Inhibition of insulin-induced glucose uptake by atypical protein kinase C isotype-specific interacting protein in 3T3-L1 adipocytes.

Atypical protein kinase C (PKC) isotype-specific interacting protein (ASIP) specifically interacts with the atypical protein kinase C isozymes PKClambda and PKCzeta. ASIP and atypical PKC, as well as their Caenorhabditis elegans counterparts (PAR-3 and PKC-3, respectively), are thought to coordinately participate in intracellular signaling that contributes to the maintenance of cellular polarity and to the formation of junctional complexes. The potential role of ASIP in other cellular functions of atypical PKC was investigated by examining the effect of overexpression of ASIP on insulin-induced glucose uptake, previously shown to be mediated through PKClambda, in 3T3-L1 adipocytes. When overexpressed in these cells, which contain PKClambda but not PKCzeta, ASIP was co-immunoprecipitated with endogenous PKClambda but not with PKCepsilon or with Akt. The subcellular localization of PKClambda was also altered in cells overexpressing ASIP. Overexpression of ASIP inhibited insulin stimulation of both glucose uptake and translocation of the glucose transporter GLUT4 to the plasma membrane, but it did not inhibit glucose uptake induced by either growth hormone or hyperosmolarity both of which promote glucose uptake in a PKClambda-independent manner. Moreover, glucose uptake stimulated by a constitutively active mutant of PKClambda, but not that induced by an active form of Akt, was inhibited by ASIP. Insulin-induced activation of PKClambda, but not that of phosphoinositide 3-kinase or Akt, was also inhibited by overexpression of ASIP. These data suggest that overexpression of ASIP inhibits insulin-induced glucose uptake by specifically interfering with signals transmitted through PKClambda.

Members of the protein kinase C (PKC) 1 family of serinethreonine kinases were originally identified as effectors of lipid-mediated intracellular signaling (1,2). Three classes of PKC isozymes have been defined: conventional PKC, consisting of PKC␣, PKC␤I, PKC␤II, and PKC␥; novel PKC, comprising PKC␦, PKC⑀, PKC, and PKC; and atypical PKC, including PKC and PKC (2,3). Atypical PKC isozymes are distinct from other members of the PKC family in that they are not activated by diacylglycerol or phorbol esters (1)(2)(3)(4). These isozymes are thought to act downstream of phosphoinositide (PI) 3-kinase. PKC is activated in vitro by phosphatidylinositol 3,4,5trisphosphate (5), a cellular product of PI 3-kinase action, whereas PKC was shown in transfected cells to contribute in a PI 3-kinase-dependent manner to trans-activation of the 12-O-tetradecanoylphorbol-13-acetate (phorbol ester)-responsive element induced by platelet-derived growth factor or epidermal growth factor (4). Furthermore, both the phosphorylation and activity of atypical PKC stimulated by growth factors were shown to be inhibited by either a pharmacological inhibitor of or a dominant negative mutant of PI 3-kinase (4, 6 -9).
Atypical PKC isotype-specific interacting protein (ASIP) was recently identified as a protein that specifically associates with atypical PKC (10). ASIP does not contain any known catalytic domain but possesses three PDZ domains (10), which mediate protein-protein interactions at the plasma membrane (11). PAR-3, the Caenorhabditis elegans ortholog of ASIP, also interacts with the worm counterpart of atypical PKC (PKC-3) (12). Worm embryos lacking PAR-3 show severe defects in asymmetrical cell division, and PKC-3-depleted cells show a phenotype highly similar to that of PAR-3-depleted cells (12). Furthermore, Bazooka, a protein essential for asymmetrical cell division in Drosophila neuroblasts, has been identified as a fly ortholog of ASIP and PAR-3 (13,14). Although a fly homolog of atypical PKC has not been identified, these observations, together with the fact that ASIP and PKC colocalize at junctional complexes in mammalian cells (10), suggest that ASIP and atypical PKC coordinately function in the maintenance of cellular polarity and cell-cell interaction.
