Insulin stimulates the serine phosphorylation of the signal transducer and activator of transcription (STAT3) isoform.

Insulin stimulation of Chinese hamster ovary cells expressing the human insulin receptor and differentiated 3T3L1 adipocytes resulted in a time-dependent reduction in the SDS-polyacrylamide gel electrophoretic mobility of STAT3. The decreased STAT3 mobility initially occurred by 2 min and was quantitative by 5 min. In addition, the change in STAT3 mobility was concentration-dependent and was detectable at 0.3 nM insulin with maximal effect between 1 and 3 nM. Although both these cell types also express the STAT1α, STAT1β, STAT5, and STAT6 isoforms, only STAT3 was observed to undergo an insulin-dependent reduction in mobility. Immunoprecipitation of STAT1 and STAT3 from P-labeled cells demonstrated that only STAT3 was phosphorylated in response to insulin whereas phosphoamino acid analysis indicated that this phosphorylation event occurred exclusively on serine residues. Furthermore, treatment of cell extracts with alkaline phosphatase reversed the insulin-stimulated decrease in STAT3 mobility. Together, these data demonstrate that insulin is a specific activator of STAT3 serine phosphorylation without affecting the other STAT isoforms.

Insulin stimulation of Chinese hamster ovary cells expressing the human insulin receptor and differentiated 3T3L1 adipocytes resulted in a time-dependent reduction in the SDS-polyacrylamide gel electrophoretic mobility of STAT3. The decreased STAT3 mobility initially occurred by 2 min and was quantitative by 5 min. In addition, the change in STAT3 mobility was concentration-dependent and was detectable at 0.3 nM insulin with maximal effect between 1 and 3 nM. Although both these cell types also express the STAT1␣, STAT1␤, STAT5, and STAT6 isoforms, only STAT3 was observed to undergo an insulin-dependent reduction in mobility. Immunoprecipitation of STAT1 and STAT3 from 32 P-labeled cells demonstrated that only STAT3 was phosphorylated in response to insulin whereas phosphoamino acid analysis indicated that this phosphorylation event occurred exclusively on serine residues. Furthermore, treatment of cell extracts with alkaline phosphatase reversed the insulin-stimulated decrease in STAT3 mobility. Together, these data demonstrate that insulin is a specific activator of STAT3 serine phosphorylation without affecting the other STAT isoforms.
Signal Transducers and Activators of Transcription (STAT) 1 proteins have emerged as important biological links between cell surface receptors and DNA transcription events (reviewed in Refs. [1][2][3][4]. Currently, there are six general members of this growing protein family (STAT1-6) with additional subtypes for several of these members (5)(6)(7). These proteins were first identified as Src Homology 2 (SH2) and Src Homology 3 (SH3) domain-containing tyrosine-phosphorylated substrates of cytokine receptor signaling pathways (8). It is now generally accepted that activation of various cytokine receptors results in the stimulation of the Janus kinase family (JAK1, -2, and -3 and Tyk2) of protein tyrosine kinase (4). The activated JAK kinases directly tyrosine phosphorylate the STAT proteins, which induce both hetero-and homodimerization through phosphotyrosine recognition by the STAT SH2 domains (9, 10).
In addition to cytokine receptors, several growth factor tyrosine kinase receptors (epidermal growth factor, platelet-derived growth factor, and colony-stimulating factor) as well as non-tyrosine kinase receptors (prolactin, growth hormone, and angiotensin II) also induce the tyrosine phosphorylation, dimerization, and transcriptional activation of the STAT proteins (12,(17)(18)(19)(20)(21)(22)(23). However, since insulin does not stimulate the JAK kinases or increase the tyrosine phosphorylation of the STAT proteins, it is generally accepted that insulin does not impinge upon this particular signal transduction pathway (18,24,25). Nevertheless, in contrast to tyrosine phosphorylation, recent studies have indicated that growth hormone, interferon ␣, and interleukin 6 can also stimulate the serine phosphorylation of the STAT proteins in a manner distinct from the JAK-mediated tyrosine phosphorylation (11, 26 -28). Furthermore, both tyrosine and serine phosphorylation of STAT1 and STAT3 is required for maximal DNA transcriptional activity of reporter genes presumably through fostering STAT homodimerization (28,29). Based upon these findings, we have re-evaluated the potential role of insulin on STAT phosphorylation. In this study, we demonstrate that insulin stimulation results in a rapid quantitative serine phosphorylation of STAT3 without any significant effect on STAT1␣, STAT1␤, STAT5, or STAT6 phosphorylation.

