JAK1-dependent Phosphorylation of Insulin Receptor Substrate-1 (IRS-1) Is Inhibited by IRS-1 Serine Phosphorylation*

Serine phosphorylation of insulin receptor substrate-1 (IRS-1) reduces its ability to act as an insulin receptor substrate and inhibits insulin receptor signal transduction. Here, we report that serine phosphorylation of IRS-1 induced by either okadaic acid (OA) or chronic insulin stimulation prevents interferon-α (IFN-α)-dependent IRS-1 tyrosine phosphorylation and IFN-α-dependent IRS-1/phosphatidylinositol 3′-kinase (PI3K) association. In addition, we demonstrate that serine phosphorylation of IRS-1 renders it a poorer substrate for JAK1 (Janus kinase-1). We found that treatment of U266 cells with OA induced serine phosphorylation of IRS-1 and completely blocked IFN-α-dependent tyrosine phosphorylation of IRS-1 and IFN-α-dependent IRS-1/PI3K association. Additionally, IRS-1 from OA-treated cells could not be phosphorylated in vitro by IFN-α-activated JAK1. Chronic treatment of U266 cells with insulin led to a 50% reduction in IFN-α-dependent tyrosine phosphorylation of IRS-1 and IRS-1/PI3K association. More importantly, serine-phosphorylated IRS-1-(511–722) could not be phosphorylated in vitro by IFN-α-activated JAK1. Taken together, these data indicate that serine phosphorylation of IRS-1 prevents its subsequent tyrosine phosphorylation by JAK1 and suggest that IRS-1 serine phosphorylation may play a counter-regulatory role in pathways outside the insulin signaling system.


Materials
Cell Culture-U266 cells were grown in growth medium (RPMI 1640 medium supplemented with 10% neonatal bovine serum, 2.0 g/liter glucose, 100,000 units/liter penicillin, and 100 mg/liter streptomycin). Cells were passaged 1:1 with fresh medium every 3 days. For OA treatment, cells were washed twice and resuspended in growth medium supplemented with 1 M OA. For insulin treatment, cells were washed twice and resuspended in growth medium supplemented with 1 nM insulin.
PI3K Assays-PI3K assays were performed as described previously (36). In brief, 20 ϫ 10 6 cells/ml were treated as indicated and lysed in 1 ml of ice-cold lysis buffer (1% Nonidet P-40, 100 mM NaCl, 50 mM NaF, 1 mM dithiothreitol, 25 mM benzamidine, 1 mM phenylmethylsulfonyl fluoride, 2 g/ml aprotinin, 2 g/ml leupeptin, 2 mM sodium orthovanadate, 250 nM okadaic acid, and 50 mM Tris, pH 7.4). IRS-1 was immunoprecipitated from lysates with 4 l of anti-IRS-1 antiserum/test, and the resultant immune complexes were washed extensively. Kinase reactions were performed in 100 l of buffer containing 0.33 mg/ml L-␣-phosphatidylinositol, 7. Western Analysis-Western analysis was performed as described previously (37). In brief, 20 ϫ 10 6 cells/ml were treated as indicated and lysed in 1 ml of ice-cold lysis buffer. Proteins of interest were immunoprecipitated with the indicated antiserum (4 l/test), and the resultant immune complexes were washed extensively. Proteins were resolved by SDS-PAGE under reducing conditions on 10% gels, electrotransferred to polyvinylidene difluoride membrane, and probed with the indicated antiserum. Immunoreactive proteins were visualized by secondary detection using an 125 I-labeled goat anti-rabbit antibody, followed by autoradiography and densitometry.
