Modulation of insulin-stimulated degradation of human insulin receptor substrate-1 by Serine 312 phosphorylation.

Ser/Thr phosphorylation of insulin receptor substrate-1 (IRS-1) is a negative regulator of insulin signaling. One potential mechanism for this is that Ser/Thr phosphorylation decreases the ability of IRS-1 to be tyrosine-phosphorylated by the insulin receptor. An additional mechanism for modulating insulin signaling is via the down-regulation of IRS-1 protein levels. Insulin-induced degradation of IRS-1 has been well documented, both in cells as well as in patients with diabetes. Ser/Thr phosphorylation of IRS-1 correlates with IRS-1 degradation, yet the details of how this occurs are still unknown. In the present study we have examined the potential role of different signaling cascades in the insulin-induced degradation of IRS-1. First, we found that inhibitors of the phosphatidylinositol 3-kinase and mammalian target of rapamycin block the degradation. Second, knockout cells lacking one of the key effectors of this cascade, the phosphoinositide-dependent kinase-1, were found to be deficient in the insulin-stimulated degradation of IRS-1. Conversely, overexpression of this enzyme potentiated insulin-stimulated IRS-1 degradation. Third, concurrent with the decrease in IRS-1 degradation, the inhibitors of the phosphatidylinositol 3-kinase and mammalian target of rapamycin also blocked the insulin-stimulated increase in Ser(312) phosphorylation. Most important, an IRS-1 mutant in which Ser(312) was changed to alanine was found to be resistant to insulin-stimulated IRS-1 degradation. Finally, an inhibitor of c-Jun N-terminal kinase, SP600125, at 10 microm did not block IRS-1 degradation and IRS-1 Ser(312) phosphorylation yet completely blocked insulin-stimulated c-Jun phosphorylation. Further, insulin-stimulated c-Jun phosphorylation was not blocked by inhibitors of the phosphatidylinositol 3-kinase and mammalian target of rapamycin, indicating that c-Jun N-terminal kinase is unlikely to be the kinase phosphorylating IRS-1 Ser(312) in response to insulin. In summary, our results indicate that the insulin-stimulated degradation of IRS-1 via the phosphatidylinositol 3-kinase pathway is in part dependent upon the Ser(312) phosphorylation of IRS-1.

The first step in insulin action is ligand stimulation of the insulin receptor tyrosine kinase. A number of endogenous sub-strates, including insulin receptor substrates (IRS) 1 1-4 (1-4) are phosphorylated on tyrosine residues. Tyrosine-phosphorylated IRS-1 and IRS-2 serve as the major docking proteins for Src homology 2 domain containing proteins (5,6). Association of the p85 regulatory subunit of phosphatidylinositol (PI) 3-kinase with tyrosine-phosphorylated IRS-1/2 results in the membrane localization and activation of the p110 catalytic subunit of PI 3-kinase, leading to the generation of phosphatidylinositol 3,4,5trisphosphate and phosphatidylinositol 3,4-bisphosphate. These lipid products induce the activation of a number of signaling kinases including the Ser/Thr kinase Akt (7,8), which is phosphorylated and activated by the upstream PDK1 kinase (9). Activation of Akt and its downstream signals have been shown to play a critical role in mediating the metabolic actions of insulin such as GLUT4 translocation and glucose transport (10 -13), glycogen synthesis kinase 3 serine phosphorylation and glycogen synthesis (14), phosphodiesterase serine phosphorylation and anti-lipolysis (15)(16)(17), and mTOR activation and protein synthesis (18 -21). However, other effects of insulin, for example its mitogenic actions, are mediated by association of Grb2 with tyrosinephosphorylated Shc and IRS-1/2, which results in the activation of the mitogen-activated protein kinase signaling pathway (5,6). In addition, other signal cascades stimulated by insulin may also participate in these responses (22). Some of the same signaling molecules that are involved in the metabolic and mitogenic actions of insulin have also been proposed to play a role in both the feedback inhibition of the insulin signal and cellular insulin resistance (23)(24)(25). It has been proposed that hyper-Ser/Thr phosphorylation of IRS proteins plays a key role in the uncoupling of the insulin signal. Hyperphosphorylation of IRS-1 on Ser/Thr residues has been shown in both cultured cells and in vivo to be associated with an insulin-resistant state (26,27). Densensitization of the insulin signal via IRS-1 Ser/Thr phosphorylation can result from counterregulatory hormone activation, proinflammatory cytokine production/cellular stress, or inhibition of protein phosphatases 1 and 2A (26, 28 -31). Several specific phosphorylation sites have been identified as targets of these counter regulatory hormones and their signaling cascades. Activation of the mitogen-activated protein kinase signaling pathway has been shown to result in an increase in the phosphorylation of Ser 616 of IRS-1 (32,33), 2 whereas activation of the JNK has been shown to result in the stimulation of Ser 3122 phosphorylation of IRS-1 (34). Other Ser/Thr phosphorylation sites have also been identified in the IRS-1 molecule (35,36). There are several potential mechanisms by which hyper-Ser/Thr phosphorylation of IRS proteins could block the insulin signal. These include uncoupling of IRS proteins from cytoskeletal/ membrane components, which causes translocation of IRS-1 to the cytosol (37,38); inhibition of the insulin receptor-IRS interaction (30,39); inhibition of the association of Src homology 2 domain-containing proteins to IRS proteins, such as PI 3-kinase (28,32,33); and inhibition of the insulin receptor phosphotransferase activity (26).