EXPERIMENTAL PROCEDURES
Materials-Monoclonal antibodies for Western blot analysis were purchased from Transduction Laboratories. Polyclonal STAT1 and STAT3 antibodies for immunoprecipitation and Protein G Plus agarose were purchased from Santa Cruz Biotechnology. [ 32 P]Orthophosphate and enhanced chemiluminescence reagents were obtained from Amersham Corp. Dephosphorylation buffer was purchased from Boehringer Mannheim. All other reagents were purchased from Sigma.
Cell Culture-Chinese hamster ovary cells stably transfected with the human insulin receptor (CHO/IR) were maintained in ␣-minimal essential medium supplemented with 10% fetal bovine serum as described previously (30). 3T3-L1 pre-adipocytes were cultured in standard medium (Dulbecco's modified Eagle's medium containing 25 mM glucose, 2 mM glutamine, and 10% calf serum). Adipocyte differentiation was induced 2 days postconfluence with differentiation medium (standard medium supplemented with 1.0 g/ml insulin, 0.1 g/ml dexamethasone, and 27.8 g/ml isobutylmethylxanthine). Following 4 days in differentiation medium, the dexamethasone and isobutylmethylxanthine were removed by switching the cells to standard medium containing 1.0 g/ml insulin. After an additional 4 days, the cells were changed back into standard medium and maintained for 1-7 days before use.
Insulin Stimulation-CHO/IR and differentiated 3T3L1 adipocytes were washed two times with phosphate-buffered saline, pH 7.4, and incubated for 3-4 h in standard medium in the absence of serum. The cells were then treated with various concentrations of insulin (0 -100 nM) and times (0 -60 min) as indicated in the individual figure legends. Insulin stimulation was terminated by two washes of ice-cold phosphate-buffered saline, pH 7.4, removal of excess liquid by aspiration, and addition of liquid nitrogen to the tissue culture plates. The snapfrozen cells were placed at Ϫ80°C until harvested.
Whole Cell Detergent Lysates-CHO/IR and differentiated 3T3L1 adipocytes were extracted in ice-cold lysis buffer (50 mM HEPES, 1% Triton X-100, 2.5 mM EDTA, 100 mM NaF, 10 mM Na 4 P 2 O 7 , pH 7.8) containing 1 mM phenylmethylsulfonyl fluoride, 2 mM Na 3 VO 4 , 1 g/ml aprotinin, 10 M leupeptin, and 1 M pepstatin A by rotation for 10 min at 4°C. Insoluble material was separated from the soluble extract by microcentrifugation for 10 min at 4°C. Protein concentration was determined, and the samples were either immunoprecipitated and/or directly subjected to SDS-polyacrylamide electrophoresis using a 32-cm long 7.5% acrylamide, 0.1% bisacrylamide running gel. The resolved proteins were then transferred to PVDF filter membranes and Western blotted using the enhanced chemiluminescence detection system.
Phosphoamino Acid Analysis-CHO/IR cells were washed twice with phosphate-buffered saline and incubated for 1 h in serum-and phosphate-free medium. The cells were then incubated for 2 h at 37°C with 6 mCi of 32 PO 4 (1.5 mCi/ml) and stimulated with and without 100 nM insulin for 5 min. The cells were rapidly washed with ice-cold phosphate-buffered saline and solubilized on ice with 500 l of radioimmune precipitation buffer (150 mM NaCl, 1% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS, 50 mM Tris, 10 mM Na 4 P 2 O 7 , 100 mM NaF, 2 mM phenylmethylsulfonyl fluoride, 2 mM Na 3 VO 4 , 2 mM pepstatin A, 1 g/ml aprotinin, 10 M leupeptin). Insoluble material was removed by microcentrifugation, and the STAT proteins were immunoprecipitated from the cell lysates using the polyclonal STAT1 and STAT3 antibodies. Following SDS-polyacrylamide gel electrophoresis and transfer to PVDF filter membranes, the samples were subjected to autoradiography for 1 h. The radiolabeled bands were excised and subjected to acid hydrolysis using 200 l of 6 N HCl at 100°C for 1 h. The hydrolysates were collected, dried, and resuspended in 10 l of water. The samples and phosphoamino acid standards were spotted on 100-mm cellulose plates and electrophoresed at 1000 V for 90 min in 0.87 M acetic acid, 0.5% (v/v) pyridine, 0.5 mM EDTA. The standards were visualized by ninhydrin staining and the radiolabeled phosphamino acids by autoradiography.