JAK1 Kinase Assays-Cells (20 ϫ 10 6 /ml) were treated as indicated and lysed in 1 ml of ice-cold lysis buffer. JAK1 was immunoprecipitated from lysates with 4 l of anti-JAK1 antiserum/test, and the resultant immune complexes were washed extensively. Kinase reactions were performed in 100 l of buffer containing 7.5 mM MgCl 2 , 2.5 mM MnCl 2 , 20 M [␥-32 P]ATP (10 Ci/nmol), and 20 mM HEPES, pH 7.5, at 22°C for 20 min with 5 g/ml IRS-1-(511-772) or eluted IRS-1 or with no substrate as indicated. In reactions using eluted IRS-1 from U266 cells, IRS-1 was immunoprecipitated from 20 ϫ 10 6 cells. The resultant immune complexes were eluted for 60 min at 37°C in 10 l of 100 mM dithiothreitol, 0.5% SDS, 1 mg/ml bovine serum, and 20 mM HEPES, pH 7.4, and used in kinase reactions at a 1:10 dilution. Kinase reactions were terminated by addition of SDS-PAGE loading buffer, and the assay conditions used were linear with respect to time and amount of kinase. Resultant phosphoproteins were resolved by SDS-PAGE under reducing conditions on 7-20% gradient gels. Serine and threonine phosphoamino acids were base-hydrolyzed (39), and phosphotyrosine-containing proteins were examined by autoradiography and densitometry.
IRS-1-(511-772) Expression-The coding sequence for amino acids 511-772 of IRS-1 was amplified from rat IRS-1 sequence (a kind gift of Morris F. White) by previously described methods (34) using the forward primer 5Ј-CAGGATCCGATCTGGATAACCGGTTTC-3Ј and the reverse primer 5Ј-GAGAATTCGCGCTGGGTGTGCTAAAAG-3Ј. This 799-base pair product was introduced into the pGex6P3 plasmid using the BamHI and EcoRI restriction sites. GST-IRS-1-(511-772) was produced as described previously (38). In brief, protein expression was induced in Escherichia coli strain BL21 by addition of isopropyl-␤-D-thiogalactopyranoside to a final concentration of 1 mM. After 1 h, bacteria were lysed by mild sonication at 4°C in phosphate-buffered saline (140 mM NaCl, 2.7 mM KCl, 10 mM Na 2 HPO 4 , and 1.8 mM KH 2 PO 4 , pH 7.4) supplemented with 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, and 10 mM dithiothreitol. GST fusion proteins were affinity-purified from clarified lysates using glutathione-Sepharose, and GST was removed by digestion with 5 units of Precission protease at 4°C.
Whole Cell Phosphorylation/Phosphoamino Acid Analysis-Whole cell phosphorylation was performed as described previously (37). In brief, 20 ϫ 10 6 cells were suspended in 1 ml of phosphate-free RPMI 1640 medium supplemented with 0.75 mCi/ml 32 P i and 20 mM HEPES, pH 7.4, at 37°C for 1.5 h. For OA treatment, 1 M OA was added at 1.5 h for 30 min. Cells were lysed, and IRS-1 was immunoprecipitated as described above. Phosphoproteins were resolved by SDS-PAGE under reducing conditions on 8% gels. For phosphoamino acid analysis, phosphoproteins were electrotransferred to polyvinylidene difluoride membrane. Bands of interest were excised and acid-hydrolyzed in 6 N constant boiling HCl for 2 h at 110°C. Samples were treated with three cycles of water resuspension and evaporation and then resuspended in H 2 O/acetic acid/pyridine (89:10:1) running buffer containing 0.3 mg/ml phosphoserine, phosphothreonine, and phosphotyrosine standards. Phosphoamino acids were separated on cellulose-coated plates by high voltage TLC, and standards were visualized with ninhydrin. Results were analyzed by autoradiography and densitometry.
PAS Kinase Assays-PAS kinase assays were performed as described previously (37). In brief, PAS kinase was purified from 20 ϫ 10 6 cells by affinity chromatography using glutathione-Sepharose-bound GST-p85 protein. Kinase reactions were performed in 100 l of reaction buffer containing 5 g/ml IRS-1-(511-772), 0.4 mM EGTA, 0.4 mM NaPO 4 , 1 M [␥-32 P]ATP (100 Ci/nmol), and 20 mM HEPES, pH 7.1, at 22°C with or without 10 mM MgCl 2 as a cofactor. For JAK1 kinase assays, IRS-1-(511-772) was prephosphorylated in the absence of [␥-32 P]ATP and recovered by filtration through minifilter columns. IFN-␣-activated JAK1 was isolated as described above, and JAK1 kinase assays using prephosphorylated IRS-1-(511-772) as a substrate were performed as described above. Reactions were terminated using SDS-PAGE loading buffer and were linear with respect to time and amount of kinase. Phosphoproteins were resolved by SDS-PAGE under reducing conditions on 7-20% gradient gels and examined by autoradiography and densitometry.
