Role of tyrosine 441 of interferon-gamma receptor subunit 1 in SOCS-1-mediated attenuation of STAT1 activation.

Suppressor of cytokine signaling (SOCS)-1, the key negative regulator of interferon (IFN)-gamma-dependent signaling, is induced in response to IFNgamma. SOCS-1 binds to and inhibits the IFNgamma receptor-associated kinase Janus-activated kinase (JAK) 2 and inhibits its function in vitro, but the mechanism by which SOCS-1 inhibits IFNgamma-dependent signaling in vivo is not clear. Upon stimulation, mouse IFNgamma receptor subunit 1 (IFNGR1) is phosphorylated on several cytoplasmic tyrosine residues, and Tyr(419) is required for signal transducer and activator of transcription (STAT) 1 activation in mouse embryo fibroblasts. However, the functions of the other three cytoplasmic tyrosine residues are not known. Here we show that Tyr(441) is required to attenuate STAT1 activation in response to IFNgamma. Several tyrosine to phenylalanine mutants of IFNGR1, expressed at normal levels in stable pools of IFNGR1-null cells, were analyzed for the phosphorylation of STAT1 during a 48-h period, and antiviral activity in response to IFNgamma was also measured. Stronger activation of STAT1 was observed in cells expressing all IFNGR1 variants mutated at Tyr(441), and, consistently, stronger antiviral activity was also observed in these cells. Furthermore, constitutive overexpression of SOCS-1 inhibited IFNgamma-dependent signaling only in cells expressing IFNGR1 variants that included the Tyr(441) mutation. Mutation of Tyr(441) also blocked the ability of SOCS-1 to bind to IFNGR1 and JAK2 in response to IFNgamma and the normal down-regulation of STAT1 activation and antiviral activity. These results, together with data from the literature, suggest a model in which, in response to IFNgamma, phosphorylation of Tyr(441) creates a docking site for SOCS-1, which then binds to JAK2 within the receptor-JAK complex to partially inhibit JAK2 phosphorylation. Furthermore, the virtually complete blockade of STAT1 phosphorylation by overexpressed SOCS-1 in this experiment suggests that the binding of SOCS-1 to Tyr(441) also blocks the access of STAT1 to Tyr(419) and that this effect may be the principal mechanism of inhibition of downstream signaling.

Interferon (IFN) 1 -␥ plays key roles in mediating antiviral and antigrowth responses and in modulating immune re-sponses (1). The major signal transduction pathway activated by IFN␥ has been elucidated through both biochemical and genetic studies. The IFN␥ receptor complex consists of two receptor subunits, IFNGR1 and IFNGR2, and the tyrosine kinases Janus-activated kinase (JAK) 1 and JAK2, which bind to IFNGR1 and IFNGR2, respectively. IFN␥ induces the oligomerization of the receptor subunits, leading to the activation of JAK1 and JAK2, which then phosphorylate tyrosine residues within the cytoplasmic domain of IFNGR1. Signal transducer and activator of transcription (STAT) 1 is then recruited to the receptor complex and phosphorylated on Tyr 701 , allowing it to be released, form homodimers, translocate to the nucleus, and bind to ␥-activated sequences to activate the transcription of interferon-stimulated genes (ISGs) (1)(2)(3).
The activation of STAT1 by IFN␥ is tightly controlled by several mechanisms (4). The SH2-containing phosphatase 2 binds to IFNGR1 and inhibits STAT1 activation without inhibiting the phosphorylation of IFNGR1 (5). Protein inhibitor of activated STAT 1 (PIAS-1) binds to STAT1 and prevents its association with target DNA (6). Both genetic and biochemical studies have shown that suppressor of cytokine signaling (SOCS)-1 is the most potent inhibitor of IFN␥ signaling (7). Mice lacking SOCS-1 develop a complex fatal neonatal disease (8 -10), and the mortality, which results from hypersensitivity to IFN␥, is largely prevented by administration of anti-IFN␥. In addition, premature death does not occur in mice lacking both SOCS-1 and IFN␥ (9,11). In response to IFN␥, STAT1 activation is much stronger in cells lacking SOCS-1 than in wild-type cells (11,12). The constitutive expression of SOCS-1 blocks IFN␥-mediated antiviral and antiproliferative activities (13), and in vitro studies have shown that the SOCS-1 protein inhibits the kinase activity of JAK2 by binding directly to the active site loop domain (14). SOCS-1 can also direct proteins to which it binds, including guanine nucleotide exchange factor VAV, insulin receptor substrates 1 and 2, and JAK2, to proteasome-mediated degradation (15)(16)(17)(18).