We have now investigated whether ASIP also contributes to other cellular functions of atypical PKC. Stimulation of glucose uptake by insulin in its target cells, such as skeletal muscle and adipocytes, is an important biological action of this hormone. Pharmacological and molecular biological evidence has revealed that PI 3-kinase plays a major role in insulin-induced glucose uptake (15)(16)(17). Although the precise mechanism by which activation of PI 3-kinase results in glucose uptake remains unclear, we and others have shown that a kinase-deficient mutant of atypical PKC inhibits insulin-induced glucose uptake (7,8) and that a constitutively active mutant of atypical PKC stimulates glucose uptake in quiescent cells (8). Furthermore, microinjection of antibodies specific to PKC inhibited insulin-induced translocation of the glucose transporter GLUT4 to the plasma membrane (18). These results support the hypothesis that atypical PKC participates in the signaling pathway by which insulin stimulates glucose uptake.
To elucidate whether ASIP contributes to insulin stimulation of glucose uptake, we have examined the effects of overexpression of ASIP on glucose uptake induced by insulin and by various other stimuli, as well as on the activation of signaling molecules by insulin, in cultured adipocytes. We now provide evidence that overexpression of ASIP inhibits insulin-induced glucose uptake by specifically interfering with signaling mediated through PKC.

EXPERIMENTAL PROCEDURES
Cells and Antibodies-3T3-L1 preadipocytes (American Type Culture Collection) were maintained and induced to differentiate into adipocytes as described previously (8). Polyclonal antibodies to PKC that were generated in response either to a peptide corresponding to amino acids 197-213 (␣197) or to a glutathione S-transferase fusion protein containing amino acids 190 -240 (␣190) of mouse PKC were as described (8). Polyclonal antibodies to ASIP (10), to GLUT4 (19), or to Akt (the latter of which recognize all three isoforms of Akt) (20) were described previously. Polyclonal antibodies to PKC⑀, a monoclonal antibody to PKC, polyclonal antibodies to Akt2, and a monoclonal antibody to the T7 epitope tag were obtained from Santa Cruz Biotechnology, Transduction Laboratories, Upstate Biotechnology, and Novagen, respectively.
Construction of and Infection with Adenovirus Vectors-Construction of an adenovirus vector encoding ASIP will be described elsewhere. 2 In short, cDNA encoding T7 epitope-tagged rat ASIP (10) was subcloned into pAxCAwt (21), and an adenovirus vector containing the cDNA was generated by transfecting 293 cells with the resulting plasmid and DNA-terminal protein complex with the use of an adenovirus expression kit (Takara, Tokyo, Japan). Adenovirus vectors encoding constitutively active mutants of PKC or Akt (AxCA⌬PD and AxCAMyr-Akt, respectively) were described previously (8,22). Fully differentiated 3T3-L1 adipocytes were infected with various adenovirus vectors as described (17) and subjected to experiments after ϳ48 h.
Protein Kinase Assays-3T3-L1 adipocytes were deprived of serum for 16 -20 h, incubated in the absence or presence of 100 nM insulin for the indicated time, and then immediately frozen with liquid nitrogen. For assay of PI 3-kinase activity, the cells were lysed and subjected to immunoprecipitation with antibodies to phosphotyrosine (PY20; Transduction Laboratories); the resulting immunoprecipitates were washed, and PI 3-kinase activity in the washed precipitates was assayed as described previously (20). For assay of Akt activity, the cells were lysed and subjected to immunoprecipitation with antibodies to Akt, and the resulting precipitates were assayed for kinase activity with histone 2B as substrate, as described (20). For assay of PKC activity, cells were lysed as described previously (4) and subjected to immunoprecipitation with ␣197. The resulting immunoprecipitates were washed twice with buffer A (50 mM MOPS-HCl (pH 7.5), 0.5% (v/v) Triton X-100, 10% (v/v) glycerol, 0.1% (w/v) bovine serum albumin, 5 mM EDTA, 5 mM EGTA, 20 mM NaF, 50 mM ␤-glycerophosphate, 2 mM