An additional mechanism of modulating the insulin signal transduction pathway is via the degradation of key components in this cascade. Degradation of IRS-1 has been shown in a number of cell lines and primary cells treated with chronic insulin (40 -47). In addition, low IRS-1 protein levels are observed in animal models of insulin resistance and in adipocytes isolated from type 2 diabetic subjects but not type 1 diabetic subjects (48,49). The molecular mechanism of IRS-1 degradation has been the focus of numerous recent studies. Selective inhibitors of PI 3-kinase and mTOR have been shown in some, but not all, systems to inhibit insulin-stimulated IRS-1 degradation, consistent with the hypothesis that the degradation of IRS-1 is initiated via the PI 3-kinase/Akt/mTOR signaling pathway (41)(42)(43)50). However, because these inhibitors may have unknown additional effects, it is important to confirm these results by other approaches. Thus, an important finding was that expression of a constitutively active, membrane-targeted PI 3-kinase, p110 CAAX , induces IRS-1 degradation and hyper-Ser/Thr phosphorylation (44,46,51). These results therefore argue that the overexpression of the PI 3-kinase cascade leads to the degradation of IRS-1. However, no studies to date have reported on the ability of insulin to induce IRS-1 degradation in cells lacking the PI 3-kinase or other downstream components of this cascade.
After activation of the PI 3-kinase cascade, IRS-1 appears to be degraded via the proteasome pathway. Evidence in favor of this model includes the finding that specific inhibitors of the 26 S proteasome block IRS-1 degradation but not Ser/Thr phosphorylation (42,44,46,47,52). More recently, a temperaturesensitive mutant of ubiquitin-activating enzyme E1 was shown to completely block insulin-stimulated IRS-1 degradation (47). In addition, several studies have reported that insulin stimulates the ubiquitination of IRS-1 (47,52).
Ser/Thr phosphorylation of IRS-1 correlates with IRS-1 degradation, suggesting that the phosphorylation triggers the subsequent degradation. The identification of regulatory phosphorylation sites required for IRS-1 degradation and the kinase that phosphorylates these sites is critical to the further understanding of the mechanism of IRS-1 degradation. The present studies were designed to examine the role of phosphorylation at specific Ser residues in IRS-1 degradation and the kinase cascade responsible for initiating the degradation process. Because both human Ser 312 phosphorylation and Ser 616 phosphorylation correlate with desensitization, we have investigated whether phosphorylation at these sites can modulate IRS-1 degradation. We have found that inhibition of IRS-1 degradation by inhibitors of the PI 3-kinase/Akt/mTOR pathway correlates with inhibition of IRS-1 Ser 312 phosphorylation. Moreover the insulin-stimulated degradation of a mutant IRS-1 in which Ser 312 was replaced with Ala was decreased compared with the wt IRS-1. In contrast, a mutant IRS-1 in which Ser 616 was changed to Ala exhibited insulin-induced degradation comparable with the wt IRS-1. Insulin-stimulated IRS-1 degradation and Ser 312 phosphorylation was partially blocked with a JNK small molecule inhibitor. However, subsequent studies showed that the insulin-stimulated phosphorylation of Ser 312 was unlikely to be due to JNK because other inhibitors (like LY294002) could also block this phosphorylation without inhibiting JNK. These results indicate that phosphorylation at Ser 312 is necessary for insulin-stimulated degradation and suggests that phosphorylation at Ser 312 has a dual mechanism of modulating insulin action: uncoupling the interaction of IRS-1 with the insulin receptor as well as the targeting of IRS-1 to the degradation pathway.
