Interaction of the Transcriptional Activator Stat-2 with the Type I Interferon Receptor*

Binding of interferon- (cid:97) (IFN (cid:97) ) to the multisubunit type I IFN receptor (IFNR) induces activation of the Tyk-2 and Jak-1 kinases and tyrosine phosphorylation of multiple signaling elements, including the Stat pro- teins that form the ISGF3 (cid:97) complex. Although Jak kinases are required for IFN (cid:97) -dependent activation of Stats, the mechanisms by which Stats interact with these kinases are not known. We report that Stat-2 associates with (cid:98) s subunit of the type I IFN receptor in an interferon-dependent manner. This association is rapid, occurring within 1 min of interferon treatment of cells, and is inducible by various type I ( (cid:97) , (cid:98) , (cid:118) ) but not type II ( (cid:103) ) IFNs. The kinetics of Stat-2-IFNR association are similar to the kinetics of phosphorylation of Stat-2, sug- gesting that during its binding to the type I IFNR, Stat-2 acts as a substrate for interferon-dependent tyrosine ki- nase activity. These findings support the hypothesis that the type I IFNR acts as an adaptor, linking Stat proteins to Jak kinases. Interaction of Stat-2 with the (cid:98) s subunit of the type I IFNR may be a critical signaling event, re- quired for the formation of the ISGF3 (cid:97) complex and downstream transcription of interferon-stimulated genes. In order for type I interferons to exert their pleiotropic bio-logical effects on cells and

I-IFN␣2-IFNR complexes with approximate molecular masses of 130 -140 kDa (␣ subunit), 110 -120 kDa (␤ subunit), 210 -230 kDa (that appears to result from an association of the ␣ and ␤ subunits), and less prominent complexes of 75 and 180 kDa (most likely an association of the ␣ subunit with the 75-kDa complex) are detected (2)(3)(4)(5)(6)(7). The variant type I IFNR, expressed in some myelomonocytic cell lines, is characterized by lack of expression of the 110 -120-and 210-kDa complexes and the presence of 125 I-IFN␣2-IFNR complexes of 130 -140 kDa (␣ subunit), 75 kDa, and 180 kDa (association of the ␣ subunit with the 75-kDa complex) (2,6). The cloning of the genes encoding two subunits of the type I IFNR has been reported (8,9). The subunit cloned by Uzé et al. (8) has been shown to correspond to the previously described ␣ subunit of the receptor (10). The relative molecular mass of the ␣ subunit appears to exhibit slight variations in different cell lines, ranging from 110 to 135 kDa (2-6, 11, 12), possibly due to differential glycosylation of the protein (4). The subunit cloned by Novick et al. (9) has been reported to encode for a 51-kDa protein. Domanski et al. (7) have recently cloned a cDNA that encodes a 100-kDa form of the type I IFNR. This receptor form and the one cloned by Novick et al. (9) have identical extracellular and transmembrane domains and the first 15 amino acids of the cytoplasmic domain but differ in the rest of the cytoplasmic region (7). In the current study, we used antibodies generated against the receptor subunits cloned by Uzé et al. (8) and by Novick et al. (9) to further characterize the structure of the type I IFNR and its interactions with other signaling molecules. To avoid confusion in the terminology of the different subunits and to be consistent with the terminology used by other groups (7), we will refer to the product of the gene cloned by Uzé et al. (8) as the ␣ subunit of the type I IFN receptor, the product of the gene cloned by Novick et al. (9) as the ␤ s subunit of the type I IFN receptor, and the product of the gene cloned by Domanski et al. (7) as the ␤ L subunit of the type I IFNR. Our findings demonstrate that during type I IFN stimulation, the transcriptional activator Stat-2 associates with the ␤ s subunit of the type I IFNR, providing direct evidence for an interaction of this member of the Stat family of proteins with a specific component of the type I IFNR. . The anti-phosphotyrosine monoclonal antibody (4G-10) was obtained from Upstate Biotechnology (Lake Placid, NY). A rabbit polyclonal antibody (IFN␣RC-1) against a synthetic peptide corresponding to the sequence DESESKTSEELQQDFV present in the C terminus of the ␣ subunit was raised in rabbits. The rabbit polyclonal antibody (IFN␣RC-2) against the subunit cloned by Novick et al. (9) (␤ s subunit) was raised against a synthetic peptide corresponding to the sequence SSWDYKRASLCPSD present in the C terminus of this protein. This sequence is not present in the form of the ␤ subunit (␤ L subunit) cloned by Domanski et al. (7). A polyclonal antibody against Stat-2 (p113, 186 -199) has been described elsewhere (6) and was used for immunoprecipitations. A polyclonal antibody raised against a peptide corresponding to amino acids 832-851 of Stat-2 was purchased from Santa Cruz Biotechnology and was used for immunoblotting. A monoclonal antibody against the tyrosine kinase Tyk-2 was purchased from Transduction Laboratories (Lexington, KY).

