Interaction between the components of the interferon gamma receptor complex.

Interferon γ (IFN-γ) signals through a multimeric receptor complex consisting of two different chains: the IFN-γ receptor binding subunit (IFN-γR, IFN-γR1), and a transmembrane accessory factor (AF-1, IFN-γR2) necessary for signal transduction. Using cell lines expressing different cloned components of the IFN-γ receptor complex, we examined the function of the receptor components in signal transduction upon IFN-γ treatment. A specific IFN-γR2:IFN-γ cross-linked complex was observed in cells expressing both IFN-γR1 and IFN-γR2 indicating that IFN-γR2 (AF-1) interacts with IFN-γ and is closely associated with IFN-γR1. We show that the intracellular domain of IFN-γR2 is necessary for signaling. Cells coexpressing IFN-γR1 and truncated IFN-γR2, lacking the COOH-terminal 51 amino acids (residues 286-337), or cells expressing IFN-γR1 alone were unresponsive to IFN-γ treatment as measured by MHC class I antigen induction. Jak1, Jak2, and Stat1α were activated, and IFN-γR1 was phosphorylated only in cells expressing both IFN-γR1 and IFN-γR2. Jak2 kinase was shown to associate with the intracellular domain of the IFN-γR2.

It has been shown that the active receptors for several cytokines consist of at least two subunits (Jung et al., 1987;Kishimoto et al., 1992;Miyajima et al., 1992;Akira et al., 1993;Taniguchi and Minami, 1993). The multicomponent structure of the interferon ␥ receptor complex was proposed based upon studies with human-rodent and mouse-hamster somatic cell hybrids. Interferon ␥ (IFN-␥) 1 binds to the IFN-␥ receptor binding subunit (IFN-␥R1), a species-specific cell-surface receptor encoded on human chromosome 6 (Rashidbaigi et al., 1986) and mouse chromosome 10 (Mariano et al., 1987;Kozak et al., 1990). However, although IFN-␥R1 is the binding subunit of the IFN-␥ receptor complex, it alone is not sufficient for signal transduction (Jung et al., 1987(Jung et al., , 1988Hibino et al., 1991). In addition to the binding subunit, an additional factor encoded on human chromosome 21 (Jung et al., 1987(Jung et al., , 1988 and mouse chromosome 16 (Hibino et al., 1991) was found to be required for responsiveness of the cells to IFN-␥ as measured by MHC class I antigen induction. This additional factor was recently cloned from human and mouse cells and has been designated as the accessory factor (AF-1) or second chain of the IFN-␥ receptor complex, IFN-␥R2 (Soh et al., 1994;Hemmi et al., 1994).
There is no intrinsic kinase motif within the intracellular domains of either IFN-␥R1 or IFN-␥R2. Two distinct regions of the intracellular domain of IFN-␥R1 have been found to be important for signal transduction (Cook et al., 1992;Farrar et al., 1991Farrar et al., , 1992. The first, proximal to the transmembrane region, is necessary for both receptor-ligand internalization and biological responses (Farrar et al., 1991). The second region, near the carboxyl terminus, includes Tyr-457 (Tyr-440, starting at the putative first amino acid of the mature chain), Asp-458, and His-461, which are required for biological responsiveness (Cook et al., 1992;Farrar et al., 1992). After phosphorylation of Tyr-457, Stat1␣ (Schindler et al., 1992) binds to this region (Stat1␣ recruitment site) due to specific interaction between the SH2 domain of Stat1␣ and phosphorylated Tyr-457 of IFN-␥R1 (Greenlund et al., 1994), with resultant phosphorylation of Tyr-701 of Stat1␣ (Shuai et al., 1993a). This phosphorylation is probably caused by the IFN-␥-activated tyrosine kinases Jak1 and/or Jak2, members of the Janus kinase (Jak) family of cytoplasmic protein tyrosine kinases (for review, see Ziemiecki et al., 1994;Ihle et al., 1994Ihle and Kerr, 1995). Jak1 and Jak2 tyrosine kinases participate in signal transduction initiated by IFN-␥ as shown in mutant cell lines defective in the IFN-␥ signal transduction pathway suggesting that the IFN-␥ receptor components and the Jak kinases interact Silvennoinen et al., 1993;Watling et al., 1993). In this report, we elucidate the structure of the IFN-␥ receptor complex and the role of the receptor components in the activation of the IFN-␥ signal transduction pathway.

