Protein Kinase C-α Regulates Insulin Action and Degradation by Interacting with Insulin Receptor Substrate-1 and 14-3-3ϵ*

Protein kinase C (PKC)-α exerts a regulatory function on insulin action. We showed by overlay blot that PKCα directly binds a 180-kDa protein, corresponding to IRS-1, and a 30-kDa molecular species, identified as 14-3-3ϵ. In intact NIH-3T3 cells overexpressing insulin receptors (3T3-hIR), insulin selectively increased PKCα co-precipitation with IRS-1, but not with IRS-2, and with 14-3-3ϵ, but not with other 14-3-3 isoforms. Overexpression of 14-3-3ϵ in 3T3-hIR cells significantly reduced IRS-1-bound PKCα activity, without altering IRS-1/PKCα co-precipitation. 14-3-3ϵ overexpression also increased insulin-stimulated insulin receptor and IRS-1 tyrosine phosphorylation, followed by increased activation of Raf1, ERK1/2, and Akt/protein kinase B. Insulin-induced glycogen synthase activity and thymidine incorporation were also augmented. Consistently, selective depletion of 14-3-3ϵ by antisense oligonucleotides caused a 3-fold increase of IRS-1-bound PKCα activity and a similarly sized reduction of insulin receptor and IRS-1 tyrosine phosphorylation and signaling. In turn, selective inhibition of PKCα expression by antisense oligonucleotides reverted the negative effect of 14-3-3ϵ depletion on insulin signaling. Moreover, PKCα inhibition was accompanied by a >2-fold decrease of insulin degradation. Similar results were also obtained by overexpressing 14-3-3ϵ. Thus, in NIH-3T3 cells, insulin induces the formation of multimolecular complexes, including IRS-1, PKCα, and 14-3-3ϵ. The presence of 14-3-3ϵ in the complex is not necessary for IRS-1/PKCα interaction but modulates PKCα activity, thereby regulating insulin signaling and degradation.

Insulin action on target tissues requires insulin receptor (IR) 2 tyrosine autophosphorylation and activation of the intrinsic tyrosine kinase (1). Phosphorylated insulin receptor substrates (IRSs) then serve as adaptor proteins for a number of signal transduction molecules (2,3). Therefore, multimolecular complexes, including IRSs, constitute a major platform through which insulin signals are generated. Indeed, it has been demonstrated that the protein/protein interaction events engaged by the IRSs are necessary for insulin induction of specific enzymatic activities and/or for the targeting of signaling molecules to the proper intracellular location (2,3).
Besides the pivotal function in transducing insulin biologic effects, IRS proteins may also serve as negative regulators of insulin signals. Several bona fide inhibitors of insulin action, such as tyrosine phosphatases and serine/threonine kinases, have been shown to interact with IRSs (3)(4)(5). Both classes of molecules are involved in turning off insulin signals. Tyrosine phosphatases may remove the phosphorylation of key tyrosines both on the IR and on intracellular substrates (4 -6), thereby inhibiting their signaling functions. Also, phosphorylation on serine/threonine residues often interferes with tyrosine phosphorylation (7,8).
Among serine/threonine kinases, a crucial role in inhibiting insulin signals is played by the protein kinase C (PKC) family members. Several PKC isoforms have been implicated as inhibitors of insulin signal transduction (8 -12). Overexpression of PKC␣, for instance, has been shown to inhibit insulin signaling in cultured cell systems (9,10,12). PKC␣ is activated by agents known to interfere with insulin action, such as angiotensin II (13), endothelin-1 (14,15), advanced glycation end products (16), and by prolonged exposure to high glucose concentrations (17). Murine models of PKC␣ gene ablation also exhibit increased sensitivity to insulin, further supporting the concept of the PKC␣ negative role in insulin signaling (18).
PKCs are present in multimolecular complexes inside the cells (19,20), and their activity is finely tuned by changes in the extracellular and/or intracellular milieu (21)(22)(23). However, the mechanisms by which PKC␣ impinges on insulin action remain poorly defined. It is known that it may phosphorylate IR (9,11), IRS-1 (10), as well as other receptor and nonreceptor tyrosine kinases (24 -26) on serine/threonine residues. These phosphorylations may block further tyrosine phosphorylation or induce intracellular degradation of the targeted proteins (9 -12, 24 -26).
