Characterization of the Interaction between Interleukin-13 and Interleukin-13 Receptors*

Interleukin-13 (IL-13) possesses two types of receptor: the heterodimer, composed of the IL-13R (cid:1) 1 chain (IL-13R (cid:1) 1) and the IL-4R (cid:1) chain (IL-4R (cid:1) ), transducing the IL-13 signals; and the IL-13R (cid:1) 2 chain (IL-13R (cid:1) 2), acting as a nonsignaling “decoy” receptor. Extracellular portions of both IL-13R (cid:1) 1 and IL-13R (cid:1) 2 are composed of three fibronectin type III domains, D1, D2, and D3, of which the last two comprise the cytokine receptor homology modules (CRHs), a common structure of the class I cytokine receptor superfamily. Thus far, there has been no information about STAT6-responsible luciferase Renilla luciferase reporter internal con- trol. 24 h after transfection, HEK 293T cells were detached by phos-phate-buffered saline containing 5 m M EDTA, washed twice with phos- phate-buffered saline, and reseeded. Adhered HEK 293T cells were stimulated with the indicated concentrations of IL-4 or IL-13 for 16 h, followed by washing once by phosphate-buffered saline, and total cell lysates were applied to the Picagene Dual Luciferase assay kit. Western Blotting and Immunoprecipitation— The procedures of Western blotting and immunoprecipitation were performed as de- scribed previously (37). DND-39 cells stably transfected with various kinds of IL-13R (cid:1) 1 were stimulated with 10 ng/ml IL-4 or IL-13 at 37 °C for 10 min. The solubilized lysates or the immunoprecipitates from the lysates with anti-TYK2 Ab (Santa Cruz Biotechnology, Santa Cruz, CA) were subjected to SDS-PAGE and transferred to polyvinylidene fluoride membrane. Membranes were incubated with either anti-FLAG Ab (Sigma), anti-HA Ab (Sigma), anti-phosphotyrosyl STAT6 Ab (Cell Signaling Inc., MA), anti-STAT6 Ab (Santa Cruz Biotechnology), anti-phosphotyrosine Ab (4G10, Upstate Placid, NY) or anti-TYK2 Ab, followed by incubation with second- ary Abs conjugated to horseradish peroxidase. The signals were visu-alized with an enhanced chemiluminescence system (ECL, Amersham Biosciences) and LAS-1000PLUS (Fuji Film,

The extracellular domains of all members of the class I cytokine receptor superfamily, including IL-13R␣1 and IL-13R␣2, contain the cytokine receptor homology module (CRH), composed of two fibronectin type III (FnIII) domains (14). Each domain consists of ϳ100 amino acid residues, generating a ␤-sandwich structure where seven ␤-strands are arranged in the Greek key topology analogous to an immunoglobulin-constant domain. Four positionally conserved cysteine residues in the first FnIII domain form two disulfide bonds, and a WSXWS sequence locates in the FЈ-GЈ loop in the second FnIII domain, both of which are critical for the receptors to position correctly and bind to ligands (14). Crystal structural analyses of the ligand-receptor complexes of growth hormone (GH), erythropoietin, IL-4, IL-6, IL-12, and the granulocyte colony-stimulating factor (G-CSF) have demonstrated that several loops of the CRHs of these receptors provide binding interfaces composed of hydrophobic and polar amino acids (15)(16)(17)(18)(19)(20).
The extracellular domains of a subgroup of the class I cytokine receptor superfamily (gp130, G-CSFR, the granulocyte/ macrophage colony-stimulating factor receptor ␣ chain, the leukemia inhibitory factor receptor (LIFR), IL-3R␣, and IL-5R␣) possess an extra FnIII domain in addition to the CRH. Although even in these receptors, it has been thought that ligand binding is driven principally by the CRH structure, several lines of evidence have shown that the extra FnIII domains of these receptors are also important for the ligand binding. Involvement of the extra FnIII domains of gp130, G-CSFR, LIFR, and IL-5R␣ in binding to the ligands has been verified by biochemical analyses (21)(22)(23)(24)(25)(26)(27). Furthermore, the x-ray structure of the IL-6⅐IL-6R␣⅐gp130 complex shows that the extra FnIII domain of gp130 has a bridging function, interacting with the binding epitope on IL-6 in the opposite trimer, generating the hexametric receptor complex (18,28).