Lysate Kinase Assays-Cells (10 ϫ 10 6 /10 ml) were treated with 1 nM insulin for 18 h and then pelleted at 500 ϫ g for 5 min. The cell pellet was lysed in 100 l of ice-cold 1 mM phenylmethylsulfonyl fluoride and 50 mM Tris, pH 7.4, by 10 passages through a 25-gauge needle. Lysates were centrifuged at 16,000 ϫ g for 10 min, and the supernatant fraction was adjusted to 0.4 mM EGTA, 0.4 mM NaPO 4 , 0.9 mM phenylmethylsulfonyl fluoride, 1 M [␥-32 P]ATP (100 Ci/nmol), 45 mM Tris, and 20 mM HEPES, pH 7.1, at 22°C with or without 10 mM MgCl 2 as a cofactor. This kinase mixture was then added to affinity-purified GST-IRS-1-(511-772) (bound to glutathione-Sepharose), and the reaction was allowed to proceed for 15 min at 22°C. Reactions were terminated by addition of phosphate-buffered saline with 1 mM EDTA and were linear with respect to time and amount of kinase. Phosphorylated IRS-1-(511-772) was then removed from the solid phase with Precision protease as described above. For JAK1 kinase assays, IRS-1-(511-772) was prephosphorylated in the absence of [␥-32 P]ATP and then used in JAK1 kinase assays as described above at a concentration of 5 g/ml.

RESULTS
OA Blocks IFN-␣-dependent IRS-1/PI3K Activation and Association-Serine phosphorylation of IRS-1 induced by OA treatment of 3T3-L1 adipocytes stops insulin-dependent activation of PI3K (26). To determine if IFN-␣-mediated PI3K activation was prevented by OA, IRS-1-associated PI3K activity was examined in U266 cells pretreated with 1 M OA for 30 min. Fig. 1A demonstrates that 1000 units/ml IFN-␣ induced 20-, 22-, and 10-fold increases in IRS-1-associated PI3K activity at 5, 10, and 30 min, respectively, and that pretreatment of cells with OA blocked this response. OA did not inhibit PI3K activity directly because PI3K activity in PI3K p85 immune complexes from OA-treated cells was no different than that from non-OA-treated cells (data not shown). To determine if this failure to activate PI3K was due to a loss of IFN-␣-dependent IRS-1/PI3K association, Western analysis was performed. Fig. 1B shows that IFN-␣-dependent IRS-1/PI3K association was increased at 5, 10, and 30 min (as measured by Western detection of PI3K p85) and that OA inhibited this association. To determine if this OA-dependent decline in IRS-1/PI3K association was due to an inhibition of IFN-␣-dependent tyrosine phosphorylation of IRS-1, Western analysis was again performed. Fig. 1C shows that IFN-␣ increased IRS-1 tyrosine phosphorylation at 5, 10, and 30 min and that OA blocked detectable tyrosine phosphorylation of IRS-1. To confirm that OA did not measurably alter IRS-1, p85, and JAK1 protein levels and the ability of these proteins to be immunoprecipitated by their respective antibodies, Western analysis was performed. Fig. 1D demonstrates that IRS-1, p85, and JAK1 protein levels and their ability to be immunoprecipitated were unaffected by OA. Finally, to show that OA did not affect JAK1 autophosphorylation or its ability to phosphorylate in vitro substrates, JAK1 kinase assays were performed. Fig. 1E demonstrates that JAK1 isolated from OA-treated cells phosphorylated recombinant IRS-1-(511-772) as well as JAK1 recovered from non-OA-treated cells and that JAK1 autophosphorylation was unchanged. Taken together, these findings indicate that OA blocks IFN-␣-dependent IRS-1/PI3K association by a mechanism that inhibits tyrosine phosphorylation of IRS-1, but does not alter JAK1 kinase activity.
