SOCS3 exerts its inhibitory function on interleukin-6 signal transduction through the SHP2 recruitment site of gp130.

Interleukin-6 is involved in the regulation of many biological activities such as gene expression, cell proliferation, and differentiation. The control of the termination of cytokine signaling is as important as the regulation of initiation of signal transduction pathways. Three families of proteins involved in the down-regulation of cytokine signaling have been described recently: (i) SH2 domain-containing protein-tyrosine phosphatases (SHP), (ii) suppressors of cytokine signaling (SOCS), and (iii) protein inhibitors of activated STATs (PIAS). We have analyzed the interplay of two inhibitors in the signal transduction pathway of interleukin-6 and demonstrate that the tyrosine phosphatase SHP2 and SOCS3 do not act independently but are functionally linked. The activation of one inhibitor modulates the activity of the other; Inhibition of SHP2 activation leads to increased SOCS3-mRNA levels, whereas increased expression of SOCS3 results in a reduction of SHP2 phosphorylation after activation of the interleukin-6 signal transduction pathway. Furthermore, we show that tyrosine 759 in gp130 is essential for both SHP2 and SOCS3 but not for SOCS1 to exert their inhibitory activities on interleukin-6 signal transduction. Besides SHP2, SOCS3 also interacts with the Tyr(P)-759 peptide of gp130. Taken together, our results suggest differences in the function of SOCS1 and SOCS3 and a link between SHP2 and SOCS3.

Interleukin-6 is involved in the regulation of many biological activities such as gene expression, cell proliferation, and differentiation. The control of the termination of cytokine signaling is as important as the regulation of initiation of signal transduction pathways. Three families of proteins involved in the down-regulation of cytokine signaling have been described recently: (i) SH2 domain-containing protein-tyrosine phosphatases (SHP), (ii) suppressors of cytokine signaling (SOCS), and (iii) protein inhibitors of activated STATs (PIAS). We have analyzed the interplay of two inhibitors in the signal transduction pathway of interleukin-6 and demonstrate that the tyrosine phosphatase SHP2 and SOCS3 do not act independently but are functionally linked. The activation of one inhibitor modulates the activity of the other; Inhibition of SHP2 activation leads to increased SOCS3-mRNA levels, whereas increased expression of SOCS3 results in a reduction of SHP2 phosphorylation after activation of the interleukin-6 signal transduction pathway. Furthermore, we show that tyrosine 759 in gp130 is essential for both SHP2 and SOCS3 but not for SOCS1 to exert their inhibitory activities on interleukin-6 signal transduction. Besides SHP2, SOCS3 also interacts with the Tyr(P)-759 peptide of gp130. Taken together, our results suggest differences in the function of SOCS1 and SOCS3 and a link between SHP2 and SOCS3.
Interleukin-6 exerts its biological activities through a receptor complex composed of the IL-6 1 binding subunit gp80 and a dimer of the signal transducing receptor subunit gp130 (for review see Ref. 1). After ligand binding and gp130 dimer formation, constitutively associated kinases of the Janus family Jak1, Jak2, and tyrosine kinase 2 become activated by autophosphorylation. gp130, subsequently tyrosine phosphorylated on its cytoplasmic tail, recruits the transcription factors of the family of signal transducers and activators of transcription (STAT1 and STAT3) (2,3) and the protein-tyrosine phospha-tase SHP2 (4) via specific phosphotyrosine-SH2 domain interactions (5,6). In turn, these signaling components become tyrosine-phosphorylated also. Jak1 has been described to be crucial for the activation of gp130, the STAT factors (7), and SHP2 (8). The tyrosine-phosphorylated STATs form homoand/or heterodimers (9) and translocate to the nucleus where they bind to enhancer elements of interleukin-6 inducible genes (10).