sodium orthovanadate, 2 mM dithiothreitol, leupeptin (1 g/ml), 2 mM phenylmethylsulfonyl fluoride), once with buffer A containing 1 M NaCl, and once with a solution containing 20 mM Tris-HCl (pH 7. All protein kinase reactions were terminated by the addition of SDS sample buffer, and the reaction mixtures were then fractionated by SDS-polyacrylamide gel electrophoresis. The radioactivity incorporated into substrates was determined with a Fuji BAS 2000 image analyzer. Glucose Uptake-Glucose uptake was assayed as described previously (8). In brief, 3T3-L1 adipocytes infected (or not) with adenovirus vectors were incubated for 16 h in Dulbecco's modified Eagle's medium containing 5.6 mM glucose and 0.5% fetal bovine serum. The cells were washed twice with DB buffer (140 mM NaCl, 2.7 mM KCl, 1 mM CaCl 2 1.5 mM KH 2 PO 4 , 8 mM Na 2 HPO 4 (pH 7.4), 0.5 mM MgCl 2 ) and then incubated in DB buffer with 100 nM insulin for 10 min, growth hormone (GH) (0.5 g/ml) for 10 min, or 300 mM sorbitol for 60 min. DB buffer (1 ml) containing bovine serum albumin (1 mg/ml) and 0.1 mM 2-deoxy-D-[1,2- Subcellular Fractionation-3T3-L1 adipocytes in 10-cm plates were deprived of serum for 16 -20 h, incubated in the absence or presence of 100 nM insulin for 15 min, washed three times with HES buffer (20 mM Hepes-NaOH (pH 7.4), 1 mM EDTA, 250 mM sucrose) on ice, and gently scraped into HES buffer containing 250 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 10 g/ml leupeptin, 10 mM NaF, 10 mM ␤-glycerophosphate, 10 mM sodium pyrophosphate, 1 mM sodium orthovanadate, and 1 mM dithiothreitol. The cells were then passed through a 22-gauge needle 10 times, after which the low density microsome (LDM), plasma membrane (PM), and cytosolic fractions were isolated by differential centrifugation as described (23). After determination of protein concentration with the use of the BCA protein assay reagent (Pierce), equal amounts of protein (ϳ20 g) from each fraction were subjected to immunoblot analysis with various antibodies.

RESULTS
Coprecipitation of ASIP with PKC from 3T3-L1 Adipocytes-ASIP was identified as a protein that specifically interacts with the atypical PKC isoforms PKC and PKC (10). We therefore first examined whether ASIP binds to atypical PKC in 3T3-L1 adipocytes, which express PKC exclusively (8). The cells were infected (or not) with an adenovirus vector that encodes T7 epitope-tagged ASIP (AxCAASIP), lysed, and subjected to immunoprecipitation with antibodies to T7 or to PKC or with control immunoglobulin G (IgG). The resulting immunoprecipitates were then subjected to immunoblot analysis with antibodies to T7 or to PKC. A protein of ϳ180 kDa, corresponding to the molecular size of ASIP, was recognized by antibodies to T7 in the immunoprecipitates prepared with antibodies to either T7 or PKC, but only when the cells had been infected with AxCAASIP (Fig. 1). Conversely, a protein of ϳ80 2 T. Onishi and S. Ohno, manuscript in preparation. kDa, corresponding to the molecular size of PKC, was recognized by antibodies to PKC in the immunoprecipitates prepared with antibodies to PKC or to T7. No marked immunoreactive bands were detected in the immunoprecipitates prepared with control IgG. These data indicate that PKC associates with ASIP in 3T3-L1 adipocytes. The immunoprecipitates with antibodies to PKC prepared from noninfected 3T3-L1 adipocytes did not react to antibodies to ASIP (data not shown). This may be because a relatively small amount of ASIP protein is expressed in the cells.
To investigate the specificity of the interaction between ASIP and PKC, we performed similar experiments with antibodies to PKC⑀ or to Akt. However, the 180-kDa protein recognized by antibodies to T7 was not evident in the immunoprecipitates prepared with antibodies to either PKC⑀ or Akt, and neither PKC⑀ nor Akt was detected in the immunoprecipitates prepared with antibodies to T7 (Fig. 1). The interaction of ASIP with PKC thus appeared to be specific, consistent with previous observations (10). Treatment of the cells with insulin did not affect the interaction between ASIP and PKC (data not shown).