JAK1-dependent Phosphorylation of IRS-1 Is Inhibited by OA-Serine phosphorylation of IRS-1 blocks insulin receptor-dependent tyrosine phosphorylation of IRS-1 (26 -35). To determine if OA inhibited JAK1-dependent IRS-1 tyrosine phosphorylation, JAK1 kinase assays were performed. Fig. 2A demonstrates that when IRS-1 isolated from OA-treated cells was used as a substrate for JAK1, IFN-␣-dependent JAK1 phosphorylation was not observed. In contrast, when IRS-1 from non-OA-treated cells was used as a substrate for JAK1, IFN-␣ induced a 5-fold increase in JAK1-dependent IRS-1 phosphorylation. To examine the phosphorylation state of IRS-1 isolated from OA-treated cells, phosphoamino acid analysis was performed. These experiments showed that IRS-1 was predominantly phosphorylated on serine residues and that no tyrosine phosphorylation was detected (Fig. 2B). Taken together, these findings indicate that serine phosphorylation of IRS-1 induced by OA renders IRS-1 a poorer substrate for JAK1.
Chronic Insulin Treatment Inhibits IFN-␣-dependent IRS-1/ PI3K Activation and Association-Chronic hyperinsulinemia induces serine phosphorylation of IRS-1 and reduces insulin signaling (33, 35, 40 -42). To determine if chronic insulin treatment inhibited IFN-␣-dependent activation of PI3K, IRS-1associated PI3K activity was examined in U266 cells pretreated with 1 nM insulin for 18 h. Fig. 3A demonstrates that 1000 units/ml IFN-␣ induced a 10-fold increase in IRS-1-associated PI3K activity at 5 min and that pretreatment of cells with insulin reduced this response by 50%. Chronic insulin treatment did not inhibit PI3K activity directly because PI3K activity in PI3K p85 immune complexes from insulin-treated cells was no different than that from non-insulin-treated cells (data not shown). To determine if this reduction in PI3K activation was due to a loss of IFN-␣-dependent IRS-1/PI3K association, Western analysis was performed. Fig. 3B shows that IFN-␣ increased IRS-1/PI3K association at 5 min (as measured by detection of PI3K p85) and that chronic insulin treatment reduced this association by 50%. To further examine the impact of chronic insulin treatment on IRS-1/PI3K association, IRS-1 present in PI3K immune complexes was examined by Western analysis (Fig. 3C). As in Fig. 3B, chronic insulin treatment reduced IFN-␣-dependent IRS-1/PI3K association, and comparable low amounts of IRS-1 were associated with PI3K before and after chronic insulin treatment in cells not treated with IFN-␣. To determine if this insulin-dependent decline in IRS-1/PI3K association after IFN-␣ treatment was due to an inhibition of IFN-␣-dependent tyrosine phosphorylation of IRS-1, Western analysis was again performed. Fig. 3D shows that IFN-␣ increased IRS-1 tyrosine phosphorylation at 5 min and that chronic insulin treatment reduced IFN-␣-dependent tyrosine phosphorylation of IRS-1 by 50%. Additionally, chronic insulin treatment did not alter JAK1 activity in that IFN-␣activated JAK1 isolated from chronically insulin-treated cells phosphorylated recombinant IRS-1-(511-772) as well as JAK1 recovered from non-insulin-treated cells (data not shown). Finally, to confirm that chronic insulin treatment did not measurably alter IRS-1, p85, and JAK1 protein levels and the ability of these proteins to be immunoprecipitated by their respective antibodies, Western analysis was performed. Fig. 3E demonstrates that IRS-1, p85, and JAK1 protein levels and the ability to be immunoprecipitated were unaffected by chronic insulin treatment. Taken together, these findings indicate that chronic insulin treatment inhibits IFN-␣-dependent IRS-1/ PI3K association by a mechanism that reduces tyrosine phosphorylation of IRS-1, but does not alter JAK1 kinase activity.

