Characterization of the Cytoplasmic Domain of Interleukin-13 Receptor-α*

Interleukin (IL)-13 and IL-4 are pleiotropic immunoregulatory cytokines that share many overlapping biological properties reflecting the fact that both can utilize a receptor complex composed of the IL-4 receptor-α (IL-4Rα) chain and the IL-13Rα chain. The cytoplasmic domain of the IL-13Rα is 60 amino acids long and is essential for IL-13-dependent growth. It contains a Pro-rich domain in the membrane-proximal region and two Tyr residues. Here we show that a truncated IL-13Rα, lacking the 38 carboxyl-terminal residues but retaining the Pro-rich region, can support IL-13-dependent proliferation, although with reduced efficiency. A Y402F mutant of the cytoplasmic domain of IL-13Rα supported normal IL-13-induced growth. However, tyrosine phosphorylation of signal transducer and activator of transcription 3 (STAT3), which we show is induced by IL-13 and IL-4 in cells that express the IL-13Rα, was significantly reduced. The cytoplasmic domain of IL-13Rα was constitutively associated with STAT3, Tyk2, and Janus kinase 1 (JAK1). IL-13-induced tyrosine phosphorylation of IL-13Rα in vivo could not be detected using anti-Tyr(P) antibodies. A glutathione S-transferase fusion protein of the cytoplasmic domain of IL-13Rα was phosphorylated on tyrosinein vitro by JAK1, JAK3, and Tyk2, although the tyrosine phosphorylation events mediated by Tyk2 and JAK3 were not detectable using anti-phosphotyrosine antibodies. These data, together with the demonstration that IL-13Rα associates constitutively with Tyk2 and that Tyr-402 is involved in IL-13-induced phosphorylation of STAT3, suggest that the latter is mediated by Tyk2. Tyrosine phosphorylation of STAT3, which was not necessary for IL-13-induced proliferation, may account for some of the effects of IL-4 and IL-13 on the function of their targets.

The overlapping activities of IL-4 and IL-13 reflect the existence of common receptor components, as revealed by receptor cross-competition studies (18,19). The best characterized form of the IL-4 receptor is composed of a 140-kDa transmembrane glycoprotein (IL-4R␣) that binds IL-4 with a K d of 50 -600 pM depending on the cell type (20 -23), and the ␥c chain of the IL-2 receptor which, upon association with the complex of IL-4 and IL-4R␣, results in a 2-3-fold increase in affinity for IL-4 (24,25). Neither IL-4R␣ nor ␥c, alone or together, binds IL-13. Novel receptor subunits that specifically bind IL-13 (IL-13R␣) have been identified in mouse (26) and human (27). IL-13R␣ has an apparent molecular mass of 60 -70 kDa and binds IL-13 with a K d of 2-10 nM when transfected alone into COS cells (26) or 293 fibroblasts (27). However, when hu-IL-13R␣ was cotransfected with hu-IL-4R␣ into 293 cells, or mu-IL-13R␣ was transfected into CTLL-2 cells that express IL-4R␣, the K d values for human or murine IL-13 were 400 and 75 pM, respectively (26,27), suggesting that IL-4R␣ and IL-13R␣ formed a higher affinity complex with IL-13. IL-13R␣ is a transmembrane glycoprotein with a short cytoplasmic domain of 60 amino acids. This region exhibits 95% amino acid identity in the murine and human species (28). It contains a membraneproximal Pro-rich region, analogous to those that occur in other members of the cytokine receptor superfamily and are thought to be involved in activation of JAK kinases. The carboxylterminal region contains two tyrosine residues, Tyr-399 and * This work was supported by The Arthritis Society of Canada. 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.
Tyr-402, the distal one in a YXXQ sequence motif that is a consensus sequence for binding of the SH2 domain of STAT3 (29). Another transmembrane protein that binds IL-13 and has a short cytoplasmic domain has been identified in humans (30) and mice (31).
