Identification of Critical Residues Required for Suppressor of Cytokine Signaling-specific Regulation of Interleukin-4 Signaling*

Suppressor of cytokine signaling (SOCS) family proteins were originally identified as cytokine-induced negative regulators of cytokine signaling. We show that SOCS-1 and SOCS-3 inhibit interleukin (IL)-4-dependent signal transducer and activator of transcription 6 (Stat6) activation of and subsequent gene induction. By contrast, SOCS-2 and cytokine-inducible Src homology domain 2 (SH2)-containing protein up-regulate these processes. IL-4 initiates transmembrane signaling through two types of receptor complexes comprising the IL-4R a subunit and the associated Janus kinase 1 (Jak1) as common essential components. We demonstrate that both SOCS-1- and SOCS-3-mediated down-regulation of IL-4 signaling is due to an inhibition of the receptor associated Jak1 activity. The SOCS proteins contain an amino-terminal region of variable length and primary structure, a central SH2 domain, and a carboxyl-termi-nal conserved motif termed SOCS-box. We show that the SH2 domains of SOCS-2, SOCS-3, and cytokine-inducible SH2-containing protein are functionally redundant in regulating the IL-4-dependent Jak-Stat signaling. The Pre-SH2 domains of SOCS-2 and SOCS-3 confer the specificity of their regulatory function. Importantly, the Pre-SH2 domain of SOCS-3 alone can inhibit IL-4 signaling. The SH2-proximal 25 amino acids of SOCS-3 are sufficient

Cytokines control a variety of cellular responses by activating intracellular signaling cascades, including the Janus kinase-signal transducer and activator of transcription (Jak-Stat) 1 pathway that leads to the transcriptional activation of specific genes. The proteins encoded by these genes mediate multiple biological functions attributed to the cytokines (1)(2)(3). Cellular response to cytokines, however, is limited in both magnitude and duration (2,4,5). Although there has been major progress in recent years in uncovering the molecular events that turn on the cytokine signaling pathways and activate gene expression, the molecular events that govern the negative regulation of cytokine signaling remained poorly defined. Recent work from a number of laboratories, including ours, has suggested that a cytokine-mediated Jak-Stat signaling may be negatively controlled at multiple levels by a number of molecules and by different means. These include inhibition of receptor function by receptor antagonists, decoy receptors, and protein-tyrosine phosphatases; inhibition of Jak activity by protein-tyrosine phosphatases and by a family of nonenzyme proteins termed suppressors of cytokine signaling (SOCS) that are induced by cytokine stimulation of cells; inhibition of Stat activity by protein-tyrosine phosphatases and a family of protein inhibitors of activated Stats; and proteolytic degradation of the signaling components (4, 6 -19).
To further understand the negative regulation of IL-4 signaling, we investigated the role of SOCS family proteins in the regulation of IL-4-dependent Jak-Stat signaling. Eight mem-bers have been identified in the SOCS family (16 -18, 34, 35). The first member of the family, termed cytokine-inducible SH2containing protein (CIS), was shown to inhibit erythropoietin (EPO) and IL-3 signaling (36). The second member was originally cloned by three independent groups of investigators and termed SOCS-1, Jak-binding protein, and Stat-inducible Stat inhibitor (16 -18). SOCS-1 can bind to the Jak kinase-domain and inhibit the kinase activity (16,17). SOCS-2 interacts with the IGF-1 receptor kinase-domain (37). Leptin induces SOCS-3, which in turn inhibits leptin-mediated signal transduction (38). The CIS, SOCS-1, SOCS-2, and SOCS-3 genes are induced by a variety of cytokines, including IL-4 (4). The SOCS family proteins are structurally characterized by three distinct domains: a region at the amino terminus that is variable in size and amino acid sequence; an SH2 domain in the middle, and a region of homology termed SOCS-box at the carboxyl terminus (4, 5, 16 -18, 39). Although the SH2 domain of SOCS-1 has been implicated in recognizing tyrosine-phosphorylated Jak proteins, its amino-terminal domain was shown to be required for the inhibition of IL-6-dependent Jak-Stat signaling (39,40). The SOCS-box has been shown to mediate interactions with elongins B and C, which in turn may couple SOCS and the SOCS-bound proteins to the proteosomal degradation pathway (41,42). Recent reports show that mice lacking the SOCS-1 or SOCS-3 gene are defective in IFN-␥ and EPO-mediated cell signaling, respectively (43)(44)(45). However, the mechanisms by which these two SOCS proteins confer the cytokine specificity in regulating the Jak-Stat signaling remained unclear.