In this study, we first built the homology modeling of the IL-13⅐hIL-13 receptor complexes and then identified critical residues in the CRHs of hIL-13R␣1 or hIL-13R␣2 by mutagenesis analyses, based on these models. Furthermore, we analyzed the roles of the D1 domains in IL-13R␣1 and IL-13R␣2 in their expression and binding to IL-13 and IL-4.
The plasmids were transfected into DND-39 cells by electroporation. Stable transfected cells were maintained with the culture medium containing 250 g/ml hygromycin B (Wako, Osaka, Japan) for the hIL-13R␣1 mutants or 1.25 mg/ml G418 (Sigma) for the hIL-13R␣2 mutants, respectively. Expression of the receptors was confirmed by flow cytometry (FACSCalibre, BD Biosciences) using anti-FLAG antibodies (Abs, Sigma) for the IL-13R␣1 mutants and anti-HA Abs (Sigma) for the IL-13R␣2 mutants. Transient transfection of the plasmids into HEK 293T cells was performed by Lipofectamine 2000 (Invitrogen), according to the manufacturer's protocol.
Sequence Alignment and Homology Model Building-Alignment of the amino acid sequences of hIL-13R␣2 and hIL-13R␣1 without signal peptides (1 to 26, and 1 to 22, respectively) and prediction of FnIII domains (D1, D2, and D3), the transmembrane regions, and several ␤-strand sequences of the D2 and D3 domains of both receptors were executed through the MyHits data base (34). The sequence alignments of extracellular portions of human prolactin (PRL) receptor-hIL-13R␣2 and human gp130-hIL-13R␣1 were performed by FUGUE (35) with slight manual modification based on experimental results.
The complex structure of IL-13 and the D2 and D3 domains of hIL-13R␣2 was modeled based on the crystal structure of the GH⅐PRLR complex (PDB code 1BP3). The hIL-13R␣1 was modeled from the crystal structure of the viral IL-6⅐gp130 complex (PDB code 1I1R), and the resulting structure was overlaid onto the IL-13⅐hIL-13R␣2 complex structure model. Homology modelings of hIL-13R␣2 and hIL-13R␣1 2 E. Honjo, unpublished data. were generated with MODELLER6v2 (36) followed by energy minimization and simulated annealing with Amber. The figures were drawn with DS Viewer Pro 5.0 (Accerlys, CA). 125 I-IL-13 Binding Assay-Iodination of IL-13 and the binding assay were performed as described previously (33). In brief, 1 g of IL-13 and 1 mCi of Na 125 I in 50 l of phosphate-buffered saline were incubated in a microcentrifuge tube coated with 2 g of IODO-GEN (Pierce) at 4°C for 10 min. Radiolabeled IL-13 was purified by a PD-10 column (Amersham Biosciences) and stored at 4°C for up to 2 weeks. The concentration of 125 I-IL-13 was quantified by self-displacement of IL-13R␣2expressing DND-39 cells. After cells were incubated with various concentrations of 125 I-IL-13 at 4°C for 2 h, bound and free ligands were separated by centrifugation through an oil gradient, and their radioactivity was measured. Nonspecific binding was determined by counts in the presence of 100-fold unlabeled IL-13. The dissociation constants (K d ) were determined from specific binding data by using the program GraphPad Prism Version 4 (GraphPad Software, Inc., San Diego).
Luciferase Assay-The luciferase assay was performed as described previously (33). DND-39/G⑀ cells transfected with the mutated hIL-13R␣1 were stimulated with the indicated concentrations of IL-4 or IL-13 for 24 h, and then total cell lysates were applied to the Picagene Dual Luciferase assay kit (Toyo, Inc., Tokyo, Japan). HEK 293T cells were cotransfected with various types of pIRES1hyg-FLAG-hIL-13R␣1, pME18S-STAT6, and pGL3-N4x8 (kind gifts of Dr. T. Sugahara, Asahi Kasei Pharma Corp., Fuji, Japan). pGL3-N4x8 has the eight tandemlined STAT6-responsible element (TTCNNNNGAA), as a promoter sequence of the firefly luciferase gene. The Renilla luciferase reporter gene (phRL-Tk; Promega, Madison, WI) was used as an internal control. 24 h after transfection, HEK 293T cells were detached by phosphate-buffered saline containing 5 mM EDTA, washed twice with phosphate-buffered saline, and reseeded. Adhered HEK 293T cells were stimulated with the indicated concentrations of IL-4 or IL-13 for 16 h, followed by washing once by phosphate-buffered saline, and total cell lysates were applied to the Picagene Dual Luciferase assay kit.