We have previously shown that a truncated IL-13R␣ encoding the extracellular and transmembrane domains but lacking the cytoplasmic domain failed to stimulate detectable biological or biochemical responses to IL-13 in terms of JAK1, Tyk2, IL-4R␣, IRS-2, and STAT6 phosphorylation and proliferation (15). Here we show that a mutant IL-13R␣ that contains only the membrane-proximal Pro-rich sequence and six carboxylterminal residues supported IL-13-induced cell proliferation, although with a diminished efficiency. Mutation of Tyr-402 of IL-13R␣ did not affect IL-13-induced mitogenesis but reduced the level of tyrosine phosphorylation of STAT3, which we show is induced by IL-13 and IL-4 when they signal through the IL-13R␣. Although we failed to detect IL-13-induced tyrosine phosphorylation of IL-13R␣ in vivo, GST protein fused to the cytoplasmic domain of IL-13R␣ was efficiently phosphorylated on tyrosine residues in vitro by JAK1, JAK3, and Tyk2. However, only Tyk2 and JAK1 associated with the IL-13R␣ in vivo, in a direct interaction not mediated by IL-4R␣. Taken together, the results suggest the existence of complex interactions between IL-13R␣, tyrosine kinases, and STAT3.
Recombinant Plasmids-pEB-13R (26), a plasmid encoding the fulllength IL-13R␣ cloned into a pEF-BOS vector (33), was digested with XbaI generating two fragments of 5.4 and 1.2 kb. The 1.2-kb fragment, encoding the mature protein sequence for IL-13R␣, was purified from 0.8% agarose gels and ligated into pBluescriptIISK, creating the plasmid pBS-13R␣. To generate the plasmid pEB-13⌬386 that encodes the extracellular, transmembrane, and 21 proximal amino acid residues of the cytoplasmic domain of IL-13R␣, a double-stranded oligonucleotide encoding a stop codon at position 386 of the protein sequence was cloned into the cDNA of IL-13R␣. Briefly, oligonucleotides 5Јp-TTAAGATCA-TTATATTTCCTCCAATTCCTGATCCCGGGAAGATTTTTAAAGAAT-AGGTAGC-3Ј and 5ЈpTTAAGCTACCTATTCTTTAAAAATCTTCCCG-GGATCAGGAATTGGAGGAAATATAATGATC-3Ј (where p indicates phosphate) were annealed and cloned into calf intestinal phosphatasetreated AflII-linearized pBS-13R␣. The new plasmid pBS-13⌬386 was digested with XbaI, and the 1.25-kb fragment was subcloned into the expression vector pEF-BOS. The plasmid pEB-13YF that encodes the full-length IL-13R␣ with a mutation of Tyr-402 to Phe was generated by polymerase chain reaction-based site-directed mutagenesis (34). Briefly, the wild-type 0.2-kb AflII-XbaI fragment of pBS-13R␣ was replaced with the 0.2-kb AflII-XbaI fragment of the polymerase chain reaction product generated using the following four primers: GGAGC-AAAACTCCACCTTCTAC, GACATCTTTGAGAAACAATCC, GGATT-GTTTCTCAAAGATGTC, and TGTAATACGACTCACTATAGGGCGA-TT. The new plasmid pBS-13YF carrying the mutated tyrosine codon was digested with XbaI, and the 1.2-kb fragment was subcloned into the expression vector pEF-BOS. The plasmid pGEX-13RCD that encodes the cytoplasmic domain of IL-13R␣ fused to GST was generated as follows: pBS-13R␣ was digested with XbaI and AflII generating three fragments of 3.0, 1.0, and 0.2 kb. The 0.2-kb fragment, encoding the cytoplasmic domain of IL-13R␣, was purified from 1.2% agarose gels, blunt-ended with Klenow polymerase, and cloned into a SmaI site in the polylinker of pGEX-4T-3. The identity of the constructs was confirmed by restriction mapping and sequencing.