EXPERIMENTAL PROCEDURES
Reagents, Restriction Endonucleases, and Other Enzymes-All restriction endonucleases were from Boehringer Mannheim Biochemicals or New England Biolabs; Sequenase 2.0 and T4 DNA ligase were from United States Biochemical Corp. The [␣-32 P]dATP and [␥-32 P]ATP were from DuPont NEN. The cross-linker bis(sulfosuccinimidyl)suberate (BS 3 ) was from Pierce Chemical Co. All other chemical reagents were analytical grade and purchased from United States Biochemical Corp.
To create an expression vector for the glutathione S-transferase/Hu-IFN-␥R2 intracellular domain fusion protein (GST/IFN-␥R2 IC ), the polymerase chain reaction was performed with two primers 5Ј-TCCTGGATCCAAATATAGAGGCCTGATT-3Ј and 5Ј-GGGAATACTG-GTCTCTGG-3Ј and plasmid p␥R2 as a template according to standard protocol (Sambrook et al., 1989). The polymerase chain reaction product was digested with BamHI restriction endonuclease and ligated into BamHI and SmaI sites of the pGEX-2T vector (Pharmacia).
Cells, Media, and Transfection-The CHO-B7 cell line is a Chinese hamster ovary cell (CHO-K1) transfected with human HLA B7 gene. The 16-9 hamster ϫ human somatic cell hybrid line is the CHO-K1 hybrid containing a translocation of the long arm of human chromosome 6 encoding the IFN-␥R1 gene and a transfected human HLA-B7 gene (Soh et al., 1993). The 16-9 cells were maintained in F-12 (Ham's) medium (Sigma) containing 5% heat-inactivated fetal bovine serum (Sigma) (complete F-12 medium). HEp-2 cells, a human epidermoid larynx carcinoma cell line, were maintained in Dulbecco's modifed Eagle's medium (Life Technologies, Inc.) with 10% heat-inactivated fetal bovine serum.
The expression vectors p␥R2, p␥R2-C174S, p␥R2t-285, and p␥R2-Y294F were stably transfected into CHO-B7 or 16-9 cells (1-3 g of supercoiled plasmid DNA per 10 5 -10 6 cells) with LipofectAMINE TM Reagent (Life Technologies, Inc.) according to the manufacturer's instructions for stable transfection of adherent cells. All cell lines transfected with plasmids carrying the neo R gene were selected and maintained in complete F-12 medium containing 450 g/ml antibiotic G418.
Cytofluorographic Analysis-Cytofluorographic analysis of cells for expression of the HLA-B7 surface antigen was performed as described previously (Cook et al., 1992;Soh et al., 1993Soh et al., , 1994. Hu-IFN-␣A/D, a chimeric human interferon active on hamster cells (Rehberg et al., 1982), was used as a control to demonstrate the integrity of the HLA-B7 gene in various cell lines.
Cross-linking of IFN-␥ to Receptors-Recombinant Hu-IFN-␥ with a specific activity of 2 ϫ 10 7 units/mg was phosphorylated as reported . The [ 32 P]IFN-␥ was bound to cells and then cross-linked as follows. Briefly, cells were released with 5 mM EDTA in phosphate-buffered saline (PBS), washed with PBS, and finally resuspended at 5 ϫ 10 6 cells/ml in PBS, 0.1% bovine serum albumin, 5 mM MgCl 2 . About 10 6 cpm of [ 32 P]IFN-␥ (370 Ci/g) were added to 0.2 ml of the cell suspension with or without a 200-fold excess of unlabeled IFN-␥ as competitor and were incubated for 75 min at 4°C. After washing with PBS, the cells were resuspended in 0.5 ml of PBS containing 5 mM MgCl 2 and cross-linked by adding 0.2 M bis(sulfosuccinimidyl)suberate (BS 3 , freshly prepared in PBS) to a final concentration of 2 mM. The cross-linking reaction was terminated after 60 min at 4°C by the addition of 10 l of 1 M Tris⅐HCl, pH 7.4. The cells were recovered by centrifugation and then extracted in 25 l of PBS containing 1% Nonidet P-40, 0.5 M NaCl, 5 mM EDTA, and 1 mM phenylmethylsulfonyl fluoride, 1 g/ml leupeptin, 1 g/ml antipain, 10 g/ml benzamidine⅐HCl, 3 g/ml aprotinin, 1 g/ml chymostatin, and 1 g/ml pepstatin. The extracts were analyzed by SDS-polyacrylamide gel electrophoresis on 7.5% polyacrylamide vertical slab gels (Laemmli, 1970). Gels were dried and exposed to Kodak XAR-5 film with an intensifying screen for 12 h at Ϫ80°C.