Western Blotting and Immunoprecipitation-The procedures of Western blotting and immunoprecipitation were performed as described previously (37). DND-39 cells stably transfected with various kinds of IL-13R␣1 were stimulated with 10 ng/ml IL-4 or IL-13 at 37°C for 10 min. The solubilized lysates or the immunoprecipitates from the lysates with anti-TYK2 Ab (Santa Cruz Biotechnology, Santa Cruz, CA) were subjected to SDS-PAGE and transferred to polyvinylidene fluoride membrane. Membranes were incubated with either anti-FLAG Ab (Sigma), anti-HA Ab (Sigma), anti-phosphotyrosyl STAT6 Ab (Cell Signaling Technology, Inc., Beverly, MA), anti-STAT6 Ab (Santa Cruz Biotechnology), anti-phosphotyrosine Ab (4G10, Upstate Biotechnology, Lake Placid, NY) or anti-TYK2 Ab, followed by incubation with secondary Abs conjugated to horseradish peroxidase. The signals were visualized with an enhanced chemiluminescence system (ECL, Amersham Biosciences) and LAS-1000PLUS (Fuji Film, Tokyo, Japan).

RESULTS
Homology Modeling of the IL-13⅐hIL-13R␣2 Complex-In IL-13R, IL-13 binds to IL-13R␣1 primarily, and this complex is further stabilized by recruiting IL-4R␣, forming at least a trimer structure of ligand-receptor complex (4,6). In contrast, IL-13R␣2 is likely to form a 1:1 complex with IL-13, which is easy to be modeled (11,38). Therefore, we first built the homology modeling of the IL-13⅐hIL-13R␣2 complex. Because no structural information was available for the IL-13 receptors, the model was built based on the x-ray structure of the GH⅐PRLR 1:1 complex (PDB code 1BP3) as a template. Because PRLR does not contain an extra FnIII domain, we built the model as for the D2 and D3 domains of hIL-13R␣2, predicting that amino acids within the CRH of hIL-13R␣2 would interact with IL-13 ( Fig. 2A). In this model, it was assumed that the E-F loop, the BЈ-CЈ loop, and the FЈ-GЈ loop of hIL-13R␣2 are composed of Leu-202 to Lys-208, Ser-259 to Arg-268, Asn-313 to Lys-328 amino acids, respectively (Fig. 1B). Furthermore, it was predicted that Tyr-207 in the E-F loop and Tyr-315 and Asp-318 in the FЈ-GЈ loop of hIL-13R␣2 would be exposed to the binding interface (Figs. 1A and 2A). It is of note that Tyr-207 in the E-F loop of hIL-13R␣2 corresponds to Trp-72 of PRLR, Trp-104 of GHR, Phe-169 of gp130, and Phe-93 of the erythropoietin receptor, known as a "hot spot" of the ligand binding (Refs. 15, 16, 18, 39, and Fig. 1B).