Transfections and Screening of Protein Expression-FD-5 and FDCP-1 cells, grown in RPMI containing 10% FCS, and either 2% LCCM or W3, respectively, were washed and resuspended in transfection buffer (25 mM Hepes; 0.75 mM Na 2 HPO 4 ; 140 mM KCl; 5 mM NaCl; 2 mM MgCl 2 ; 0.5% Ficoll 400) at a concentration of 1.3 ϫ 10 7 cells/ml. For each transfection, 1 ϫ 10 7 cells were mixed with 1 g of pPGK/Neo, a plasmid conferring neomycin resistance, alone or together with 10 g of either pEB-13R, pEB-13⌬386, or pEB-13YF cDNA, and were subjected to electroporation using a Bio-Rad gene-pulsar at 960 microfarads and 280 or 300 V for FD-5 or FDCP-1, respectively. In parallel, 10 7 cells were electroporated without DNA, to subsequently monitor neomycin-induced death, or with 1 g of pPGK/Neo alone, to monitor potential transfection-induced phenotypic alterations. After transfection, these groups of cells were cultured in the appropriate media for 48 h and then transferred to selection media in 96 well-plates at 10 4 cells/well. FD-5 cells were selected in RPMI containing 10% FCS, 200 g/ml G418, and 2% LCCM. FCDP-1 cells were selected in RPMI containing 10% FCS, 300 g/ml G418, and 2% W3. Individual colonies of neomycin-resistant cells were cloned and propagated for assays (cell proliferation and protein tyrosine phosphorylation of whole cell lysates). Cells were tested for expression of wild-type (FD-13R or FDCP-13R), partially truncated IL-13R␣ (FD-13⌬386 or FDCP-13⌬386), and mutated IL-13R␣ (FD-13YF or FDCP-13YF) by FACScan analysis using M2, a mAb against FLAG, and goat anti-mouse IgG-FITC. A representative clone from each cell population was recloned by limit dilution and used for further experiments. Expression of the second functional chain of the receptor was tested by FACScan analysis using anti-IL-4R␣ Ab and goat anti-rat IgG-FITC.
Biological Assays-FD-5 and FDCP-1 transfectants were washed three times with Hanks' buffered salts supplemented with 2% (v/v) FCS and plated at 10,000 or 25,000 cells per well, respectively, in 96-well tissue culture plates in the presence of various concentrations of recombinant mu-IL-4 and recombinant mu-IL-13. As positive control, cells were incubated with 1 g/ml synthetic mu-GM-CSF. After 5 days' incubation for FD-5 transfectants and 2 days' incubation for FDCP-1 transfectants, MTT (3,4,5-dimethyltiazole-2,5-diphenyltetrazolium bromide) was added at a concentration of 375 g/ml for 4 h. Cells were lysed in 6% SDS, 20% N,N-dimethylformamide, and the optical density was measured at 550 nm (35). The results were expressed as a percentage of the maximal MTT incorporation observed in cultures stimulated with GM-CSF. Experiments were done in triplicate, and standard errors were less than 10%. Results are representative of several independent experiments.
In Vitro Kinase Assay-FD5-13R cells (3-6 ϫ 10 6 ) were cultured overnight in a suboptimal concentration of LCCM, stimulated with IL-4 or IL-13, and lysed as described previously. Lysates of cells stimulated with IL-4 were immunoprecipitated with JAK1 or JAK3 antisera; lysates of cells stimulated with IL-13 were immunoprecipitated with Tyk2 antisera, as described previously. The beads were washed three times with cold lysis buffer containing decreasing concentrations of detergent, twice more with 100 mM Tris (pH 7.4), 500 mM LiCl 2 , and once with kinase buffer (20 mM Tris (pH 7.4), 10 mM MgCl 2 , 10 mM MnCl 2 ). Immunoprecipitates were incubated with 20 l of kinase buffer containing 10 Ci [␥-32 P]ATP and 1 g of either IL-13R␣ cytoplasmic domain-GST fusion protein or GST protein alone. After 20 min at room temperature, SDS sample buffer was added to stop the reaction. Samples were electrophoresed onto 10% SDS-polyacrylamide gel electrophoresis and transferred to PVDF membranes, and phosphorylation was detected by autoradiography. After 32 P-protein decay, membranes were washed with TBST and proteins immunoblotted with anti-Tyr(P) mAbs, as described previously. A duplicate membrane was treated with 1 M KOH at 55°C for 2 h to dephosphorylate Thr/Ser residues (38).