Effect of ASIP on the Subcellular Localization of PKC-We next investigated the effect of overexpression of ASIP on the subcellular localization of PKC. 3T3-L1 adipocytes infected (or not) with AxCAASIP were lysed, and LDM, PM, and cytosolic fractions were prepared by differential centrifugation. Equal amounts of protein from each fraction were then subjected to immunoblot analysis with antibodies to either ASIP or PKC. ASIP was detected only in the LDM fraction of noninfected cells, indicating that endogenous ASIP is abundant in this fraction (Fig. 2). In cells infected with AxCAASIP, however, ASIP was apparent not only in the LDM fraction but also in the PM and cytosolic fractions. Endogenous PKC was abundant in the LDM and cytosolic fractions of noninfected cells, with a relatively small amount also apparent in the PM fraction. Overexpression of ASIP reduced the amount of PKC in the cytosolic fraction and increased it in both the LDM and PM fractions. Exposure of infected or noninfected cells to insulin had no marked effect on the intracellular distribution of ASIP or of PKC. Thus, overexpression of ASIP altered the intracellular localization of endogenous PKC.
Inhibition of Insulin-induced Glucose Uptake by ASIP-Given that PKC is thought to participate in insulin-induced glucose uptake and GLUT4 translocation in 3T3-L1 adipocytes (8, 18), we examined the effects of overexpression of ASIP on these actions of insulin. Infection of the cells with AxCAASIP resulted in a dose-dependent increase in the amount of ASIP protein as assessed by immunoblot analysis; the amount of ASIP in cells infected at a multiplicity of infection (m.o.i.) of 150 plaque-forming units (pfu)/cell was ϳ20 times that of endogenous ASIP (Fig. 3A). Insulin induced an ϳ7-fold increase in glucose uptake in 3T3-L1 adipocytes within 10 min (Fig. 3A). Glucose uptake data are expressed as fold stimulation relative to the extent of uptake observed with noninfected cells not exposed to insulin. B, 3T3-L1 adipocytes were infected (or not) with AxCAASIP at an m.o.i. of 150 pfu/cell, incubated in the absence or presence of 100 nM insulin for 15 min, and then subjected to subcellular fractionation for the isolation of LDM and PM fractions. Equal amounts of protein (ϳ20 g) from each fraction were subjected to immunoblot analysis with antibodies to GLUT4. C, 3T3-L1 adipocytes were infected (or not) with AxCAASIP at an m.o.i. of 150 pfu/cell, incubated in the absence or presence of GH (0.5 g/ml) for 10 min or 300 mM sorbitol for 60 min, and then assayed for glucose uptake. Data are expressed as fold stimulation relative to the uptake apparent in noninfected, nonstimulated cells. Quantitative data in A and C are mean Ϯ S.E. from three experiments. Immunoblot data are representative of three independent experiments.
Expression of ASIP inhibited insulin stimulation of glucose uptake in a dose-dependent manner, with ϳ50% inhibition apparent at an m.o.i. of 150 pfu/cell; ASIP had little effect on the basal level of glucose uptake apparent in the absence of insulin. We next investigated the effect of ASIP on the translocation of GLUT4 by subcellular fractionation. Treatment of noninfected cells with insulin reduced the amount of GLUT4 in the LDM fraction and increased the amount of GLUT4 in the PM fraction (Fig. 3B). Infection of cells with AxCAASIP, at an m.o.i. of 150 pfu/cell, markedly inhibited the insulin-induced decrease in the amount of GLUT4 in the LDM fraction as well as the increase in its abundance in the PM fraction. The extent of ASIP-induced inhibition of GLUT4 translocation was similar to that of ASIP-induced inhibition of glucose uptake. The amount of GLUT4 in total cell lysates was not affected by overexpression of ASIP (data not shown). These results suggest that overexpression of ASIP inhibited insulin-induced glucose uptake by preventing the intracellular translocation of GLUT4.
Effects of ASIP on Glucose Uptake Induced by Various Stimuli-GH and hyperosmolarity each stimulate glucose uptake by promoting translocation of GLUT4 (24,25). The observation that glucose uptake induced by these stimuli was not inhibited by a dominant negative mutant of PKC (8) suggests that it is independent of PKC. Infection of 3T3-L1 adipocytes with Ax-CAASIP at an m.o.i. of 150 pfu/cell, a virus dose sufficient to inhibit insulin-induced glucose uptake by ϳ50%, had no effect on hyperosmolarity induced glucose uptake and actually potentiated GH-induced glucose uptake (Fig. 3C), suggesting that expression of ASIP inhibits glucose uptake by specifically interfering with signaling through PKC.