The Jak/STAT signal transduction pathway is under negative control by several different mechanisms. The presence of a nuclear phosphatase leading to dephosphorylation of activated STAT1 has been proposed by Haspel et al. (11). These authors observed a quantitative recycling of dephosphorylated STAT1 from the nucleus to the cytoplasm implicating a circulation of STAT factors between the cytoplasm and the nucleus. These data contradict those of Kim and Maniatis (12) who demonstrated a proteasome-dependent loss of activated STAT1 in the nucleus. Recently, another group of IL-6 signaling inhibitors has been described, STAT-binding proteins, known as protein inhibitors of activated STATs (PIAS) (13,14). Although the PIAS do not contain phosphotyrosine binding domains such as SH2 or PTB domains, they associate with activated, tyrosinephosphorylated STATs, leading to a loss of STAT-DNA binding activity. The mechanism of this highly specific interaction of protein inhibitor of activated STATs with activated STAT factors remains to be elucidated. Another new family of inhibitors of cytokine signaling has been discovered in three different laboratories, recently. These proteins are referred to as suppressors of cytokine signaling (SOCS) (15), Jak-binding proteins (16), or STAT-induced STAT inhibitors (17). The members of this family contain a central SH2 domain as well as a carboxyl-terminal domain called the SOCS box. Depending on the cell type examined, SOCS1, SOCS2, and SOCS3 expression was found to be rapidly induced by IL-6. Because the SOCS proteins inhibit the IL-6-induced phosphorylation of Janus kinases, gp130 and STAT factors, they are regarded as feedback inhibitors of IL-6 signaling. SOCS1 inhibits the kinase activity of the three Janus kinases Jak1, Jak2, and tyrosine kinase 2 involved in IL-6 signaling (15)(16)(17). Recently, it has been described that SOCS1 binds to phosphotyrosine 1007 within the kinase domain of activated Jak2 (18). Also, the protein-tyrosine phosphatase SHP2 was found to inhibit IL-6 signal transduction. Activation of the IL-6 receptor complex leads to a recruitment of SHP2 to tyrosine 759 in gp130 and to its subsequent tyrosine phosphorylation (4). SHP2 activation is a crucial event for the induction of the mitogen-activated protein kinase (MAPK) pathway upon IL-6 stimulation (19). Mutation of Tyr-759 in gp130 results in an enhanced and prolonged STAT1 and STAT3 activation and in an increased gene induction (8,20,21). Because SOCS3 is induced by IL-6 (15-17) and SHP2 is simultaneously activated (4), we asked whether these proteins influence each other in respect to expression (SOCS3) or tyrosine phosphorylation (SHP2). We observed that an inhibition of SHP2 activation led to an enhanced induction of SOCS3 mRNA. On the other hand the expression of the SOCS3 protein decreased the level of tyrosine-phosphorylated SHP2 after IL-6 stimulation. Furthermore, we found that SOCS3, but not SOCS1, requires the SHP2 recruitment site in gp130 to exert its negative function on the IL-6 signal transduction pathway. Finally, it is demonstrated in the present study that both SHP2 and SOCS3 interact with a phosphotyrosine peptide containing the Tyr-759 motif of gp130. Although we demonstrate an SHP2-SOCS3 protein-protein interaction, we were also able to show that binding of SOCS3 to Tyr-759 of gp130 does not depend on the presence of SHP2.
Construction of Expression Vectors-Constructions were carried out by standard procedures (24). pGL3␣2M-215Luc contains the promoter region Ϫ215 to ϩ8 of the rat ␣2-macroglobulin gene fused to the luciferase encoding sequence and was described previously (8). The expression vector pSVL-EG encoding the chimeric EpoR/gp130 receptor (25) was modified by polymerase chain reaction mutagenesis to code for a chimeric EpoR/gp130 receptor (pSVL-EG(YYYYYY)) and a variant where tyrosine 759 in the cytoplasmic domain of gp130 was exchanged for phenylalanine (pSVL-EG(YFYYYY)) (8). The DNA fragments encoding the EpoR/gp130 chimeric receptors were also transferred into pRc/ CMV-EG and used for expression in HepG2 cells (pRc/CMV-EG(YYYYYY) and pRc/CMV-EG(YFYYYY)). The DNA fragments coding for the transmembrane and cytoplasmic domains of the chimeric receptors were introduced into pSVL-gp130 resulting in expression vectors for wild type gp130 (pSVL-gp130(YYYYYY)) and for gp130 with a tyrosine to phenylalanine exchange at position 759 (pSVL-gp130(YFYYYY)) (8). pSVL-gp130(FYFFFF) was generated by mutation of all but not Tyr-759 cytoplasmic tyrosine residues to phenylalanine. The latter constructs were used for generating stably transfected Ba/F3 cells. The expression vectors for the SOCS or CIS proteins were pEF-Flag-I/mSOCS1, pEF-Flag-I/mSOCS2, pEF-Flag-I/mSOCS3, and pEF-Flag-I/hCIS (15). The expression vector for GST-SOCS3-(23-151) was generated by insertion of the DNA fragment encoding amino acids 23-151 of SOCS3 into the GST fusion protein expression vector pGEX-5X-3 (Amersham Pharmacia Biotech).
Northern Blot Analysis-Total RNA was isolated from cultured Ba/F3 cells by using the Qiagen total RNA kit (Qiagan, Hilden, Germany) according to manufacturer's instructions. Gel electrophoresis and Northern blot analysis were performed with 10 g of total RNA as described previously (26) using a 32 P-labeled murine SOCS3-cDNA (15).
