Common Inhibitory Serine Sites Phosphorylated by IRS-1 Kinases, Triggered by Insulin and Inducers of Insulin Resistance*

The Insulin Receptor Substrate (IRS) proteins are key players in insulin signal transduction and are the best studied targets of the insulin receptor. Ser/Thr phosphorylation of IRS proteins negatively modulates insulin signaling; therefore, the identification of IRS kinases and their target Ser phosphorylation sites is of physiological importance. Here we show that in Fao rat hepatoma cells, the IκB kinase β (IKKβ) is an IRS-1 kinase activated by selected inducers of insulin resistance, including sphingomyelinase, ceramide, and free fatty acids. Moreover, IKKβ shares a repertoire of seven potential target sites on IRS-1 with protein kinase C ζ (PKCζ), an IRS-1 kinase activated both by insulin and by inducers of insulin resistance. We further show that mutation of these seven sites (Ser-265, Ser-302, Ser-325, Ser-336, Ser-358, Ser-407, and Ser-408) confers protection from the action of IKKβ and PKCζ when they are overexpressed in Fao cells or primary hepatocytes. This enables the mutated IRS proteins to better propagate insulin signaling. These findings suggest that insulin-stimulated IRS kinases such as PKCζ overlap with IRS kinases triggered by inducers of insulin resistance, such as IKKβ, to phosphorylate IRS-1 on common Ser sites.

The Insulin Receptor Substrate (IRS) proteins are key players in insulin signal transduction and are the best studied targets of the insulin receptor. Ser/Thr phosphorylation of IRS proteins negatively modulates insulin signaling; therefore, the identification of IRS kinases and their target Ser phosphorylation sites is of physiological importance. Here we show that in Fao rat hepatoma cells, the IB kinase ␤ (IKK␤) is an IRS-1 kinase activated by selected inducers of insulin resistance, including sphingomyelinase, ceramide, and free fatty acids. Moreover, IKK␤ shares a repertoire of seven potential target sites on IRS-1 with protein kinase C (PKC), an IRS-1 kinase activated both by insulin and by inducers of insulin resistance. We further show that mutation of these seven sites (Ser-265, Ser-302, Ser-325, Ser-336, Ser-358, Ser-407, and Ser-408) confers protection from the action of IKK␤ and PKC when they are overexpressed in Fao cells or primary hepatocytes. This enables the mutated IRS proteins to better propagate insulin signaling. These findings suggest that insulin-stimulated IRS kinases such as PKC overlap with IRS kinases triggered by inducers of insulin resistance, such as IKK␤, to phosphorylate IRS-1 on common Ser sites.
The Insulin Receptor Substrate (IRS) 2 proteins are key players in insulin signal transduction and are the best studied targets of the insulin receptor (reviewed in Refs. 1 and 2). They contain a conserved pleckstrin homology domain, located at the amino terminus, that serves to anchor the IRS proteins to membrane phosphoinositides in close proximity to the insulin receptor (3). The pleckstrin homology domain is flanked by a P-Tyr binding (PTB) domain that functions as a binding site to the NPXY motif at the juxtamembrane domain of the insulin receptor (4). The C-terminal region of IRS proteins is poorly conserved. It contains multiple Tyr phosphorylation motifs that serve as a signaling scaffold, providing a docking interface for Src homology 2 domain-containing proteins like the p85␣ regulatory subunit of phosphatidylinositol 3-kinase, Grb2, Nck, Crk, Fyn, and SHP-2, which further propagate the metabolic and growth-promoting effects of insulin (5,6).