To examine further this hypothesis, we investigated the effect of ASIP on glucose uptake stimulated by a constitutively active mutant of PKC. The ⌬PD mutant lacks the pseudosubstrate domain and exhibits a kinase activity markedly greater than that of the wild-type enzyme (8). Given that ASIP binds to the kinase domain of atypical PKC (10), it is likely that ASIP also interacts with ⌬PD. Consistent with our previous observation (8), expression of ⌬PD stimulated glucose uptake in quiescent 3T3-L1 adipocytes. Overexpression of ASIP inhibited ⌬PD-induced glucose uptake in a dose-dependent manner (Fig. 4A). Expression of constitutively active mutants of Akt also stimulates glucose uptake and translocation of GLUT4 in adipocytes (26,27). Indeed, infection of 3T3-L1 adipocytes with AxCAMyr-Akt, an adenovirus vector that encodes a constitutively active mutant of Akt (22), stimulated glucose uptake in 3T3-L1 adipocytes in the absence of insulin (Fig. 4B). However, expression of ASIP did not inhibit Myr-Akt stimulation of glucose uptake, suggesting that ASIP does not interrupt signaling through Akt.
Effects of ASIP on Various Signaling Molecules Activated by Insulin-To elucidate further the mechanism by which ASIP inhibits insulin-induced glucose uptake, we investigated the effects of overexpression of this protein on the activation of various signaling molecules by insulin. Insulin-induced activation of PI 3-kinase or of Akt was not affected by ASIP (Fig. 5, A  and B). Furthermore, overexpression of ASIP had no effect on the insulin-induced translocation of Akt to the PM fraction (Fig. 5C).
Finally, we examined the effect of ASIP overexpression on kinase activity immunoprecipitated with antibodies to PKC. Exposure of cells to insulin increased the amount of kinase activity, measured with myelin basic protein as substrate, in such immunoprecipitates by a factor of ϳ1.8 (Fig. 6A). Overexpression of ASIP inhibited the insulin-induced kinase activity in a dose-dependent manner, with complete inhibition apparent at an m.o.i. of 150 pfu/cell (Fig. 6A). The amount of PKC protein in the immunoprecipitates prepared with antibodies to PKC was increased by expression of ASIP in a dose-dependent manner (data not shown), with an ϳ2-fold increase apparent at an m.o.i. of 150 pfu/cell (Fig. 6B). DISCUSSION Given that cell permeable pharmacological inhibitors or activators of atypical isoforms of PKC are not available, the cellular functions of these isozymes have been explored with the use either of pseudosubstrate peptides that specifically compete with endogenous substrates for kinase activity or of mutant enzymes that either are constitutively active or act in a dominant negative manner. Such tools have revealed putative roles for atypical PKC in oocyte maturation (28), mitogenesis (29), protection of cells from apoptosis (30), cell differentiation (31), and insulin-induced glucose uptake (7,8). Another approach to investigating the cellular functions of signaling molecules is to identify proteins that directly interact with them. Several proteins that associate with atypical PKC have been identified, including p62/ZIP (32), LIP (33), PAR-4 (34), UNC-76 (35), and the small GTP-binding protein RAS (36). The cDNA that encodes ASIP was also isolated as a result of screening an expression library with PKC as a probe (10).
With the use of adenovirus-mediated gene transfer, we have now shown that overexpression of ASIP in 3T3-L1 adipocytes inhibited insulin stimulation of both glucose uptake and GLUT4 translocation. GH and hyperosmolarity (sorbitol) also stimulate glucose uptake but in a manner insensitive to a dominant negative mutant of PKC (8), suggesting that these stimuli induce glucose uptake through a PKC-independent mechanism. Our observation that overexpression of ASIP did not inhibit glucose uptake in response to either GH or hyperosmolarity suggests that ASIP attenuates glucose uptake not by an effect on a common component of the transport machinery but by interfering with specific signals transmitted through PKC. This hypothesis was further supported by the observation that ASIP inhibited glucose uptake induced by a constitutively active mutant of PKC but not that induced by an active form of Akt. It is not clear why overexpression of ASIP enhanced glucose uptake induced by GH. Given that a dominant negative mutant of PKC did not exhibit such an effect (8), it is not likely that prevention of signaling through PKC was responsible for this action of ASIP. We therefore cannot exclude the possibility that overexpression of ASIP affects not only PKC-mediated signaling but also an unidentified signaling pathway that contributes to GH-induced glucose uptake.
