New Activation Modus of STAT3

Cell Culture and Cytokines

BW5147 lymphoma cells and HepG2 human hepatoma cells were grown as described previously (15). HEK293-EBNA cells and COS cells were grown in Dulbecco's modified Eagle's medium supplemented with 10% fetal calf serum. HT-29 and colo205 human epithelial cell lines were grown in Iscove-Dulbecco medium supplemented with 10% fetal calf serum, 0.55 mm l-arginine, 0.24 mm l-asparagine, and 1.25 mm l-glutamine. Recombinant human IL-22 was produced in Escherichia coli as described previously (15). Recombinant hIL-9 and mouse IL-9 were produced in the baculovirus system and purified as described previously (16). To produce anti-hIL-22R polyclonal antibodies, P815 mastocytoma cells were transfected with the human IL-22R cDNA cloned into the pEF-BOS plasmid before injection into DBA/2 mice. After rejection of the tumors, the sera of these mice had high titers of neutralizing anti-IL-22R antibodies, and the spleens were used to generate hIL-22R-specific monoclonal antibodies.

Plasmid Constructions, Stable Transfections, and Analysis of Transfected Cells

The IL-22R cDNA was amplified by reverse transcription-PCR from the HepG2 hepatoma cell line before cloning into the pEF-BOSpuro expression vector. For the IL-22R 574ΔF receptor, the cDNA sequence corresponding to nucleotides 1–1050 of the IL-22R open reading frame was fused to nucleotides 1209–1725 using mutated primers to allow direct cloning into pEF-BOSpuro. IL-22R Tr490 and Tr490ΔF receptors were generated by using an antisense primer introducing a stop codon after amino acid 490. Mutagenesis was performed by changing tyrosine into phenylalanine residues using the GeneEditor in vitro site-directed mutagenesis system (Promega, Madison, WI). Clones obtained were sequenced using the DYEnamic ET dye terminator kit (Amersham Biosciences). BW5147 stable transfection was performed as described previously (11). Cell surface expression of the different IL-22R variants was checked by flow cytometry using an anti-IL-22R monoclonal antibody.

For the GST fusion protein, the intracytoplasmic part of IL-22R or fragment of IL-22R corresponding to amino acids located between 491 and 574 was cloned into pGMEX3-T vector (GE Healthcare). The pBS-STAT3 vector encoding the wild-type form of mouse STAT3 (provided by T. Hirano, Osaka University, Osaka, Japan) was subcloned into pCEP4 vector (Invitrogen). Truncated STAT3 isoforms were cloned into the same vector using antisense primers introducing a stop codon after amino acid 685, 585, 466, 326, or 130, according to the murine STAT3 amino acid numbers. The construct encoding the chimeric STAT1-STAT3 protein was generated by amplifying the cDNA sequences corresponding to amino acids 1–351 from STAT1 (N-terminal and coiled-coil domains) and to amino acids 356–770 of STAT3 (from DNA-binding to C-terminal domain) using mutated primers to allow direct cloning into pCEP4. A chimeric STAT3-STAT5 construct was generated similarly by amplifying the cDNA sequences corresponding to amino acids 1–325 from STAT3 and to amino acids 337–793 from STAT5. The chimeric IL-9RF116–22R was composed of a complete human IL-9R cDNA sequence with a Tyr-to-Phe mutation of residue 116 of the cytoplasmic domain and a mutation of the stop codon introducing a BstBI restriction site to insert the sequence coding for the C-terminal part of the IL-22R (amino acids 487–574, with Tyr⁴⁹¹ mutated into Phe).

Western Blotting

BW5147 stable transfectants (10⁶) were stimulated for 20 min with control medium or medium containing mouse IL-9 or human IL-9 (50 units/ml). Cells were lysed in 200 μl of Laemmli buffer (Bio-Rad), and anti-phospho-STAT3 Western blotting was performed as described previously (17).

Fluorescence-activated Cell Sorter Analysis

BW5147 cells (2 × 10⁵) were left unstimulated or stimulated with IL-22 for 15 min at 37 °C. Cells were fixed in 2% paraformaldehyde for 10 min at 37 °C and then permeabilized in 90% methanol for 30 min on ice. Cells were washed twice with Hanks' medium and stained with anti-phosphorylated STAT3-Y705 (BD Biosciences). Cells were analyzed on a BD Biosciences FACSCalibur^TM flow cytometer.