GST Fusion Protein and Affinity Purification-The IL-13R␣ cytoplasmic domain-GST fusion protein as well as GST protein alone were expressed in Escherichia coli DH5␣ cells under isopropyl-1-thio-␤-Dgalactopyranoside induction. GST fusion proteins were loaded onto glutathione-Sepharose 4B columns and eluted with 20 mM reduced glutathione. Affinity purified proteins were typically 90 -95% homogeneous as determined by SDS-polyacrylamide gel electrophoresis/Coomassie Blue staining and immunoblotting with anti-GST antibodies.

Generation of Transfectants Expressing the Wild-type or Mutant IL-13R␣-
The expression plasmid pEB-13R (Fig. 1A, top) contains the coding sequence for the full-length IL-13R␣ cDNA. The sequence encoding the carboxyl-terminal region of IL-13R␣ was deleted by introduction of a stop codon at position 1159 of the cDNA. The new expression plasmid pEB-13⌬386 (Fig. 1A, middle) encodes the extracellular, transmembrane, and 22 proximal amino acid residues of the cytoplasmic domain including the Pro-rich region. A point mutation at position 1204 of the cDNA was introduced by polymerase chain reaction-mediated site-directed mutagenesis. The plasmid generated in this way, pEB-13YF (Fig. 1A, bottom), encodes a full-length IL-13R␣ in which Tyr-402 has been substituted by Phe.
The mutated and full-length receptor expression plasmids were used to transfect FD-5 and FCDP-1 cells. Both cell lines are unresponsive to IL-13 stimulation, but respond to IL-4, in the case of FD-5 growing continuously and in the case of FDCP-1 exhibiting prolonged survival but not long term growth. Establishment of FD-5 cells expressing the wild-type IL-13R␣ (FD-13R) and propagation conditions were previously described (15). For each mutated form of IL-13R␣ transfected into FD-5 and FDCP-1 cells, neomycin-resistant colonies were individually cloned and analyzed for IL-13R␣ expression by fluorometric studies using anti-FLAG mAb and FITC conjugated to goat anti-mouse antibody. FACS analysis of FD-5 clones showed that 4 out of 6 clones analyzed expressed IL-13⌬386 and 12 out of 12 clones analyzed expressed IL-13YF, all of them exhibiting the same level of receptor expression for each transfected population (results not shown). In the case of FDCP-1 cells, 6 out of 6 and 7 out of 12 clones analyzed expressed IL-13⌬386 and IL-13YF receptor, respectively, with similar levels of expression for each different population. Within each transfected cell type, each clone responded equally in terms of proliferation or survival in response to specific cytokines. Therefore, only one clone of each transfected type was used for further experiments. FD-5 and FDCP-1 cells transfected with pPGK/Neo alone responded normally to IL-4 in terms of proliferation and survival, indicating that expression of this plasmid did not interfere with IL-4 signaling. The transfected receptors were expressed as surface proteins as detected by FACScan analysis. Fig. 1B shows that whereas FD-13R and FD-13YF cells express equivalent amounts of receptor protein, the levels of expression were 10-fold higher in FD-13⌬386 cells. This differential level of expression between wild-type and IL-13⌬386 receptor was also observed in FDCP-1 cells (Fig. 1D). Importantly, endogenous murine IL-4R␣ was expressed at equivalent levels on the surface of parental or transfected FD-5 (Fig. 1C) and FDCP-1 (Fig. 1E) cells, as determined by staining with anti-IL-4R␣ Ab and goat anti-rat IgG-FITC and flow cytometry.
Biological Responses to IL-13-To elucidate whether the Prorich region of IL-13R␣ was sufficient to mediate responses to IL-13, we measured the effect of IL-13 on cellular proliferation and/or survival as determined by MTT incorporation. Fig. 2, A and C, shows that both FD-13⌬386 and FDCP-13⌬386 cells responded to IL-13 in a dose-dependent manner, although 30% less efficiently than cells expressing the full-length receptor, FD-13R and FDCP-13R, respectively. As expected, neither FD-5 ( Fig. 2A) nor FDCP-1 (Fig. 2C) cells responded to IL-13. The ability of FDCP-1, FDCP-13R, and FDCP-13⌬386 cells to respond to IL-4 was very similar (Fig. 2D). In the case of FD-5 cells and transfectants derived from them, the dose-response curves showed similar slopes at high and intermediate concentrations of IL-4, but a higher basal level at low concentrations of IL-4 for FD-13⌬386 cells was observed (Fig. 2B).