Generation of Stable Cell Lines Expressing the Wild Type and Mutant IRS-1 and PDK1-H4IIErat hepatoma cells were infected with pWZLexpressing IRS-1 or PDK1 constructs as described (55), with slight modifications. Briefly, 85% confluent Phoenix packaging cells were transfected using FuGENE 6 (Roche Molecular Biochemicals) or Lipo-fectAMINE (Invitrogen) according to the manufacturer's instructions. 24 h post-transfection, fresh medium was added and allowed to incubate at 37°C for 24 h to generate viral supernatant. 50 -60% confluent H4IIE cells were incubated with viral supernatant for 24 h at 37°C. Stable cell pools expressing IRS-1 were generated by selection for 48 h with 5 g/ml blasticidine HCl and then maintained in 1 g/ml blasticidine HCl. Stable cell pools expressing PDK1 were selected with 500 g/ml hygromycin.
Cell Culture and Treatments-H4IIE cells were maintained in Dulbecco's modified Eagle's medium containing 5% fetal bovine serum, 5% newborn calf serum at 37°C and 5% CO 2 and then serum deprived for 10 -18 h in Dulbecco's modified Eagle's medium containing 0.2% bovine serum albumin. The cells were pretreated with various inhibitors or Me 2 SO for 30 min prior to insulin stimulation. Insulin or 0.01 N HCl was added at the time of serum deprivation as described in the figure legends. Mouse embryonic stem (ES) cells lacking PDK1 and control ES cells were cultured as previously described (56,57). ES cells were serum deprived for 3 h in Dulbecco's modified Eagle's medium containing 0.5% bovine serum albumin and 20 mM Hepes prior to insulin stimulation as described in the figure legends.
Immunoprecipitation and Immunoblotting-H4IIE cells were lysed by shaking on ice for 20 min with 50 mM Hepes, pH 7.5, 150 mM NaCl, 1% Nonidet P-40, 1 mM NA 3 VO 4 , 100 nM okadaic acid, and 1ϫ protease inhibitor set I mixture (Calbiochem). Cellular debris was removed by centrifugation at 15,000 rpm for 15 min at 4°C. The protein content was determined by the BCA assay. Approximately 20 -30 and 200 -300 g of protein were used for analyses of total cell lysates and immunoprecipitations, respectively. Immunoprecipitations were performed with 30 l of anti-FLAG agarose at 4°C for 3 h. Immunoprecipitated IRS-1 proteins were washed with 20 mM Tris, pH 7.4, 200 mM NaCl, 0.1% Triton X-100; then with 50 mM Tris, pH 7.4, 500 mM NaCl; then with 50 mM Tris, pH 7.4, 150 mM NaCl; and then with 20 mM Tris, pH 7.4, 200 mM NaCl. Agarose beads were resuspended in 1ϫ sample buffer (58 mM Tris, pH 6.8, 1% SDS, 40% glycerol, and 0.1 M dithiothreitol). H4PDK1 and ES cells were lysed on ice with buffer B (50 mM Tris, pH 7.5, 150 mM NaCl, 1% Triton X-100, 10% glycerol, 1 mM EDTA, 1 mM dithiothreitol, 1 mM benzamidine, 1 mM phenylmethylsulfonyl fluoride, 1 mM leupeptin, 10 g/ml aprotinin, 30 mM NaPP i , 1 mM NA 3 VO 4 , and 100 nM okadaic acid). Cellular debris was removed by centrifugation at 15,000 rpm for 15 min at 4°C. The immunoprecipitations were performed with 2 l of anti-IRS-1 antibody (ID6) at 4°C for 3 h, followed by collection on protein G-Sepharose. Immunoprecipitated IRS-1 proteins were washed three times with buffer B. Sepharose beads were resuspended in 1ϫ sample buffer. The samples were boiled for 5 min and subjected to SDS-PAGE using 6%, 7.5%, or 10% Tris-glycine gels, then transferred to nitrocellulose (Schleicher & Schü ll). The membranes were incubated with various antibodies overnight at 4°C. Following washing, incubation with horseradish peroxidase-conjugated secondary antibody (1:5000) for 1 h at room temperature, and washing, detection was carried out with enhanced chemiluminescent substrate. Analysis of c-Jun Phosphorylation-H4IIE cells in 6-well plates were lysed with 1ϫ sample buffer. The samples were boiled for 5 min, and cellular debris was pelleted by centrifugation at 15,000 rpm for 5 min at room temperature. 30 l were used for SDS-PAGE on 10% Tris-glycine gels and processed as described above.
Statistical Analysis-Chemiluminescent signals were directly quantitated using the Kodak 440CF Image Station and Kodak 1D version 3.5.3 software or by scanning densitometry and image analysis using NIH Image software (rsb.info.nih.gov/nih-image/). The absolute integration value of the immunoreactive bands minus background was determined. The statistical significance was determined by Students t test (␣ ϭ 0.05).