Immunoprecipitation

Twenty million cells were stimulated with hIL-22 (50 nm) or control medium for 5 min and washed, and lysates were obtained as described previously (17). Five micrograms of anti-IL-22R monoclonal antibody was added to the supernatant and incubated overnight at 4 °C. Lysates were then incubated with protein G-agarose for 2 h. Beads were washed, resuspended in Laemmli buffer (25 μl), and boiled. Proteins were separated in a 10% SDS-polyacrylamide gel (Novex, Carlsbad, CA) and transferred onto a nitrocellulose membrane. The membrane was blocked in 1% bovine serum albumin solution before overnight incubation with anti-STAT3 polyclonal antibody (Cell Signaling, Beverly, MA). Proteins were detected by chemiluminescence (SuperSignal West Pico, Pierce), and membranes were reprobed with anti-IL-22R antibody as a control.

BW5147 Bioassays

Cell proliferation assays and dexamethasone-induced apoptosis assay were performed as described previously (11). Then, cell proliferation was assessed either by a hexosaminidase assay or by thymidine proliferation as described previously (11).

Transient Cell Transfection and Luciferase Assay

The pGL3-Pap1 luciferase reporter plasmid was used as a specific promoter for STAT3 activation, whereas the pRL-TK plasmid (Promega), encoding Renilla luciferase, was cotransfected as an internal control of the transfection process. Transient COS cell transfections and luciferase assays were carried out as described previously (18). Briefly, 5 h after transfection, cells were stimulated with control medium or with human IL-22 (50 nm) for 16 h.

GST Purification Assay

To measure the interaction between GST fusion protein and STAT3 truncated mutants, the different proteins were produced independently in HEK293 cells by transient transfection using the Lipofectamine 2000 method (Invitrogen), according to the manufacturer's recommendations. Two days later, cells were lysed in 1 ml of lysis buffer (1% Triton X-100, 10% glycerol, 10 mm Tris (pH 7.6), 150 mm sodium chloride, 5 mm EDTA, 1 mm dithiothreitol, 1 mm sodium vanadate, 1 mm sodium fluoride, and inhibitor mixture (Roche Applied Science)). Lysates were homogenized as described above. Samples containing a GST fusion protein or a STAT3 protein were mixed, kept at 4 °C for 16 h, and purified on glutathione-Sepharose beads (Amersham Biosciences). After washing, specifically bound products were eluted from the beads with 20 mm reduced glutathione (Sigma). Eluted samples were subjected to SDS-PAGE, and proteins were transferred to nitrocellulose filters, probed with STAT3-specific antibodies, either specific for the C terminus (Cell Signaling Technology) or for the N terminus (Transduction Laboratories) depending on the truncated form of STAT3, and then reprobed with anti-GST monoclonal antibody.

Tyrosine Phosphorylation of IL-22R Is Not Required for STAT3 Activation

To determine whether STAT3 activation in response to IL-22 depends on the phosphorylation of IL-22R, two different types of IL-22R mutants were generated, either by mutating all of its cytoplasmic tyrosine residues into phenylalanine (574F) (Fig. 1) or by deleting a fragment of IL-22R containing five tyrosines from a repeated 13-amino acid motif located between Tyr³⁵¹ and Tyr⁴⁰³ and mutating the three other tyrosines of the cytoplasmic domain (574ΔF). Unexpectedly, as shown in Fig. 2, A and C, both receptors lacking cytoplasmic tyrosines could still induce phosphorylation of STAT3 but not STAT5 or STAT1. Receptors lacking tyrosines (574F and 574ΔF) still activate STAT-dependent reporter gene transcription, and this activity requires the C terminus of IL-22R because deletion of the last 84 amino acids (Tr490ΔF) abolished the IL-22 response (Fig. 2B). As shown in Fig. 2C, a receptor lacking this C-terminal domain but keeping the cytoplasmic tyrosines (Tr490) still induced some STAT3 phosphorylation but to a lesser extent than receptors lacking tyrosines (574ΔF or 574F).

The C-terminal Domain of IL-22R Is Sufficient for STAT3 Recruitment

To determine whether these 84 C-terminal residues are sufficient to recruit and activate STAT3 in the context of another receptor, we took advantage of a mutant form of human IL-9R that had lost the ability to activate STAT factors (h9RF116) (11). Different human IL-9 receptor constructs were transfected in BW5147 lymphoma cells, which endogenously respond to murine IL-9 but not to human IL-9. Transfection of wild-type (BWh9R) but not mutated BWh9RF116 IL-9R allowed hIL-9-dependent STAT3 phosphorylation. However, fusion of the C-terminal domain (residues 491–574) of IL-22R after the last amino acid of the mutated hIL-9R restored its ability to activate STAT3 at the same level as the wild-type hIL-9R (Fig. 3A). At the functional level, this correlated with a restored ability of this receptor to protect BW5147 transfectant cells against dexamethasone-induced apoptosis (Fig. 3B), an activity previously shown to be mediated by STAT3 (19). Taken together, these data indicate that the C-terminal domain of the IL-22R is sufficient to confer the ability to recruit STAT3 to an unrelated cytokine receptor, leading to functional STAT3 activation.