In this study, we demonstrate that SOCS family proteins differentially regulate IL-4-dependent signal transduction. We also describe the structure-function analyses of the SOCS proteins by swapping domains among the family members and by introducing mutations in SOCS-3 to define the structural basis for the specificity of SOCS-mediated regulation of IL-4-activated Jak-Stat signal transduction.

EXPERIMENTAL PROCEDURES
Plasmid Constructions-The cDNAs for murine CIS, SOCS-1, SOCS-2, and SOCS-3, cloned into the vector pEF-BOS, were expressed under the control of the human elongation factor-1␣ promoter (18,46). The chimeric SOCS constructs were prepared by generating polymerase chain reaction (PCR) fragments encoding the appropriate domains of SOCS-2, SOCS-3, and CIS using the respective cDNA as the template (see Fig. 5A). The fragment generated from each template has a nucleotide sequence overlapping that of the junction of the chimeric SOCS protein. Two cognate PCR fragments were mixed and subjected to PCR amplification in the absence of any added primers so that these partially overlapping (at the junction) fragments could function as both templates and primers to generate the full-length chimeric cDNA. PCR fragments encoding the chimeric cDNAs were subcloned into the MluI site of pEF-BOS (18).
The deletion mutants of SOCS-3 were constructed by PCR using specific primers and the SOCS-3 cDNA as the template. Site-directed mutagenesis was performed by using a pair of complementary mutant primers and two wild type terminal primers to generate overlapping PCR fragments. The full-length cDNAs encoding the mutant SOCS-3 were generated as described above for the chimeric SOCS proteins. The green fluorescent protein (GFP)-SOCS (Pre-SH2) fusion proteins were constructed by generating PCR fragments using specific primer pairs and SOCS cDNA as templates and subcloned into EcoRI-BamHI sites of the vector pEGFP-C1 (47).
The chimeric receptor containing the extracellular and transmembrane domains of murine EPO receptor and the cytoplasmic domain of human IL-4R␣ was prepared using PCR as described for the chimeric SOCS proteins (48,49). The chimeric receptor cDNA was subcloned into the HindIII-EcoRI sites of the vector pcDNA3.1/V5-His A (Invitrogen).
The glutathione S-transferase (GST)-SOCS-1 construct was prepared by subcloning the murine SOCS-1 coding sequence (PCR fragment) into the EcoRI-XhoI sites of the vector pGEX-5X-3. GST-SOCS-3 was constructed by subcloning the filled-in MluI fragment containing the coding sequence of murine SOCS-3 derived from the pEF-BOS-SOCS-3 plasmid (18, 46) into the SmaI site of pGEX-5X-3. GST and GST-SOCS fusion proteins were expressed in Escherichia coli and purified as described (10).
Cell Culture and Transfection-293T cells were grown in Dulbecco's modified Eagle's medium containing 10% bovine serum and 50 mg/liter of penicillin and streptomycin. Cells (10 6 cells/10-cm plate) were transfected with 2 g of Stat6 expression plasmid and 5 g of SOCS expression plasmids by calcium phosphate coprecipitation method as described (50).