Insulin Stimulates IRS-1 Degradation-Insulin-stimulated
degradation of IRS-1 has been reported in a number of cellular models of insulin resistance (42,46,47,58). To determine whether IRS-1 protein levels are regulated by insulin in the highly responsive rat hepatomas H4IIE cells (59), cells stably expressing FLAG-tagged wild type human IRS-1 (H4 wt IRS-1) were stimulated with various concentrations of insulin for 18 h. The cell lysates were analyzed by SDS-PAGE followed by transfer to nitrocellulose and immunoblotting with anti-IRS-1 antibodies. A dose-dependent decrease in IRS-1 protein levels was detected with a 60 Ϯ 4% decrease at 10 nM insulin (Fig.  1A). A control blot of the p85 subunit of PI 3-kinase in total cell lysates showed no decrease under the same conditions (Fig.  1A). To confirm the role of intracellular proteolysis in this insulin-stimulated IRS-1 degradation, H4 wt IRS-1 cells were pretreated with or without MG132 and lactacystin, specific inhibitors of the 26 S proteasome. Consistent with other cellular models of insulin-stimulated IRS-1 degradation (42,46,47,58), MG132 and lactacystin completely blocked insulin-stimulated degradation (Fig. 1B, lanes 5 and 7 versus lane 3). The ability of MG132 alone to stimulate a shift in the molecular weight of the IRS-1 (Fig. 1B) indicates that it is likely activating the stress-activated kinases in these cells. The lack of an IRS-1 band in the anti-FLAG immunoprecipitates from the parental H4IIE cells Fig. 1B, lane 1) verifies that the band being detected is the expressed human IRS-1.
Modulation of Insulin-stimulated IRS-1 Degradation by the PI3 Kinase/Akt/mTOR Pathway-To investigate the signaling pathways mediating insulin-stimulated IRS-1 degradation in the H4IIE cells, H4 wt IRS-1 cells were pretreated with various inhibitors for 30 min and then stimulated with 10 nM insulin for 18 h. The cell lysates were analyzed by SDS-PAGE followed by transfer to nitrocellulose and immunoblotting. In the absence of any inhibitor, a 60 Ϯ 4% (n ϭ 4) reduction in IRS-1 protein levels was detected with insulin stimulation. In the presence of the PI 3-kinase inhibitor LY294002, no IRS-1 degradation occurred (Fig. 2, A and C). Treatment of the H4 cells with this drug actually increased the IRS-1 levels, suggesting that even the basal turnover of IRS-1 was being inhibited. To determine whether downstream members of this pathway are involved, rapamycin, an inhibitor of mammalian target of rapamycin (mTOR), was tested for its effect. Rapamycin also completely blocked insulin-stimulated IRS-1 degradation (116 Ϯ 2% of nontreated, n ϭ 4). Controls verified that at the concentrations used, rapamycin and LY294002 completely inhibited the insulin-stimulated mobility shift in p70 S6 kinase ( Fig. 2A). In contrast to these results, the MEK1/2 inhibitor UO126 did not inhibit (p ϭ 0.48, n ϭ 3) insulin-stimulated IRS-1 degradation (Fig. 2, A and C). As a control, the effect of pretreatment with UO126 on insulin-stimulated ERK1/2 acti-vation was assessed. Insulin stimulated a ϳ18-fold increase in ERK1/2 (Thr 202 /Tyr 204 ) phosphorylation that was completely blocked by pretreatment with UO126 (Fig. 2B), whereas no inhibition of the insulin-stimulated shift in p70 S6 kinase was observed.
To assess what effects these inhibitors had on the specific phosphorylation sites in IRS-1, the IRS-1 from these cells was probed with anti-Ser(P) 312 and anti-Ser(P) 616 . Intriguingly, the insulin-stimulated increase in Ser 312 phosphorylation was reduced 90 Ϯ 2 and 78 Ϯ 2% by LY294002 and rapamycin pretreatment, respectively (Fig. 2, A and D). On the other hand, the MEK inhibitor, UO126, only partially reduced Ser 312 phosphorylation (48 Ϯ 1%). In contrast, the MEK inhibitor had a greater ability to inhibit Ser 616 phosphorylation (66 Ϯ 17%). This is consistent with the prior data that identified this site as a mitogen-activated protein kinase-regulated Ser phosphorylation site (32,33). In contrast to the results on Ser 312 , the insulinstimulated increase in Ser 616 phosphorylation (ϳ5-fold) was not significantly reduced by LY294002 and rapamycin pretreatment (74 Ϯ 35 and 75 Ϯ 16% of nontreated, respectively) (Fig. 2D). Because chemical inhibitors may act on undefined targets in cells in addition to their intended targets, we sought to further test the role of the PI 3-kinase cascade in the insulin-stimulated degradation of IRS-1. As a first step, we overexpressed PDK1 in the H4IIE cells. As expected, these cells showed an increase in response to the insulin-stimulated phosphorylation of Akt on Thr 308 , the PDK1 phosphorylation site (60, 61). More importantly for the purposes of these studies, the H4IIE cells overexpressing PDK1 showed an increased sensitivity in the insulin-stimulated degradation of IRS-1 (Fig. 3A). To further test the role of this pathway in the insulin-stimulated degradation of IRS-1, we then examined the ability of insulin (and insulin-like growth factor-I) to stimulate IRS-1 degradation in ES cells lacking PDK1 (57). In comparison with the parental PDK1 cells, insulin and insulin-like growth factor I stimulated a much lower amount of IRS-1 degradation in these cells (Fig. 3B).