Tyrosine phosphorylation is used by most cell surface receptors to initiate downstream signaling in response to cytokines and growth factors (19 -23). Receptor phosphorylation creates docking sites for downstream signaling components and mutation of specific tyrosine residues blocks signal transduction (24 -28). The situation for IFN␥-dependent signaling has been well studied. Activation of JAK1 and JAK2 is mediated by trans-and auto-phosphorylation. Phosphorylated Tyr 419 of IFNGR1 creates the docking site for STAT1 in mouse embryo fibroblasts (MEFs), and phosphorylation of Tyr 440 of human IFNGR1 has a major role in activating STAT1 and mediating antiviral activity. Tyrosine to phenylalanine mutation of this motif impairs STAT1 activation as well as the expression of ISGs (29,30).
Tyrosine phosphorylation also provides the basis for negative regulation of receptor-dependent signaling (31)(32)(33)(34). For the interleukin-6 family of cytokines, the gp130 receptor subunit is phosphorylated, and four of its cytoplasmic tyrosine residues are involved in the activation of STATs 1 and 3 (35). Furthermore, Tyr 759 is required for the binding of SH2-containing phosphatase 2 and SOCS-3 (36,37). SH2-containing phosphatase 2 mediates the activation of mitogen-activated protein kinase in pro-B cell lines (38) and negatively regulates STAT activation in gp130-dependent signaling (39); SOCS-3 is the key negative regulator of gp130-dependent signaling (40 -42). Mutation of Tyr 759 to phenylalanine enhances signal transduction in response to gp130-linked cytokines, including interleukin-6, leukemia inhibitory factor, and oncostatin M (43,44). Furthermore, mice expressing a gp130 mutant lacking Tyr 759 have splenomegaly, lymphadenopathy, and an enhanced acute phase reaction (45). Here, we investigate the involvement of specific tyrosine residues of IFNGR1 in IFN␥-dependent signaling and find that Tyr 441 is required for attenuation. Tyrosine to phenylalanine mutation of this residue leads to stronger STAT1 activation and antiviral activity, and inhibition of signaling in response to SOCS-1 requires Tyr 441 , as does the IFN␥-dependent binding of SOCS-1 to IFNGR1 and JAK2.

EXPERIMENTAL PROCEDURES
Constructs-Plasmids expressing murine IFNGR1 and the IFNGR1 mutant Y419F were kindly provided by Dr. Robert Schreiber (Washington University, St. Louis, MO). IFNGR1 and IFNGR1 Y419F cDNAs were subcloned into pBABEpuro3. Tyrosine to phenylalanine mutations of IFNGR1 were generated by PCR-splicing overlapping extension (46). The identity of each plasmid was confirmed by DNA sequencing. The plasmid encoding SOCS-1, kindly provided by Dr. Ke Shuai (University of California Los Angeles, Los Angeles, CA), was used to subclone this gene into pBluescript at the XbaI site and then into pLHCX at the HindIII and HpaI sites.
Antiviral Assays-Cells seeded into 96-well plates at 2 ϫ 10 4 cells/ well were incubated overnight at 37°C and treated with serial 10-fold dilutions of IFN␥ for 18 h. Where indicated, the cells were challenged for 20 h with encephalomyocarditis virus (0.5 plaque-forming unit/cell), fixed, and stained with Giemsa.
Coimmunoprecipitation-After treatment, cells at 80% confluence in 150-mm dishes were washed once with phosphate-buffered saline, and cell pellets were lysed for 20 min at 4°C in 1 ml of buffer containing 0.5% Triton X-100, 50 mM HEPES, pH 7.4, 150 mM NaCl, 15 mM MgCl 2 , 0.1 mM EGTA, 10 mM sodium fluoride, 1 mM sodium orthovanadate, 1 mM phenylmethanesulfonyl fluoride, 3 g/ml aprotinin, 2 g/ml pepstatin, and 1 g/ml leupeptin. Cellular debris was pelleted by centrifugation at 16,000 ϫ g at 4°C for 10 min. Cell lysates were incubated with anti-FLAG antibody and protein G-Sepharose (Amersham Biosciences) overnight. Immunoprecipitates were washed four times with ice-cold lysis buffer and analyzed by the Western method.