Electrophoretic Mobility Shift Assay, Luciferase Assay, Immunoprecipitation, and Western Blot Analyses-Electrophoretic mobility shift assay (EMSA) was performed using 10 g of whole cell extract (WCE) proteins and 0.2 ng of radiolabeled N6-GAS oligonucleotide probe as described (11,12). Luciferase activity was determined and normalized as described (11,12). For immunoprecipitation, extracts were prepared by lysing the cells in ice-cold buffer containing 50 mM Tris, pH 7.9, 150 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol, 1% Nonidet P-40, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, 2 g/ml leupeptin, 2 g/ml pepstatin, and 5 g/ml aprotinin on ice for 30 min. The cleared supernatant containing 500 g proteins was incubated with 2-5 g of the appropriate antibody immobilized on agarose beads for 12 h at 4°C. The captured beads were boiled in denaturing buffer, and the released proteins were analyzed by Western blotting as described (12).

SOCS Family Proteins Differentially Regulate IL-4-activated
Jak-Stat Signaling-The SOCS-1, SOCS-2, SOCS-3, and CIS genes are induced by a variety of cytokines, including IL-4 and IL-13, and the SOCS family proteins have been implicated in the negative regulation of cytokine signaling (4, 16 -18). These findings prompted us to examine whether these proteins could inhibit the IL-4-dependent Jak-Stat signal transduction and the subsequent gene induction. To address this, we expressed the SOCS-1, SOCS-2, SOCS-3, and CIS proteins under a constitutive heterologous promoter in 293T cells. These cells exhibit very high transfection efficiency (Ͼ90%) and do not express the endogenous Stat6 protein (31). This results in the elimination of background IL-4 signaling derived from the nontransfected cells (Ͻ10%) that are exposed to IL-4, which is a potential problem in cells transfected at lower efficiency. 293T cells were transiently transfected with a Stat6 expression plasmid and an individual SOCS expression plasmid, and Stat6 activation in response to IL-4 was measured by EMSA using a Stat6-specific N6-GAS oligonucleotide probe. (11,12). The results show that expression of SOCS-1 and SOCS-3 completely blocked the IL-4-mediated activation of Stat6 (Fig. 1A). By contrast, under the same experimental conditions, both SOCS-2 and CIS up-regulated the IL-4-mediated activation of Stat6 (Fig. 1A). To ensure that the SOCS proteins were expressed in the transfected cells at comparable levels, extracts derived from the exogenous SOCS-expressing cells were subjected to immunoprecipitation with a monoclonal antibody against the Flag epitope encoded at the amino terminus of the SOCS constructs (18), and immunoblots were probed with a polyclonal anti-Flag antibody. The results show that all four SOCS proteins were expressed at comparable levels in transfected 293T cells (Fig. 1B).
To examine whether the expression of SOCS proteins has any effect on the expression of the Stat6 transgene (tagged with V5 epitope), immunoblot analysis was performed using whole cell extracts derived from Stat6-expressing and individual SOCS-expressing 293T cells. The results show that the expression of SOCS-1, SOCS-2, SOCS-3, or CIS did not alter the expression of the Stat6 transgene in 293T cells (Fig. 1C). Therefore, we conclude that despite overexpression, the SOCS proteins differentially regulate IL-4-dependent activation of Stat6 in 293T cells.