The Role of Ser 616 in Insulin-stimulated IRS-1 Degradation-The above results strongly support the role of the PI 3-kinase/ Akt/mTOR pathway in the insulin-stimulated degradation of IRS-1. To determine the role of specific Ser phosphorylation sites in the insulin-stimulated degradation of IRS-1, we then assessed the ability of insulin to stimulate the degradation of various mutant IRS-1 molecules. As noted above, insulin greatly stimulated the phosphorylation of Ser 616 . The human IRS-1 Ser 616 was mutated to Ala (S616A), and stable H4IIE cells expressing S616A IRS-1 (H4 S616A IRS-1) were generated as described under "Experimental Procedures." H4IIE cells stably expressing either the wt IRS-1 or S616A IRS-1 were stimulated with 10 nM insulin for 18 h. Anti-FLAG immunoprecipitates and cell lysates were analyzed by SDS-PAGE followed by transfer to nitrocellulose and immunoblotting. Both wild type IRS-1 and the S616A mutant were comparably degraded after stimulation with insulin (62 Ϯ 4 and 49 Ϯ 9% of nontreated, respectively) (Fig. 4). The specificity of the anti-Ser(P) 616 antibody was confirmed by the finding that the IRS-1 S616A mutant protein only weakly reacted with this antibody compared with wild type IRS-1 (Fig. 4A). These results suggest that phosphorylation at Ser 616 does not play a major role in insulin-stimulated IRS-1 degradation.
The Role of Ser 312 in Insulin-stimulated IRS-1 Degradation-In addition to Ser 616 , insulin also stimulates the phosphorylation of IRS Ser 312 . However, in contrast to the insulin-stimulated phosphorylation of IRS-1 on Ser 616 , IRS-1 Ser 312 phosphorylation was inhibited by the same agents (LY294002 and rapamycin) that inhibited the insulin-stimulated IRS-1 degradation. To directly test the potential role of Ser 312 phosphorylation in insulin-stimulated IRS-1 degradation, this Ser was mutated to Ala, and H4IIE cells stably overexpressing S312A IRS-1 (H4 S312A IRS-1) were generated as described under "Experimental Procedures." H4IIE cells stably expressing either wt or the mutant S312A IRS-1 were stimulated with 10 nM insulin for 18 h. Anti-FLAG immunoprecipitates and cell lysates were analyzed by SDS-PAGE followed by transfer to nitrocellulose and immunoblotting. 70 Ϯ 2 and 68 Ϯ 2% (n ϭ 3) reductions in wt IRS-1 protein levels were detected in anti-FLAG immunoprecipitates and total cell lysates, respectively. In contrast, the degradation of S312A IRS-1 was greatly reduced, whether the IRS-1 was detected in anti-FLAG immunoprecipitates with anti-IRS-1 immunoblotting or in total cell lysates with anti-FLAG immunoblotting (Fig. 5, A and C). The specificity of the anti-Ser(P) 312 antibody used in these studies was confirmed by the finding that the IRS-1 S312A mutant protein only weakly reacted with this antibody compared with wild type IRS-1 (Fig. 5A).
To test the possibility that there is a general defect in the insulin-stimulated protein degradation in the H4 S312A IRS-1-expressing cells, H4 wt and S312A IRS-1 cells were stimulated with 10 nM insulin for 18 h, and the extent of degradation of the endogenous IRS-2 was measured. No difference was detected in the insulin-stimulated degradation of IRS-2 in H4 wt and S312A IRS-1 cells (Fig. 5B), indicating that there was no general decrease in the ability of insulin to stimulate IRS degradation in these cells expressing the mutant IRS-1.