IRS-1 contains 232 Ser/Thr residues (7), many of which could be subjected to phosphorylation. Ser/Thr phosphorylation has been increasingly recognized as a negative counterbalance to positive IRS signaling through tyrosine phosphorylation (1). Ser/Thr phosphorylation reduces IRS-1 ability to undergo Tyr phosphorylation by the insulin receptor kinase and might serve as a physiological negative feedback control mechanism to turn off insulin's signaling by uncoupling the IRS proteins from their upstream and downstream effectors (8 -10). Furthermore, this mechanism can be utilized by inducers of insulin resistance under pathological conditions. Thus, Ser/Thr phosphorylation could be a generalized mechanism for insulin resistance (1). Several candidate Ser residues were identified as potential targets for IRS-1 kinases. These include Ser-24 (11), -302 (12), -307 (13), -318 (14), -408 (10), -612 (15), -636 and -639 (16), -731 (17), and -789 (18). 3 Similarly, a number of kinases, including extracellular signal-regulated kinase, protein kinase C (PKC), IKK␤, JNK, S6K1, AMP kinase, and mTOR were implicated as potential IRS kinases (reviewed in Ref. 19). Still, from a molecular perspective it has been difficult to identify discrete sites that are both phosphorylated in vivo and, when phosphorylated, have relevant functional consequences.
In the present study we undertook to address this question in an attempt to identify some of the common Ser sites subjected to phosphorylation by IRS kinases induced both by insulin and by inducers of insulin resistance. We have chosen to concentrate upon IRS-1 kinases activated by sphingomyelinase and ceramide, as their target phosphorylation sites on IRS-1 were not yet fully characterized. Our findings indicate that in Fao rat hepatoma cells, IKK␤ is activated by sphingomyelinase, ceramide, and free fatty acids (FFAs). Moreover, IKK␤ shares a repertoire of seven potential target sites on IRS-1 with PKC, an This article has been withdrawn by Erez Ilan, Sigalit Boura-Halfon, and Yehiel Zick. The withdrawing authors have become aware of several errors in the way images were presented in this manuscript. Because the original data are no longer available, the authors wish to withdraw the article in the interests of maintaining their publication standards and those of the journal. Lanes 3-6 of the IRS-1 immunoblot in Fig. 3A was reused in lanes 2-5 of the IRS-1 immunoblot in Fig. 3B. Fig. 7, A and B, was inappropriately manipulated. The withdrawing authors state that these presentational errors do not impact the underlying scientific findings of the article, which are also presented as quantitative line/bar graphs that essentially summarize data of a number of experiments. Therefore, the withdrawing authors stand by the original scientific results as described, which in the authors' opinion has been confirmed in other laboratories (Zhang, J., Gao, Z., Yin, J., Quon, M. J., and Ye, J. IRS-1 kinase activated both by insulin and by inducers of insulin resistance. We further show that mutation of these seven sites confers protection from the action of IKK␤ and PKC, thus enabling the mutated IRS proteins to better propagate insulin signaling. These findings suggest that insulin-stimulated IRS kinases such as PKC overlap with IRS kinases triggered by inducers of insulin resistance, such as IKK␤, to phosphorylate IRS-1 on common Ser sites.
Preparation of Primary Rat Hepatocytes-Isolated hepatocytes were prepared from fed male Wistar rats (260 -340 gram) according to the method of Berry et al. (20). In brief, a 20-gauge BD Venflon intravenous cannula (BD Biosciences) was inserted into the portal vein of anesthetized animals. The thorax was exposed, and the inferior vena cava was severed. The livers were injected with 1 ml (100 units) of heparin solution in Hanks buffer followed by sequential perfusions with Hanks/EGTA and Hanks/collagenase buffers. Livers were removed, minced, and placed in Hanks buffer containing collagenase at 37°C for 10 min with constant agitation. The digested tissue was passed through serial nylon mesh filters, and the resultant cell suspension was washed in ice-cold Hanks buffer. Viability was measured by Trypan blue exclusion before plating, and it exceeded 90%. Cells were suspended in Dulbecco's modified Eagle's medium containing 10% fetal calf serum at ϳ10 6 cells/ml and were seeded on culture plates precoated with fibronectin. The next day, hepatocytes were starved in serum-free Dulbecco's modified Eagle's medium for 16 h, and then the indicated treat-ments were applied. Alternatively, the cells were infected with adenoviruses harboring the gene of interest as described below.
