A Gain-of-function Mutation in STAT6*

Interleukin-4 (IL-4) is a cytokine that plays a crucial role in the pathophysiology of asthma and allergic dis-eases. IL-4-induced gene expression is largely mediated through the activation of the latent transcription factor STAT6. We identified a STAT6 mutant (STAT6VT)) that is activated independently of IL-4 stimulation. STAT6VT carries two amino acid changes in the SH2 domain that affect the overall structure and stability of the monomeric and dimeric protein. When overexpressed in mammalian cells, STAT6VT undergoes tyrosine phosphorylation, binds DNA, and activates transcription in the absence of IL-4 stimulation. Using the Jak1- and Jak3-deficient fibroblast line U4A, we demonstrate that phosphorylation is mediated by an IL-4-independent tyrosine kinase that is not able to activate the wild-type STAT6 protein. These results suggest that small changes in STAT6 could result in hyperactivation of the protein and constitutive expression of STAT6-dependent genes. Such a mutation, if found in vivo , could cause genetic predisposition for atopic diseases. IL-4 (10 ng/ml) 6 h before harvesting. Luciferase and b -galactosidase activities were determined 2 days post-transfection using the Promega assay system. Mobility Shift Assay and Immunoprecipitations— Probes correspond-ing to the STAT6 binding site (N4) have been described previously (8). Nuclear extracts were prepared, and mobility shift assays were performed as described (12, 24). Peptide binding studies using the tyrosine- phosphorylated IL-4 receptor peptide, ASSGEEGPY*KPFQDLI, have been described elsewhere (12, 24). The same nuclear extracts were used for immunoprecipitations with an anti-STAT6-specific polyclonal antibody. The precipitated material was electrophoretically separated and subjected to anti-phosphotyrosine Western blots using the 4G10 mono-clonal anitbody (Upstate Biotechnology). Limited Proteolysis— Wild-type STAT6 and STAT6VT were expressed in insect cells and purified to homogeneity as described previ- ously (13). 30 m g of each protein were exposed to 3 m g of activated trypsin (Sigma) protease. The reaction was incubated at 37 °C in 1 3 digestion buffer (20 m M Tris, pH 7.2, 1 m M dithiothreitol). 10 m l aliquots were removed every 20 min for analysis by SDS-polyacrylamide gel electrophoresis. For each reaction, a negative control was performed without protease.

IL-4 1 is a multifunctional cytokine that stimulates changes in many cell types and is involved in the regulation of immune and inflammatory responses (1,2). Among the many physiological responses mediated by IL-4 is its ability to promote the differentiation of T-helper (Th) precursors toward the Th2 lineage while inhibiting Th1 development (3)(4)(5). Furthermore, IL-4 stimulation of B-cells triggers immunoglobulin class switching to the IgE isotype (6). This Ig recombination is thought to be initiated following the transcriptional activation of the germline ⑀ promoter, which leads to the generation of the sterile ⑀ transcript. Numerous studies have shown that the IL-4-induced activation of the ⑀ promoter requires STAT6, a transcription factor activated upon IL-4 stimulation (7)(8)(9)(10)(11).
STAT6 is activated by the Jak/STAT pathway (12). Upon IL-4 binding to its cognate receptor, the latent, monomeric STAT6 protein is recruited to the ␣ chain of the IL-4 receptor, where it specifically interacts with two phosphotyrosine-containing regions (12)(13)(14)(15). This interaction requires the integrity of the STAT6 SH2 domain. Subsequently, STAT6 becomes phosphorylated on tyrosine 641 by Jak1 and/or Jak3, two members of the Jak family of tyrosine kinases (8). The activated protein dimerizes, migrates to the nucleus, binds specific cis-acting elements, and activates transcription of IL-4 responsive genes (16 -18).
One of the best characterized IL-4 responsive elements is located in the germline ⑀ promoter (7,8). B-cells derived from STAT6 knock-out mice do not produce IgE in response to IL-4 treatment, illustrating that STAT6 is absolutely required for this event (19 -21). These observations suggest that STAT6 is an excellent target for therapeutic intervention in cases of allergic rhinitis and allergic asthma, which are often associated with hyper-IgE production.