In order to determine whether the expression of SOCS transgenes regulates the transcription of IL-4-responsive genes, a Stat6-inducible luciferase reporter construct was cotransfected into 293T cells with the expression plasmid for Stat6 and individual SOCS (12,31). IL-4 treatment (20 ng/ml) of the transfected cells for 15 h resulted in a 7-fold increase in luciferase activity in the absence of any exogenous SOCS expression (Fig. 2). IL-4-responsive luciferase activity was completely ablated by SOCS-1 and was inhibited by 50% by SOCS-3 expression (Fig. 2). These results are consistent with the recent report by Losman et al. (51) that SOCS-1 is a potent inhibitor of IL-4 signaling. We measured the time course (0 -15 h) of SOCSmediated regulation of IL-4-dependent Stat6 activation. Stat6 activation by IL-4 was completely inhibited for 15 h in SOCS-1-expressing cells; however, in SOCS-3-transfected cells, Stat6 activation by IL-4 remained completely inhibited at an early phase (up to 4 h) of cytokine stimulation, and the inhibition was reduced to ϳ50% by 15 h (data not shown). In contrast, Stat6 activation by IL-4 remained up-regulated for 15 h in SOCS-2 and CIS-transfected cells (data not shown). Consistent with the IL-4-dependent Stat6 activation profile, SOCS-2 and CIS remarkably up-regulated the IL-4-dependent transcription compared with the vector-transfected cells (Fig. 2). Thus, the Stat6 activation kinetics exhibited correlation with the gene induction profile in SOCS-overexpressing cells. Taken together, these results clearly indicate that IL-4-dependent Jak-Stat activation is inhibited by SOCS-1 and SOCS-3 but upregulated by SOCS-2 and CIS.
Jak1 Is a Target of SOCS-1 and SOCS-3-mediated Inhibition of IL-4 Signaling-IL-4 triggers transmembrane signaling through two types of receptors, both of which utilize the IL-4R␣ chain and the associated Jak1 as unique components (21,22). SOCS-1 was originally identified as a Jak-binding protein (16). To examine whether SOCS-1 and SOCS-3 could inhibit the activation of both type I and type II IL-4 receptors, we took the following biochemical and genetic approaches. 293T cells were cotransfected with the expression plasmids for murine Jak1 (tagged with the c-Myc epitope) and SOCS-1 or SOCS-3 (both tagged with the Flag epitope). An anti-c-Myc monoclonal antibody coprecipitated Jak1-associated SOCS-1 or SOCS-3 from 293T cells that were treated with IL-4 for 30 min or left untreated. This indicates that Jak1 physically associates with both SOCS-1 and SOCS-3, even in the absence of IL-4 receptor activation (Fig. 3A). This result was confirmed by an in vitro GST pull-down experiment. Lysates derived from c-Myc-tagged Jak1-expressing 293T cells treated with IL-4 or left untreated were incubated with GST-, GST-SOCS-1-, or GST-SOCS-3bound glutathione-Sepharose beads, and the captured protein complex was resolved by SDS-polyacrylamide gel electrophoresis and subjected to immunoblot analysis using the anti-c-Myc monoclonal antibody. The results show that Jak1 from both IL-4-treated and control cells physically associated with SOCS-1 and SOCS-3 in vitro (Fig. 3B). Taken together, these data suggest that Jak1 is a target of SOCS-1 and SOCS-3. However, because SOCS-1 and SOCS-3 bind to the activated as well as nonactivated Jak1, the above experiments did not address whether SOCS-dependent down-regulation of IL-4 signaling is mediated through Jak1 inhibition.
In order to define the role of Jak1 in SOCS-mediated inhibition of IL-4 signaling, we constructed a chimeric cytokine receptor that contains the extracellular and transmembrane domains from murine EPO receptor and the cytoplasmic domain from human IL-4 receptor ␣-chain. Unlike IL-4 receptors, the EPO receptor forms a homodimer upon ligand binding (Fig.  4A). Therefore, expression of this chimeric receptor in appropriate cells is expected to activate Jak1 upon EPO treatment, and this, in turn, would phosphorylate Stat6. Accordingly, we cotransfected 293T cells with the plasmids encoding chimeric cytokine receptor (EPO receptor-IL-4R␣), Stat6, and SOCS-1 or SOCS-3 and measured Stat6 activation in response to EPO by EMSA. The results show that SOCS-1 and SOCS-3 both inhibited an EPO-dependent chimeric receptor-bound Jak1-mediated activation of Stat6 (Fig. 4B). These data indicate that Jak1 is a target of inhibition by both SOCS-1 and SOCS-3. Therefore, these two SOCS proteins can inhibit the ligand-dependent activation of both type I and type II IL-4 receptors.