Cell Treatment-Fao rat hepatoma cells were grown in RPMI medium supplemented with 10% fetal calf serum. At 70 -80% confluence, cells were deprived of serum for 16 h prior to each experiment and subjected to different stimuli in serum-free medium. Cells were then incubated with or without 100 nM insulin for the indicated times at 37°C. Cells were washed three times with ice-cold phosphate-buffered saline and were harvested in buffer A (25 mM Tris-HCl, 2 mM sodium orthovanadate, 0.5 mM EGTA, 10 mM NaF, 10 mM sodium pyrophosphate, 80 mM ␤-glycerophosphate, 25 mM NaCl, 1% Triton X-100, and protease inhibitor mixture 1:1000, pH 7.4). The supernatants were collected, samples of 50 -150 g of protein were mixed with 5ϫ Laemmli sample buffer (21), resolved by SDS-PAGE under reducing conditions, and transferred into nitrocellulose membrane for Western blot with the indicated antibodies. The intensity of the bands under study was quantified using the NIH Image densitometry program (developed at the U.S. National Institutes of Health and available on the Internet at rsb.info. nih.gov/nih-image). Each experiment was carried out two to three times in duplicate or triplicate samples.
Preparation of FFA Solutions-Solutions containing FFA complexed with FFA-free bovine serum albumin (BSA) were prepared as described (22). In brief, 100-mM FFA (palmitic acid or oleic acid) stock solutions were prepared in 0.1 M NaOH at 70°C. Next, the appropriate amount was complexed with 10% FFA-free BSA solution at 55°C for 30 min. The FAA⅐BSA complexes were allowed to cool to room temperature and were filtered with a 0.45-m pore size membrane filter before administering to the cells at the indicated concentrations. Alternatively, the FFA⅐BSA complexes were frozen at Ϫ20°C for up to 4 weeks for future use. Stored FFA⅐BSA was heated for 15 min at 55°C and cooled down to room temperature before use.
Thymidine Incorporation-Fao cells at 70% confluence were incubated in serum-free medium for 10 h. The medium was removed, and the cells were further incubated with or without insulin for 14 h at 37°C in serum-free medium. The medium was then replaced with serum-free medium containing 0.5 Ci/ml 3 H-thymidine for 2 h at 37°C. Cells were washed three times with phosphate-buffered saline and incubated with 0.5 ml of ice-cold 7.5% trichloroacetic acid at 4°C for 30 min. The precipitates were washed three times with 98% ice-cold ethanol and dissolved in 0.5 ml of 0.1 M NaOH. The cell precipitates were suspended in a scintillation mixture and counted in a scintillation counter.
Plasmid Construction-Hemagglutinin-tagged human PKC and Myc-tagged-IRS-1 were generated as we described previously (10). The following constructs were generated as described below: FLAG-tagged-IKK␤-pcDNA3-IKK␤ (courtesy of D. Wallach, Weizmann Institute) was restricted using HindIII and NotI, and a double-stranded synthetic oligonucleotide, containing matching overhangs of HindIII and NotI, coding for the immunogenic sequence of the FLAG tag (N-Met-Asp-Tyr-Lys-Asp-Asp-Asp-Asp-Lys-C) was ligated into the restriction sites. Correctness of the construct was verified by restriction mapping and sequencing, as well as by transient transfection into Chinese hamster ovary-T cells.
Construction of Recombinant Adenoviruses and Cell Infection-Adenoviruses expressing IRS-1 wild type (IRS-1 WT ) and an IRS-1 mutant in which Ser-265, Ser-302, Ser-325, Ser-336, Ser-358, Ser-407, and Ser-408 were mutated to Ala (IRS-1 7A ) were generated as we previously described (10). Additional constructs harboring genes of interest were generated according to the protocol provided with the AdEasy vector system (Quantum) as we described (10). In brief, the constructs of wild-type PKC WT , IKK␤ WT , and IKK␤ MUT (S177A,S181A) in the pcDNA-3 expression plasmid were ligated into the shuttle plasmid pAdTrack-CMV. This plasmid contains a green fluorescent protein cassette whose expression is driven by an independent promoter and serves as a tracing marker as well as a kanamycin-resistance coding sequence. Thus, the successful generation of viruses can be tracked by amplification of green fluorescent protein-expressing cells. The pAdTrack-CMV carrying the target genes was co-transformed with pAdEasy-1 containing the adenovirus genome. The recombination took place in a specific Escherichia coli strain, BJ5183, and positive kanamycin-resistant colonies were linearized by PacI and transfected into human embryonic kidney 293A cells. Viral plaques were isolated, and the viruses were amplified to generate high titer viral stocks (ϳ10 9 -10 10 pfu/ml). For infection, 7 ϫ 10 7 viral pfus were added to cultured Fao cells. 2 h following infection the viral medium was diluted 1:5 with fresh RPMI medium, and incubation was continued for 24 h. The cells were then transferred to virus-free medium. 48 h post-infection the cells were starved in serum-free RPMI for 16 h and then the indicated treatments were applied.