We recently conducted an extensive alanine scan mutagenesis of the STAT6 SH2 domain (22). These studies revealed amino acids that are critical for dimer formation and subsequent DNA binding. Consistent with the model that the STAT SH2 domain is also required for receptor interaction, the same amino acids were shown to be important for binding a tyrosinephosphorylated peptide derived from the IL-4 receptor signaling chain. Furthermore, STAT6 mutants that were unable to bind DNA were also unable to activate transcription in response to IL-4 stimulation, whereas STAT6 derivatives that could bind DNA were transcriptionally active (22).
In this study we have characterized one of these mutants (STAT6VT) in greater detail. STAT6VT differs from the wildtype in that it carries two amino acid changes in the SH2 domain. These mutations alter the confirmation of the protein such that it becomes phosphorylated, binds DNA, and activates transcription independent of IL-4 stimulation. This hyperactivation of STAT6, if found in vivo, could lead to increased IgE production. Hence, it will be interesting to see whether such a mutation occurs in patients suffering from allergic rhinitis and/or allergic asthma.

MATERIALS AND METHODS
Cell Culture-HEK293 cells were cultured in Dulbecco's modified Eagle's medium/F12 (Meditech, Herndon, VA) supplemented with 10% fetal bovine serum. The human fibrosarcoma U4A cells (a gift from Dr. George Stark) were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum.
Plasmid Construction-The IL-4 inducible luciferase reporter construct carries four copies of the IL-4 response element (C/EBP-N4 site) found in the human germline ⑀ promoter. Preparation of this construct has been described previously (8). The mammalian expression plasmids for Jak1, the wild-type STAT6 protein as well as the mutant STAT6 derivatives, carrying two amino acid changes in the SH2 domain, have been described elsewhere (22). The triple mutant STAT6VT/Y was prepared using polymerase chain reaction to introduce two alanine residues at amino acid positions 547/548 and a phenylalanine at amino acid position 641. The fragment containing the indicated mutations was then used to replace the BglII/SacI fragment of TPU 389 (22). The presence of the mutations was confirmed by DNA sequence analysis. The baculovirus expression plasmids, wild-type STAT6 and STAT6VT, have also been described (8,22).
Transfection-HEK293 or U4A cells were transfected using calcium phosphate precipitation (Promega, Madison, WI). A control plasmid carrying the ␤-galactosidase gene under the cytomegalovirus promoter was co-transfected in each experiment. The cells were stimulated with * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Mobility Shift Assay and Immunoprecipitations-Probes corresponding to the STAT6 binding site (N4) have been described previously (8). Nuclear extracts were prepared, and mobility shift assays were performed as described (12,24). Peptide binding studies using the tyrosinephosphorylated IL-4 receptor peptide, ASSGEEGPY*KPFQDLI, have been described elsewhere (12,24). The same nuclear extracts were used for immunoprecipitations with an anti-STAT6-specific polyclonal antibody. The precipitated material was electrophoretically separated and subjected to anti-phosphotyrosine Western blots using the 4G10 monoclonal anitbody (Upstate Biotechnology).
Limited Proteolysis-Wild-type STAT6 and STAT6VT were expressed in insect cells and purified to homogeneity as described previously (13). 30 g of each protein were exposed to 3 g of activated trypsin (Sigma) protease. The reaction was incubated at 37°C in 1ϫ digestion buffer (20 mM Tris, pH 7.2, 1 mM dithiothreitol). 10 l aliquots were removed every 20 min for analysis by SDS-polyacrylamide gel electrophoresis. For each reaction, a negative control was performed without protease.

Identification of STAT6VT, a Constitutively Active STAT6
Derivative-Previously, we generated a series of STAT6 mutants that differ from the wild type in that they carry two amino acid substitutions in the SH2 domain (22). In the present study, we have compared the transcription potential of these mutants in the presence and absence of cytokine. The proteins were co-expressed in HEK293 cells with an IL-4 inducible luciferase reporter carrying four copies of the STAT6dependent IL-4 response element derived from the human germline ⑀ promoter. HEK293 cells lack endogenous STAT6, but they are able to activate an IL-4 responsive promoter upon ectopic expression of STAT6, suggesting that this is the only component missing in the IL-4/STAT6 signaling cascade (8). As observed previously, all STAT6 derivatives that were known to bind DNA also activated transcription (Fig. 1). With one exception, all proteins were active only in the presence of IL-4 and were not active without cytokine stimulation. Strikingly, the substitution of two residues at positions 547 (V) and 548 (T) in the SH2 domain resulted in a STAT6 mutant that was transcriptionally active even in the absence of IL-4 stimulation (VT, Fig. 1). In addition, this protein was about three times more active than wild-type STAT6 in the induced state. These data suggested that mutations of these two amino acids (VT) dramatically affect the IL-4-dependent regulation of STAT6.