Identification of Critical Residues of the D2 and D3 Domains of hIL-13R␣2 for Binding to IL-13-To address the importance of the amino acids in the D2 and D3 domains of hIL-13R␣2, predicted based on the homology modeling for binding of IL-13, we generated several kinds of hIL-13R␣2 mutated in these amino acids and analyzed their binding affinity for IL-13. We first analyzed the mutated hIL-13R␣2 in which Tyr-207, Asp-271, Tyr-315, and Asp-318 were replaced with Ala (4A-mut, Fig. 3A). In the modeling, Asp-271 on the strand CЈ was assumed to locate close to Asp-318 ( Fig. 2A). When 4A-mut was transfected into DND-39 cells, although the expression level of the receptor on the cell surface was invariable with that of the wild type of hIL-13R␣2 (Fig. 3B), no binding activity was detected (Fig. 3, A and C), demonstrating the critical roles of these four amino acids in binding to IL-13. To delineate the contribution of each of these four residues in binding to IL-13, we next analyzed the mutants in which Tyr-207, Asp-271, Tyr-315, or Asp-318 was replaced with Ala (Fig. 3A). When any of these four amino acids was mutated with Ala, the affinity of the mutated hIL-13R␣2 was significantly decreased. Expression of these four single mutated hIL-13R␣2 on the cell surface was invariable with the wild type (Fig. 3B). When Asp-318 was replaced with Ala together with Asp-271, the affinity was lowered more than in the single exchange of Asp-271 or Asp-318 (Fig. 3A). The replacement of amino acids that were either adjacent or close to these four critical amino acids (Asp-206, Ala-267, Arg-268, Ser-317, and Asp-319) affected the affinity with IL-13 (Fig. 3A). Double mutations of Glu-289 and Glu-291 with Ala, but not with Gln, slightly decreased the affinity, indicating that a hydrogen bond may be formed between Glu-289 and/or Glu-291 and IL-13. These results coincided with the molecular model in that the E-F loop and the FЈ-GЈ loop of hIL-13R␣2 generates the main binding surface to IL-13 and that the critical amino acids in these loops (Tyr-207, Asp-271, Tyr-315, and Asp-318) are involved in comprising the binding interface at focal contacts. Particularly, it turned out that Tyr-207 is a hot spot for the ligand binding as well as Trp-72 of generated a truncated type of hIL-13R␣2 lacking its D1 domain (hIL-13R␣2⌬D1, Fig. 4A). When hIL-13R␣2⌬D1 was transfected in HEK 293T cells, its expression was not detected by either Western blotting or flow cytometry analysis (Fig. 4, B  and C). These results showed that the D1 domain of hIL-13R␣2 is critical for its expression.
Homology Modeling of the IL-13⅐hIL-13R␣1 Complex-Because the extracellular portions of IL-13R␣1 and IL-13R␣2 are homologous ( Fig. 1A; ϳ33%) and several common amino acids in ␣-helix D of IL-13 are important for binding to both receptors (30), it is reasoned that amino acid residues involved in IL-13 binding are topologically conserved between these two receptors. We built a homology model of the IL-13⅐hIL-13R␣1 complex using the x-ray structures of the viral IL-6⅐gp130 complex (PDB code 1I1R) and the IL-13⅐hIL-13R␣2 modeling as templates (Fig. 2B). In this model, it was assumed that the E-F loop, the BЈ-CЈ loop, and the FЈ-GЈ loop are composed of Thr-190 to Glu-198, Glu-248 to Arg-256, and Asn-317 to Glu-333 amino acids, respectively (Fig. 1C). Furthermore, it was predicted that Phe-197 in the E-F loop would correspond to Tyr-207 of hIL-13R␣2 and that Leu-319 and Asp-323 in the FЈ-GЈ loop would correspond to Tyr-315 and Asp-318 of hIL-13R␣2, respectively (Fig. 1A). The E-F loop of hIL-13R␣1 was two residues longer than that of hIL-13R␣2, and Ser-195 to Ser-196 was assumed to be the unique sequence to hIL-13R␣1 because no amino acid in hIL-13R␣2 corresponds to these amino acids (Fig. 1A).