Inducers of Insulin Resistance Promote Phosphorylation of Ser
Residues of IRS-1 with Different Kinetics-Ser/Thr phosphorylation plays a pivotal role in the modulation of IRS protein function, mostly serving as a negative feedback control mechanism to turn off insulin signaling following prolonged exposure to insulin under physiological conditions. Furthermore, agents and conditions that induce insulin resistance utilize a similar strategy, leading to the activation of IRS-1 kinases and the phosphorylation of Ser inhibitory sites under pathological conditions (1). To illustrate this concept we focused on two Ser residues, Ser-307 and Ser-408, implicated in the regulation of IRS-1 function (10,23). As shown in Fig. 1, prolonged (60 min) insulin treatment as well as a variety of inducers of insulin resistance such as TNF␣, okadaic acid, calyculin A, palmitic acid, anisomycin, TPA, and sphingomyelinase (SMase) were all capable of inducing the phosphorylation of these Ser sites, albeit to different extents.
To study the effects of these stimuli in somewhat more detail we compared the kinetics of Ser-307 and Ser-408 phosphorylation in response to either insulin or treatment with inducers of insulin resistance. As shown in Fig. 2A, treatment with insulin of rat hepatoma Fao cells, overexpressing IRS-1, resulted in a rapid rise followed by a slower decline in the extent of Tyr phosphorylation of IRS-1. Insulin also induced, albeit with a delayed onset, the phosphorylation of IRS-1 both on Ser-307 and Ser-408, revealed by using P-Ser-specific antibodies directed to these sites. This sequence of events is in accordance with the concept that Ser/Thr phosphorylation serves as a negative feedback control mechanism and should therefore follow Tyr phosphorylation. The different kinetics of Ser-307 and Ser-408 phosphorylation in response to insulin treatment suggests that different IRS-1 kinases presumably mediate the phosphorylation of these sites. Furthermore, as shown in Fig. 2, B-E, several inducers of insulin resistance such as SMase, TPA, or okadaic acid also activated IRS kinases that phosphorylated each of these sites. Some of the inducers (e.g. SMase or okadaic acid) promoted the phosphorylation of Ser-408 at a faster rate, whereas others (e.g. anisomycin) induced the phosphorylation of Ser-307 and Ser-408 at comparable rates. These findings lend support to the hypothesis that these Ser sites might be targets for different IRS kinases. Experiments were therefore initiated to identify the IRS-1 kinases activated by SMase and related inducers of insulin resistance.