Role of the C Terminus of IL-22R in IL-22 Bioactivity

To investigate whether the C-terminal domain of IL-22R is required for STAT3-dependent biological activities, we took advantage of BW5147 lymphoma cells in which prolonged STAT3 activation induces growth inhibition (19). In this model, IL-22 inhibited more than 90% of the proliferation of BW cells expressing the wild-type receptor (574), but deletion of the C-terminal domain of the receptor (Tr490) dramatically reduced this activity (Fig. 4B). Deletion or mutation of the tyrosines (574ΔF and 574F) decreased the activity of IL-22, but to a lesser extent, indicating that tyrosines were less critical than the C-terminal domain. Finally, a receptor mutant lacking both the C-terminal domain and the cytoplasmic tyrosines (Tr490ΔF) was completely devoid of activity as expected from the luciferase experiments described before. All transfectants had a similar cell surface expression of IL-22R, ruling out the possibility that these mutations affect the stability or trafficking of the protein (Fig. 4A).

STAT3 Is Preassociated with IL-22R

To understand better this unconventional mechanism of STAT3 activation, we analyzed the interaction between IL-22R and STAT3 by immunoprecipitation. As shown in Fig. 5, STAT3 coprecipitated with IL-22R even in the absence of the ligand. Similar results were obtained in stable transfectants (Fig. 5A) and in cells endogenously expressing this receptor such as HepG2, HT-29, and colo205 (Fig. 5B). Deletion or mutation of all tyrosines in the IL-22R 574ΔF mutant did not affect the association with STAT3, but deletion of the C-terminal domain in the 490 and Tr490ΔF mutant abolished the preassociation with STAT3.

To characterize further the interaction between the cytoplasmic domain of IL-22R and STAT3, we screened a series of GST fusion proteins that contained variants of the cytoplasmic domain of IL-22R for the binding of STAT3 (Fig. 6A). After purification on Sepharose-glutathione beads, STAT3 was found to copurify specifically with the GST fusion proteins that contained the C-terminal domain of IL-22R, irrespective of the presence of tyrosines. Thus, both immunoprecipitation and GST pulldown experiments indicate that STAT3 constitutively associates with the cytoplasmic domain of IL-22R.

The Coiled-coil Domain of STAT3 Is Responsible for Its Association with IL-22

To identify which region of STAT3 is involved in this interaction, we generated STAT3 truncated mutants, successively deleting STAT3 domains from the C terminus, and analyzed their ability to interact with IL-22R-GST fusion proteins. As shown in Fig. 6B, a STAT3 variant lacking the C-terminal, SH2 and DNA-binding domain interacted like the wild type STAT3 with a tyrosine-less IL-22R-GST fusion protein (GST-574ΔF, lanes 2), but not with GST alone (lanes 1) or a fusion protein that lacked the C-terminal domain (GST-Tr491ΔF, lanes 3). STAT3 mutants that were truncated from the N terminus were used to dissect the role of coiled-coil and N-terminal domains (Fig. 6C). Although deletion of the N-terminal domain did not affect the interaction with IL-22R (STAT3₇₇₀-N130), further deletion of the coiled-coil domain completely abolished the association of STAT3 with IL-22R (STAT3₇₇₀-N326). To confirm these results in the context of a full-length STAT protein, we generated a chimeric STAT1-STAT3 protein in which the N-terminal and coiled-coil domains of STAT3 were replaced by those of STAT1. As shown in Fig. 6D, this chimeric protein failed to interact with GST fused to IL-22R, further demonstrating that the coiled-coil domain of STAT3 is specifically required for its interaction with C-terminal region of IL-22R.

The Coiled-coil Domain of STAT3 Is Sufficient to Confer Recruitment to IL-22R

To determine whether the coiled-coil domain of STAT3 might be sufficient to mediate association with IL-22R in vivo, we took advantage of HEK293 cells, which express very low levels of endogenous STAT5. As a result, IL-22 did not induce any detectable STAT5 phosphorylation in these cells, even when the wild-type IL-22R (574) is transfected (Fig. 7A). When STAT5 was cotransfected with different IL-22R mutants (Fig. 7B), STAT-5 phosphorylation was observed only with the wild-type receptor but not with the receptors lacking tyrosines (574ΔF, 574F, and Tr490ΔF), confirming that IL-22R tyrosines are required for STAT5 phosphorylation. However, when we replaced the N-terminal and coiled-coil domains of STAT5 with those of STAT3 (Fig. 7C), the resulting chimeric protein could be activated even by tyrosine-less receptors (such as 574ΔF and 574F) but not by the receptor lacking the C-terminal domain (Tr490ΔF). These data demonstrated that the coiled-coil domain of STAT3 conferred the ability to be recruited to the IL-22R C-terminal region independently of the SH2 domain of the STAT protein.