In this work, we have shown that PKC␣ directly binds insulin-activated IRS-1 and provided evidence that adaptors of the 14-3-3 family participate in the complex formation and down-regulate PKC␣ activity. This, in turn, removes the inhibitory constraint exerted by PKC␣, enhancing IR tyrosine kinase activity and signaling. In addition, we have shown that 14-3-3 modulation of PKC␣ also controls insulin intracellular sorting toward degradation.
Cell Culture and Transfection-NIH-3T3 fibroblasts overexpressing human insulin receptor have been described previously (27). Cells were cultured at 37°C in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum, 2% L-glutamine, 10,000 units/ml penicillin, 10,000 g/ml streptomycin. Transient transfection of the phosphorothioate oligonucleotides and of the 14-3-3⑀ plasmid was performed by the Lipofectamine method according to the manufacturer's instructions. For these studies, 60 -80% confluent cells were washed twice with Opti-MEM (Invitrogen) and incubated for 8 h with 5 g of plasmid construct or with 12 g of oligonucleotides and 45-60 l of the Lipofectamine reagent. The medium was then replaced with DMEM with 10% fetal calf serum, and the cells were further incubated for 15 h before being assayed.
Overlay Blot Assay-Protein lysates were obtained by cells stimulated with insulin for different times as described previously (28). Equal amounts (500 g) of proteins were immunoprecipitated with anti-IR or IRS-1 antibodies, separated by 7 or 12% SDS-PAGE, and transferred on 0.45-m Immobilon-P membranes (Millipore, Bedford, MA). The filters were incubated 2 h at room temperature in 5% (w/v) nonfat dry milk/NaCl/Tris (20 mM Tris, 140 mM NaCl, pH 7.6), 0.1% Tween 20 (blocking solution). At this time, recombinant biotinylated PKC␣ was supplemented with the lipid activators (10 mM phorbol 12-myristate 13-acetate, 0.28 mg/ml phosphatidylserine, and 4 mg/ml dioleine, final concentrations). The phosphorylation solution (20 M ATP, 1 mM CaCl 2 , 20 mM MgCl 2 , and 4 mM Tris, pH 7.5, final concentrations) was then added to those mixtures. The reaction mixtures were further incubated for 10 min at room temperature, rapidly cooled on ice, and added to the blocking solution. After 1 h of incubation at room temperature, the blots were washed twice with NaCl, Tris, 0.1% Tween 20 and then incubated with 0.5% formaldehyde for 10 min. The filters were washed three times with 2% glycine and once with NaCl, Tris, 0.1% Tween 20. PKC␣-bound proteins were detected by horseradish peroxidase-bound streptavidin blotting and autoradiography.
Western Blot Analysis and Immunoprecipitation Procedure-For these studies, cells were solubilized in lysis buffer (50 mM Hepes, pH 7.5, 150 mM NaCl, 10 mM EDTA, 10 mM Na 4 P 2 O 7 , 2 mM Na 3 VO 4 , 100 mM NaF, 10% glycerol, 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin) for 4 h at 4°C. Cell lysates were clarified by centrifugation at 5,000 ϫ g for 20 min, separated by SDS-PAGE, and transferred on 0.45-m Immobilon-P membranes. Upon incubation with primary and secondary antibodies, immunoreactive bands were detected by ECL according to the manufacturer's instructions. Immunoprecipitation assays were accomplished as described previously (29). Briefly, for these assays the cells were solubilized in 20 mM Tris, pH 7.5, 0.5 mM EDTA, 0.5 mM EGTA, 25 g/ml aprotinin, and 25 g/ml leupeptin (extraction buffer). Supernatants were further centrifuged at 60,000 ϫ g for 2 h, and pellets were solubilized with extraction buffer supplemented with 0.5% Triton X-100. Soluble pellets were immunoprecipitated for 18 h with PKC␣ or IRS-1 antibody and incubated with protein A-Sepharose for 2 h. Immunocomplexes were supplemented with the lipid activators (0.32 mg/ml phosphatidylserine and 0.032 mg/ml diacylglycerol, final concentrations), and the phosphorylation reaction was initiated by adding the substrate solution (50 mM PKC␣ substrate peptide, 0.5 mM ATP, 0.25 mM EGTA, 0.4 mM CaCl 2 , 0.1 mg/ml bovine serum albumin, 20 mM Tris, pH 7.5, 10 mM MgCl 2 , and 10 Ci/ml (3000 Ci/mmol) [␥-32 P]ATP, final concentrations). The reaction mixture was further incubated for 30 min at room temperature, and the phosphorylation reaction was terminated by adding 7.5 M guanidine hydrochloride and spotting on phosphocellulose discs. Disc-bound radioactivity was quantitated by liquid scintillation counting.