Antibodies-Rabbit anti-Jak1 and anti-Jak2 antibodies were developed against synthetic peptides (KTLIEKERFYESRCRPVTPSC and DSQRKLQFYEDKHQLPAPKC, respectively) corresponding to the end of the second kinase-like domains of murine Jak1 and Jak2, respectively. Rabbit anti-Stat1␣ antibody, raised against the carboxyl terminus of Stat1␣, was a gift from James Darnell. Monoclonal anti-phosphotyrosine antibody was from Sigma (catalog No. P3300). Rabbit anti-IFN-␥R1 antibody was prepared with the extracellular domain of Hu-IFN-␥R1 as antigen.
Immunoprecipitations and Blottings-Cells were starved overnight in serum-free media and subsequently stimulated with IFN-␥ (1000 units/ml) for 15 min at 37°C. Cells were solubilized in a lysis buffer consisting of 0.1% Nonidet P-40, 50 mM Tris⅐HCl, pH 8.0, 0.15 M NaCl, 10 mM NaF, 5 mM sodium pyrophosphate, 1 mM orthovanadate, 1 mM phenylmethylsulfonyl fluoride, leupeptin (1 g/ml), antipain (1 g/ml), benzamidine-HCl (10 g/ml), aprotinin (3 g/ml), chymostatin (1 g/ ml), and pepstatin (1 g/ml). Following a 1-h incubation at 4°C, the samples were centrifuged at 10,000 ϫ g for 10 min. The supernatants were cleared with a suspension of rabbit serum-agarose (Sigma, catalog No. R-6755) for 1 h at 4°C. After separation of the agarose by centrifugation for 1 min, the lysates were incubated with the indicated antibodies for 2-16 h at 4°C. Protein A-agarose was then added for 1 h at 4°C to bind the immune complexes. The protein A-agarose was then washed 3 times in the lysis buffer and boiled in SDS sample buffer to remove bound components. The samples were resolved by SDS-polyacrylamide gel electrophoresis (SDS-PAGE), transferred with the Trans-Blot S.D. Semi-Dry Transfer Cell  according to the manufacturer's instructions to polyvinylidene difluoride (PVDF) membranes (Bio-Rad). The membranes were then blocked with 3% (w/v) blocking agent (Amersham, catalog No. NIP551) and 1% bovine serum albumin in PBS with 0.5% Tween-20 (PBST), 10% glycerol, 1 M glucose for 1 h at 22°C. The membranes were washed in PBST twice for 15 min and probed with the indicated primary antibodies for 1-18 h at 22°C. The blots were then washed in PBST, incubated for 1 h with donkey anti-rabbit or anti-mouse horseradish peroxidase-conjugated secondary antibodies and developed with the enhanced chemiluminescence method (Amersham, ECL TM Western blotting analysis system, catalog No. RPN2108), according to the manufacturer's protocol.

Glutathione S-Transferase/Hu-IFN-␥R2 Intracellular Domain Fusion Protein (GST/IFN-␥R2 IC ) for Affinity Binding of Proteins-Sf9
insect cells were infected with baculovirus encoding Jak1 or Jak2 kinases, and cell lysates were prepared and incubated with the GST/IFN-␥R2 IC fusion protein or with GST alone immobilized on glutathione-Sepharose as described previously . After extensive washing, the proteins associated with the fusion proteins were eluted in SDS sample buffer and resolved by SDS-PAGE, transferred to nitrocellulose, and probed with antisera to Jak1 and Jak2 kinases. The Jak1 and Jak2 protein concentrations were normalized from total cellular lysates probed with anti-Jak1 or anti-Jak2 antibodies as described .