Further studies were performed on both the dose and time dependence of wt and S312A mutant IRS-1 degradation. As above, controls were performed for total cellular protein by measuring the total levels of p85 in all the lysates. Insulin stimulated a dose-dependent decrease in wt IRS-1 levels with an ED 50 value of ϳ1 nM. In contrast, the S312A IRS-1 required ϳ100 times as much insulin to induce a comparable amount of degradation (ED 50 ϭ ϳ100 nM) (Fig. 6A). The time dependence of the insulin-stimulated decrease in IRS-1 levels was also impaired in the S312A mutant as compared with the wt IRS-1. The IRS-1 degradation shifted from having a t1 ⁄2 value of ϳ4 h for the wt IRS-1 to a t1 ⁄2 value of ϳ14 h with the mutant IRS-1 (Fig. 6B).
Degradation of IRS-1 Inhibits Insulin Signaling-To test whether IRS-1 degradation inhibits insulin action, H4 wt IRS-1 cells were stimulated with or without insulin for 15 min, after pretreatment with 5 nM insulin for 24 h to induce IRS-1 degradation. The cell lysates were analyzed by SDS-PAGE followed by transfer to nitrocellulose and immunoblotting. FIG. 4. Insulin-stimulated degradation of a S616A mutant IRS-1. H4IIE cells were infected with a retrovirus encoding a FLAGtagged wild type human IRS-1 (H4 wt IRS-1) or a S616A mutant (H4 S616A IRS-1) and drug-selected to isolate a pool population of cells stably expressing the two constructs. The cells were insulin-treated or not and lysed, and the total cell lysate or anti-FLAG immunoprecipitates were separated on 6% SDS-PAGE gels, transferred to nitrocellulose, and immunoblotted (IB) with the indicated antibodies. A representative experiment is shown (A) as well as the means Ϯ S.E. of three independent experiments (B). ERK1/2 in vivo activation was assessed by immunoblotting with an anti-phospho-ERK1/2 antibody. No differences were detected in ERK1/2 protein levels in H4 wt IRS-1 cells pretreated with 5 nM insulin for 24 h (Fig. 7A). However, a 65 Ϯ 4% (n ϭ 3) reduction in insulin-stimulated ERK1/2 activation was detected in H4 wt IRS-1 cells pretreated with 5 nM insulin for 24 h (Fig. 7A). To determine whether insulinstimulated ERK1/2 activation was impaired to a similar extent in cells expressing the S312A mutant that is resistant to insulin-induced degradation, H4IIE cells stably expressing either wt or the mutant S312A IRS-1 were pretreated with 5 nM insulin for 24 h to stimulate IRS-1 degradation and then stimulated with various doses of insulin for 15 min. ERK1/2 was activated to a greater extent in H4IIE cells expressing the mutant S312A IRS-1 compared with wt expressing cells (Fig. 7B). These results indicate that S312A mutant IRS-1 that is more resistant to insulin-induced degradation confers on these cells a greater insulin response after a 24-h insulin pretreatment than the wild type IRS-1.
Modulation of Insulin-stimulated IRS-1 Degradation by the JNK Inhibitor SP600125-To further characterize the signaling pathway leading to insulin-stimulated IRS-1 degradation and Ser 312 phosphorylation, H4 wt IRS-1 cells were pretreated with the JNK inhibitor SP600125 (62) for 30 min and then stimulated with 10 nM insulin for 18 h. The cell lysates were analyzed by SDS-PAGE followed by transfer to nitrocellulose and immunoblotting. In contrast to the 61 Ϯ 6% degradation induced by insulin in the absence of inhibitor, insulin only induced a 27 Ϯ 11% degradation in the presence of 25 M SP600125 (Fig. 8, A and B). When 10 M SP600125 was present, the ability of insulin to stimulate IRS-1 degradation was unaffected. In these experiments, insulin stimulated a ϳ7-fold increase in Ser 312 phosphorylation, that was reduced by 53 Ϯ 4% with SP600125 pretreatment at 25 M (Fig. 8). However, pretreatment with SP600125 at 10 M did not block Ser 312 phosphorylation induced after an 18-h treatment with insulin.

Modulation of Insulin-stimulated c-Jun Phosphorylation by the JNK Inhibitor SP600125 and by the PI3 Kinase/Akt/mTOR
Pathway-To correlate the ability of the various inhibitors used to block insulin-stimulated IRS-1 degradation and inhibit JNK activation, H4 wt IRS-1 cells were pretreated with the inhibitors for 30 min and then stimulated with 100 nM insulin for 15 min. The cells were lysed by the addition of 1ϫ sample buffer and analyzed by SDS-PAGE followed by transfer to nitrocellulose and immunoblotting. JNK in vivo activity was assessed by immunoblotting with an anti-pS63 c-Jun antibody. Insulin stimulated a ϳ2-fold increase in c-Jun phosphorylation that was completely blocked by pretreatment with either 25 or 10 M SP600125 (Fig. 9A). In contrast, pretreatment with UO126 resulted in a 25 Ϯ 21% reduction in c-Jun phosphorylation that was not significantly different from nontreated cells.