To analyze the contribution of Tyr-402 of IL-13R␣ to IL-13induced responses, proliferation was tested in FD-13YF and FDCP-13YF cells. Fig. 2, A and C, showed the same level of response to IL-13 in FD-13YF and FDCP-13YF as compared with their wild-type counterparts. Thus, Tyr-402 did not appear to play a crucial role in IL-13-induced proliferation or increased survival. Moreover, responses to IL-4 were similar in FD-13YF (Fig. 2B) and FDCP-13YF (Fig. 2D) and cells expressing the full-length IL-13R␣, FD-13R, and FDCP-13R.
Phosphorylation of STAT3-We noted that Tyr-402 of the IL-13R␣ is part of a YXXQ sequence motif that has been pre- viously identified as a consensus sequence for binding the SH2 domain of STAT3 binding (29). To determine whether this Tyr residue plays a role in STAT3 activation, FD-5, FD-13R, and FD-13YF cells were stimulated with IL-4, IL-13, or left untreated, and tyrosine phosphorylation of STAT3 was determined. Fig. 3A, shows that STAT3 was phosphorylated on tyrosine residues in response to either IL-4 or IL-13 in FD5-13R cells. In contrast in the parental FD-5 cells, neither IL-13 nor IL-4 induced tyrosine phosphorylation of STAT3, indicating that the IL-13R␣ was required for IL-13 and IL-4 to induce phosphorylation of STAT3. The level of IL-13-induced phosphorylation of STAT3 was greatly reduced in FD-13YF cells. In both, FD-13R and FD-13YF cells, IL-13 induced higher levels of tyrosine phosphorylation of STAT3 than IL-4. The much lower levels of IL-13-induced phosphorylation of STAT3 in FD-13YF cells suggests that STAT3 phosphorylation is greatly enhanced in the presence of Tyr-402. The decreased levels of IL-4-or IL-13-induced tyrosine phosphorylation of STAT3 in FD-13YF cells was specific for STAT3 as IL-4-or IL-13-induced phosphorylation of STAT6 was not reduced in FD-13YF cells (Fig. 3E).
It has been reported that phosphorylation of Ser-727 on STAT3 might potentiate phosphorylation on Tyr-705. Fig. 5C shows that STAT3 was constitutively phosphorylated on Ser-727 in these cells. Constitutive phosphorylation of STAT3 on serine in NJBC cells has been previously reported (39).
Association of STAT3 with IL-13R␣-To investigate the association of STAT3 with IL-13R␣, we used anti-FLAG antibodies to immunoprecipitate IL-13R␣ from FD-13R and FD-13YF cells. The precipitates were immunoblotted with anti-STAT3 antibodies. We observed that STAT3 was constitutively associated with both wild-type and mutated IL-13R␣ (Fig. 3D); however, the interaction with IL-13YF was slightly decreased indicating that even though Tyr-402 plays some role in stabilizing the interaction, it is not the major domain responsible for binding.

IL-13-and IL-4-induced Phosphorylation of Other Proteins-
We have previously shown that FD5-13R cells respond to IL-13 with tyrosine phosphorylation of JAK1 and Tyk2 and to IL-4 with tyrosine phosphorylation of JAK1, JAK3, and Tyk2. Moreover, both IL-4 and IL-13 induce tyrosine phosphorylation of IL-4R␣, STAT6, and IRS-2 (15). In FD-13⌬386 cells stimulated with IL-4, the levels of tyrosine phosphorylation of all the above proteins were reduced (results not shown). Levels of tyrosine phosphorylation of these proteins were even lower in FD-13⌬386 cells stimulated with IL-13. In that these cells grew continuously in IL-4 or IL-13, these data indicate that only a minor level of IL-4/IL-13-induced phosphorylation of intermediate substrates was sufficient to support proliferation of FD-13⌬386 ( Fig. 2A). Neither IL-13 nor IL-4 induced detect-able levels of tyrosine phosphorylation of STAT3 in FD-13⌬386 cells (results not shown).