Transfection and Reporter Gene Analysis-Human hepatoma cells HepG2 were grown and transiently transfected by the calcium phosphate coprecipitation method as described previously (27). Transfections were adjusted with control vectors to equal amounts of DNA. Cell lysis and luciferase assays were carried out using the luciferase kit (Promega, Madison, WI) as described by the manufacturer's instructions. All transient expression experiments were done at least in trip-licate. Luciferase activity values were normalized to transfection efficiency monitored by the cotransfected ␤-galactosidase expression vector (pCR3lacZ, Amersham Pharmacia Biotech) (1.5 g). COS7 cells were grown in Dulbecco's modified Eagle's medium with 10% fetal calf serum, 100 mg/liter streptomycin, and 60 mg/liter penicillin. Approximately 1.5 ϫ 10 7 COS7 cells were transiently transfected with 6 -25 g of DNA using the DEAE-dextran method. Briefly, cells were incubated in medium containing the DNA, 80 M chloroquine, and 0.4 mg/ml DEAE-dextran for 80 min avoiding gas exchange. Afterward, cells were incubated for 1 min in phosphate-buffered saline containing 10% Me 2 SO. After 24 h cells were split 1:2, and after an additional 24 h in culture medium cells were stimulated. Ba/F3 cells were grown and stably transfected as described previously (28). Equal surface expression levels of gp130 were verified by fluorescence-activated cell sorter analysis with the B-P4 antibody specific for the extracellular domain of gp130.
Nuclear Extract Preparation and Electrophoretic Mobility Shift Assay-The preparation of nuclear extracts, measurements of protein concentrations, and electrophoretic mobility shift assays have been described (28). We used a STAT1-and STAT3-specific double-stranded 32 P-labeled probe, a mutated SIE oligonucleotide of the c-fos promoter (m67 SIE, 5Ј-GATCCGGGAGGGATTTACGGGAAATGCTG-3Ј). Protein-DNA complexes were separated on a 4.5% polyacrylamide gel containing 7.5% glycerol in TBE (23 mM Tris, 23 mM boric acid, 0.5 mM EDTA, pH 8.0) at 20 V/cm for 4 h. Gels were fixed in 10% methanol, 10% acetic acid, and 80% water for 30 min, dried, and autoradiographed.
Expression of Glutathione S-transferase Fusion Proteins-The expression GST-SOCS3-(23-151) was performed in Escherichia coli (BL21). Bacteria were grown at 37°C in LB medium with ampicillin to an A 600 of 1.2 and treated with isopropyl-1-thio-␤-D-galactopyranoside (1 mM) for 5 h at 20°C to induce expression to the GST-SOCS3-(23-151) fusion protein. Cells were lysed by sonication, and purification was performed according to the manufacturer's instructions (Amersham Pharmacia Biotech).

SOCS1 and SOCS3 but Not SOCS2 or CIS Are Potent Inhibitors of Acute Phase Protein Induction by Interleukin-6 -To
find out which of the IL-6-induced SOCS proteins interfere with the IL-6-stimulated induction of acute phase protein (APP) synthesis in liver cells, we tested whether SOCS1, SOCS2, SOCS3, or CIS expression affects APP gene promoter induction in human HepG2 hepatoma cells (Fig. 1). The respective SOCS or CIS cDNAs were cotransfected together with a reporter gene construct harboring the promoter of the ␣2-macroglobulin gene linked to the luciferase reporter gene (pGL3␣2M-215Luc) and an expression vector for a chimeric receptor containing the extracellular domain of the EpoR and the transmembrane and cytoplasmic domains of gp130 (EG(YYYYYY)), which allowed us to study the IL-6 signal transduction pathway independently from endogenous gp130 (25). Stimulation with erythropoietin led to a 20-fold induction of the reporter gene in the transiently transfected cells expressing the chimeric receptor in the absence of SOCS/CIS. Coexpression of SOCS1 or SOCS3 led to a dramatic reduction in APP gene promoter induction. Even the background level of reporter gene expression was reduced indicating that SOCS1 and SOCS3 also influence basal transcription in unstimulated cells. In contrast, SOCS2 and CIS had only moderate effects. Thus, SOCS1 and SOCS3 but neither SOCS2 nor CIS are potent inhibitors for acute phase protein gene induction by interleukin-6.