Thus, SP600125 appears to be a potent inhibitor of JNK. In contrast to the complete inhibition of insulin-stimulated c-Jun phosphorylation at even 10 M, SP600125 treatment of cells resulted in only a partial inhibition of insulin-stimulated IRS-1 Ser 312 phosphorylation observed after 15 min (Fig. 9B). For example, at 10 M, SP600125 caused a 42 Ϯ 2% reduction in insulin-stimulated phosphorylation of Ser 312 (Fig. 8B). Surprisingly, SP600125 also caused a similar level of inhibition in insulin-stimulated IRS-1 Ser 616 phosphorylation (64 Ϯ 18 and 41 Ϯ 7% reductions in Ser 612 phosphorylation were observed at 25 and 10 M, respectively) (Fig. 9B). To determine whether the PI 3-kinase/Akt/mTOR pathway mediates insulin-stimulated c-Jun phosphorylation, the cells were pretreated with LY294002 and rapamycin prior to insulin stimulation. In contrast to the results with the JNK inhibitor, SP600125, LY294002, and rapamycin did not inhibit insulin-stimulated c-Jun phosphorylation (Fig. 9C). DISCUSSION Numerous studies have demonstrated that Ser/Thr phosphorylation of IRS-1 inhibits insulin receptor-catalyzed IRS-1 tyrosine phosphorylation and the subsequent downstream signaling actions of insulin (24,25,63). One of the specific Ser phosphorylation sites in IRS-1 that has been proposed to negatively modulate the insulin signal is Ser 312 (numbered according to the human sequence). Prior studies have demonstrated that IRS-1 associates with and is phosphorylated by JNK in vitro on Ser 312 (34). Using a yeast tri-hybrid assay, it was shown that an active JNK blocked the interaction between the insulin receptor and IRS-1, thus suggesting that Ser 312 phosphorylation may interfere with IRS-1 PTB domain function (64). Phosphorylation of IRS-1 at Ser 312 also mediates the inhibitory effect of tumor necrosis factor ␣ on insulin signaling, although the kinase responsible for this phosphorylation was inhibited by the MEK inhibitor PD98059 and thus is presum- ably not JNK (34,64,65). Insulin also stimulates the phosphorylation of IRS-1 at Ser 312 by JNK-dependent and -independent pathways (65,66). It has also been suggested that the effect of free fatty acids on IRS-1 tyrosine phosphorylation and associated PI 3-kinase activity in rats is mediated by activation of PKC leading to IRS-1 Ser 312 phosphorylation (67). Thus, phosphorylation of IRS-1 at Ser 312 by a variety of kinases may represent a general mechanism of inhibiting the insulin signal pathway by uncoupling IRS-1 from the insulin receptor.
In the present studies we have used a recently described inhibitor of JNK, SP600125 (62). Treatment of cells with 10 M SP600125 completely blocked insulin-stimulated in vivo c-Jun phosphorylation. However, even 25 M SP600125 only partially inhibited insulin induced Ser 312 phosphorylation. Further, inhibitors of the PI 3-kinase/Akt/mTOR pathway that block insulin-stimulated IRS-1 degradation and Ser 312 phosphorylation had no effect on insulin-stimulated in vivo c-Jun phosphorylation. These results are most consistent with a kinase other than JNK being responsible for the insulin-stimulated Ser 312 phosphorylation of IRS-1. In contrast to our results in rat hepatoma cells, Lee et al. (66) recently reported that the PI 3-kinase inhibitor LY294002 blocked insulin-stimulated JNK activation and c-Jun phosphorylation in 32D cells overexpressing the insulin receptor. The finding that the JNK inhib- itor partially blocks Ser 616 phosphorylation of IRS-1 suggests that this inhibitor is not specific for JNK. Another recent report also concluded that this inhibitor is not specific for JNK (68). In the original report on this inhibitor (62), it was suggested that SP600125 may have activity toward several mitogen-activated protein kinase kinase isoforms and Akt based on in vitro screening and the partial inhibition of phospho-p38 and ATF in Jurkat T cells. In any case, our results indicate that JNK is unlikely to be the kinase phosphorylating IRS-1 Ser 312 in response to insulin and are consistent with the hypothesis that multiple kinases mediate IRS-1 Ser 312 phosphorylation. In agreement with this hypothesis is the recent evidence that IKK2 directly phosphorylated IRS-1 Ser 312 (69).