In Vivo and in Vitro Phosphorylation of the IL-13R␣-To determine whether phosphorylation of IL-13R␣ was induced in response to IL-13 in vivo, FD-13R cells were stimulated and IL-13R␣ was precipitated with anti-FLAG-mAb. Fig. 4A shows that at least two major proteins of Ϯ75 and Ϯ65 kDa were precipitated, consistent with the broad form of IL-13R␣ (65-85 kDa) observed in chemical cross-linking studies (22,23,27). However, tyrosine phosphorylation of these proteins could not be detected by immunoblotting with anti-Tyr(P) mAb 4G10 (Fig. 4B). Furthermore, three other anti-phosphotyrosine Abs failed to detect tyrosine phosphorylation of IL-13R␣ (results not shown).
Phosphorylation of the IL-13R␣ was also tested in an in vitro kinase assay. JAK kinases were precipitated, and their ability to phosphorylate the IL-13R␣ was tested using GST protein fused to the cytoplasmic domain of IL-13R␣ or GST alone as substrates. Fig. 4, C-E (upper), shows that JAK1, JAK3, and Tyk2 can phosphorylate IL-13R␣ in vitro. However, when the membranes were immunoblotted with anti-Tyr(P) mAb, only IL-13R␣ phosphorylated by JAK1 was detected (Fig. 4C, bottom) but not IL-13R␣ phosphorylated in vitro by Jak3 (Fig. 4D, bottom) or Tyk2 (Fig. 4E, bottom). Membranes were subjected to treatment at high pH with KOH to dephosphorylate serine/ threonine residues (38). This treatment resulted in only a marginal decrease in the level of phosphorylation of GST-IL-13R␣ fusion protein, indicating that 32 P labeling occurred mainly on Tyr residues (results not shown). Control experiments showed that the treatment with KOH successfully dephosphorylated other proteins phosphorylated on Ser/Thr residues. JAK2 did not phosphorylate the IL-13R␣ as detected either by in vitro kinase assay or anti-Tyr(P) mAb (results not shown).
Association of JAK Kinases with IL-13R␣-We have previously shown that FD-13R cells responded to IL-13 with tyrosine phosphorylation of JAK1 and Tyk2 but not JAK3 or JAK2 and to IL-4 with tyrosine phosphorylation of JAK1, JAK3, and Tyk2 but not JAK2 (15). To determine whether any of the JAK kinases that phosphorylated the IL-13R␣ in vitro associate with IL-13R␣, we performed coprecipitation experiments. Fig.  5A shows that JAK1 and Tyk2 (Fig. 5B) but not JAK3 (Fig. 5C) were constitutively associated with the IL-13R␣. Nor was the IL-4R␣ constitutively associated with IL-13R␣ (Fig. 5D), demonstrating that the IL-13R␣ interacted directly with JAK1 and Tyk2 and not through IL-4R␣. As expected, stimulation of the cells with IL-13 resulted in phosphorylation of JAK1 (Fig. 5A) and Tyk2 (Fig. 5B) but not JAK3 (Fig. 5C), whereas stimulation with IL-4 resulted in phosphorylation of all three kinases. The association of Tyk2 (but not JAK1) with IL-13R␣ was increased following stimulation with IL-13 or IL-4. JAK3 was not tyrosine-phosphorylated in response to IL-13, but it was in response to IL-4; however, it did not associate with the IL-13R␣ even upon IL-4 stimulation (Fig. 5C), ruling out the possibility of a trimeric (IL-4R␣⅐IL-13R␣⅐␥c) receptor complex. Fig. 5D shows that both IL-4 and IL-13 can induce phosphorylation of the IL-4R␣ as detected by anti-Tyr(P) mAb; however, IL-4R␣ was only precipitated by anti-FLAG mAb after stimulation of cells with IL-13. These results might reflect higher stability of the complex in the presence of IL-13 and also the engagement and sequestration of IL-4R␣ with ␥c upon IL-4 stimulation. As expected, the IL-4R␣ was detected when precipitated and immunoblotted with IL-4R␣ antibodies, independently of the stimulation conditions. The results indicate that JAK1 and Tyk2 bind directly to IL-13R␣ under these experimental conditions. DISCUSSION We have previously shown that the cytoplasmic domain of IL-13R␣ was necessary for biological and biochemical responses to IL-13 (15). Here we have performed a more detailed analysis of structure-function relationships in this relatively short 60 amino acid cytoplasmic domain. We show that the 22 membrane-proximal amino acids are sufficient to support IL-13-dependent proliferation albeit with reduced efficiency. We also show that Tyr-402, while not required for proliferative responses, did contribute to IL-13-induced tyrosine phosphorylation of STAT3. The cytoplasmic domain of IL-13R␣ was shown to be constitutively associated with Tyk2, JAK1, and STAT3 and was phosphorylated in vitro by Tyk2, JAK1, and JAK3.