Tyrosine Phosphorylation Is Required for STAT6VT Activity-To explore the mechanism underlying the constitutive activity of STAT6VT, we asked whether the protein is able to dimerize and bind DNA without IL-4 stimulation. STAT6 or STAT6VT were expressed in HEK293 cells, and nuclear extracts were prepared from IL-4-treated or untreated cells and subjected to DNA binding studies using the IL-4 response element. Fig. 2 shows that the DNA binding activity of wildtype STAT6 can be seen only upon IL-4 stimulation (lane 2), but not under unstimulated conditions (lane 1). In contrast, STAT6VT was able to bind the IL-4 response element even in the absence of cytokine (lane 5). Some increase in binding was observed after IL-4 treatment (lane 6). These results show that the double amino acid substitution yields a protein that binds DNA and activates transcription in an IL-4-independent manner.
Phosphorylation of Tyr-641 is critical for STAT6 dimerization and DNA binding (8). To explore whether Tyr-641 phosphorylation is required in the context of the V547A/T548A mutation, we substituted tyrosine 641 with phenylalanine (STAT6VT/Y) and analyzed the protein using DNA binding. Fig. 2A shows that the tyrosine mutation abolished DNA binding activity even in the presence of IL-4 stimulation (lanes 7 and 8). STAT6Y carrying the Tyr-641 mutation in the context of the wild-type was used as control (lanes 3 and 4).
We also investigated the phosphorylation status of these proteins in the absence and presence of IL-4 stimulation using immunoprecipitations followed by anti-phosphotyrosine immunoblotting. Fig. 2B shows that STAT6VT is tyrosine-phosphorylated in the absence of IL-4 treatment (lane 5). Consistent with the DNA binding studies, some increase in tyrosine phosphorylation was seen upon IL-4 stimulation (lane 6). In contrast, no phosphorylation of the wild-type protein was seen in the uninduced state, whereas strong phosphorylation was seen following IL-4 treatment (lanes 1 and 2). Phosphorylation was completely abolished when Tyr-641 was replaced by phenylalanine in the context of the wild-type (lanes 3 and 4) or the mutant protein (lanes 7 and 8).
We also asked whether these proteins were able to activate transcription when expressed in HEK293 cells (Fig. 2C). Both STAT6Y and STAT6VT/Y were unable to activate transcription even after IL-4 stimulation. In contrast, STAT6 and STAT6VT were fully active. These studies clearly show that Tyr-641 is

IL-4-independent Activation of a STAT6 Mutant
absolutely essential for STAT6 function even in the context of the V547A/T548A mutation. STAT6VT is not able to utilize an alternative tyrosine residue for dimerization nor to bind DNA or activate transcription in the absence of tyrosine phosphorylation.
The V547A/T548A Mutation Alters the Confirmation of Monomeric STAT6 and Stabilizes the STAT6:DNA Complex-IL-4 binding to its receptor triggers the phosphorylation of tyrosines in the intracellular domain of the IL-4 receptor ␣ chain, which allow the specific recruitment of STAT6 (8, 12, 13). So far, we have demonstrated that phosphorylation of STAT6VT is re-quired but also that this event is independent of IL-4 binding to its receptor. Hence, one could argue that the gain-of-function phenotype seen with STAT6VT is due to a structural change mediated by the two amino acid substitutions. To address the possibility of conformational change, we compared the stability of STAT6 and STAT6VT by limited proteolysis (Fig. 3A). Recombinant proteins were expressed in insect cells, purified to homogeneity (lane 1), and exposed to trypsin protease for increasing incubation periods (lanes 2-9). The results clearly demonstrate that STAT6VT is more resistant to protease digestion. Even after prolonged exposure to the protease, STAT6VT did not generate the complex digestion pattern obtained for wild-type STAT6. This finding suggests that the two amino acid changes in the SH2 domain alter the conformation of the protein.