FIG. 3. Chronic insulin treatment inhibits IFN-␣-dependent IRS-1/PI3K activation and association.
A, U266 cells were pretreated with (ϩ) or without (Ϫ) 1 nM insulin for 18 h as indicated and then treated with (closed bars) or without (open bars) 1000 units/ml IFN-␣ for 5 min. PI3K activity was measured in IRS-1 immunoprecipitates. Data are representative of triplicate experiments Ϯ S.E. B, cells were treated as described for A, and Western analysis was used to detect PI3K p85 (p85) in IRS-1 immunoprecipitates (IP) using an anti-p85 antibody. C, cells were treated as described for A, and Western analysis was used to detect IRS-1 in PI3K p85 (p85) immunoprecipitates using an anti-IRS-1 antibody. D, cells were treated as described for A, and Western analysis was used to detect IRS-1 tyrosine phosphorylation using an anti-phosphotyrosine antibody (pY). E, cells were treated with (ϩ) or without (Ϫ) 1 nM insulin for 18 h as indicated, and Western analysis was performed on IRS-1, p85, and JAK1 immunoprecipitates using the antibodies indicated. Data in B-E are representative of triplicate experiments.  ). B, IRS-1-(511-772) was prephosphorylated with (Insulin ϩ) or without (Insulin Ϫ) serine kinase activity generated in A and then used as a substrate for JAK1 isolated from U266 cells treated with (ϩ) or without (Ϫ) 1000 units/ml IFN-␣ for 5 min. C, PAS kinase was affinity-purified from U266 cells using GST-p85. PAS kinase assays were performed using IRS-1-(511-772) as a substrate in the presence (ϩ) or absence (Ϫ) of 10 mM MgCl 2 . D, IRS-1-(511-772) was prephosphorylated as described for C without [␥-32 P]ATP and then used as a substrate for IFN-␣-activated JAK1 isolated from U266 cells treated with 1000 units/ml IFN-␣ for 5 min. All data are representative of triplicate experiments.

DISCUSSION
These data establish that serine phosphorylation of IRS-1 renders it a poorer substrate for JAK1. Western analysis and PI3K assays demonstrated that IFN-␣-dependent IRS-1 tyrosine phosphorylation and IRS-1/PI3K association and activity were blocked by OA treatment and that this was not due to an effect of OA on JAK1 (Fig. 1, A-D). Isolation of IRS-1 from OA-treated cells showed that OA-dependent serine phosphorylation of IRS-1 completely inhibited the ability of IFN-␣-activated JAK1 to phosphorylate IRS-1 (Fig. 2). Likewise, chronic insulin stimulation reduced by 50% the ability of IFN-␣ to stimulate IRS-1 tyrosine phosphorylation and IRS-1/PI3K association and activity (Fig. 3). More importantly, serine phosphorylation of IRS-1-(511-772) by serine kinases derived from chronically insulin-stimulated cells and by PAS kinase reduced by 50 and 75%, respectively, the ability of IFN-␣-activated JAK1 to phosphorylate IRS-1-(511-772) (Fig. 4). Taken together, these findings indicate that serine phosphorylation of IRS-1 reduces the ability of IRS-1 to serve as a JAK1 substrate, that IRS-1 serine phosphorylation inhibits signal transduction in pathways outside the insulin system, and that hyperinsulinemia may alter signaling of JAK1-dependent cytokine receptors.