Here we formally demonstrate that association of IL-13R␣ and IL-4R␣ was not constitutive and occurred only after ligand binding (Fig. 5). Our demonstration that the IL-13R␣ coprecipitated with JAK1 and Tyk2 from unstimulated cells thus suggests that IL-13R␣ associates directly with these kinases and not through IL-4R␣. It has been reported that IL-4 and IL-13 can bind the trimeric complex of IL-4R␣⅐IL-13R␣⅐␥c (49,50). Here we show that this is unlikely since JAK3, which is constitutively associated with ␥c (51), did not coprecipitate with IL-13R␣ under any conditions.
The truncated IL-13R⌬386 retained the membrane-proximal Pro-rich region but lacked Tyr or Ser residues. The fact that cells expressing this mutant (FD-13⌬386 and FDCP-13⌬386) responded to IL-13 with 30% lower efficiency than cells expressing the wild-type receptor (Fig. 2) correlated with observations that IL-13-induced tyrosine phosphorylation of JAK1, Tyk2, STAT3, STAT6, and IRS-2 was greatly diminished in FD-13⌬386 cells in comparison to FD-13R. These data are consistent with our results 2 showing that receptors able to mediate stimulation of only minor levels of phosphorylation nevertheless support ligand-dependent growth. Our results indicate that the Pro-rich region and 6 downstream amino acids, which based on experiments with other receptors of the family are likely to govern interactions with JAK kinases (40), are sufficient to support cell proliferation. The fact that the truncated receptor signaled with reduced efficiency could indicate that the carboxyl-terminal region of IL-13R␣ mediates interactions with additional signaling molecules that enhance proliferation. Alternatively, the carboxyl-terminal 38 amino acids could be required for proper folding of the cytoplasmic domain and thus for appropriate presentation of the Pro-rich region or stabilization of the receptor complex. Sequences downstream of the Pro-rich region in the ligand-binding chain of other heterodimeric cytokine receptors are critical for function (41)(42)(43)(44)(45).
The IL-13R␣ has two tyrosine residues in its cytoplasmic domain, one (amino acids 402-405) we noted is in a consensus  (13) or GST alone (G) as alternative substrates. Phosphorylated samples were electrophoresed, transferred to PVDF membranes, and detected by autoradiography. After 32 P decay, the membranes were immunoblotted with anti-Tyr(P) 4G10 mAb. The arrows denote the position of phosphorylated IL-13R␣ cytoplasmic domain-GST fusion protein (13-GST) and the putative position of GST alone (GST).