Lack of SHP2 Activation Leads to an Enhanced SOCS3 Expression-SHP2 counteracts the IL-6-induced acute phase protein gene induction (8,20,25). To study the interplay of the IL-6 signal transduction inhibitors SOCS1, SOCS3, and SHP2, we asked whether SHP2 might also negatively regulate the IL-6-induced expression of SOCS1 and SOCS3 (Fig. 2, A and  B). Therefore, IL-3-dependent Ba/F3 cells, which do not express endogenous gp130 (29), were stably transfected with gp130 receptor mutant cDNAs. Receptor surface expression of the transfected Ba/F3 cells was monitored by fluorescence-activated cell sorter analysis with an antibody raised against the extracellular domain of gp130 and was found to be comparable for all transfectants (Fig. 2C). Stimulation of cells expressing the wild type receptor gp130(YYYYYY) led to a rapid induction of SOCS3 mRNA as determined in Northern blot analysis ( Fig.  2A). The data were normalized to glyceraldehyde-3-phosphate dehydrogenase-mRNA levels (Fig. 2B). There was no detectable induction of SOCS1 mRNA in Ba/F3 cells (data not shown). Stimulation of cells carrying a mutation of the SHP2 recruitment site in gp130 by a substitution of Tyr-759 to Phe (gp130(YFYYYY)) resulted in increased SOCS3 mRNA levels compared with the wild type receptor. The single tyrosine 759 in the cytoplasmic part of gp130 (gp130(FYFFFF)) was not sufficient to mediate induction of the SOCS3 gene. These observations indicate that SHP2 activation counteracts SOCS3 gene expression.
SOCS3 Is a Potent Inhibitor of SHP2 Phosphorylation-To test whether SOCS3 counteracts SHP2 phosphorylation, COS7 cells were transiently transfected with SOCS3 expression vectors. SHP2 phosphorylation after stimulation with IL-6⅐sIL6R-complexes was analyzed by Western blotting (Fig. 3). Coexpression of SOCS3 led to a reduced SHP2 phosphorylation compared with cells not transfected with SOCS3-cDNA. We conclude from these data that SOCS3 also regulates signaling components, which themselves are negative regulators of the Ba/F3 cells stably expressing gp130(YYYYYY), gp130(YFYYYY), or gp130(FYFFFF) were stimulated with 20 ng/ml IL-6 and 1 g/ml sIL6R for the times indicated. Total RNA was isolated and analyzed in Northern blots with specific probes for SOCS3 and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) (A). The amount of radioactivity was quantified using a PhosphorImager (Storm840, Molecular Dynamics), and the data for the SOCS3-mRNA signals were normalized to the glyceraldehyde-3-phosphate dehydrogenase expression of the corresponding RNA preparation (B). Cell surface expression of the receptors was monitored by fluorescence-activated cell sorter analysis with the B-P4 antibody raised against the extracellular domain of gp130 (C) and was found to be similar for the different cell lines used. The data with stably transfected Ba/F3 cells were confirmed by the use of two independently transfected cell lines. IL-6 signal transduction pathway.
SOCS3 but Not SOCS1 Requires the SHP2 Recruitment Site in gp130 to Exert Its Inhibitory Activity on Acute Phase Protein Gene Induction-SOCS1 interacts with the kinase domain of activated Jak2 (16,18,30). On the other hand, only a weak association of Jak2 with SOCS3 has been described by Suzuki et al. (31). Unlike SOCS1, SOCS3 does not inhibit Jak kinase activity in vitro (30). To learn more about the mechanism of action of SOCS3, we tested whether SOCS3 activity depends on the activation of SHP2 and examined the potential of SOCS1 and SOCS3 to inhibit acute phase protein gene induction in the presence or absence of tyrosine 759 in gp130. Therefore, chimeric EpoR/gp130 wild type or mutant receptors were expressed in HepG2 cells together with or without SOCS1/ SOCS3. Reporter gene assays similar to those described above were performed (Fig. 4). Mutation of Tyr-759 in gp130 to Phe led to an enhanced APP promoter/reporter gene induction as expected from previously described experiments (8) (compare YYYYYY with YFYYYY). Both the expression of SOCS1 (Fig.  4A, left part) or SOCS3 (Fig. 4B, left part) led to reduced levels of the reporter luciferase activity upon stimulation of the wild type receptor EG(YYYYYY). Interestingly, unlike SOCS1, SOCS3 was unable to reduce APP gene promoter induction in the absence of Tyr-759 (EG(YFYYYY)) ( Fig. 4B, right part). Rather a slightly enhanced luciferase activity was measured after expression of SOCS3. Thus, SOCS3 but not SOCS1 requires the SHP2 recruitment site, i.e. Tyr-759 in gp130 to exert its inhibitory effect on acute phase protein gene induction.