Jak1 and Jak2 Activation Assay-Cells were starved overnight in serum-free media and subsequently stimulated with IFN-␥ (200 units/ ml) for 10 min at 37°C. The cells were lysed, and Jak1 and Jak2 proteins were immunoprecipitated as described above. Activation of Jak1 and Jak2 was determined by an in vitro kinase assay with a peptide corresponding to the putative phosphorylation site of Jak2 as a substrate. Specific phosphorylation of the Jak2 peptide (VLPQD-KEYYKVKEPGE) by activated Jak1 and Jak2 was performed as previously reported . Equal loading of Jak proteins in the assay was demonstrated by using one-third of the samples in Western blotting with antibodies against Jak1 and Jak2.

RESULTS
Biological Assay-We measured class I MHC antigen induction to evaluate the ability of the cloned components of the human IFN-␥ receptor complex to transduce a signal upon Hu-IFN-␥ treatment in hamster cells. The 16-9 cells, expressing only the IFN-␥R1 chain of the human IFN-␥ receptor complex, exhibit little or no response to Hu-IFN-␥ ( Fig. 1A; see also Soh et al. (1994)). The 16-9/IFN-␥R2 cells, 16-9 cells stably transfected with an expression vector encoding the intact IFN-␥R2 under the control of the cytomegalovirus promoter, exhibited a substantial response to Hu-IFN-␥ (Fig. 1C). To test if the intracellular domain of IFN-␥R2 is necessary for signal transduction, we introduced a stop codon after amino acid 285 of IFN-␥R2 resulting in a deletion of the COOH-terminal 51 amino acids. Hu-IFN-␥ did not induce MHC class I antigens in 16-9 cells stably transfected with the expression vector encoding the truncated IFN-␥R2t (16-9/IFN-␥R2t) (Fig. 1E). As a control, it was shown that all cells responded to Hu-IFN-␣A/D demonstrating that the MHC class I antigen could be induced in all cell lines (Fig. 1, B, D, and F). We therefore conclude that the intracellular domain of the IFN-␥R2 plays an important role in IFN-␥ signal transduction. The portion of the intracellular domain of IFN-␥R2, which was removed by a premature termination codon, had only one Tyr residue at position 294. Because phosphorylation of tyrosine residues in various receptors (including IFN-␥R1) was shown to be important for signal transduction, we mutated this Tyr residue to Phe. Cells expressing this mutated IFN-␥R2-Y294F were active in MHC class I antigen induction (data not shown), indicating that this Tyr is not necessary for the activity of the IFN-␥R2 chain.
The formation of an intermolecular disulfide bridge between a Cys residue at position 174 of the human IFN-␥R2 and the corresponding Cys residue at position 167 of the human IFN-␥R1 (Soh et al., 1994) could account for the interaction between these two subunits of the IFN-␥ receptor complex. To test this hypothesis, we mutated IFN-␥R2 at position 174 from Cys to Ser. However, the mutated IFN-␥R2-C174S remained fully active in class I MHC induction (data not shown), indicating that Cys-174 is not essential for formation of an active IFN-␥ receptor complex.
To determine the contribution of each chain of the IFN-␥ receptor complex in activation of intracellular signaling, we examined Jak1 and Jak2 activation and activation of Stat1␣ in cells expressing various components of the receptor. In cells expressing both chains of the human IFN-␥ receptor complex (16-9/IFN-␥R2), IFN-␥ activated Jak1 and Jak2 as measured by the in vitro phosphorylation of Jak1 and Jak2 substrate peptide by immunoprecipitated Jak1 and Jak2 kinases, respectively (Fig. 3). However, we did not observe activation of Jak1 and Jak2 in the 16-9 cells expressing only one chain of the human IFN-␥ receptor complex (Fig. 3). The 16-9/ IFN-␥R2t cells, expressing the IFN-␥R1 and the truncated IFN-␥R2t, failed to show activation of the Jak kinases (Fig. 3). In addition, we observed ligand-induced phosphorylation of Jak1 and Jak2 only in the cell line 16-9/IFN-␥R2 expressing both chains of the human IFN-␥ receptor complex (Fig. 4). Phosphorylation of Jak1 and Jak2 upon IFN-␥ treatment was not detected in 16-9 and 16-9/IFN-␥R2t cells (Fig. 4).