RESULTS
Tyrosine Residues of IFNGR1 Provide Negative Regulation of IFN␥-dependent Signaling-Mouse IFNGR1 has four cytoplasmic tyrosine residues, which are phosphorylated upon stimulation with IFN␥. Tyr 419 is required for the activation of STAT1 (46,47), but the functions of the other tyrosines are not known. Previous studies have shown that Tyr 759 of gp130, which is not needed for STAT3 activation, is required for the negative regulation of this process (36,37,39). To reveal whether a tyrosine residue of IFNGR1 other than Tyr 419 is similarly required for the negative regulation of IFN␥-dependent signaling, IFNGR1null cells expressing either wild-type or 3F/419Y IFNGR1 (three of the four cytoplasmic tyrosine residues were mutated to phenylalanines) were used to examine the phosphorylation on tyrosine of STAT1, which was stronger in cells expressing 3F/419Y than in cells expressing wild-type IFNGR1 (Fig. 1a). Because phosphorylated STAT1 is essential for the transcription of most IFN␥-induced genes, the expression of three ISGs was also examined. After treatment with IFN␥, the levels of irf-1, ip-10, and socs-3 mRNAs were increased by 2-fold or more in cells expressing 3F/419Y compared with cells expressing wild-type IFNGR1 (Fig. 1b). Therefore, STAT1 activation is negatively regulated by a tyrosine residue other than Tyr 419 of IFNGR1.
Tyr 441 of IFNGR1 Is Required for Negative Regulation of STAT1 Activation in Response to IFN␥-To determine which tyrosine(s) might be required, several different tyrosine to phenylalanine mutants were generated (Fig. 2). In mutant 2F/ 441Y/419Y, STAT1 phosphorylation was attenuated normally, and the attenuation was abolished in all mutants that include Y441F (Fig. 3a). Note that the basal levels of STAT1 in these cells are very low. Detailed time courses with single tyrosine to phenylalanine mutants confirmed that Tyr 441 is the primary residue that mediates negative regulation of STAT1 activation in response to IFN␥ (Fig. 3b). To test whether Tyr 441 is also involved in modulating the antiviral response, an assay was performed to assess the ability of IFN␥ to protect cells from the cytopathic effects of encephalomyocarditis virus in cells expressing wild-type IFNGR1 or several mutants. Cells expressing the 3F/419Y, 2F/285Y/419Y, or 2F/370Y/419Y mutants of IFNGR1 were protected at much lower doses of IFN␥ than were cells expressing wild-type IFNGR1 or the 2F/441Y/419Y mutant (Fig. 3c). These data provide additional evidence that Tyr 441 of IFNGR1 mediates the negative regulation of IFN␥dependent signaling, including antiviral activity. (7); SOCS-1 was expressed constitutively in cells with wild-type IFNGR1 or the 3F/419Y, 2F/441Y/419Y, and 3Y/441F mutants, and responses to IFN␥ were examined. SOCS-1 blocked STAT1 phosphorylation virtually completely, but only in cells expressing wild-type IFNGR1 or the 2F/441Y/419Y mutant, and not in cells expressing the 3F/419Y or 3Y/441F mutants (Fig. 4a). Constitutive overexpression of SOCS-1 attenuated the IFN␥-dependent phosphorylation of JAK2 by about 2-fold in cells expressing wild-type IFNGR1 or the 2F/441Y/419Y mutant, but not in cells expressing the 3F/419Y or 3Y/441F mutants, and the phosphorylation of JAK1 was intact (Fig. 4b). Protein levels of JAK1 and JAK2 did not change during the time course investigated (data not shown).

Inhibition of IFN␥-dependent Signaling by SOCS-1 Is Mediated by Tyr 441 -Previous studies identified SOCS-1 as a critical inhibitor of IFN␥-dependent signaling
Previous studies have shown that SOCS-1-null mice are more resistant to viral infection than are wild-type mice (11) and, conversely, that overexpression of SOCS-1 completely blocks the antiviral activity of IFN␥ (13). Constitutively overexpressed SOCS-1 abrogated antiviral responses in cells ex-pressing wild-type IFNGR1 or the 2F/441Y/419Y mutant but did not inhibit this response in cells carrying the 3F/419Y or 3Y/441F mutants of IFNGR1. Consistent with these results, IFN␥ protected cells expressing the 3F/419Y or 3Y/441F mutants of IFNGR1 at lower doses than cells with wild-type IFNGR1 or the 2F/441Y/419Y mutant (Fig. 4c) 2. Expression levels of IFNGR1 mutants. a, diagram of the cytoplasmic domain of IFNGR1 and of the tyrosine to phenylalanine mutants studied. The four cytoplasmic tyrosines of IFNGR1, which are conserved between human and mouse, were mutagenized to produce 16 different mutants, 8 of which were used in this study. b, expression levels of IFNGR1 tyrosine mutants. IFNGR1-null MEFs were infected with empty vector pBABEpuro3 or with vector including DNA for expression of various mutants of IFNGR1, and pools of cells stably expressing the different IFNGR1s were selected. Total cell lysates, together with lysates from wild-type (WT) MEFs, were analyzed by the Western method using antibodies against IFNGR1 and actin.