Alkaline Phosphatase Treatment-CHO/IR cells unstimulated or stimulated with 100 nM insulin for 5 min were harvested in alkaline phosphatase lysis buffer (30 mM Tris, pH 7.4, 100 mM NaCl, 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 1 g/ml pepstatin A, 5 g/ml leupeptin). Insoluble material was removed by microcentrifugation for 10 min at 4°C and assayed for protein content. Each alkaline phosphatase reaction contained 300 g of cell lysate, 10 l of 10 ϫ dephosphorylation buffer (Boehringer Mannheim), 50 g/ml aprotinin, 500 M leupeptin, and 1000 units of calf intestinal phosphatase (Sigma), and the volume was increased to 100 l with buffer (20 mM HEPES, pH 7.4, 100 mM NaCl, 1 mM dithiothreitol, 1 mM MgCl 2 , 0.1 mM ZnCl 2 ). Reactions were incubated at room temperature for 1 h and terminated by the addition of SDS sample buffer.

RESULTS AND DISCUSSION
Insulin Stimulation Results in a Reduction in STAT3 Electrophoretic Mobility-Previous studies have reported that although several growth factors can stimulate the tyrosine phosphorylation of STAT proteins (12, 17, 19 -23), insulin does not induce STAT tyrosine phosphorylation or activation of JAKs (18,24,25). However, recent studies have also demonstrated that STAT1␣ and STAT3 are also serine-phosphorylated, which appears to enhance their transcriptional activity in concert with tyrosine phosphorylation (29). To evaluate the poten-tial role for insulin in mediating the serine phosphorylation of STAT proteins, CHO/IR and 3T3L1 adipocytes were incubated with and without insulin, and cell extracts were subjected to STAT Western blotting (Fig. 1). Using STAT isoform-specific antibodies both cell types were found to express STAT1␣, STAT1␤, STAT3, STAT5, and STAT6. In contrast, we were unable to detect any significant levels of STAT2 or STAT4 (data not shown). The absence of STAT4 is consistent with the observation of its limited tissue distribution with expression limited to the testis, thymus, and spleen (31). However, the absence of STAT2 was somewhat surprising as it has been observed in several tissues and cultured cell lines (4).
Nevertheless, insulin stimulation had no effect on the apparent mobility of STAT1␣, STAT1␤, STAT5, and STAT6 (Fig. 1,  A, C, and D). In contrast, insulin treatment for 5 min of either CHO/IR or 3T3L1 adipocytes resulted in a marked reduction of SDS-polyacrylamide gel electrophoretic mobility of STAT3 (Fig. 1B). This insulin-stimulated decrease in STAT3 SDSpolyacrylamide gel electrophoretic mobility was a typical characteristic of post-translational serine/threonine phosphorylation similar to that observed for a number of intracellular signaling proteins including STAT3 (32)(33)(34).
Insulin-stimulated Reduction in STAT3 Mobility Is Timeand Concentration-dependent-To determine whether this modification of STAT3 was an early or late event in insulin action, we next examined the time dependence of STAT3 gel shift in CHO/IR and 3T3L1 adipocytes (Fig. 2). In both cell types, insulin stimulation for 2 min resulted in a small but detectable reduction in STAT3 mobility, which was fully shifted by 5 min (Fig. 2, A and B). The insulin-stimulated decreased mobility of STAT3 in CHO/IR cells was persistent for up to 60 min. In contrast, following 60 min of insulin stimulation in 3T3L1 adipocytes there was a partial recovery of STAT3 mobility back toward its basal state. This difference in kinetics probably reflects the greater amount of insulin receptors expressed in the CHO/IR cells compared with the endogenous levels in 3T3L1 adipocytes. In any case, the effect of insulin on STAT3 was quantitative since the entire immunoreactive population of STAT3 was gel shifted between 5 and 30 min. Similarly, stimulation of CHO/IR cells with 3 nM insulin was sufficient to induce a complete gel shift of STAT3 with the halfmaximal response occurring at approximately 0.3 nM (Fig. 2C). These data demonstrate that the insulin-induced modification of STAT3 was a relatively rapid and sensitive event in both CHO/IR and 3T3L1 adipocytes. STAT3 Is Serine-phosphorylated in Response to Insulin-Post-translational modification of proteins by phosphorylation is commonly associated with changes in SDS-polyacrylamide gel electrophoretic mobilities (11,30,33,34). To determine the basis of the insulin-stimulated reduction in STAT3 mobility was, in fact, due to phosphorylation, we immunoprecipitated both STAT3 and STAT1 from 32 P-labeled cells (Fig. 3A). In the absence of insulin, there was a relatively low level of STAT3 phosphorylation, which was markedly increased following 5 min of insulin stimulation (Fig. 3A, lanes 1 and 2). In contrast, we were unable to detect any significant phosphorylation of STAT1 in either unstimulated or insulin-stimulated cells (Fig.  3A, lanes 3 and 4). As controls for the immunoprecipitation, STAT1 and STAT3 were clearly immunoprecipitated by their respective antibodies, but only STAT3 displayed the characteristic insulin-stimulated reduction in mobility (Fig. 3A, lanes  5-8).