Analogous results were obtained for activation of Stat1␣. Only in 16-9/IFN-␥R2 cells expressing both receptor chains did IFN-␥ produce an active Stat1␣ as measured by the electrophoretic mobility shift assay (Fig. 5A). The formation of the Stat1␣ DNA-binding complex upon IFN-␥ treatment was suppressed after addition of an excess of unlabeled oligonucleotides as a competitor (Fig. 5A). To show that this complex was formed by Stat1␣ proteins, anti-Stat1␣ antibodies were added to the nuclear extracts from IFN-␥-treated 16-9/IFN-␥R2 cells, and the extracts were incubated with the same radiolabeled probe. The specific DNA-binding complex was supershifted after addition of anti-Stat1␣ antibodies (Fig. 5B).
It was shown that overexpression of Jak1 and Jak2 by transient transfection leads to tyrosine autophosphorylation of Jak1 and Jak2 and activates Stat1␣, as measured by DNA binding . However, overexpression of Jak1 and/or Jak2 by stable transfection of 16-9 cells, expressing only Hu-IFN-␥R1, did not permit IFN-␥ to induce MHC class I antigens. 2 Furthermore, the activation of Stat1␣ was not detected in these cells. Thus, both receptor chains are necessary and sufficient for activation of Jak1, Jak2, and Stat1␣ upon IFN-␥ induction.

FIG. 3. Activation of Jak1 and Jak2 tyrosine kinases upon IFN-␥ treatment.
Top panel, the activation of Jak1 and Jak2 was determined with an in vitro kinase assay as described under "Experimental Procedures" with a peptide substrate corresponding to the putative phosphorylation site of Jak2. The cell lines were 16-9, 16-9/IFN-␥R2,and 16-9/IFN-␥R2t as defined in the legend to Fig. 1. Bottom two panels, equal loading of Jak proteins in the assay was demonstrated by using one-third of the samples in Western blotting with Jak1 (middle panel) and Jak2 (bottom panel) antibodies.

FIG. 5. Electrophoretic mobility shift assay (EMSA).
A, EMSAs were performed as described under "Experimental Procedures" with the 22-base-pair labeled sequence containing the Stat1␣ binding site corresponding to the GAS element in the promoter region of the human IRF-1 gene (Yuan et al., 1994) with nuclear extracts from following cells: 16-9, 16-9/IFN-␥R2, 16-9/IFN-␥R2t, CHO-B7, and CHO-B7/IFN-␥R2 cells as defined in the legends to Figs. 1 and 2. In addition, HEp-2 cells, a human epidermoid larynx carcinoma cell line, were used as a positive control. B, the supershift assays were performed as described under "Experimental Procedures" with the 16-9/IFN-␥R2 cell line. The position of the Stat1␣ DNA-binding complex and supershifted complex are indicated by the arrows. The same unlabeled oligonucleotides were used as a competitor in 100-fold excess. cells expressing both chains of the IFN-␥ receptor complex (Fig.  6). The truncated IFN-␥R2t did not support the ability of IFN-␥ to induce tyrosine phosphorylation of the IFN-␥R1 chain. This demonstrates that the IFN-␥R2 chain is necessary for tyrosine phosphorylation of the IFN-␥R1 chain.