The Effects of TNF␣ on Insulin-stimulated Tyr Phosphorylation of IRS-1 Are Mimicked by SMase and Ceramides-We have previously shown that treatment of Fao cells with TNF␣ inhibits insulin-stimulated Tyr phosphorylation of IRS-1 while increasing its Ser/Thr phosphorylation (8). Consistent with these observations, TNF␣ decreased the electrophoretic mobility (increases Ser/Thr phosphorylation) of IRS proteins as illus- trated in Fig. 3A. Ceramide has been reported to induce insulin resistance (24) that could be the consequence of SMase activation (25). Indeed, treating the cells with either SMase (Fig. 3B) or with the ceramide analog C2 (Fig. 3C) resulted in a dose-dependent decrease in Tyr phosphorylation of IRS-1 and its mobility, similar to the effects induced by TNF␣. These and previous studies (8,25) suggest that TNF␣ may utilize the SMase pathway leading to the generation of ceramides, to induce ceramideactivated protein kinase(s) that phosphorylate IRS-1 and impair its function. To facilitate the identification of IRS-1 kinases activated by SMase, several known kinase inhibitors were applied. As shown in Fig.  4, wortmannin attenuated SMaseinduced Ser phosphorylation of IRS-1 and its mobility shift. This was accompanied by enhanced Tyr phosphorylation of IRS-1 following 2 min of insulin treatment. Rapamycin, which inhibits mTOR, was similarly an effective inhibitor. Inhibition of IKK␤ activity with sodium salicylate partially prevented the effects of SMase (Fig. 4). In contrast, inhibition of conventional PKCs using Go6983 (Fig. 4) or mitogenactivated protein kinase/extracellular signal-regulated kinase kinase (MEK) using PD98059 (not shown) failed to attenuate SMase effects on IRS-1. These results suggest that IRS-1 kinase(s) acting downstream of phosphatidylinositol 3-kinase, such as mTOR or IKK␤, could mediate SMase effects on IRS-1.
IKK␤ Mediates the Effects of FFA on IRS-1-Similar to TNF␣ and SMase, FFAs are potent inducers of IRS-1 kinases and insulin resistance (compare Ref. 10). Saturated FFAs like palmitic acid (C16:0) may lead to ceramide formation by de novo synthesis (26,27) and in such a way converge upon the mode of action of SMase. In contrast, unsaturated fatty acids like oleic acid cannot induce ceramide synthesis. Indeed, treating Fao cells with oleic acid failed to inhibit insulin-stimulated Tyr phosphorylation of IRS-1, whereas palmitic acid readily exerted such an inhibitory effect (Fig. 5A). These results suggest that FFA and TNF␣ may trigger common ceramide-activated Ser/Thr kinase(s) that attenuate IRS-1 function. To further investigate this possibility we examined the effects of different kinase inhibitors on the action of palmitic acid. As shown in Fig. 5B, wortmannin attenuated the inhibitory effects of palmitic acids on Tyr phosphorylation of IRS-1 and restored its mobility, while a specific IKK␤ inhibitor 15dPGJ2 (28) was even more potent in inhibiting the effects of palmitic acid. These findings suggest that IKK␤ could mediate both FFA and SMase effects on IRS-1, serving a common IRS-1 kinase activated by these obesity-related inducers of insulin resistance.
SMase Utilizes p38 MAPK, but Not JNK, as an Upstream Activator of IKK␤-Recent studies have implicated p38 MAPK or JNK in mediating the effects of FFAs on IRS-1 (28,29). These studies prompted us to explore the possible involvement of these enzymes as the upstream kinases, mediating FFA-and SMase-induced activation of IKK␤. Treating Fao cells with increasing doses of SMase increased both JNK and p38 MAPK phosphorylation (Fig. 6A). However, whereas the p38 MAPK inhibitor SB202190 attenuated the effects of SMase to a similar extent as that of wortmannin and 15dPGJ2 (a selective IKK␤ inhibitor) (Fig. 6B), SP600125, which inhibits JNK activation, failed to attenuate the effects of SMase on IRS-1 Tyr phosphorylation. These results are consistent with previous findings (28,29) implicating p38 MAPK, but not JNK, as being an upstream activator of IKK␤ following SMase treatment (Scheme I). Similar results were obtained when the effects of these inhibitors were examined on FFAs-induced phosphorylation of IRS-1 (not shown). These findings suggest that both SMase and FFAs might utilize IKK␤ as an IRS-1 kinase, with p38 MAPK as its upstream regulator.