Determination of Thymidine Incorporation and of Glycogen Synthase Activity-Twenty four-well plates were seeded with 10 6 cells/plate in 1 ml of DMEM supplemented with 10% fetal calf serum. After the transient transfection with plasmid construct or specific antisense oligonucleotides, the medium was replaced with DMEM with 10% fetal calf serum for 24 h. Then the cells were starved with DMEM containing 0.25% BSA and no serum at 37°C. After an additional 24 h, the medium was removed again and replaced with DMEM, 0.25% BSA with or without 100 nM insulin. Incubation was prolonged for an additional 16 h, and FIGURE 1. PKC␣ interaction with IRS proteins. A, 3T3-hIR cells were stimulated with 100 nM insulin for the indicated times, and the cellular extracts (500 g) were immunoprecipitated (IP) with anti-IR antibodies and loaded on SDS-PAGE. Gels were then transferred on nitrocellulose filters and incubated with recombinant biotinylated PKC␣ (OB rPKC␣) as described under "Experimental Procedures." B, lysates (300 g), obtained from 3T3-hIR cells treated or not with 100 nM insulin for 30 min were immunoprecipitated with IRS-1 or IRS-2 antibodies followed by blotting with IRS-1, IRS-2, or PKC␣ antibodies, ECL, and autoradiography. The autoradiographs shown are representative of three (A) and four (B) independent experiments.

IRS-1 Interaction with PKC␣ and 14-3-3⑀
DECEMBER 9, 2005 • VOLUME 280 • NUMBER 49 incubation media were replaced with the same media supplemented with [ 3 H]thymidine (500 nCi/ml). 4 h later, the cells were washed with ice-cold 0.9% NaCl and then with ice-cold 20% trichloroacetic acid followed by solubilization with 1 N NaOH. Radioactivity was quantitated by liquid scintillation counting. Glycogen synthase activity was assayed in cell extracts as described by Miele et al. (16).
Insulin Internalization and Degradation-Cells were washed three times with ice-cold KRP solution (120 mM NaCl, 5 mM KCl, 1.2 mM MgSO 4 , 1.2 mM KH 2 PO 4 , 100 mM Na 2 HPO 4 ), pH 7.8, and incubated at 4°C for 16 h with 30 pM 125 I-insulin diluted in binding buffer (KRP, pH 7.8, containing 2% BSA and 1.3 mM CaCl 2 ). Insulin internalization and degradation were measured as described previously (29). Briefly, cells were rinsed with KRP, pH 7.8, in order to remove the unbound ligand and further incubated at 37°C in binding buffer. After ice-cold acid washes, cellular radioactivity was considered internalized insulin. Alternatively, cells were re-incubated at 37°C, and aliquots of the incubation medium were removed to determine the amount of 125 I-insulin released from the cells. Degraded insulin was assessed by determining the trichloroacetic acid solubility of 125 I-insulin products.