The results of the present and previous (26,27) studies have shown that constitutively active mutants of Akt stimulate both glucose uptake and the translocation of GLUT4 in quiescent cells. However, studies on the effects of kinase defective Akt mutants on insulin-induced glucose uptake or translocation of GLUT4 have produced conflicting results (20,37,38). It is thus unclear whether Akt plays a role in insulin-stimulated glucose uptake. We have now shown that the activation of Akt by insulin was not affected in cells overexpressing ASIP, whereas insulin-induced glucose uptake in these cells was inhibited by ϳ50%. Moreover, insulin-induced translocation of Akt to the plasma membrane, thought to be an important step in Akt activation (39), was not inhibited by ASIP, consistent with the observation that ASIP did not associate with Akt in intact cells.
It is not likely that ASIP inhibits insulin-induced glucose uptake by interfering with signaling downstream of Akt, because ASIP had no effect on glucose uptake induced by a constitutively active mutant of Akt. Although our data do not exclude a role for Akt in insulin stimulation of glucose uptake, insulininduced activation of Akt alone is likely not sufficient to fully stimulate glucose uptake, at least in 3T3-L1 adipocytes. Immunocytochemical analysis of MDCKII and NIH 3T3 cells has revealed that endogenous ASIP is specifically localized to regions of cell-cell contact, most likely at tight junctions, whereas PKC not only colocalizes with ASIP in these regions but is also distributed throughout the cytoplasm (10). With the use of subcellular fractionation analysis, we have now shown that overexpression of ASIP alters the subcellular localization of endogenous PKC in 3T3-L1 adipocytes. Whereas PKC was abundant in the LDM and cytosolic fractions of control cells, overexpression of ASIP resulted in a decrease in the amount of PKC in the cytosolic fraction and in marked and small in- creases in the amounts of the isozyme in the LDM and PM fractions, respectively. The observation that recombinant ASIP was also abundant in the LDM and PM fractions is consistent with the notion that the altered subcellular localization of PKC is attributable to its direct interaction with ASIP.
The mechanism by which ASIP inhibits PKC signaling in the pathway that leads to glucose uptake remains unclear. Our observation that the amount of PKC immunoprecipitated by antibodies to this isozyme from cells overexpressing ASIP was about twice that precipitated from control cells might be explained if one ASIP molecule interacts with multiple PKC molecules or if ASIP forms a multimeric complex in cells. Despite the increase in the amount of PKC protein, the kinase activity in the immunoprecipitates prepared from cells overexpressing ASIP was lower than that in precipitates prepared from control cells, suggesting that PKC activity is suppressed when it is bound to ASIP. It is thus possible that overexpression of ASIP inhibits PKC activity in the cells, thereby resulting in inhibition of insulin-induced glucose uptake. At present, we do not know the mechanism how overexpression of ASIP reduces PKC activity. Because the activity of PKC isoenzymes is regulated by its phosphorylation (3), it is possible that ASIP affects phosphorylation of specific residues of PKC.
Biochemical and genetic evidence suggests that ASIP, as well as its worm (PAR-3) and fly (Bazooka) counterparts, participates in asymmetrical cell division or maintenance of cellular polarity (10,(12)(13)(14). Because ASIP does not contain any known catalytic domain but does possess PDZ domains, which are thought to mediate protein-protein interactions, it is possible that ASIP acts as a cellular scaffold that provides a location for atypical PKC to transmit signals that establish cell polarity. Moreover, ASIP is phosphorylated in vitro (10), and our recent experiments showed that ASIP is phosphorylated in intact cells via a PKC-dependent manner. 3 It is possible that overexpression of ASIP may displace PKC from a compartment in which it plays a role in glucose uptake or may compete with an unidentified substrate that is involved in glucose uptake. Identification of the cellular compartment in which PKC acts, or of molecules with which it interacts, to mediate signaling that leads to glucose uptake should help to clarify the molecular mechanism of this important biological effect of insulin.