FIG. 3. Tyr 441 is required to attenuate IFN␥ signaling. a, cells expressing various mutants of IFNGR1 were treated with IFN␥ for 24 h. Total cell lysates were prepared and analyzed by the Western method, using antibodies against phospho-STAT1 (pYSTAT1), total STAT1, and actin. Note that the basal level of STAT1 is very low in these cells. b, kinetics of STAT1 activation in response to IFN␥ in cells expressing wild-type (WT) IFNGR1 or the 3F/419Y, 2F/441Y/419Y, and 3Y/441F mutants. c, IFN␥-induced antiviral responses in cells expressing wild-type (WT) IFNGR1 or the 2F/285Y/419Y, 2F/370Y/419Y, and 2F/441Y/419Y mutants. Cells were incubated with the indicated concentrations of IFN␥ for 18 h before infection with encephalomyocarditis virus for 20 h, and live cells were stained. C, control (medium only); V, virus-infected, no IFN␥. analysis showed that, in response to IFN␥, association was observed only in cells expressing wild-type IFNGR1 or the 2F/441Y/ 419Y mutant, and not in cells with the 3F/419Y or 3Y/441F mutants. In addition, association of SOCS-1 and JAK2 was also observed only in cells with wild-type IFNGR1 or the 2F/441Y/ 419Y mutant (Fig. 5). These data indicate that SOCS-1 is recruited to the IFN␥ receptor complex in a ligand-dependent manner and that Tyr 441 is required for this association.

Contribution of Specific Tyrosine
Residues of IFNGR1 to IFN␥-dependent Signaling-In response to IFN␥, the phospho-rylation of Tyr 419 of murine IFNGR1 is absolutely required for STAT1 activation (46), the antiviral response (Fig. 3c), and the transcription of ISGs in reconstituted IFNGR1-null MEFs. 2 It is interesting that the requirement for the corresponding tyrosine residue of human IFNGR1, Tyr 440 , appears to be different: in primary human fibroblasts lacking IFNGR1 and reconstituted with the Y440F mutant, the expression of ISGs in response to IFN␥ was substantial, although the antiviral effect was inhibited. 3 The hybrid murine fibroblast cell line SCC16-5, containing a single copy of human chromosome 21 encoding human IFNGR2, was used initially to characterize human IF-NGR1 (49,50). The expression of various mutants of IFNGR1 in these cells revealed that Tyr 440 is required for STAT1 activation and antiviral activity (29,51). These results, taken together, suggest that cellular backgrounds affect the nature of IFNGR1-dependent signals as well as the responses to IFN␥.
When the Y440F mutant of human IFNGR1 was expressed in SCC16-5 cells, or when the Y419F mutant of mouse IFNGR1 was expressed in IFNGR1-null MEFs, substantial tyrosine phosphorylation of IFNGR1 in response to IFN␥ was still observed (29) (data not shown). The functions of the other phosphorylated tyrosine residues of IFNGR1 have not been investigated until now. Here we show that an important function of one of these residues is attenuation of IFN␥-dependent responses, which are stronger in IFNGR1-null MEFs expressing the 3F/419Y mutant than in cells expressing wild-type IFNGR1 ( Figs. 1 and 3c). Stronger STAT1 activation in response to IFN␥ was also observed in SCC16-5 cells expressing the 4F/440Y mutant, in which four of the five tyrosines in the cytoplasmic domain of IFNGR1 were mutated to phenylalanines, than in cells with wild-type human IFNGR1 (data not shown). Moreover, Tyr 441 is the primary residue that mediates this attenuation (Fig. 3). Therefore, IFNGR1 uses different tyrosine residues to provide activation and feedback inhibition signals. Mutant IFNGR1 constructs were made in a retroviral expression vector, and pools of IFNGR1-null cells were generated by infection, followed by selection with puromycin. By titering the viruses, relatively even expression of different mutant IFNGR1s was achieved, at levels matching that of IFNGR1 in wild-type cells (Fig. 2). We believe that it is advantageous to use stable pools of cells, allowing one to observe an average response and to avoid being misled by differences among different cell clones.