PKC␣ Interaction with IRS-1 in 3T3-hIR Cells-NIH-3T3
fibroblasts expressing human wild-type insulin receptors (3T3-hIR cells) were treated with 100 nM insulin for 5, 30, and 60 min. Cell lysates were precipitated with insulin receptor antibodies and tested in overlay assays for their ability to bind recombinant biotinylated PKC␣. A major 180-kDa band was revealed upon insulin stimulation of the cells (Fig.  1A). Both the apparent molecular weight and mass spectrometric analysis indicated that the 180-kDa species corresponded to IRS-1. The observation that the maximal interaction occurred 5 min after insulin exposure was also consistent with the timing of the IR interaction with IRS proteins. To verify the possibility that IRSs could interact with PKC␣ in intact cells, we performed selective co-precipitation experi-ments. PKC␣ was detectable in IRS-1 and, at very low levels, in IRS-2 precipitates (Fig. 1B). Moreover, the exposure of 3T3-hIR cells to insulin increased PKC␣/IRS-1 co-precipitation in a time-dependent manner ( Fig. 2A). It was enhanced by 2.5-fold in 5 min, rising to 4-fold in 1 h, and then returning to basal levels after 16 h of insulin treatment. PKC␣ activity was then measured in IRS-1 precipitates. Again, phosphorylation of the specific PKC substrate peptide was dependent on the length of insulin exposure. It featured a rapid increase that peaked at 15 min followed by a slower decrease (Fig. 2B).
14-3-3⑀ Interaction with PKC␣ and IRS-1 in 3T3-hIR Cells-We hypothesized that the apparent discrepancy between the relative amounts of IRS-1-precipitated PKC␣ activity and PKC␣ protein (Fig.  2B) could be due, at least in part, to the presence of PKC modulators within the IRS-1 complexes.
In order to analyze this possibility, IRS-1 precipitates were probed with recombinant PKC␣ in overlay assays (Fig. 3A). The expected 180-kDa IRS-1 band was observed, and an additional intensely stained 30-kDa band was detectable in IRS-1 precipitates obtained from insulinstimulated cells. No specific band was detected in IRS-2 precipitates (data not shown). Microsequencing of the isolated 30-kDa species revealed an amino acid composition consistent with that of the adaptor proteins of the 14-3-3 family.
To investigate whether a specific 14-3-3 isoform could interact with IRS-1, cell lysates from insulin-stimulated and unstimulated cells were precipitated with selective antibodies and blotted with IRS-1 antibodies (Fig. 3B). We used 14-3-3␤, -, and -⑀ antibodies because prior immunoblot experiments of whole 3T3-hIR cell lysates did not indicate other 14-3-3 isoforms (data not shown). Although a basal signal was detected in all the samples, no change of IRS-1 levels in 14-3-3␤ and -precipi-   Fig. 1. B, lysates (300 g) from cells exposed to 100 nM insulin for 30 min were immunoprecipitated with 14-3-3␤, -⑀, and -antibodies and blotted with IRS-1 antibodies. WB, Western blot. C, co-precipitation between IRS-1 or IRS-2 and 14-3-3⑀ was evaluated in 3T3-hIR cellular extracts before and after 100 nM insulin stimulation for 30 min. The same filter was subsequently re-probed with IRS-1or IRS-2 antibodies. D, cells were treated or not treated with 100 nM bisindolylmaleimide (BIM) before insulin stimulation (100 nM insulin for 30 min). Cell extracts were then precipitated with IRS-1 antibodies and immunoblotted with 14-3-3⑀ antibodies. E, 3T3-hIR cells lysates obtained as described previously were immunoprecipitated with antibodies against different isoforms of PKC (␣, ␦, and ) and blotted with 14-3-3⑀ antibodies. The autoradiographs shown are representative of at least three independent experiments.
Overexpression of 14-3-3⑀ cells did not change the protein levels of IRS-1 and PKC␣ when assayed in whole cell lysates (Fig. 5A). Also, no difference in PKC␣ levels was detected in IRS-1 precipitates from 14-3-3⑀-overexpressing and control cells, whether in the absence or in the presence of insulin (Fig. 5B). At variance, although in control cells insulin increased the amount of IRS-1-bound phosphorylated PKC␣ by Ͼ2-fold, no such effect was elicited in the 14-3-3⑀-overexpressing cells (Fig. 5B). In parallel with phosphorylation, insulin exposure caused a 2-fold increase of PKC activity in IRS-1 precipitates in the 3T3-hIR cells but not in 14-3-3⑀-overexpressing cells (Fig. 5C). Therefore, it appears that 14-3-3⑀ plays a role in inhibiting IRS-1-bound PKC␣ activity and in increasing IR/IRS-1 tyrosine phosphorylation and downstream signaling.