A role for the PI 3-kinase cascade in the insulin-stimulated degradation of IRS-1 is supported by the finding that overexpression of a constitutively active form of the p110 subunit of PI 3-kinase is sufficient to induce IRS-1 degradation, whereas inhibition of PI 3-kinase by LY294002 inhibits insulin-stimulated IRS-1 degradation (43,44,46,51). In H4IIE cells stably expressing a wt IRS-1, inhibition of PI 3-kinase by LY294002 also completely blocked insulin-induced IRS-1 degradation. Further, inhibition of mTOR, a downstream effector of PI 3-kinase, by rapamycin also blocked insulin-stimulated IRS-1 degradation in these cells. The mechanism by which mTOR regulates the phosphorylation and subsequent degradation of IRS-1 is not known. It is possible that mTOR is associated with and regulates a phosphatase (70). It is also possible that mTOR directly phosphorylates IRS-1 Ser 312 , although phosphorylation at this Ser residue in IRS-1 was not previously identified (35).
The inhibitors used in the present work may act on targets in addition to their intended enzymes. Therefore, we sought to confirm the role of the PI 3-kinase cascade in the insulinstimulated degradation of IRS-1. We first overexpressed PDK1, the upstream activator of the Akt pathway. We were able to show that overexpression of this kinase potentiated the ability of insulin to stimulate IRS-1 degradation. Even more important, we found that ES cells lacking PDK1 showed a dramatic decrease in the ability of insulin and insulin-like growth factor-I to stimulate IRS-1 degradation. These results therefore further document the hypothesized role of the PI 3-kinase cascade in the insulin-induced degradation of IRS-1 (42)(43)(44)46). However, these studies are not consistent with a recent study in Chinese hamster ovary cells overexpressing the insulin receptor in which the mTOR inhibitor rapamycin was found not to block insulin induced IRS-1 degradation, possibly suggesting that there may be cell-specific pathways for IRS-1. To further clarify the steps whereby activation of the PI 3-kinase pathway leads to the degradation of IRS-1, we have examined the role of specific IRS-1 phosphorylation sites in this process. We found that inhibitors of the PI 3-kinase/Akt/mTOR pathway block insulin-stimulated IRS-1 degradation as well as inhibit Ser 312 phosphorylation of IRS-1. We therefore produced an IRS-1 mutant in which Ser 312 was changed to Ala and examined the ability of insulin to stimulate the degradation of this molecule. We found that this mutant IRS-1 was resistant to insulin-stimulated IRS-1 degradation. In contrast, a mutant IRS-1 in which Ser 616 was changed to Ala exhibited a comparable insulin-stimulated degradation as the wt IRS-1. We next sought to investigate insulin action in H4IIE cells in which IRS-1 was degraded.
Our results strongly suggest a novel dual mechanism for the inhibition of the insulin signal, whereby Ser 312 phosphorylation uncouples IRS-1 from the insulin receptor as well as regulating its degradation. However, the finding that the IRS-1 mutant was degraded at high concentrations of insulin during a long term stimulation suggests that Ser 312 phosphorylation is not the sole determinant for degradation.
The mechanism by which Ser 312 phosphorylation promotes degradation is not known. It is possible that Ser 312 phosphorylation promotes ubiquitination of IRS-1. However, the ability of insulin to stimulate ubiquitination of IRS-1 has not been consistently observed (47,58). An alternative possibility for the mechanism by which Ser 312 phosphorylation promotes degradation is that when phosphorylated, Ser 312 facilitates the subcellular redistribution of IRS-1 from the low density microsomal or high speed pellet fraction to the cytosol. Such a redistribution would occur, for example, if the phosphorylation of Ser 312 interferes with PTB function of the IRS-1 molecule (64). In support of such a model is the finding that rapamycin blocks the insulin-stimulated redistribution of IRS-1 from the low density microsomal fraction/high speed pellet to the cytosol (46). In the present studies rapamycin was also found to inhibit Ser 312 phosphorylation. It is possible that these results are linked; that is, the phosphorylation of Ser 312 may participate in the insulin-stimulated redistribution of IRS-1, which could subsequently facilitate IRS-1 degradation.
In conclusion, we have found that inhibition of IRS-1 degradation by inhibitors of the PI 3-kinase/Akt/mTOR pathway correlates with inhibition of IRS-1 Ser 312 phosphorylation. Direct evidence for a participation of Ser 312 phosphorylation in insulin-stimulated IRS-1 degradation was obtained because Ser 312 mutant IRS-1 was found to be resistant to insulinstimulated degradation. This study therefore provides the first evidence that a specific Ser/Thr phosphorylation site can modulate IRS-1 degradation.