Association of Jak2 with the IFN-␥ Receptor Complex-Ligand-independent association of Jak1 to the IFN-␥R1 chain and recruitment of Jak2 to the IFN-␥ receptor complex after IFN-␥ treatment have been demonstrated (Igarashi et al., 1994). To investigate the role of the IFN-␥R2 chain in recruitment of Jak2 into the IFN-␥ receptor complex, we performed immunoprecipitation of IFN-␥R1 from lysates of various cells with anti-IFN-␥R1 antibodies. After immunoprecipitation, the samples were resolved by SDS-PAGE, transferred to a PVDF membrane, and the blot was probed with anti-Jak2 antibodies. The band recognizable by anti-Jak2 antibodies was observed in immunoprecipitates from cells expressing both receptor subunits (16-9/IFN-␥R2) and to a lesser extent in 16-9 cells expressing only the IFN-␥R1 and in both cases only after IFN-␥ treatment (Fig. 6). The coprecipitation of Jak2 with antibodies to IFN-␥R1 after IFN-␥ induction in the cells expressing only IFN-␥R1 may be explained by weak association of Jak2 with dimerized IFN-␥R1 formed upon IFN-␥ binding or by ligandinduced association with Jak1 which is bound to the IFN-␥R1 chain.
From this observation we hypothesized that the intracellular domain of IFN-␥R2 associates with Jak2 and is responsible for recruitment of Jak2 to the IFN-␥ receptor complex after IFN-␥ binding. To examine the interaction between Jak2 and the intracellular domain of IFN-␥R2, we prepared a glutathione S-transferase/IFN-␥R2 intracellular domain fusion protein (GST/IFN-␥R2 IC ). Lysates from Sf9 insect cells infected with baculovirus producing Jak1 and Jak2 were incubated with the GST/IFN-␥R2 IC or GST proteins immobilized on glutathione-Sepharose. The material bound to the glutathione-Sepharose was eluted, analyzed by SDS-PAGE, blotted, and then probed with anti-Jak1 and anti-Jak2 antibodies (Fig. 7). The Jak1 and Jak2 protein concentrations were equalized and monitored from total cellular baculovirus lysates . The GST/IFN-␥R2 IC fusion protein was found to bind Jak2 kinase (Fig. 7), demonstrating that the intracellular domain of the IFN-␥R2 chain of the IFN-␥ receptor complex associates directly with Jak2.

DISCUSSION
It has been shown that binding of IFN-␥ to the receptor leads to formation of a receptor-ligand complex consisting of two molecules of IFN-␥ (a dimer of IFN-␥) and two molecules of IFN-␥R1 (Fountoulakis et al., 1992;Greenlund et al., 1993). No other receptor components were found in cross-linked complexes in previous experiments (Greenlund et al., 1993). However, the interaction between the extracellular domains of IFN-␥R1 and IFN-␥R2 was suggested in studies with mouse-human and hamster-mouse IFN-␥R1 chimeras (Hibino et al., 1992;Hemmi et al., 1992;Gibbs et al., 1991;Kalina et al., 1993). To investigate the interaction between the chains of the IFN-␥ receptor complex and IFN-␥, we carried out cross-linking experiments with a number of cell lines expressing different cloned components of the human IFN-␥ receptor complex (Fig.  2). Our cross-linking experiments, showing that IFN-␥ can be cross-linked to IFN-␥R2, can be explained most plausibly by postulating a close physical association between the IFN-␥R1 and IFN-␥R2 chains of the IFN-␥ receptor complex.
In response to treatment with IFN-␥, the IFN-␥R1 chain can dimerize and internalize in the absence of the IFN-␥R2 chain (Greenlund et al., 1993;Farrar et al, 1991), but these events are insufficient for signal transduction. To investigate the functional importance of the intracellular domain of IFN-␥R2 for IFN-␥ signal transduction, we introduced a stop codon after amino acid 285 of IFN-␥R2. The IFN-␥⅐IFN-␥R2t complex was still observed in cross-linking experiments (Fig. 2). Similar behavior of the intact and truncated IFN-␥R2 in cross-linking experiments demonstrated that the intracellular domain of IFN-␥R2 is not necessary for ligand binding or for association of the extracellular domains of IFN-␥R2 and IFN-␥R1. However, truncated IFN-␥R2t lacking the COOH-terminal 51 amino acid residues was unable to elicit signal transduction for class I MHC antigen induction in respond to IFN-␥ treatment (Fig. 1), indicating that the intracellular domain of IFN-␥R2 is involved in signal transduction. The fact that the expression of human IFN-␥R2 and IFN-␥R1 in hamster cells, and mouse IFN-␥R2 and IFN-␥R1 in human cells, is sufficient to reconstitute the FIG. 6. Tyrosine phosphorylation of IFN-␥R1 and coimmunoprecipitation of Jak2 kinase with antibody to IFN-␥R1. The immunocomplexes were precipitated with anti-IFN-␥R1 antibodies from lysates prepared from untreated and IFN-␥-treated cells as described under "Experimental Procedures." The cell lines were: 16-9, lanes 1 and 2; 16-9/IFN-␥R2, lanes 3 and 4; 16-9/IFN-␥R2t, lanes 5 and 6. The cell lines are defined in the legend to Fig. 1. Immunoprecipitates were resolved on SDS-PAGE, transferred to PVDF membranes, and Western blots were probed with anti-phosphotyrosine antibodies, first panel; with anti-Jak2 antibodies, second panel; and with anti-IFN-␥R1 antibodies, third panel.
