimulated sequestration of ₂ARs, the answer seems very interesting to ponder. We speculate that the ₂ARs, upon tyrosyl phosphorylation in response to insulin, provide a docking site for Src that may facilitate activation of Src itself (Figure 9). Some Src associated with the ₂AR signaling complex in the basal state does not seem to be fully activated, perhaps representing Src binding to other elements in the ₂AR signaling complex (Fan et al., 2001a,b). Displacement of the intramolecular interactions of the SH2 domain by an SH2 docking site like that in the phosphorylated ₂ARs and subsequent autophosphorylation of Src Y416, however, are both well-known mechanisms that promote activation of Src phosphotyrosine kinases. Mutagenesis studies of the ₂AR Y350 site support the first tenet, whereas the studies performed with antibodies specific for the phosphorylated Y416 motif demonstrate the second tenet. The activation of Src seems essential for ₂AR sequestration based upon three findings: the ability of PP2 to inhibit insulin-induced ₂AR sequestration; the ability of depletion of Src with antisense morpholinos to block insulin-stimulated ₂AR sequestration; and the ability of the constitutively active version of Src to provoke ₂AR sequestration in the absence of insulin stimulation.

In A431 cells, PP2 blocked not only the activation of Src but also the activation of PI3-kinase in response to insulin (Figure 9). If Src activity were required for insulin activation of PI3-kinase, one might expect that PP2 would block all PI3-kinase-dependent downstream events. The recent observation that PP2 does not block insulin-stimulated glucose uptake in mouse 3T3-L1 adipocytes would seem to argue against this obligate role of Src (Pessin and Saltiel, 2000; Imamura et al., 2001). It has been shown for at least the GLUT4 translocation response that PI3-kinase activation is not required for insulin action, and the CAP/Cbl pathway might provide an alternative means to translocate GLUT4 in response to insulin (Pessin and Saltiel, 2000). In this regard, it is of interest that for the A431 cells not only does PP2 block insulin-stimulated ₂AR sequestration from the plasma membrane but also GLUT4 translocation to the cell membrane in response to insulin. The cellular "context" of insulin signaling obviously plays an important role and the current results demonstrate that Src activation may be important for insulin signaling in cellular events other than ₂AR sequestration, for example, GLUT4 translocation. Much further work will be required to define the full family of GPCRs whose trafficking is influenced by insulin and/or other growth factors. Clearly, the well-known counterregulatory effects of insulin on catecholamine action include a central role for Src.