Tyrosine-phosphorylated peptides derived from the IL-4 receptor ␣ chain inhibit STAT6 DNA binding because they interact with the SH2 domain, which is required for dimer formation (12,13). The transcriptional activity of STAT6VT is the same in the induced and uninduced state even though its phosphorylation status and DNA binding activity can still be increased upon IL-4 stimulation. Hence, we wished to determine whether the V547A/T548A mutation stabilizes the dimer upon DNA binding. Proteins expressed in HEK293 cells were subjected to DNA binding in the presence or absence of tyrosine-phosphorylated receptor peptide (Fig. 3B). In agreement to previous results (13), the peptide inhibited DNA binding of the wild-type STAT6 protein in a dose-dependent manner (lanes 2-4). Interestingly, the peptide had no effect on DNA binding of the mutant STAT6VT protein (lanes 7-9). Using fluorescence polarization experiments (13) we demonstrated that the monomeric wild-type and mutant proteins bound this peptide with the same affinity (K d ϳ 0.3 M). The failure of IL-4 receptorderived peptides to displace the STAT6VT dimer as it is seen for the wild-type protein strongly suggests that the STAT6VT dimer is more stable. These results help explain the high tran- Proteins were expressed in HEG293 cells, and nuclear extracts prepared from IL-4-treated cells were subjected to DNA binding. As indicated above the lanes, increasing amounts of tyrosine-phosphorylated peptides derived from the IL-4 receptor ␣ chain were included in the binding reactions. scriptional activity in unstimulated cells (Fig. 1).
The Activity of STAT6VT Is Independent of Jak1 and Jak3-The data indicate that receptor ligation is not required for the activation of STAT6VT. However, tyrosine phosphorylation at residue Tyr-641 is absolutely essential. HEK293 cells do not express Jak3, demonstrating that this kinase is not necessary for STAT6 phosphorylation. We next determined whether the function of Jak1, which is associated with the IL-4 receptor ␣ chain, is required for STAT6VT activation. In these studies we characterized the behavior of wild-type STAT6 and STAT6VT in U4A cells, a Jak1-deficient fibroblast cell line (25,26). The parental line 2fTGH served as positive control. First, we compared the transcription activation potentials of STAT6 and STAT6VT in both 2fTGH and U4A cells. Fig. 4A shows that IL-4-induced luciferase activity could be detected in 2fTGH but not in U4A. Overexpression of wild-type STAT6 increased the activity in 2fTGH about 2-fold but had no effect in U4A cells, supporting the notion that Jak1 is absolutely essential for STAT6 function. In contrast, strong luciferase activity was observed in 2fTGH and U4A cells upon overexpression of STAT6VT. Again, transcription activation was seen in the absence of cytokine stimulation. For reasons we do not under-stand, STAT6VT was even more active in U4A cells than in the parental line, 2fTGH. However, the data shows that STAT6VT is constitutively active even in the absence of Jak1 and Jak3.
Next we wished to determine the DNA binding activity of both proteins when expressed in U4A or 2fTGH cells. Both cell lines were transfected with constructs encoding either wildtype STAT6 or STAT6VT. Nuclear extracts were prepared and assayed for their ability to associate with a STAT6-specific DNA binding element (Fig. 4B). In agreement with our activation studies and earlier data, wild-type STAT6 protein could not be activated in U4A cells because of the lack of Jak1 (top panel, lane 4) (26,27). However, the protein was active upon IL-4 stimulation in the parental line 2fTGH (bottom panel, lane  4). Furthermore, when Jak1 was co-expressed, IL-4-induced activation of wild-type STAT6 was reconstituted in U4A cells (top panel, lane 10). In contrast, STAT6VT was active in both the parental line 2fTGH as well as the mutant U4A line independent of whether Jak1 was co-transfected (lanes 6 and 12). As observed in HEK293 cells, activation of STAT6VT did not require cytokine stimulation. Taken together, these findings clearly demonstrate that STAT6VT is phosphorylated by a kinase that does not need to be activated by IL-4 stimulation. DISCUSSION We have identified two amino acids in the STAT6 SH2 domain that, when mutated, uncouple the activity of the STAT6 protein from its regulatory cytokine, IL-4. Strikingly, this mutant, STAT6VT, is tyrosine-phosphorylated and is able to bind DNA and activate transcription in the absence of IL-4. Using limited proteolysis, we show that STAT6VT expressed in insect cells is more resistant to protease digestion than the