The Inhibitory Effect of SOCS3 on STAT3 Activation Is Mediated by Tyrosine 759 in gp130 -To further confirm the requirement of Tyr-759 in gp130 for SOCS3 action, we asked whether the SOCS3-mediated reduction of STAT1/STAT3 activation (15)(16)(17)30) depends on Tyr-759 in gp130. Therefore, COS7 cells were transfected with wild type EG(YYYYYY) or mutant EG(YFYYYY) chimeric receptor-cDNAs together with increasing amounts of SOCS3-cDNA. STAT tyrosine phosphorylation in Western blots (Fig. 5, A and B) and DNA binding in electrophoretic mobility shift assays (Fig. 5C) were analyzed after receptor stimulation. In these assays, SOCS3 turned out to be more potent in counteracting STAT3 (Fig. 5A) and STAT1 (Fig. 5B) phosphorylation in the presence of (YYYYYY) than in the absence (YFYYYY) of Tyr-759 in the cytoplasmic part of gp130. Similar results were obtained with electrophoretic mobility shift assays (Fig. 5C). STAT activation upon stimulation of the wild type cytoplasmic part of gp130 is quite sensitive to SOCS3 expression. In contrast, STAT activation through the YFYYYY receptor mutant was significantly less sensitive to SOCS3 cotransfection. However, transfection of large amounts of SOCS3-cDNA also led to a reduced STAT activation in eryth-ropoietin-stimulated cells expressing the Y759F receptor mutant, probably mediated by a receptor-independent mechanism. For comparison, similar experiments were performed in the presence of SOCS1. Again, increasing amounts of SOCS1-cDNA led to a reduction of STAT3 and STAT1 tyrosine phosphorylation (Fig. 5, D and E) and DNA binding activity (Fig. 5F) after stimulation. The potential of SOCS1 to inhibit STAT activation was not affected by a mutation of tyrosine 759 in gp130. From these observations we conclude that the efficient inhibition of STAT1 and STAT3 activation by SOCS3, but not by SOCS1, requires Tyr-759 in gp130.
SOCS3 Binds to SHP2-Because the inhibitory activities of both SHP2 and SOCS3 depend on Tyr-759 in gp130, it was interesting to examine whether SOCS3 interacts with SHP2. For this purpose SHP2 was immunoprecipitated from cellular extracts of COS7 cells transfected with SOCS3-cDNA. Coprecipitation of SOCS3 shows the interaction of SHP2 with SOCS3 (Fig. 6, right lane). When the reverse precipitation was performed with antibodies to the Flag-tag of SOCS3 only a very faint band of co-precipitated SHP2 was observed after long exposure times (data not shown). It is likely that the antibody used for the precipitation of SOCS3 interferes with the SOCS3⅐SHP2 complex formation.
SOCS3 but Not SOCS1 Binds to the Phosphotyrosine Peptides Corresponding to Tyr-759 in gp130 -Because SOCS3 con-FIG. 3. SHP2 phophorylation is impaired by SOCS3 expression. COS7 cells were transiently transfected with expression vectors for SOCS3 as indicated. 48 h after transfection, cells were stimulated with 20 ng/ml IL6 and 1 g/ml sIL6R for 15 min and harvested, and cellular extracts were prepared for immunoprecipitation (IP) with antibodies raised against SHP2. The proteins bound were separated by SDS-polyacrylamide gel electrophoresis. Western blots (IB) were developed using phosphotyrosine antibodies (upper part) and after stripping of the membrane reprobed with SHP2 antibodies to verify equal loading (lower part). Arrowheads indicate the position of SHP2. tains an SH2 domain we tested whether this protein is capable to interact with phosphotyrosine 759 in gp130. Therefore, whole cell extracts of mock or SOCS3-transfected COS7 cells were incubated with biotin-conjugated peptides corresponding to the Tyr-759 motif of gp130 in its phosphorylated or unphosphorylated form (Tyr-759, biotin-␤A-TSSTVQYSTVVHSG and Tyr(P)-759, biotin-␤A-TSSTVQpYSTVVHSG). Precipitation of the peptides was performed with NeutrAvidin-coupled agarose. The precipitates were analyzed for SOCS3 and SHP2 by Western blotting (Fig. 7A). Essentially, neither SOCS3 nor SHP2 precipitation was found with the unphosphorylated peptide (Tyr-759). SOCS3 was only precipitated with the phosphotyrosine peptide (Tyr(P)-759) from extracts of cells transfected with SOCS3-cDNA (Fig. 7A, upper panel). Endogenous SHP2 was precipitated with the Tyr(P)-759 from both cell extracts (lower panel). Thus, SOCS3 is able to form a complex with the peptide containing the phosphorylated Tyr-759 of gp130. In contrast to SOCS3, SOCS1 does not interact with Tyr(P)-759. In a similar experiment as the one for SOCS3, we only detected SHP2 binding to the phosphopeptide even though SOCS1 was present in the cellular extracts of the transfected COS7 cells as demonstrated by immunoprecipitation of SOCS1 (Fig. 7B, upper panel, right lane). Similar experiments, which were performed with all the phosphotyrosine peptides corresponding to the six cytoplasmic tyrosine motifs of gp130, demonstrate that only the Tyr(P)-759 peptide is capable to bind SOCS3 and SHP2 (Fig. 7C). No binding of SOCS1 to any of the phosphotyrosine peptides was observed (Fig. 7D).