Inhibition of PP1 and PP2A serine phosphatases by OA and calyculin A increase IRS-1 serine phosphorylation and leads to decreased insulin receptor-mediated IRS-1 tyrosine phosphorylation (26,33). Chronic insulin treatment has also been shown to induce serine phosphorylation of IRS-1 and to inhibit insulin receptor-dependent phosphorylation of IRS-1 (33, 35, 40 -42). Recently, the region of IRS-1 susceptible to chronic insulin treatment-dependent serine phosphorylation has been reported, and it appears to reside between amino acids 530 and 843 (35). The kinase responsible for this phosphorylation is unknown, but appears to be insensitive to inhibitors of protein kinases C and A, PI3K, and mitogen-activated protein kinase (35). We have identified a kinase (PAS kinase) that can serine phosphorylate IRS-1 and inhibit the ability of IRS-1 to serve as an insulin receptor substrate (34,37). This kinase associates with the p85 subunit of PI3K through SH2 domain interactions and phosphorylates IRS-1 in IRS-1/PI3K complexes after insulin stimulation (37). Here, we show that PAS kinase can phosphorylate IRS-1-(511-772) and that this phosphorylation inhibits the ability of IFN-␣-activated JAK1 to subsequently phosphorylate IRS-1-(511-772).
Although serine phosphorylation of IRS-1 decreases the ability of the insulin receptor and now JAK1 to phosphorylate IRS-1, the mechanism of this effect is not clearly delineated. In the insulin signaling system, serine phosphorylation of IRS-1 within the IH1 phosphotyrosine-binding domain appears to impair NPXY-mediated IRS-1/insulin receptor association (33), thus abrogating direct IRS-1/insulin receptor interaction. Like the insulin receptor, the IL-4 receptor contains an NPXY motif, and this motif appears to coordinate the formation of receptor/ JAK/IRS-1 complexes, which result in IRS-1 tyrosine phosphorylation (43). The IFN-␣ receptor does not contain an NPXY motif and may rely on the IRS-1 IH1 pleckstrin homology domain to coordinate receptor/JAK/IRS-1 association and subsequent IRS-1 phosphorylation (8). This suggests that serine phosphorylation within the IRS-1 IH1 pleckstrin homology domain might be important for preventing IFN-␣-activated JAK1dependent tyrosine phosphorylation of IRS-1. We show here, however, that IFN-␣-activated JAK1 can phosphorylate IRS-1-(511-772) and that serine phosphorylation of IRS-1-(511-772) inhibits this effect. This is important in that IRS-1-(511-772) does not contain either the IRS-1 IH1 pleckstrin homology or IH2 phosphotyrosine-binding domain and suggests that other regions of IRS-1 may be important in IRS-1/JAK1 interactions.
Hyperinsulinemia and insulin resistance are prominent features in both syndrome X and the development of type 2 diabetes mellitus (44). However, the pathogenesis of the multiple complications and conditions associated with these diseases is not yet understood (45). We show here that chronic insulin treatment and IRS-1 serine phosphorylation decrease JAK1mediated IRS-1 tyrosine phosphorylation and IRS-1/PI3K association, suggesting that cytokine signal transduction may be altered during hyperinsulinemia. Currently, a rapidly growing number of hormone and cytokine receptors appear to signal through JAK and IRS family members, and this appears to be critical to hormone/cytokine function (1). This is most clearly understood in IL-4 signaling, where IRS function has been shown to be critical to IL-4-dependent mitogenesis and antiapoptosis (4,46). Additionally, site-specific mutagenesis of the phosphotyrosine-binding domain-binding motif in the IL-4 receptor reduces both IRS and STAT6 tyrosine phosphorylation and abolishes the effect of IL-4 on the induction of DNA binding activity and CD23 induction (47). Thus, by inducing IRS serine phosphorylation, hyperinsulinemia may potentially contribute to the pathogenesis of syndrome X/type 2 diabetes mellitus complications by disrupting JAK-mediated cytokine and hormone signaling pathways that use IRS. In summary, we show that OA and chronic insulin treatment inhibit IFN-␣-dependent IRS-1 tyrosine phosphorylation and IRS-1/PI3K association and activity. More importantly, we show that these effects are mediated by serine phosphorylation of IRS-1. We conclude that IRS-1 serine phosphorylation plays an inhibitory role in signaling pathways outside the insulin system and suggest that hyperinsulinemia may alter signaling of JAK1-dependent cytokine receptors through serine phosphorylation of IRS-1.