Interaction of SOCS-1 and IFNGR1-Analyses of the mechanisms of the inhibitory effects of SOCS-1 have focused on the ability of SOCS-1 to bind to the active loop of JAKs, and interaction of SOCS-1 with the interleukin-2 receptor ␤ chain FIG. 4. SOCS-1-mediated inhibition of IFN␥ signaling requires Tyr 441 . a, cells expressing wild-type (WT) IFNGR1 or the 3F/419Y, 2F/441Y/419Y, and 3Y/441F mutants were infected with empty vector pLHCX or with pLHCX containing DNA for expression of FLAG-tagged SOCS-1, and stable pools of cells were generated and treated with IFN␥ for 30 min. Total cell lysates were prepared and analyzed by the Western method, using antibodies against phospho-STAT1 (pYSTAT1), FLAG, and actin. b, kinetics of JAK1 and JAK2 phosphorylation in response to IFN␥. c, effect of SOCS-1 on IFN␥-induced antiviral responses in cells expressing wild-type (WT) IFNGR1 or the 3F/419Y, 2F/441Y/419Y, and 3Y/441F mutants. Antiviral assays were performed as described in the legend to Fig. 3c.   FIG. 5. Association of SOCS-1 and IFNGR1 or JAK2 requires Tyr 441 . Cells with wild-type (WT) IFNGR1 or the 3F/419Y, 2F/441Y/ 419Y, and 3Y/441F mutants, all overexpressing FLAG-tagged SOCS-1, were treated with IFN␥ for 10 min. Total cell lysates were prepared and incubated with anti-FLAG antibody and protein G-Sepharose beads. Immunocomplexes were separated by SDS-PAGE and analyzed by the Western method, using antibodies against IFNGR1, phospho-JAK2 (pYJAK2), and FLAG. has been shown not to be required for inhibitory effects (52). Our results show that, in IFN␥-dependent signaling, constitutive overexpression of SOCS-1 partially inhibited JAK2 phosphorylation (Fig. 4b), and this inhibition was observed only in cells retaining Tyr 441 of IFNGR1. We were also able to show that SOCS-1 binds to IFNGR1 at Tyr 441 in a ligand-dependent manner, presumably through an interaction between its SH2 domain and phospho-tyrosine 441, because tyrosine to phenylalanine mutation of Tyr 441 abolishes the interaction (Fig. 5). These results, together with others in the literature, suggest that SOCS-1, like SOCS-3 (the most homologous member of the SOCS family) (53), is likely to inhibit cytokine-dependent signaling though its interaction with a receptor (7,31,34), although SOCS-1 does interact with and inhibit JAKs in vitro (14,48).
How SOCS-1 Inhibits IFN␥-dependent Signaling-Studies in vitro have shown that SOCS-1 inhibits the kinase activity of JAK2, probably by binding to the active site loop (14,53), and the binding of SOCS-1 may also target JAK2 for proteasomedependent degradation (7). Our results show that SOCS-1 binds to JAK2 only in cells expressing wild-type IFNGR1 or the 2F/441Y/419Y mutant and that mutation of Tyr 441 completely abrogates the IFN␥-induced binding of SOCS-1 to IFNGR1 and JAK2 (Fig. 5). Consistently, in cells expressing the 3Y/441F or 3F/419Y mutant of IFNGR1, IFN␥-dependent signaling was prolonged, revealing that the interaction of SOCS-1 and Tyr 441 of IFNGR1 is required for negative regulation. A likely scenario is that, in response to IFN␥, SOCS-1 expression is induced, and SOCS-1 is recruited to IFNGR1 through phospho-tyrosine 441, which brings it close enough to JAK2 to enable it to bind to the active site loop, thus inhibiting the kinase activity and possibly also catalyzing proteasome-mediated degradation of JAK2, leading to negative feedback of IFN␥-dependent signaling. However, the relatively small effect on JAK2 phosphorylation contrasts with the dramatic effect on STAT1 phosphorylation. Therefore, it seems likely that a second mechanism is more important in vivo, namely, the binding of SOCS-1 to Tyr 441 of IFNGR1 blocks the access of STAT1 to Tyr 419 , thus preventing STAT1 activation.