Silencing of 14-3-3⑀ in 3T3-hIR Cells-To address the functional significance of the endogenous 14-3-3⑀ protein in insulin signaling, we used a specific phosphorothioate antisense oligonucleotide (AS⑀) to block 14-3-3⑀ expression. The AS⑀ caused a Ͼ70% reduction of 14-3-3⑀ expression compared with a scrambled oligonucleotide (SC⑀) with no effect on the 14-3-3␤ and -isoforms (Fig. 6A). Antisense reduction of 14-3-3⑀ levels was accompanied by a constitutive increase of IRS-1 co-   DECEMBER 9, 2005 • VOLUME 280 • NUMBER 49 precipitation with the phosphorylated form of PKC␣, which was further augmented upon insulin exposure of the cells (Fig. 6B). In parallel, total and IRS-1-associated PKC␣ activities were constitutively increased by 1.5-and 2.5-fold, respectively, and further enhanced by insulin (Fig. 6C). At variance, PKC␣ levels in IRS-1 precipitates were almost identical in SC⑀-and in AS⑀-treated cells, in basal conditions, and were stimulated by insulin to a similar extent (Fig. 6B). Also, PKC␣ immunodetection in whole cell lysates was not changed by the treatment with both oligonucleotides (Fig. 6D).
PKC␣ and 14-3-3⑀ Roles in Insulin Degradation-It has been reported previously in 3T3-hIR cells that inhibition of PKC activity   D). A, aliquots of the cell lysates were precipitated with IR or IRS-1 antibodies followed by blotting with IR, IRS-1, and phosphotyrosine antibodies (A and B). Alternatively, lysates were blotted with the phosphorylated or the total form of Raf, ERK1/2, Akt/PKB, and Foxo1 antibodies as indicated (C and D). The autoradiographs shown are representative of at least four independent experiments. I.P., immunoprecipitation. FIGURE 8. Effect of simultaneous 14-3-3⑀ and PKC␣ inhibition on IR and IRS-1 tyrosine phosphorylation and on Akt/PKB and MAPK. 3T3hIR cells were transiently transfected with a 14-3-3⑀ (AS⑀) and/or with a specific PKC␣ phosphorothioate antisense oligonucleotide (AS-PKC␣) or scrambled oligonucleotide (SC-PKC␣). Cells were treated with 100 nM insulin for 5 min, and cell lysates were precipitated with IR or IRS-1 antibodies followed by blotting with phosphotyrosine antibodies. Alternatively, lysates were blotted with 14-3-3⑀, PKC␣, or with the phosphorylated forms of Akt/PKB and ERK1/2 antibodies. The autoradiographs shown are representative of at least three independent experiments. I.P., immunoprecipitation; W.B., Western blot. reduced the degradation of insulin (27). To investigate further a possible role for PKC␣ in controlling insulin degradation, 3T3-hIR fibroblasts were transfected with AS-PKC␣. (Fig. 10A). Selective inhibition of PKC␣ was accompanied by a 64% decrease of insulin degradation compared with the control untransfected cells or to cells transfected with the scrambled control oligonucleotide (SC-PKC␣) (Fig. 10B). In contrast, insulin internalization was increased by Ͼ50% (Fig 10B). Specific PKC␦ antisense (AS-PKC␦) and scrambled oligonucleotides (SC-PKC␦) did not change insulin internalization and degradation (Fig. 10, B and  C). Most interestingly, the overexpression of 14-3-3⑀ also induced a 65% reduction in the insulin degradation, although the AS⑀ block of 14-3-3⑀ expression caused a 70% increase in the amount of the degraded insulin, as compared with either untransfected or cells transfected with the SC⑀ control oligonucleotide. In parallel with PKC␣ changes, insulin internalization was increased by about 40% and reduced by about 35%, respectively, in AS⑀-transfected cells and in those overexpressing 14-3-3⑀.