Structural Basis of SOCS-3-mediated Inhibition of IL-4 Signaling-To define domains or motifs of the SOCS proteins that confer the selective regulation of IL-4-dependent Jak-Stat signaling, we constructed chimeric SOCS proteins by interchanging domains among SOCS-2, SOCS-3, and CIS (Fig. 5A). In 293T cells, the expression of the chimeric SOCS protein (SOCS-n3/c2), in which the amino-terminal (Pre-SH2) 47 amino acids of murine SOCS-2 were replaced by the amino-terminal (Pre- SH2) 45 amino acids of murine SOCS-3, completely inhibited IL-4-dependent activation of Stat6 (Fig. 5B). In contrast, the chimeric SOCS protein (SOCS-n2/c3), containing the aminoterminal 47 amino acids of SOCS-2 and the SH2 domain (96 amino acids) plus the carboxyl-terminal 84 amino acids of SOCS-3, failed to inhibit Stat6 activation by IL-4; rather, it markedly up-regulated the Stat6 activation (Fig. 5B). These results indicate that the Pre-SH2 domains of SOCS-2 and SOCS-3 confer the specificity of up-and down-regulation of IL-4 signaling, respectively, whereas the contribution of their SH2 domains appears to be redundant in these regulatory functions (Fig. 5B).
CIS was found to interfere with IL-3 signaling by masking the Stat5 docking site on the IL-3R␤ chain following cytokine stimulation of cells (4,5,36). We constructed a SOCS-3/CIS chimeric protein (SOCS-n3/cC) in which the amino-terminal 81 amino acids encoding the Pre-SH2 domain of murine CIS were replaced by the 45-amino acid Pre-SH2 region of SOCS-3 (Fig.  5A). Expression of SOCS-n3/cC in 293T cells completely prevented the IL-4-dependent Stat6 activation (Fig. 5B), which confirms that the Pre-SH2 region of SOCS-3 confers the specificity of inhibition of IL-4 signaling. The expression levels of the chimeric SOCS proteins were comparable in the trans- 293T cells were cotransfected with 2.0 g of Stat6 expression plasmid and 5.0 g of the indicated chimeric SOCS expression plasmid or the empty vector, using the calcium phosphate coprecipitation method. 48 h posttransfection, the cells were treated with 20 ng/ml of IL-4 for 30 min, and WCE was prepared. EMSA was performed as described in Fig. 1A. C, expression of chimeric SOCS proteins in transfected 293T cells. Cells were transfected as described above. 48 h posttransfection, the Flagtagged chimeric SOCS proteins were measured as described in Fig. 1B. IP, immunoprecipitation; IB, immunoblotting. fected 293T cells, as determined by immunoprecipitation and Western blot analysis (Fig. 5C).
Mapping of the Inhibitory Motif in the Pre-SH2 Domain of SOCS-3-To identify the amino acid motifs in the Pre-SH2 region of SOCS-3 that are responsible for conferring the inhibition of IL-4 signaling, we constructed deletion mutants of SOCS-3, which were cotransfected with the Stat6 expression plasmid in 293T cells. Deletion of the amino-terminal 6, 15, and 20 amino acids did not affect the SOCS-3-mediated inhibition of IL-4-dependent Stat6 activation, whereas deletion of the amino-terminal 26 or 32 amino acids of SOCS-3 completely failed to inhibit Stat6 activation by IL-4 (Fig. 6A). Expression of the mutant proteins in transfected 293T cells was at comparable levels as measured by immunoprecipitation and Western blot analysis (Fig. 6B). These results indicate that the SH2proximal 25 amino acids are responsible for conferring the SOCS-3-mediated inhibition of IL-4 signaling. These mutant SOCS proteins did not alter the expression levels of Stat6 (data not shown).