wild-type protein. This observation does not reflect the difference in tyrosine phosphorylation, because STAT6VT is not differentially phosphorylated in insect cells when compared with the wild-type protein (data not shown). Hence, we argue that the two amino acid changes stabilize STAT6VT so that it is less susceptible to protease digestion. In mammalian cells, this change in conformation seems to give rise to a protein that is phosphorylated by an IL-4-independent tyrosine kinase. These results are supported by the observation that Jak1-and Jak3deficient cells are able to phosphorylate STAT6VT but fail to phosphorylate the wild-type protein. Furthermore, STAT6VT forms a more stable protein:DNA complex than the wild-type protein. These two mechanisms, phosphorylation in the absence of IL-4 and increased DNA binding stability, help to explain the constitutive transcription activity seen with STAT6VT. Obviously, these studies do not rule out the possibility that wild-type STAT6 is also phosphorylated by this IL-4-independent kinase. In this scenario, STAT6 may be rapidly dephosphorylated in the absence of IL-4 stimulation. This phosphatase may be unable to recognize STAT6VT because of the two amino acid changes. Our results do not allow us to discriminate between these two potential mechanisms.
Constitutively active versions of STAT1 and STAT5 have been identified (28,29). In the case of STAT1, it was shown that a mutation of amino acid 31 (Arg to Ala) inhibits the dephosphorylation of STAT1. Furthermore, a deletion of the N terminus resulted in a protein that was constitutively phosphorylated (28). A similar mechanism does not apply for STAT6 because deletion of the N terminus has no effect on the activation/phosphorylation properties of STAT6 (8). Two amino acids have been identified in STAT5 that, when mutated simultaneously, yield a constitutively active STAT5 protein (29). One of these resides is upstream of the DNA binding domain and the other is located in the C-terminal activation domain. Also for STAT6VT, both amino acid changes were required to observed the full constitutive phenotype. Hence, it appears that Nuclear extracts were prepared 48 h post-transfection from IL-4-stimulated or unstimulated cells as indicated above the lanes. Extracts shown in lanes 7-12 were prepared from cells that had also been transfected with an expression construct encoding Jak1. The proteins were subjected to DNA binding using a STAT6-specific DNA probe (18). different mutations can yield STAT proteins that are either phosphorylated in a ligand-independent manner or that render them less susceptible to dephosphorylation.
Inappropriate activation of STAT proteins has been described for some pathological conditions. For example, STAT3 and STAT5 are constitutively active in HTLV-1 transformed T-cells (30). STAT3 has also been shown to be activated in a cytokine-independent manner by Src, resulting in cell transformation (31)(32)(33). Furthermore, constitutively active STAT3 is found in bone marrow mononuclear cells from patients with multiple myelomas and is thought to contribute to pathogenesis by preventing apoptosis (34). Also, the oncogene v-abl can activate the Jak/STAT pathway (35,36). More recently, constitutively active STAT1 has been observed in airway epithelial cells from asthmatic individuals (37). These observations illustrate that the inappropriate activation/tyrosine phosphorylation of a STAT protein, in the absence of cytokine signaling, can have a profound effect on the status of a cell. Our finding that the mutation of two key residues within the STAT6 SH2 domain results in a protein that is constitutively active provides the first reported evidence of a hyperactive form of STAT6.
Asthma is familial, and many genetic loci predispose individuals to the disease (38). Furthermore, asthma as well as allergy are multifactorial in that they are influenced by many genetic and environmental factors. IL-4 is a critical cytokine in the development of atopy, the IgE-mediated syndrome of allergic asthma and rhinitis (39). Gain-of-function polymorphisms in the IL-4 gene as well as the gene encoding the IL-4 receptor ␣ chain are associated with higher serum IgE levels and asthma (40 -46). A gain-of-function mutation in STAT6 could lead to a similar phenotype, because IgE production in response to IL-4 is STAT6-dependent (11, 19 -21, 47) and allergic rhinitis or asthma are characterized by increased IgE levels (23,48,49). Hence, one could speculate that a gain-of-function mutation in STAT6 could predispose patients to allergic asthma and atopic diseases.