Cells and Reagents-The
Immunoprecipitations and Immunoblotting-Cells were stimulated * This work was supported by a grant from the Department of Veterans Affairs (to L. C. P.). 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.
‡ Recipient of a Career Development Award from the American Cancer Society. To whom correspondence should be addressed: Division of Hematology-Oncology, Loyola University Chicago, Bldg. 112, 2160 South First Ave., Maywood, IL 60153. Tel.: 708-327-3304; Fax: 708-216-2319. 1 The abbreviations used are: IFN, interferon; ISG, interferon-stimulated gene; PAGE, polyacrylamide gel electrophoresis; IFNR, interferon receptor; Stat, signal transducer and activator of transcription; HMWC, high molecular weight complex. with the indicated amounts of different interferons for the indicated time periods. After stimulation, the cells were rapidly centrifuged in a microcentrifuge and lysed in a phosphorylation lysis buffer (0.5% Triton X-100, 150 mM NaCl, 200 M sodium orthovanadate, 10 mM sodium pyrophosphate, 100 mM sodium fluoride, 1 mM EDTA, 50 mM Hepes, 1.5 mM magnesium chloride, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, and 10 g/ml aprotinin) for 60 min at 4°C. Insoluble material was removed by centrifugation, and cell lysates were immunoprecipitated with the indicated antibodies using protein G-Sepharose (Pharmacia Biotech Inc.). After five washes with phosphorylation lysis buffer containing 0.1% Triton X-100, proteins were analyzed by SDS-PAGE and transferred onto polyvinylidene difluoride filters (Immobilon, Millipore). The residual binding sites on the filters were blocked by incubating with TBST (10 mM Tris, pH 8.0, 150 mM NaCl, 0.05% Tween 20)/10% bovine serum albumin for 1-3 h at room temperature or overnight at 4°C. The filters were subsequently incubated with the antiphosphotyrosine monoclonal antibody and developed using an enhanced chemiluminescence (ECL) kit following the manufacturer's recommended procedure (Amersham Corp.).
Affinity Cross-linking of Iodinated IFN␣ to Its Receptor-Affinity cross-linking of 125 I-IFN␣2 to its receptor using the homobifunctional cross-linker disuccinimidyl suberate was performed essentially as described previously (4,13).

RESULTS AND DISCUSSION
We initially sought to determine the specificity of the antibodies raised against the different type I IFN receptor subunits. We performed experiments in which 125 I-IFN␣2 was cross-linked to its receptor on human cells, and after cell lysis, the lysates were immunoprecipitated with the IFN␣RC-1 or IFN␣RC-2 antibodies and analyzed by SDS-PAGE. Fig. 1A shows such an experiment using the U-266 human myeloma cell line. The IFN␣RC-1 antibody immunoprecipitates the ␣ subunit of the receptor, which migrates as a doublet at 130 -140 kDa (Fig. 1A). It also co-immunoprecipitates an associated high molecular weight complex (HMWC-1) migrating at approximately 210 -230 kDa, which most likely results from an association of the ␣ subunit with the 100-kDa form of the ␤ subunit (5, 6) (Fig. 1A). The IFN␣RC-2 antibody immunoprecipitates a 70 -75-kDa complex (Fig. 1A), corresponding to the ␤ s subunit of the type I IFNR (the expected M r of the ␤ s subunit in affinity cross-linking studies is approximately 71 kDa when the M r of the cross-linked IFN␣2 molecule is taken into ac-count). This antibody also co-immunoprecipitates a 130 -140-kDa doublet corresponding to the ␣ subunit, but it does not co-immunoprecipitate the HMWC-1 complex (Fig. 1A). HMWC-1, the 130 -140-kDa (␣ subunit), and 70 -75 kDa (␤ s subunit) complexes were also detectable when total cell lysates from affinity cross-linked cells were analyzed in parallel (Fig.  1A). A 110 -120-kDa complex seen in total lysates from affinity cross-linked cells, which corresponds to the 100-kDa ␤ subunit (␤ L subunit) (5), could not be detected by any of the antireceptor antibodies studied here. A complex at approximately 180 kDa (HMWC-2) was also seen in total cell lysates and was weakly immunoprecipitated by both the IFN␣RC-1 and IFN␣RC-2 antibodies. Similar results were obtained when the Molt-4 human cell line was studied, except that the receptor complexes migrated slightly slower in these cells (approximately 15-20-kDa difference), a finding consistent with the reported variations in the mobility of type I IFNR components in different cell lines (11). Taken altogether, the results of the affinity cross-linking experiments strongly suggested that two distinct forms of the type I IFN receptor are co-expressed on the surface of human cells. The form of the receptor immunoprecipitated by the IFN␣RC-2 antibody is consistent with the previously described variant form of the type I IFNR (2,6). Further studies are required, however, to characterize the exact interactions between different receptor subunits and to establish that the form precipitated by the IFN␣RC-2 antibody corresponds to the previously described variant receptor form (2,6).