FIG. 5. Tyrosine phosphorylation of JAK kinases and IL-4R␣
and association with IL-13R␣. FD5-13R cells were stimulated with either synthetic IL-4 (4), IL-13 (13), or left untreated (Ϫ). Cells were lysed and immunoprecipitated (I.P.) with specific antibodies, and membranes were immunoblotted (I.B.). Lysates of the equivalent of 5 ϫ 10 7 cells were immunoprecipitated with either anti-JAK1 Ab or anti-FLAG mAb and immunoblotted with either anti-Tyr(P) mAb or anti-JAK1 Ab (A). Lysates of the equivalent of 5 ϫ 10 7 cells were immunoprecipitated with either anti-Tyk2 Ab or anti-FLAG mAb and immunoblotted with either anti-Tyr(P) mAb or anti-Tyk2 Ab (B). Lysates of the equivalent of 3 ϫ 10 7 cells were immunoprecipitated with either anti-JAK3 Ab or anti-FLAG mAb and immunoblotted with either anti-Tyr(P) mAb or anti-FLAG mAb (C). Lysates of the equivalent of 5 ϫ 10 7 cells were immunoprecipitated with either anti-IL-4R␣ Ab or anti-FLAG mAb and immunoblotted with either anti-Tyr(P) mAb or anti-IL-4R␣ Ab (D). Membranes probed with anti-Tyr(P) mAb were stripped and reprobed with the specific immunoprecipitating Ab. YXXQ, which when phosphorylated is recognized by the SH2 domain of STAT3 (29). We have shown that both IL-13 and IL-4 induced tyrosine phosphorylation of STAT3 in FD-13R cells (Fig. 3). IL-4 was unable to induce phosphorylation of STAT3 in parental FD-5 cells indicating that the IL-13R␣ chain was necessary for IL-4-induced activation of STAT3. Replacement of Tyr-402 with Phe resulted in decreased levels of IL-4-or IL-13-induced phosphorylation of STAT3, suggesting that Tyr-402 was phosphorylated in response to IL-4 and IL-13 and was important in activation of STAT3 by these cytokines. This decrease in IL-13-or IL-4-induced tyrosine phosphorylation of STAT3 mediated by the Y402F mutant of the IL-13R␣ was specific for STAT3, as it was accompanied by normal levels of IL-4-or IL-13-induced tyrosine phosphorylation of STAT6. We also observed that IL-13R␣ was constitutively associated with STAT3. This was specific as STAT6 did not associate with the IL-13R␣. In FD-13YF cells, association of STAT3 with the IL-13R␣ was only slightly diminished, indicating that the constitutive interaction of STAT3 with the IL-13R␣ did not require interaction of its SH2 domain with phosphorylated Tyr-402. Collectively, our data suggest a model where STAT3 is constitutively associated with the IL-13R␣ through a non-SH2 domain interaction. When cells are stimulated with IL-13 or IL-4, a kinase is activated that phosphorylates IL-13R␣ on Tyr-402, allowing the SH2 domain of STAT3 to interact with the phosphorylated Tyr-402. STAT3 could be phosphorylated on tyrosine residues by the same or another kinase. An alternative model proposes that the constitutive interaction of STATs with receptors occurs indirectly through JAK proteins (46 -47). However, we saw no evidence that STAT3 coprecipitated with Tyk2 or JAK1 (results no shown), and thus we favor the notion that IL-13 R␣ interacts directly with STAT3.
The kinase involved in phosphorylation of Tyr-402 of the IL-13R␣ is probably Tyk2, as Tyk2 was associated with the IL-13R␣ in vivo (Fig. 5) and Tyk2 could phosphorylate the IL-13R␣ in vitro (Fig. 4). Our failure to detect tyrosine phosphorylation of IL-13R␣ in IL-13-stimulated cells using antiphosphotyrosine antibodies is consistent with the failure of a series of anti-Tyr(P) antibodies to detect tyrosine phosphorylation of IL-13R␣ by Tyk2 in vitro (Fig. 4). In contrast, when IL-13R␣ was phosphorylated in vitro by JAK1, the resultant phosphotyrosine was readily detected by anti-Tyr(P) antibodies (Fig. 4). Certainly our results demonstrating that the Y402F mutant is defective in mediated IL-13-induced phosphorylation of STAT3 suggest that Tyr-402 is phosphorylated in vivo in response to IL-13. We exclude JAK3 because it is not phosphorylated in response to IL-13 in vivo (Fig. 5).
We do not know the functional relevance of phosphorylation of STAT3 in response to IL-4 and IL-13. Certainly, evidence that the Y402F mutation of the IL-13R␣ failed to affect IL-13induced proliferation (Fig. 2) suggests that phosphorylation of STAT3 is not important in growth. We speculate that activation of STAT3 may account for other activities of IL-13 and IL-4 such as inhibition of the growth in carcinoma cells (52) or of the production of proinflammatory cytokines such as IL-12 or tumor necrosis factor-␣ by macrophages, which is only partially dependent upon STAT6 (48).