Identification of Critical Residues of the D2 and D3 Domains of hIL-13R␣1 for Binding to IL-13-We explored whether the E-F loop and the FЈ-GЈ loop of hIL-13R␣1 generate the main binding interface to IL-13 as well as hIL-13R␣2 and whether the amino acids in those loops of hIL-13R␣1 corresponding to the critical amino acids in hIL-13R␣2 contribute to the binding. To address this possibility, we generated several kinds of mutated hIL-13R␣1 and analyzed their binding affinity for IL-13. Because DND-39/G⑀ cells express endogenous hIL-4R␣, but not hIL-13R␣1, the mutated hIL-13R␣1 transfected on the cells comprises IL-13R/type II IL-4R together with endogenous hIL-4R␣ (1). We first analyzed the involvement of the E-F loop of hIL-13R␣1 for binding to IL-13. When both Ser-195 and Ser-196, unique amino acids in hIL-13R␣1, were exchanged with Ala, the affinity was decreased, although the deleted mutant of both amino acids showed only a slight decrease (Fig. 5A). Mutations of Phe-197 corresponding to Tyr-207 in hIL-13R␣2 or adjacent Glu-198 did not show any difference. We next analyzed the involvement of the FЈ-GЈ loop. When either Leu-319 or Asp-323 corresponding to Tyr-315 and Asp-318 in hIL-13R␣2 was replaced with Ala, the affinity of the mutated hIL-13R␣1 to IL-13 was dramatically decreased in L319A, but there was no change in D323A (Fig. 5, A and C). Replacement of Asp-324 adjacent to Asp-323 showed only a slight decrease of the affinity. In contrast, when Tyr-321 was exchanged with Ala, the affinity was decreased significantly. Replacement of Glu-322 also attenuated the affinity, but less than Leu-319 or Tyr-321. Expression of these mutated types of hIL-13R␣1 was invariable with the wild type ( Fig. 5B and data not shown). Substitution at Phe-259 corresponded to Asp-271 in hIL-13R␣2 on the strand CЈ and did not influence the affinity.
We next tested whether the lowered affinities of the mutated hIL-13R␣1 would lead to reduction in the IL-13 signal. DND-39/G⑀ cells expressing endogenous hIL-4R␣ and the transfected hIL-13R␣1 are able to transduce the IL-13 signal by engagement of the ligand, augmenting expression of the reporter gene. When we performed a reporter gene assay using DND-39/G⑀ cells expressing all kinds of the mutated hIL-13R␣1 investigated for the binding assay, only the mutant types of Leu-319 and Tyr-321 significantly impaired the IL-13 response (Fig. 5D and data not shown). The double mutation type at Ser-195 and Ser-196 and the single mutation types at Glu-322 or Asp-324 showed normal IL-13 responses. The responses of all transfectants to IL-4 were invariable. To identify the amino acid residues critical for binding to hIL-13R␣1, we generated five more mutants and analyzed their IL-13 responses in HEK 293T cells. In this experiment, we stimulated HEK 293T cells with 0.01 ng/ml IL-13 in which expression of the reporter gene through transfected hIL-13R␣1, but not endogenous hIL-13R␣1, was detected (Fig. 5D). We confirmed that the responses to IL-13 could be detected as the same as in the system using DND-39/G⑀ cells and that again L319A and Y321A showed lower activities. However, none of the additionally investigated mutants, single mutations of or Asp-194 in E-F loop and Lys-318 or Lys-325 in FЈ-GЈ loop with Ala, changed the IL-13 responses. These results suggested that both the E-F loop and more dominantly the FЈ-GЈ loop contribute to binding to IL-13 in hIL-13R␣1 as well as hIL-13R␣2 and that particularly, Leu-319 and Tyr-321 in the FЈ-GЈ loop are critical residues for the binding to transduce the IL-13 signal. Leu-319 in hIL-13R␣1 corresponding to Tyr-315 in hIL-13R␣2 is a positionally conserved hydrophobic residue for binding to IL-13.
Critical Role of the D1 Domain of hIL-13R␣1 in Its Binding to IL-13 but Not to IL-4 -We next analyzed the functional role of the D1 domain of hIL-13R␣1, which could not be modeled because of a lack of information about homologous structure as well as hIL-13R␣2. For this purpose, we generated a truncated type of hIL-13R␣1 lacking its D1 domain (hIL-13R␣1⌬D1; Fig.  6A). When hIL-13R␣1⌬D1 was transfected in HEK 293T cells, its expression was detected by both Western blotting and flow cytometry analysis at the same level as the wild type in contrast to hIL-13R␣2⌬D1, confirming the ability of this mutated type to be expressed on the cell surface (Fig. 6, B and C).