DISCUSSION
Insulin stimulates the activation of PKC isoforms in several tissues and cell types (27, 30 -35). It has also been reported that PKC isoforms may form complexes with the IR and phosphorylate several molecules involved in IR-initiated signaling, inhibiting their function (9 -12, 27, 36, 37). For instance, PKC␣ has been reported to inhibit insulin action in both cellular and animal models (17,18). Here we have used NIH-3T3 fibroblasts expressing human wild-type IR (3T3-hIR cells) and investigated the mechanism of PKC␣ complex formation with molecules involved in early events of insulin signaling. By overlay blot and coprecipitation experiments, we show that PKC␣ selectively binds IRS-1 in an insulin-stimulated manner. This finding is consistent with previous reports indicating that IRS-1 expression is required for insulin activation of PKC␣ in L6 rat skeletal muscle cells (30). Most interestingly, in these same cells, the depletion of IRS-1 also inhibits PKC␣ co-precipitation with the IR, 3 suggesting the existence of a multimolecular complex, where IRS-1 docks PKC␣ to the IR. Indeed, overlay blots revealed a direct interaction of PKC␣ with IRS-1 but not with the IR.
Our experiments also indicated the presence of the scaffold protein 14-3-3⑀ in IRS-1 precipitates from insulin-stimulated cells. No other 14-3-3 protein, among those expressed in NIH-3T3 cells, co-precipitated with IRS-1 in an insulin-stimulated manner. At variance with these cells, it has been reported that insulin stimulates IRS-1 association with 14-3-3␤ in 3T3-L1 adipocytes (38). Although abundant in several tissues (39 -42), 14-3-3⑀ is not expressed in the adipocytes. It is possible, therefore, that individual 14-3-3 isoforms play specific roles in different cell types and/or at different stages of differentiation.
14-3-3 proteins often function as molecular adaptors and regulate the

IRS-1 Interaction with PKC␣ and 14-3-3⑀
DECEMBER 9, 2005 • VOLUME 280 • NUMBER 49 intracellular localization of signaling proteins (43). Mounting evidence also indicates that these proteins modulate the activity of a number of serine/threonine kinases, including members of the PKC family (44 -48). In the present report, we show that 14-3-3⑀ directly interacts with PKC␣ in the NIH-3T3 cells. Under basal conditions, 14-3-3⑀ also binds to other PKC isoforms. Insulin, however, selectively induced the coprecipitation of 14-3-3⑀ with PKC␣. In order to elucidate the role of 14-3-3⑀ in insulin signaling, we have analyzed the consequences of 14-3-3⑀ overexpression in the 3T3hIR cells. We found that it significantly reduced IRS-1-bound PKC␣ activity. Conversely, antisense suppression of 14-3-3⑀ expression was accompanied by an increase of PKC␣ activity. These changes in enzymatic activity were not paralleled by changes in the amount of PKC␣ in IRS-1 precipitates, indicating that 14-3-3⑀ does not regulate the formation of the IRS-1-PKC␣ complex. The finding that PKC␣ independently binds both IRS-1 and 14-3-3⑀ in overlay assays further supports this hypothesis.
We have shown previously (27) that pharmacological inhibition of PKC activity reduces insulin degradation. We now provide evidence that PKC␣ is involved in targeting insulin toward degradation. Indeed, possibly because of the increased receptor autophosphorylation and/or signaling, insulin endocytosis is increased in the cells with reduced PKC␣ activity (either by antisense depletion or by 14-3-3⑀ overexpression). The increase of internalization, however, is not followed by a similarly sized hormone degradation, as the selective depletion of PKC␣ actually reduces the amount of degraded insulin. Moreover, when 14-3-3⑀ is overexpressed in 3T3hIR cells, the insulin degradation is markedly reduced, paralleling the decrease of PKC␣ activity. Thus, although IR kinase activation and increased PKC signaling may be required for the internalization of the insulin-receptor complex (29,50), PKC␣ activation represents a further step necessary for the progression of incoming insulin toward degradation.
In conclusion, in this work we provide the following evidence: (i) insulin stimulates the formation of a multimolecular complex, including IRS-1 and PKC␣; (ii) the presence of 14-3-3⑀ is not necessary for the formation of this complex, but it regulates the activity of PKC␣ within the complex; and (iii) PKC␣ inhibits IR/IRS-1 signaling and regulates insulin degradation. This may provide a novel mechanism for the auto-regulation of insulin sensitivity and for insulin desensitization in response to a number of PKC-activating agents.