We subsequently performed studies in which cells were treated with IFN␣ and cell lysates were immunoprecipitated with the anti-receptor antibodies, analyzed by SDS-PAGE, and immunoblotted with anti-phosphotyrosine. Fig. 2, A and B, shows that the ␣ subunit of the receptor is tyrosine-phosphorylated in response to IFN␣ treatment of cells, in agreement with our previous findings using a monoclonal antibody against this subunit (14,15). In addition, the IFN␣RC-1 antibody coimmunoprecipitated an interferon-dependent tyrosine-phosphorylated protein with an M r of 135 kDa, corresponding to the phosphorylated form of the tyrosine kinase Tyk-2 (16). Immunoblotting of anti-IFN␣RC-1 immunoprecipitates with a monoclonal anti-Tyk-2 antibody demonstrated that Tyk-2 is associated with the ␣ subunit of the type I IFNR prior to and after IFN␣ stimulation (Fig. 2C), confirming the findings of a previous study (16) that had established an association of the ␣ subunit with Tyk-2. Tyk-2 was not detectable in immunoprecipitates obtained with the anti-IFN␣RC-2 antibody (Fig. 2C), suggesting that this kinase does not associate with the ␤ s subunit of the receptor. Fig. 3A shows an experiment in which cell lysates from IFN␣-treated cells were immunoprecipitated with the IFN␣RC-2 antibody and immunoblotted with antiphosphotyrosine. A band corresponding to the 51-kDa ␤ s subunit could not be detected in such immunoblots, perhaps because it co-migrates with the heavy chain of rabbit immunoglobulin. Also no bands migrating at 102 kDa that would correspond to a phosphorylated receptor dimer were detectable. A 113-kDa tyrosine-phosphorylated protein, however, was clearly co-immunoprecipitated by this antibody upon treatment of cells with IFN␣. As the M r of this protein was identical to the M r of the transcriptional activator Stat-2, we sought to determine whether it corresponds to Stat-2. Fig. 3B shows an anti-Stat-2 immunoblot on immunoprecipitates obtained with the IFN␣RC-1 or IFN␣RC-2 antibodies. Stat-2 is not present in IFN␣RC-1 immunoprecipitates, but it is clearly detectable in IFN␣RC-2 immunoprecipitates from IFN␣treated cells. Thus, Stat-2 appears to specifically associate with the ␤ s but not the ␣ subunit of the type I IFNR. The kinetics of the association of Stat-2 with the ␤ s subunit were subsequently studied. Fig. 4A shows an experiment in which Daudi cells were treated for different times with IFN␣, and after cell lysis, the lysates were immunoprecipitated with the IFN␣RC-2 antibody and immunoblotted with ␣Stat-2. IFN␣-dependent association of Stat-2 with the ␤ s subunit was detectable within 1 min of treatment of cells; the signal peaked at 5-30 min and decreased, although it was still clearly detectable after 90 min of IFN␣ treatment. When the time course of phosphorylation of the ␤ s subunit-associated form of Stat-2 was studied, we noticed that the signal peaked at 5-30 min and diminished by 90 min of IFN␣ treatment (Fig. 4B). When the tyrosine phosphorylation of Stat-2 directly immunoprecipitated by an ␣Stat-2 antibody was studied, the signal was more intense at all times but also declined at 90 min (Fig. 4B).