We next analyzed the involvement of the D1 domain of hIL-13R␣1 in binding to IL-13. hIL-13R␣1⌬D1 completely lost the binding affinity to IL-13, although it was expressed on the cell surface at the same level as the wild type (Fig. 7, A and B). In concordance with the results of the binding assay, hIL-13R␣1⌬D1 failed to induce the transcription of the reporter gene and activate both STAT6 and TYK2 by engagement of IL-13 (Fig. 7, C and D). In contrast, expression of hIL-13R␣1⌬D1 did not prevent the reporter gene activity or STAT6 activation by IL-4 ( Fig. 7, C and D). IL-4 could activate STAT6 through either type I IL-4R composed of IL-4R␣ and the common ␥ chain (␥c) or type II IL-4R composed of IL-4R␣ and IL-13R␣1. It would be possible that even though the type II IL-4R composed of IL-4R␣ and hIL-13R␣1⌬D1 was nonfunctional, IL-4 could activate STAT6 through the type I IL-4R. However, activation of TYK2, a specific signaling event of type II IL-4R/IL-13R, was detected, when hIL-13R␣1⌬D1-expressed cells were stimulated with IL-4 (Fig. 7D), indicating that the type II IL-4R composed of hIL-4R␣ and hIL-13R␣1⌬D1 was functional for the IL-4 binding and its signaling. These results suggested that the D1 domain of hIL-13R␣1 is critical for binding to IL-13, but not to IL-4.

DISCUSSION
In this article, we identified for the first time critical residues in the CRHs of hIL-13R␣1 and hIL-13R␣2 in binding to IL-13 by the mutagenesis approach based on homology modeling of the IL-13⅐hIL-13 receptor complexes. In our findings, Tyr-207, Asp-271, Tyr-315, and Asp-318 in the CRH of hIL-13R␣2, and Leu-319 and Tyr-321 in the CRH of hIL-13R␣1 are critical residues for binding to . Leu-319 in hIL-13R␣1 and Tyr-315 in hIL-13R␣2 are positionally conserved hydrophobic amino acids. Tyr-207, Tyr-315, and Asp-318 in hIL-13R␣2, and Leu-319 in hIL-13R␣1 are conserved in all known species, whereas Tyr-321 in hIL-13R␣1 is conserved in porcine and canine IL-13R␣1 but is replaced by Phe in rat and mouse IL-13R␣1. It has been assumed that the binding site on IL-13 to IL-13R␣1 is ␣-helix A and ␣-helix D because the ␣-helix A and ␣-helix D of IL-4 interact with either ␥c or IL-13R␣1, based on the structure of the IL-4⅐IL-4R␣ complex (20), and IL-13 has a significant similarity in folding topology with IL-4 (40,41). Consistent with this assumption, it has been demonstrated that several amino acid residues in ␣-helix D of IL-13 (Lys-90, Ile-91, His-103, Leu-104, Lys-105, Lys-106, Arg-109, Glu-110, and Arg-112) are important for binding to hIL-13R␣1 and/or hIL-13R␣2, although ␣-helix A of IL-13 is predicted to interact with IL-4R␣ (29 -31). Some of these amino acids would interact with those identified in our present study, involved in the binding between IL-13 and hIL-13R␣1/hIL-13R␣2. It is of note that Asp-318 of hIL-13R␣2, contributing most to the binding among the investigated amino acid residues, was assumed to interact with Lys-105 of IL-13, forming a salt bridge, in our present model (data not shown).
We furthermore demonstrated that the D1 domain is necessary for expression of hIL-13R␣2, but not for that of hIL-13R␣1 ( Figs. 4 and 6), whereas the D1 domain of hIL-13R␣1 is critical for binding to IL-13 (Fig. 7). Thus far, it is unclear how the D1 domain of hIL-13R␣2 is involved in the expression mechanism of the receptor. When the D1 domain is deleted, the D2/D3 domains of hIL-13R␣2 may be unable to keep their conformations. Mutagenesis analyses of the extra FnIII domain have already shown its importance in binding to ligands and the signal transduction in gp130 (21,22), G-CSFR (23,24), LIFR (25,26), and IL-5R␣ (27). Our present finding is the first evidence showing involvement of the D1 domain of hIL-13R␣1 in binding to IL-13. Structural analyses of the IL-6⅐IL-6R␣⅐gp130 complex show that this complex is a 2:2:2 hexamer, in which the extra FnIII domains interact with the binding epitopes on IL-6 in the opposite trimers (18,28). Although the precise structure of the IL-13⅐hIL-13R␣1⅐hIL-4R␣ complex remains undetermined, this complex may also form a 2:2:2 hexamer via the D1 domain of hIL-13R␣1 (Fig. 8A).