The Inhibitory Effects of FFA and SMase on Insulin-stimulated Tyr Phosphorylation of IRS-1 Are Prevented by Overexpression of a Mutant Form of IKK␤-
To further explore the role of IKK␤, a wild type or an inactive mutant of IKK␤, lacking two Ser residues required for its activation by upstream kinases (30), was overexpressed in Fao cells treated either with FFA or SMase or subjected to prolonged (60 min) insulin treatment. As shown in Fig. 7, cells overexpressing the mutant IKK␤ better maintained the extent of Tyr phosphorylation of IRS-1 when compared with cells overexpressing WT-IKK␤ following treatment with either palmitic acid (Fig. 7A) or SMase (Fig. 7B). These results together with our observations utilizing IKK␤ inhibitors (Figs. 3-5) support the hypothesis that IKK␤ negatively regulates IRS-1 function in response to obesity-related inducers of insulin resistance such as FFA and SMase. Of interest, the mutant IKK␤ also conferred partial protection from the reduction in Tyr phosphorylation of IRS-1 following prolonged (60 min) treatment  with insulin (Fig. 7, A and B). These findings suggest that IKK␤ could also be activated in an insulin-dependent manner. Consistent with these findings, extracts from insulintreated Fao cells induced a modest, albeit significant, increase (40%) over basal in IB phosphorylation in vitro (not shown).

Mutation of Ser Sites Surrounding the PTB Domain of IRS-1 Confers Resistance from the Inhibitory Effects of IKK-␤-We
have previously shown that mutation of seven Ser sites within or in close proximity to the PTB domain of IRS-1 results in an IRS-1 mutant, denoted IRS-1 7A , that is more resistant to the inhibitory effects of selected inducers of insulin resistance (10). The example depicted in Fig. 8 shows primary rat hepatocytes infected with adenoviral constructs expressing either IRS-1 WT or IRS-1 7A and treated with inducers of insulin resistance such as TPA or anisomycin. Pretreatment with TPA or anisomycin significantly inhibited the insulin-stimulated Tyr phosphorylation of IRS-1 WT , whereas the Tyr phosphorylation of IRS-1 7A was not affected. These results suggest that mutation of these sites indeed protects IRS-1 from the inhibitory effects of inducers of insulin resistance.
To assess the possibility that these sites could be targets for IKK␤, we compared the susceptibility of IRS-1 WT versus IRS-1 7A to the inhibitory effects of this kinase. To this end, Fao cells were infected, alone or in combination, with adenoviral constructs encoding IRS-1 WT , IRS-1 7A , and IKK␤. As a control, some of the cells were infected with IRS-1 (wild type or mutant) and PKC, which acts, as we have previously shown (9), as a potential IRS-1 kinase. Part of the infected cells were subjected to 12 h of treatment with palmitic acid, followed by a brief (2 min) insulin treatment.  As shown in Fig. 9, adenoviral infection of Fao cells with IRS-1 WT or IRS-1 7A resulted in similar levels of Tyr phosphorylation in response to acute (2 min) insulin treatment. This suggests that the 7A mutation per se did not impair the ability of IRS-1 7A to localize in close proximity to the insulin receptor and undergo Tyr phosphorylation. Preincubation of the cells with palmitic acid inhibited the extent of their insulin-stimulated Tyr phosphorylation; however, cells expressing the IRS-1 7A were more resistant to the inhibitory effects of palmitic acid. As expected (9), the inhibitory effects of palmitic acid were exacerbated in cells that also overexpress PKC in addition to IRS-1 WT ; however, the presence of IRS-1 7A conferred almost full protection from the inhibitory effects of palmitic acid and PKC (Fig. 9, upper panel). Similarly, overexpression of IKK␤ potentiated the inhibitory effects of palmitic acid while the presence of IRS-1 7A , but not IRS-1 WT , afforded protection from this inhibition (Fig. 9, lower panel). These results support our hypothesis that Ser residues among the seven mutated sites are the targets of these two kinases.
IRS-1 7A Better Propagates Insulin Action-The Ser sites mutated in IRS-1 7A are not only targets for IRS-1 kinases triggered by inducers of insulin resistance as described above but are also targets of IRS-1 kinases triggered by insulin itself, as part of a physiological negative feedback control mechanism exerted by this hormone. Therefore, mutation of these sites (as in IRS-1 7A ) generates an IRS-1 protein that is less susceptible to this negative feedback inhibition. To further establish this hypothesis we compared insulin action in Fao cells infected with equal amounts of either IRS-1 WT or IRS-1 7A . As shown in   These results lend further support to the hypothesis that IRS-1 kinases activated by insulin or by inducers of insulin resistance can phosphorylate the same repertoire of Ser sites of IRS-1.