We have previously shown that different type I IFNs induce tyrosine phosphorylation of a common set of signaling proteins, including the ␣ and ␤ (100 kDa) subunits of the type I IFNR (14,15), the Tyk-2 and Jak-1 kinases (15), Stat-2 and Stat-1 (15), p95 vav (17), and insulin receptor substrate (IRS) proteins (18). 2 These data have suggested that all type I IFNs activate common signaling cascades. However, differences among the signaling pathways of different type I IFNs also exist, as suggested by our finding that IFN␤ selectively phosphorylates p100, a protein that associates with the ␣ subunit of the type I IFNR (15). To determine whether different IFNs induce an  1 and 2). B, ␣Stat-2 immunoblot. U-266 cells (4.2 ϫ 10 7 /lane) were treated for 5 min with 10 4 units/ml IFN␣ as indicated, and cell lysates were immunoprecipitated with the IFN␣RC-1 antibody (lanes 1 and 2) or the IFN␣RC-2 antibody (lanes 3 and 4) or preimmune rabbit serum (PIRS) (lane 5). association of Stat-2 with the ␤ s subunit, Daudi cell lysates were immunoprecipitated with the IFN␣RC-2 antibody and immunoblotted with anti-phosphotyrosine or ␣Stat-2. Association of the phosphorylated form of Stat-2 with the ␤ s subunit was clearly inducible during treatment of cells with IFN␤ or IFN (Fig. 5, A and B). In contrast, IFN␥ failed to induce such an association (Fig. 5, A and B), a finding consistent with the lack of involvement of Stat-2 in IFN␥ signaling (19). Interestingly, no protein of the size of the IFN␤-specific p100 protein was seen in IFN␣RC-2 immunoprecipitates (Fig. 5A), suggesting that this protein specifically associates with the ␣ but not the ␤ s subunit. On the other hand, p100 was clearly detectable in IFN␣RC-1 immunoprecipitates from IFN␤-treated cells, 3 in agreement with our original observation (15).
Significant progress has been made recently on our understanding of the mechanisms of activation of the ISGF3 transcriptional activator during IFN␣ treatment of target cells. During IFN␣ stimulation, three Stat proteins (Stat-2, Stat-1␣, and Stat-1␤) are phosphorylated on tyrosine and associate with a 48-kDa protein (ISGF3␥) to form an active complex (19 -22). This complex translocates to the nucleus and activates gene transcription during binding to interferon-stimulated response elements (19). The exact mechanisms, however, by which Stat proteins interact with IFN-dependent Jak kinases to act as substrates for their kinase activity remain unknown.
In the current report we present evidence that Stat-2 specifically associates with the subunit of the type I IFN receptor cloned by Novick et al. (9) (␤ s subunit). This association is rapid, is induced by various type I but not type II IFNs, and has similar kinetics with the IFN␣-induced phosphorylation of Stat-2. A previous study (10) had suggested that Stat-2 may associate with the the type I IFN receptor, as evidenced by the weak co-precipitation of 125 I-IFN␣2 cross-linked complexes by an ␣Stat-2 antibody. By using this methodology (affinity crosslinking), however, it is not possible to distinguish the specific receptor component that interacts with Stat-2 nor is it possible determine whether such an association is IFN␣-dependent. The results of our studies demonstrate that the association of Stat-2 with the type I IFNR is IFN␣-dependent, occurs specifically with the ␤ s but not the ␣ subunit, and appears to be of relatively high stoichiometry as evidenced by the intensity of the detected signal.
Our findings also provide some hints on the kinase activity responsible for Stat-2 phosphorylation. Colamonici et al. (10,16) have reported that the ␣ subunit of the receptor forms a complex with the tyrosine kinase Tyk-2, a finding confirmed by us using the IFN␣RC-1 antibody. Novick et al. (9) used an antibody that apparently detects both forms of the ␤ subunit (51 and 102 kDa) and were able to demonstrate an association with the tyrosine kinase Jak-1. As Stat-2 appears to interact specifically with the ␤ s but not the ␣ subunit of the receptor, it is tempting to hypothesize that Stat-2 acts as a specific substrate for Jak-1 but not Tyk-2. Furthermore, as the Stat-2-IFNR association is IFN␣-dependent, it is possible that it involves binding of the SH2 domain of Stat-2 to the ␤ s subunit of the type I IFNR. Such a model for an interaction of Stat-2 with the type I IFNR would be also consistent with the findings of a recent study that demonstrated that the SH2 domain of Stat-2 is the determinant of signaling specificity, while Tyk-2 is not specifically required for Stat-2 phosphorylation (23). It remains to be determined whether Stat-1 also utilizes components of the type I IFNR for its interaction with Jaks. Interestingly, a recent study has demonstrated that phosphorylation of Stat-2 is required for activation of Stat-1, but not vice versa, suggesting that one binding site necessary for activation of Stat-1 may be the phosphotyrosine of Stat-2 itself (24). Taken together with our data, these findings raise the possibility that binding of Stat-2 to the ␤ s subunit of the type I IFNR is the critical event required for the formation of the ISGF3␣ complex and downstream transcription of ISGs.