Previous studies have reported that insulin neither activates the JAK kinases nor stimulates the tyrosine phosphorylation of STAT proteins (18,24,25). Consistent with these data, we were unable to detect any tyrosine phosphorylation of STAT3 by phosphotyrosine immunoblotting (data not shown). Thus, we next performed phosphoamino acid analysis of STAT3 isolated from unstimulated and insulin-stimulated cells (Fig. 3B). These data directly demonstrate that insulin stimulates the serine phosphorylation of STAT3 without any significant tyrosine or threonine phosphorylation.
Even though the insulin-stimulated serine phosphorylation of STAT3 correlated with its reduction of electrophoretic mobility, this does not prove cause and effect. Thus, to confirm that the serine phosphorylation of STAT3 was responsible for the reduction in mobility, extracts from unstimulated and insulin-stimulated cells were treated with alkaline phosphatase (Fig. 3C). As previously observed, STAT3 displayed the characteristic decreased mobility following insulin stimulation (Fig.  3C, lanes 1 and 2). Alkaline phosphatase treatment of extracts from unstimulated cells did not alter the mobility of STAT3 (Fig. 3C, lane 3). In contrast, alkaline phosphatase treatment of extracts from insulin-stimulated cells displayed an increase in mobility to the same extent as STAT3 from control cells (Fig.  3C, lane 4). These results demonstrate that the insulin-stimulated serine phosphorylation of STAT3 was directly responsible for the reduction in SDS-polyacrylamide gel electrophoretic mobility.
In summary, insulin-stimulated serine phosphorylation of STAT3 is unique in several ways compared with other hormones that activate the STAT pathway. For example, ciliary neurotrophic factor, oncostatin M, leukemia-inhibitory factor, interleukin 6, growth hormone, and interferon ␥ have all been found to induce both tyrosine and serine phosphorylation of STAT proteins (11,28,29,35). However, the data presented in this study demonstrate that insulin is the only ligand to date that mediates the exclusive serine phosphorylation of a STAT protein. Although it has recently been suggested that both tyrosine and serine phosphorylation were due to a bifurcation of the JAK pathway (27), our data indicate the presence of at least two distinct tyrosine and serine kinase pathways. Furthermore, our data demonstrate that the insulin-stimulated serine kinase pathway responsible for STAT protein phosphorylation was apparently specific for STAT3. This supports previous reports, which demonstrated that the interleukin-6 receptor selectively serine-phosphorylated STAT3 but not STAT1␣ (36). Currently, the identification of the serine kinase and molecular pathways mediating these differences remains to be determined.  1 and 3) or presence of 100 nM insulin (I, lanes 2 and 4) for 5 min at 37°C. Whole cell detergent extracts were prepared and immunoprecipitated using the STAT3 antibody (lanes 1 and 2) or the STAT1 antibody (lanes 3 and 4) as described under "Experimental Procedures." The immunoprecipitates (IP) were then resolved by SDSpolyacrylamide gel electrophoresis, transferred to PVDF filter membranes, and subjected to autoradiography. Unlabeled CHO/IR cells were treated identically as described above and immunoprecipitated with the STAT3 antibody (lanes 5 and 6) or the STAT1 antibody (lanes 7 and 8) and subjected to STAT3 and STAT1 Western blotting as indicated. IB, immunoblot. B, the immunoprecipitated 32 P-labeled STAT3 bands from control and insulin-stimulated cells described in A were cut from the PVDF filter membranes and subjected to one-dimensional phosphoamino acid analysis. Phosphoamino acid standards are indicated. P-Ser, phosphoserine; P-Thr, phosphothreonine; P-Tyr, phosphotyrosine. C, 150 g of cell lysate harvested from CHO/IR cells either untreated (C) or incubated with 100 nM insulin for 5 min (I) were subjected to treatment with alkaline phosphatase (PPase, 1000 units/ ml) for 1 h at room temperature. The samples were resolved by SDSpolyacrylamide gel electrophoresis and Western blotted with a STAT3 antibody.