An alignment of the Pre-SH2 domain of murine SOCS-1 containing 79 amino acids with the 45-amino acid Pre-SH2 domain of murine SOCS-3 shows that a two-amino acid motif, Thr-Phe, is conserved between these two proteins; this motif is preceded by a Lys residue at position 23 in SOCS-3 and by an Arg residue at position 57 in SOCS-1 (52) (Fig. 7A). To determine whether the Lys-Thr-Phe (KTF) motif contributes to the inhibitory function of SOCS-3, it was replaced by three alanine residues and the expression of this mutant SOCS-3 in 293T cells failed to inhibit IL-4-dependent Stat6 activation (Fig. 7B). This result suggests that the KTF motif is required for the SOCS-3-mediated inhibition of IL-4 signaling.

Mutation in Either Thr-24 or Phe-25 Completely Prevents SOCS-3-mediated Inhibition of IL-4
Signaling-To identify which individual amino acids in the KTF motif are critical for SOCS-3-mediated inhibition of IL-4 signaling, we introduced individual mutations in this motif. The K23E mutant SOCS-3, in which the positively charged Lys residue at position 23 was replaced by a negatively charged glutamic acid, had a marginal effect on the SOCS-3-mediated inhibition of IL-4-dependent activation of Stat6 (Fig. 7B). The SOCS-3 molecule completely retained its inhibitory function when the Thr-24 was replaced by a serine, another uncharged polar amino acid. However, when Thr-24 was substituted by an aspartic acid, SOCS-3 completely failed to inhibit IL-4 signaling (Fig. 7B). Replacement of Phe at position 25 in SOCS-3 by a glutamic acid resulted in the complete loss of inhibitory function of this protein (Fig. 7B). The expression levels of the mutant SOCS-3 were comparable, as shown in Fig. 7C. Clearly, the amino acids Thr-24 and Phe-25 are critical in conferring the SOCS-3-mediated inhibition of IL-4-activated Jak-Stat signaling pathway, and mutation in either Thr-24 or Phe-25 completely abolishes the inhibitory function of SOCS-3.
The Pre-SH2 Domain of SOCS-3 Alone Confers Inhibition of IL-4 Signaling-To examine whether the Pre-SH2 domain of SOCS-3 inhibits IL-4 signaling independently of the adjacent SH2 plus the SOCS-box domain, we constructed chimeric proteins encoding GFP fused with the wild type or mutant Pre-SH2 domain of SOCS-3 or wild type Pre-SH2 domain of SOCS-2. The results (Fig. 8A) show that GFP fusion of SOCS-3 Pre-SH2 domain inhibited IL-4-dependent activation of Stat6 in 293T cells. To examine the specificity of this inhibition, we have demonstrated that the GFP fusion of SOCS-3 Pre-SH2 domain in which Thr-24 was replaced by an aspartic acid failed to inhibit IL-4-dependent Stat6 activation, but the replacement of Lys-23 by a glutamic acid retained its inhibitory function (Fig. 8A). Interestingly, GFP fusion of the Pre-SH2 domain of SOCS-2 up-regulated the IL-4-mediated activation of Stat6. The expression constructs were tagged with a Flag epitope, and the expression levels of the GFP fusion proteins were comparable, as determined by immunoprecipitation and Western blot analysis using anti-Flag antibodies (Fig. 8B). These data indicate that Pre-SH2 domain of SOCS-3 can function independently of the SH2 or the SOCS-box domain.