SOCS3 Binding to Tyr(P)-759 of gp130 Is Independent of SHP2-Our data suggest that SOCS3 is precipitated with the phophotryosine peptide Tyr(P)-759 by either direct binding to the peptide or by binding through SHP2. Therefore, we tested whether the SOCS3-Tyr(P)-759 interaction depends on the presence of SHP2 (Fig. 8). Cellular extracts of cells expressing SOCS3 or extracts from untransfected cells were prepared and incubated with the peptides Tyr-759, Tyr(P)-759 (lanes 3-8), and as further controls with Tyr(P)-683 (lane 1) or without peptide to exclude unspecific binding to the NeutrAvidine matrix (lane 2). The precipitates were analyzed by Western blotting. Only peptide Tyr(P)-759 was able to bind SHP2 and SOCS3 (lane 6) reflecting the specificity for the phosphorylated tyrosine motif Tyr-759 of gp130. In a similar approach we first precipitated SHP2 with the Tyr(P)-759 peptide and incubated this precipitate subsequently with cellular extracts of SOCS3expressing cells (lane 8) or control cells (lane 7). Again, an interaction of SOCS3 with the peptide Tyr(P)-759 could be shown. SHP2 binding to the peptide was not affected by the incubation with SOCS3.
To test the possible requirement of SHP2 for the SOCS3-Tyr(P)-759 interaction, SHP2 was depleted from cellular extracts of COS7 cells expressing SOCS3 by immunoprecipitation with SHP2 antibodies. As a control the antibody-bound SHP2 was monitored (lane 10). The coprecipitated SOCS3 is visible in the upper panel (lane 10). The analysis of the remaining supernatant is shown in lane 9. The amounts of SHP2 and SOCS3 in the SHP2-depleted extract were estimated by a second immunoprecipitation with antibodies against SHP2 (lane 12) and against the Flag-tag of SOCS3 (lane 11) demonstrating the lack of SHP2 and the presence of SOCS3 in the SHP2-depleted extract. Finally, this SHP2-depleted extract was used for the precipitation with the phosphotyrosine peptide 759 of gp130 (lane 13). It is concluded from this experiment that SHP2 is not required for SOCS3 binding to the Tyr(P)-759 peptide.
Binding of an SH2-Domain-containing SOCS3 Fragment to the Tyr(P)-759 Peptide of gp130 -The specific binding of SOCS3 to the phosphotyrosine peptide Tyr(P)-759 implicates that this interaction is because of the SOCS3-SH2 domain. To examine this interaction in more detail, we expressed a GST-SOCS3 fusion protein comprising the amino acids 23-155 of SOCS3 (GST-SOCS3-(23-151)) in E. coli. From structural predictions this part of SOCS3 has been suggested to contain the extended SH2 domain (15)(16)(17)(18). The purified fusion protein was incubated with biotin-conjugated peptides corresponding to the Tyr-759 motif of gp130 in phosphorylated or unphosphorylated form. The peptide precipitates were analyzed for GST-SOCS3-(23-151) by Western blotting using a GST-specific antibody (Fig. 9). A strong binding of the fusion protein was only observed with the tyrosine-phosphorylated peptide Tyr(P)-759. DISCUSSION SHP2, SOCS1, and SOCS3 are regulators of cytokine signaling (8, 15-17, 20, 21). In the present study, we have focused on the interplay of these proteins in IL-6 signal transduction. We describe for the first time a physical and functional link between SHP2 and SOCS3 and differences in the mode of action of SOCS1 and SOCS3. The inhibitory action of SHP2 on the IL-6 signal transduction pathway depends on Tyr-759 in gp130 (8,20,21). Surprisingly, we found that the presence of Tyr-759 in gp130 is also crucial for the inhibitory function of SOCS3 (Figs. 4 and 5). An exchange of Tyr-759 by Phe impairs the inhibitory activity of SOCS3 on both STAT1 and STAT3 activation (Fig. 5) as well as on acute phase protein gene promoter induction (Fig. 4). The function of SOCS1, however, is not affected by this point mutation in gp130, demonstrating the specific requirement of Tyr-759 for SOCS3 activity. The simplest explanation for this finding is that the inhibitory activity of SOCS3 depends on the interaction of its SH2 domain with Tyr(P)-759 in the receptor protein. Indeed, we were able to demonstrate that SOCS3, but not SOCS1, specifically associates with the phosphopeptide Tyr(P)-759 of gp130. Interestingly, SHP2 also binds to this peptide (Fig. 7). The interaction of SOCS3 with the cytoplasmic tail of gp130 could be direct or via other proteins. We have presented evidence for the existence of SHP2⅐SOCS3 complexes (Fig. 6), which could reflect an adaptor