We found that the D1 domain of hIL-13R␣1 is critical for the binding and signal transduction of IL-13, but not IL-4 (Fig. 7). This finding suggests that the binding modes of IL-4 and IL-13 with IL-13R␣1 are different, although both ligands utilize the common heterodimeric complex composed of IL-4R␣ and IL-13R␣1. Similarly, it has been already shown that the D1 domain of gp130 is needed for binding to IL-6, but not to LIF, IL-11, ciliary neurotrophic factor, or oncostatin-M (21,42). IL-4 first binds to IL-4R␣ and then recruits IL-13R␣1 or ␥c to the complex, forming the high affinity receptor. IL-13R␣1 alone has almost no binding activity to IL-4 (4,43). Mutagenesis experiments have suggested that Arg-121 and Tyr-124 located at the ␣-helix D of IL-4 are important for the interaction with IL-13R␣1 (44,45), which overlaps with the interaction site to ␥c (46,47). Our present finding indicates that ␣-helix D of IL-4 probably interacts with the D2 and D3 domains of IL-13R␣1, independently from its D1 domain (Fig. 8B). Involvement of the D1 domain of hIL-13R␣1 in binding to IL-13, but not to IL-4, at The shaded and open areas represent the count with or without the first Ab, respectively. B, the specific bindings of 125 I-IL-13 to DND-39/G⑀ cells stably transfected with the wild (squares) and D1deleted (triangles) types of hIL-13R␣1. C, luciferase assay using DND-39/G⑀ cells expressing wild and D1-deleted types of hIL-13R␣1. The cells were stimulated with IL-4 or IL-13 (0, 0.1, 0.4, 1, and 10 ng/ml) for 24 h. D, activation of STAT6 and TYK2 using DND-39/G⑀ cells expressing wild and D1-deleted types of hIL-13R␣1. The cells were stimulated with 10 ng/ml IL-4 or IL-13 for 10 min. least partially explains the different affinities of hIL-13R␣1 to IL-13 and IL-4.
We found previously that there exists a variant of the IL13 gene, in which arginine residue at 110 (Arg-110; numbering from the starting residue of the mature protein at Gly-1) is replaced by glutamine (Gln-110); this variant is associated with bronchial asthma in both Japanese and British populations (48). The same variant was thereafter reported to be positively correlated with high IgE levels and atopic dermatitis (49 -51). We furthermore demonstrated that the Gln-110 type has a lower affinity with hIL-13R␣2 than the Arg-110 type, whereas both types show the same affinity with hIL-13R␣1, which would cause up-regulation of the IL-13 concentration in the body (33). We assumed that the interaction between Arg-110 and hIL-13R␣2 might be disrupted by the substitution of the glutamine residue, although the alanine scanning approach showed only a slight involvement of this residue in binding to hIL-13R␣2 (30). If such an amino acid in hIL-13R␣2 interacting with Arg-110 in IL-13 were displaced by another amino acid, the affinities of the mutant hIL-13R␣2 with the Arg-110 and Gln-110 types would become the same. All of the investigated mutant hIL-13R␣2 showed lower affinities with the Gln-110 type than the Arg-110 type or the wild type (data not shown). These results implied the possibility that R110Q may change the conformation of IL-13 itself, not the direct interaction with IL-13R␣2.
Considering the importance of IL-13 in the pathogenesis of allergic diseases, particularly bronchial asthma, several IL-13 antagonists have been developed as means of improving allergic states (1). Our present finding would be useful in these strategies. The D1 domain is a particularly good target to develop a neutralizing Ab or a low molecular weight compound to block specifically the interaction between hIL-13R␣1 and IL-13, but not IL-4. It has been already shown that the monoclonal Abs against the D1 domains of gp130 (21,22,42) or G-CSFR (24) inhibit binding to ligands and their signals. It will be of great interest to analyze the effects of a neutralizing Ab or a low molecular weight compound targeting the D1 domain of IL-13R␣1 on the development of allergic diseases.
In conclusion, we for the first time identified the critical residues in the CRH of hIL-13R␣1 and hIL-13R␣2 for binding to IL-13 and clarified the roles of the D1 domains of these receptors by the mutagenesis approach. These results provide the basis for a precise understanding of the interaction between IL-13 and its receptors.