DISCUSSION
In the present study we provide evidence that inducers of insulin resistance such as TNF␣ (which activates SMase) and FFAs trigger IKK␤, which phosphorylate IRS-1 on Ser sites within or in close proximity to its PTB domain. Mutation of these sites generates an IRS-1 that can better propagate insulin action, can better sustain the effects of inducers of insulin resistance, and can better resist the inhibitory actions of IKK␤ as well as PKC. These results suggest that IRS-1 kinases activated by insulin, such as PKC, and those activated by inducers of insulin resistance, such as IKK␤, can phosphorylate the same repertoire of inhibitory Ser sites that might serve as points of convergence where insulin-stimulated IRS kinases overlap with IRS kinases triggered by inducers of insulin resistance.
Several lines of evidence support our conclusions. First we demonstrated that SMase and ceramide analogs mimic the effects of TNF␣ on Ser phosphorylation of IRS-1, suggesting that they can act as downstream effectors of the TNF receptor (8,25,31). The effects of SMase were inhibited by phosphatidylinositol 3-kinase inhibitors, indicating that SMase activates IRS-1 kinases acting downstream of phosphatidylinositol 3-kinase, such as mTOR or IKK␤. Similar to TNF␣ and SMase, saturated free fatty acids that induce de novo synthesis of ceramide (26,27) inhibited Tyr phosphorylation of IRS-1, suggesting they can converge upon the mode of action of SMase and may trigger common ceramide-activated Ser/Thr kinase(s) such as PKC (32) and IKK␤ (28) that attenuate IRS-1 function. Indeed, we could show that a specific IKK␤ inhibitor was even more potent than wortmannin in inhibiting the effects of palmitic acid. These observations support previous findings implicating IKK␤ as a mediator of FFAs (28,33) and TNF␣-induced phosphorylation of IRS-1 (30,34,35). We could further demonstrate that p38 MAPK, but not JNK, is an upstream regulator of IKK␤ following SMase treatment. Combined with previous findings (28,29) these results suggest that both SMase and FFAs might utilize IKK␤ as an IRS-1 kinase, with PKC (30) and p38  . IRS-1 7A potentiates insulin-stimulated thymidine incorporation. Fao cells at 70% confluence were infected with Adeno-Myc-IRS-1 either wild type (WT) or 7A mutant (7A) at a titer of 7 ϫ 10 7 pfu/ml. After 24 h, the cells were split into 6-well tissue culture plates and were incubated in serumfree medium for 10 h. The medium was removed, and the cells were incubated with or without the indicated concentrations of insulin for 14 h at 37°C in serum-free medium. The medium was then replaced with serum-free medium containing 0.5 Ci/ml 3 H-thymidine for 2 h at 37°C. Cells were washed three times with phosphate-buffered saline, and the extent of 3 H-thymidine incorporation was determined as described under "Experimental Procedures." Results are mean Ϯ S.D. of duplicate experiments. In parallel, cytosolic extracts were prepared and samples were resolved by means of 7.5% SDS-PAGE and immunoblotted with anti-IRS-1 antibody. Results were normalized to protein expression levels. Results are means and S.D. of duplicate samples.
MAPK as its upstream kinases (Scheme I). This conclusion was supported by studies with cells overexpressing either wild type or a mutant IKK␤ lacking two Ser residues required for its activation by upstream kinases (30). Cells overexpressing the mutant IKK␤ better maintained the extent of Tyr phosphorylation of IRS-1 than cells overexpressing WT-IKK␤ following treatment with either palmitic acid, SMase, or insulin. These results support the hypothesis that IKK␤ negatively regulates IRS-1 function in response to obesity-related inducers of insulin resistance such as FFA and SMase. They further suggest that IKK␤ could also serve as a downstream kinase along the insulin-signaling pathway. Along this line, IKK␤ has already been shown to serve as downstream effector of PKC (30), a known downstream effector of IRS proteins (9) (Scheme 1).