function of SHP2 for SOCS3. Therefore, we tested whether SOCS3 binds directly or through SHP2 to the Tyr(P)extracts (20 g) were analyzed for STAT3 (A) and STAT1 (B) phosphorylation by Western blotting (IB) using antibodies for the tyrosinephosphorylated forms of STAT3 and STAT1, respectively. After stripping the blots were reprobed with STAT3 or STAT1 antibodies (lower panels). STAT DNA binding was analyzed by electrophoretic mobility shift assay with STAT1/STAT3-specific DNA probes (C). Bands resulting from STAT1 and STAT3 homo-and heterodimers are indicated. The experiments were also performed with the indicated amounts of SOCS1-cDNA. STAT1 and STAT3 tyrosine phosphorylation (D and E) as well as DNA binding activity (F) were determined as described in A-C, respectively. SOCS3 Acts via Tyr-759 of gp130 on the IL-6 Signal Transduction 759 peptide. The depletion of SHP2 did not impair SOCS3 binding to the Tyr(P)-759 receptor peptide (Fig. 8). Furthermore, we were able to demonstrate that the recombinant GST-SOCS3-(23-151) fusion protein containing the SOCS3-SH2 domain binds to phosphotyrosine Tyr(P)-759 of gp130 (Fig. 9). These results show that SHP2 is not required for binding of SOCS3 to the gp130 receptor motif.
In previous reports it has been described that SHP2 inhibits IL-6-induced acute phase protein gene promoter activation by down-regulation of STAT phosphorylation (8,20,21). In the present paper we have shown that SOCS1 and SOCS3, but neither SOCS2 nor CIS, are potent inhibitors of APP gene promoter induction (Fig. 1). This observation is consistent with data of Nicholson et al. (30) who described a powerful inhibitory activity of SOCS1 and SOCS3 on leukemia inhibitory factor (LIF) signaling. SOCS1 and SOCS3 exhibit similar effects on STAT3 phosphorylation and APP gene induction (18,30). It has been shown by several investigators that SOCS1 and SOCS3 bind to the kinase domain of activated Jak1 and Jak2 (18,30,32). An attractive model for Jak inhibition by SOCS1 suggests that the kinase activation loop of Jak2 interacts with the SH2 domain of SOCS1. This allows SOCS to present its kinase inhibitory region, which is quite homologous to the kinase activation loop, to the pocket in the activation site, which in turn might prevent the access of substrates and/or ATP (18).
However, SOCS1 and SOCS3 seem to exert their feedback inhibitory action on the Jak/STAT pathway by different mechanisms as indicated by our results and the observation that SOCS1, but not SOCS3, inhibits Jak autophosphorylation in an in vitro kinase assay (18,30). In contrast, Sasaki et al. (32) did not find these differences but a higher affinity of SOCS1 than SOCS3 to bind Jak2. Furthermore, the kinase inhibitory region of SOCS3 was more potent in inhibiting of Jak2 than the kinase inhibitory region of SOCS1 (32).
The novel aspect of the data presented in this paper is the fact that SOCS3, in contrast to SOCS1, has to be recruited to the receptor complex to inhibit IL-6 signal transduction. Thus, the inhibition of IL-6 signaling by SOCS3 could be because of different mechanisms. First, SOCS3 could inhibit Jak activity FIG. 7. SOCS3, but not SOCS1, binds to a tyrosine-phosphorylated peptide, which corresponds to the Tyr-759 motif of gp130. COS7 cells were mock transfected or transfected with expression vectors for SOCS3 (20 g) (A) or SOCS1 (20 g) (B). 48 h after transfection cellular extracts were prepared and incubated with the biotinylated peptides Tyr-759 or Tyr(P)-759 or without peptide as indicated (PP). After precipitation with NeutrAvidin-coupled agarose the proteins were subjected to Western blot analysis (IB) using antibodies to SHP2 (lower panels) or to the Flag-tag of the SOCS proteins (A, upper panel, SOCS3; B, upper panel, SOCS1). SOCS1 expression in B was monitored by direct precipitation (IP) with an antibody against the Flag-tag of the transfected SOCS1 protein. C, cellular extracts containing SOCS3 were prepared and incubated with the indicated peptides corresponding to the six cytoplasmic tyrosine motifs of gp130 as described in A. SOCS3 and SHP2 association with the peptides was detected by immunoblotting with antibodies to the Flag-tag of SOCS3 or to SHP2, respectively. D, the experiment in C was also performed for SOCS1. The expression of SOCS1 was monitored by direct precipitation (IP) with an antibody against the Flag-tag of the transfected SOCS1 protein (far right lane).