FIG. 7. Association of Jak2 kinase with the IFN-␥R2 intracellular domain. The Sf9 cell lysates with baculovirus-produced Jak1 and Jak2 were incubated with the GST/IFN-␥R2 intracellular domain fusion protein or with GST alone immobilized on glutathione-Sepharose. The proteins associated with GST or GST/IFN-␥R2 IC were resolved on the SDS-PAGE, transferred to PVDF membranes, and Western blots were probed with anti-Jak1 and anti-Jak2 antibodies. The Jak1 and Jak2 protein concentrations were equalized with total cellular lysates (TCL). signaling pathway upon treatment with human and mouse IFN-␥, respectively, for MHC class I induction (Soh et al., 1994;Hemmi et al., 1994) suggests that the IFN-␥R2 intracellular domain interacts with the hamster and mouse signal transduction components.
The participation of the protein kinases Jak1 and Jak2 as well as Stat1␣ in the IFN-␥ signal activation pathway has been described Ihle et al., 1994;Ziemiecki et al., 1994). It has been shown that, upon IFN-␥ treatment, four participants of the IFN-␥ signaling pathway (the IFN-␥R1, Jak1, Jak2, and Stat1␣) are phosphorylated on tyrosine and that the tyrosine phosphorylation occurs rapidly (less then 1 min) after IFN-␥ treatment (Igarashi et al., 1994;Greenlund et al., 1994). However, the sequence of events leading to activation of all components and the particular role of each participant in the IFN-␥ signaling pathway is still unknown. Interaction between intracellular domains of the subunits of the IFN-␥ receptor complex and components of the signaling pathway was proposed Watling et al., 1993). Indeed, association of Jak1 tyrosine kinase with IFN-␥R1 prior to IFN-␥ treatment and recruitment of Jak2 only upon IFN-␥ binding has been shown by coimmunoprecipitation experiments (Igarashi et al., 1994). We elucidated the role of the second receptor chain IFN-␥R2 in signal transduction with cell lines expressing different components of the human IFN-␥ receptor complex.
The IFN-␥R2 chain renders cells expressing the IFN-␥R1 chain responsive to IFN-␥. None of the components of the signal transduction machinery (Jak1, Jak2, Stat1␣) were activated upon IFN-␥ treatment in cells expressing only one chain of IFN-␥ receptor complex or IFN-␥R1 and the truncated . No phosphorylation of the IFN-␥R1, Jak1, or Jak2 was observed in these cell lines even after a longer film exposure (Figs. 4 and 6). However, we observed coimmunoprecipitation of Jak2 with antibodies to IFN-␥R1 in cells expressing both chains of the IFN-␥ receptor complex after IFN-␥ treatment. To a lesser extent, we observed coprecipitation of Jak2 with antibodies to the IFN-␥R1 in cells expressing only IFN-␥R1 and only after IFN-␥ treatment (Fig. 6).