DISCUSSION
In this investigation we found that whereas SOCS-1 and SOCS-3 inhibit IL-4-dependent Stat6 activation and IL-4-responsive reporter gene expression, SOCS-2 and CIS up-regulate these processes (Figs. 1 and 2). IL-4-dependent cell signaling is mediated through two types of receptor complexes, type I and type II (21,22). Both receptor types utilize the IL-4R␣ subunit as an essential ligand-binding transmembrane subunit that physically associates with Jak1 in a specific manner (21,22). Using a chimeric receptor encoding the extracellular and transmembrane domains of the EPO receptor and the cytoplas- mic domain of IL-4R␣ that associates with Jak1 and forms a homodimer upon EPO binding, we have demonstrated that Jak1 is a target for inhibition of IL-4 signaling by both SOCS-1 and SOCS-3 (Fig. 4).
To understand the structural basis of the SOCS-mediated differential regulation of IL-4 signaling, we conducted domain swapping experiments (Fig. 5). The amino-terminal Pre-SH2 domains of SOCS family proteins are variable in length and amino acid sequence; however, SOCS-2 and SOCS-3 have amino-terminal domains of similar size (4,18). The exchange of amino-terminal domains between SOCS-2 and SOCS-3 reverses their mode of regulation of IL-4 signaling (Fig. 5). The Pre-SH2 domain of CIS is relatively longer than that of SOCS-2 and SOCS-3 (4, 18), but when this domain was replaced by that of SOCS-3, CIS became an inhibitor of IL-4 signaling (Fig. 5). Therefore, the specificity of regulation lies in the Pre-SH2 domain of these SOCS molecules. It is likely that the SH2 domains of SOCS proteins are involved in recruiting the SOCS molecules to the cognate phosphotyrosine residue on the target proteins.
Two different mechanisms of SOCS-mediated inhibition of cytokine signaling have been attributed to the function of the SH2 domain (4,5). SOCS-1 attenuates gp130-mediated signal transduction both by binding to activated Jaks via the SH2 domain and by inhibiting Jak kinase activity via the pre-SH2 region (39,40). SOCS-3 was also reported to inhibit gp130mediated LIF signaling; however, the mechanisms of this SOCS-3 action remained unclear (40). Mutations in SH2 domains have been shown to abrogate both SOCS-1-and SOCS-3-mediated inhibition of LIF signaling (40). By contrast, CIS appears to bind directly to the phosphorylated receptors, competing with the signaling intermediates, such as Stat5, for receptor binding (4,5,36,53). Therefore, both the Jaks and the cytokine receptors upon tyrosyl phosphorylation could potentially serve as docking targets of the SOCS-SH2 domains.
Overexpression of the Pre-SH2 domain of SOCS-3 inhibited IL-4-dependent Stat6 activation (Fig. 8), raising the possibility that, independent of the SH2 domain, the Pre-SH2 domain of SOCS-3 could bind to the activated Jaks and inhibit their kinase activity or bind to the activated cytokine receptors to compete for the Stat recruitment inhibiting the IL-4 signaling. The SH2-domain of SOCS-3 may be required to serve the docking function for the inhibition of IL-4 signaling when SOCS-3 is induced or present in the cell at physiological concentrations. But clearly, when SOCS-3 protein is overexpressed in cells, its SH2-mediated docking function is dispensable (Fig.  8). Further studies are necessary to precisely define the func-tions of the SOCS SH2-domains in the regulation of the cytokine signaling pathway. To map the inhibitory motif in the Pre-SH2 domain of SOCS-3, we introduced mutations in this region and showed that two adjacent amino acids, Thr-24 and Phe-25, are critical for the SOCS-3-mediated inhibition of IL-4 signaling. The SH2-proximal 25 amino acids are sufficient for the inhibitory function (Fig. 6). Importantly, mutation of Thr-24 and Phe-25 also resulted in the loss of function of the isolated Pre-SH2 domain (Fig. 7). Thus, we have demonstrated that the specificity of SOCS-mediated differential regulation of IL-4 signaling is conferred by the Pre-SH2 domains that function independently of other domains. Experiments are under way to address the mechanisms of action of the SOCS Pre-SH2 domains in the regulation of cytokine-mediated Jak-Stat signaling.