via binding of the SOCS3-SH2 domain to the activation loop of the kinase. Second, SOCS3 could be recruited to gp130, directly to Tyr(P)-759 or by binding to the receptor-associated SHP2, leading to the inhibition of Jaks through the kinase inhibitory region of SOCS3, which is in line with the observation of Sasaki et al. (32).
Because SOCS3 gene expression is induced by IL-6 and because SOCS3 inhibits IL-6-signal transduction, SOCS3 functions as a feedback inhibitor (15)(16)(17). We did not observe a significant SOCS1-mRNA induction after stimulation with IL-6⅐sIL-6R complexes (Fig. 2). However, this does not exclude that SOCS1 also influences IL-6 signaling when induced via another pathway. It has been proposed that SHP2 phosphorylation results in an increase in enzymatic (phosphotyrosine phosphatase) activity (33), which in turn might also negatively influence signal transduction by dephosphorylating signaling components. Interestingly, an impaired SHP2 activation by the mutation of Tyr-759 in the gp130 receptor correlates with an enhanced SOCS3 gene induction (Fig. 2). Furthermore, SOCS3 itself can not act on this receptor mutant as an inhibitor of its own expression. Thus, a reduction in SHP2 activation might be compensated by an increase in SOCS3 expression. On the other hand, an enhanced SOCS3 expression leads to a reduced level of SHP2 phosphorylation (Fig. 3). This might be because of the inhibition of Jaks by SOCS3 or by direct competition of SHP2 and SOCS3 for Tyr(P)-759 of gp130. Therefore, an enhanced expression of SOCS3 might be compensated by a reduced level of phosphorylated SHP2. Further experiments are required to show whether tyrosine phosphorylation of SHP2 modulates SOCS3 activity. It is not yet clear which tyrosine residue(s) in SHP2 are phosphorylated after stimulation of the IL-6 signal transduction pathway.
The MAPK activator phorbol 12-myristate 13-acetate has recently been shown to inhibit the Jak/STAT pathway (34,35). Work from our laboratory (36) demonstrates that the inhibitory function of MAPK on IL-6 signal transduction also depends on the presence of Tyr-759 in gp130. Mutation of Tyr-759 in gp130 to Phe abolishes the potential of phorbol 12-myristate 13-acetate to inhibit IL-6-induced STAT activation. Because phorbol 12-myristate 13-acetate also induces SOCS3 gene transcription, it is very likely that phorbol 12-myristate 13-acetate exerts its action via SOCS3 whose activity also depends on Tyr(P)-759 in gp130 as described in the present report.
It is intriguing to assume that SHP2 exerts at least part of its negative regulatory function on the Jak/STAT pathway through the recruitment of SOCS3 to the activated receptor complex. To test this idea, experiments are presently in progress. . Furthermore, extracts of SOCS3transfected cells were depleted from SHP2 by immunoprecipitation using SHP2 antibodies. The antibody-bound SHP2 and the coprecipitated SOCS3 were monitored (lane 10). The analysis of the remaining supernatant is shown in lane 9. The amounts of SHP2 and SOCS3 in the SHP2-depleted extract were estimated by a second immunoprecipitation with antibodies against SHP2 (lane 12) and the Flag-tag of SOCS3 (lane 11). A peptide precipitation with Tyr(P)-759 peptide was performed with the SHP2-depleted extract (lane 13). After precipitation the proteins were separated on SDS-polyacrylamide gel electrophoresis and analyzed for SHP2 and SOCS3 as described in the legend to Fig. 6.   FIG. 9. The SOCS3-SH2-containg fusion protein GST-SOCS3-(23-151) binds to Tyr(P)-759. GST-SOCS3-(23-151) was expressed in E. coli. As detailed under "Experimental Procedures" purified fusion protein was incubated without peptide or with the biotinylated peptides Tyr-759 or Tyr(P)-759 as indicated (PP). After precipitation with Neu-trAvidin-coupled agarose the proteins were subjected to Western blot analysis (IB) using antibodies to GST.