We proposed that IFN-␥R2 may associate directly with Jak2 and that the major IFN-␥-induced recruitment of Jak2 to the IFN-␥ receptor complex is the result of association of IFN-␥R2 with IFN-␥R1 after IFN-␥ binding. To investigate the possibility of association of IFN-␥R2 with Jak2, we used the GST/IFN-␥R2 intracellular domain fusion protein. The specific association of Jak2 with GST/IFN-␥R2 IC was observed (Fig. 7) indicating that Jak2 associates with the intracellular domain of the IFN-␥R2 directly in the absence of the IFN-␥ ligand. The coprecipitation of Jak2 with antibodies to IFN-␥R1 after IFN-␥ induction in cells expressing only IFN-␥R1 may be explained by the existence of a low affinity Jak2 binding site on dimerized IFN-␥R1 chains upon IFN-␥ binding or by the ligand-induced interaction between Jak2 and Jak1, which is bound to the IFN-␥R1 chain. We propose that one region of Jak2 interacts with IFN-␥R2 and a second region of Jak2 with the IFN-␥R1 dimer or with Jak1 attached to the IFN-␥R1. An analogous situation has been observed with the IL-2-induced association of the IL-2R␤ chain with Jak3: it was shown that Jak3 primarily associates with the ␥ c chain of the IL-2 receptor complex, but after IL-2 stimulation is weakly coprecipitated with the IL-2R␤ chain which primarily associates with Jak1 (Russell et al., 1994). In the presence of the truncated IFN-␥R2t chain, IFN-␥-dependent coprecipitation of Jak2 with IFN-␥R1 chain was not seen (Fig. 6). This may be due to different conformations of the IFN-␥ receptor complex formed upon IFN-␥ binding with and without the IFN-␥R2 chain. Another explanation may be that the endogenous hamster IFN-␥R2 in the cell lines expressing the human IFN-␥R1 chain can still perform some function, such as bringing Jak2 to the complex, but is not sufficient for signal transduction. To test this possibility, cells with a deletion of the IFN-␥R2 gene will be required. It should be noted, however, that we have not been able to detect binding of Jak2 in vitro to a GST/Hu-IFN-␥R1 cytoplasmic domain fusion protein (data not shown).
Based on the results, we propose a model for the cascade of events in IFN-␥ signaling (Fig. 8). There are at least two receptor subunits: IFN-␥R1, the primary ligand binding subunit, and IFN-␥R2, the second chain of the IFN-␥ receptor complex. Both are required for signal transduction. Upon primary binding of the IFN-␥ dimer to the IFN-␥R1, all the components of the receptor signaling complex are brought together due to association of IFN-␥ with the IFN-␥R1 and the association of IFN-␥R1 with IFN-␥R2 (Hibino et al., 1992). The most likely stoichiometry of this complex is two molecules of IFN-␥R1, two of IFN-␥R2, and one of the IFN-␥ dimer. Jak1 is associated with IFN-␥R1 before IFN-␥ treatment (Igarashi et al., 1994). IFN-␥R2 associates primarily with Jak2 and brings Jak2 kinase to the IFN-␥ receptor complex upon IFN-␥ treatment. After oligomerization of the receptor chains, Jak1 and Jak2 tyrosine kinases associated with the intracellular domains of IFN-␥R1 and IFN-␥R2, respectively, are brought together and reciprocally activated by phosphorylation. This interaction of Jak1 and Jak2 kinases results in their activation, probably via heterodimerization of the kinases, and consequently in phosphorylation of Tyr-457 of the IFN-␥R1 chain, which comprises the Stat1␣ recruitment site (Greenlund et al., 1994). Recruitment of Stat1␣ to the receptor complex due to the specific interaction between the Stat1␣ SH2 domain and the phosphorylated Tyr-457 of the IFN-␥R1 (Greenlund et al., 1994) results in phosphorylation of Tyr-701 of Stat1␣ (Shuai et al., 1993a) and in subsequent Stat1␣ homodimerization (Shuai et al., 1994). Most likely, the proximity of two recruitment sites after IFN-␥R1 dimerization facilitates Stat1␣ dimerization and dissociation of the dimer from the receptor complex. Existence of two recruitment sites for Stat IL-4 in close proximity on one receptor chain of the IL-4 receptor complex substitutes for the necessity to bring two separate chains, such as IFN-␥R1, together to form a STAT dimer (Hou et al., 1994). The Stat1␣ dimer formed in response to IFN-␥ then translocates to the nucleus and interacts with the GAS element in the promoter regions of IFN-␥-inducible genes, the process that begins the induction of this family of genes.