IRS-1 mutated at seven Ser phosphorylation sites, C-terminal to its PTB domain (10), was used in the present study. This mutant was used based on the concept that IRS-1 undergoes phosphorylation on a number of sites, a concept already documented in previous studies. For example, phosphorylation of IRS-1 by JNK1 at both Ser-302 and Ser-307 was found to be necessary but insufficient for disruption of insulin receptor-IRS-1 interactions, and phosphorylation of additional sites was required (12). Similarly, a single mutation to Ala of Ser-408 provided certain protection from the action of inducers of insulin resistance while a much stronger effect was obtained when an additional six Ser phosphorylation sites were mutated as well (10). The mutated sites conform to phosphorylation sites for IKK␤ (RXXSXXS, Ser-265, -302, -336, -408), protein kinase A (RRXS, Ser-408), protein kinase B (RXRXXS, Ser-265, -302, -325, -358), and PKC (RXXS, Ser-265, -302, -325, -336, -358, -408); hence they could be targets for a number of Ser/Thr kinases.
Here we show that mutation of these seven Ser sites turns the mutant IRS-1 resistant to the inhibitory effects of IKK-␤. This was evident when Fao cells were infected with combinations of adenoviral constructs encoding IRS-1 WT , IRS-1 7A , and IKK␤.
Overexpression of IRS-1 7A , but not IRS-1 WT , afforded protection from the inhibitory effects of FFAs that were exuberated by overexpression of IKK␤. Consistent with the role of IKK␤ as a mediator of the inhibitory effects of FFAs, we could show that overexpression of a mutant IKK␤, but not WT-IKK␤, partially protected the endogenous IRS from the inhibitory effects of palmitic acid. As expected (9), the inhibitory effects of palmitic acid were also enhanced in cells overexpressing PKC WT . Here again, the presence of IRS-1 7A , but not IRS-1 WT , conferred almost full protection from the combined inhibitory effects of palmitic acid and PKC WT . Taken together, these results suggest that Ser residues among the seven mutated sites are common targets of both IKK␤ and PKC.
An important outcome of the present study is the observation that inducers of insulin resistance promote phosphorylation of Ser residues of IRS-1 with different kinetics. This concept could be illustrated when we compared the rate of phosphorylation of two Ser residues, Ser-307 and Ser-408, implicated as negative regulators of IRS-1 function (10,23). The faster rate of phosphorylation of Ser-408 by selected inducers (SMase and okadaic acid) when compared with the kinetics of Ser-307 phosphorylation suggests that different IRS-1 kinases could presumably mediate the phosphorylation of these sites. Alternatively, these sites could be phosphorylated by the same kinase (e.g. IKK␤) having different affinities to the two targets. For example, phosphorylation at Ser-408 could prime IRS-1 for phosphorylation at Ser-307. This concept is in accordance with the findings of Werner et al. (12) that phosphorylation of Ser-307 could serve to prime phosphorylation of Ser-302. Hence, a tentative phosphorylation cascade might involve sequential phosphorylation of Ser-408, Ser-307, and Ser-302. Still, it needs to be determined why Ser phosphorylation of IRS proteins occurs on multiple sites and what the specific functions are of each of these sites.
In summary, Ser/Thr phosphorylation plays a pivotal role in the modulation of IRS protein function, serving as a negative feedback control mechanism to turn off insulin signaling following prolonged exposure to insulin (1). Inducers of insulin resistance utilize a similar strategy, leading to the activation of IRS-1 kinases and the phosphorylation of similar clusters of Ser inhibitory sites under pathological conditions (1). The present study suggests that IRS-1 kinases phosphorylate overlapping, but clearly not identical, Ser sites. At least two such inducers, TNF␣/SMase and FFAs, might utilize the same kinase, IKK␤, to exert their inhibitory action through the phosphorylation of Ser sites that are also targets for insulin-stimulated IRS-1 kinases such as PKC. Further studies are still required to figure out how phosphorylation of each of these sites affects IRS-1 function and insulin action.