A Region of the β Subunit of the Interferon α Receptor Different from Box 1 Interacts with Jak1 and Is Sufficient to Activate the Jak-Stat Pathway and Induce an Antiviral State*

Coexpression of the α and βL subunits of the human interferon α (IFNα) receptor is required for the induction of an antiviral state by human IFNα. To explore the role of the different domains of the βL subunit in IFNα signaling, we coexpressed wild-type α subunit and truncated forms of the βL chain in L-929 cells. Our results demonstrated that the first 82 amino acids (AAs) (AAs 265–346) of the cytoplasmic domain of the βL chain are sufficient to activate the Jak-Stat pathway and trigger an antiviral state after IFNα2 binding to the receptor. This region of the βL chain, required for Jak1 binding and activation, contains the Box 1 motif that is important for the interaction of some cytokine receptors with Jak kinases. However, using glutathioneS-transferase fusion proteins containing amino- and carboxyl-terminal deletions of the βL cytoplasmic domain, we demonstrate that the main Jak1-binding region (corresponding to AAs 300–346 on the β subunit) is distinct from the Box 1 domain (AAs 287–295).

Type I interferons (IFNs) 1 bind to a multimeric receptor (type I IFN-R, IFN␣R, or IFN␣␤R) composed of at least two subunits designated as ␣ (1)(2)(3)(4)(5), and ␤ (6, 7). The IFN␣/␤R cDNA (8) corresponds to the short form of the ␤ subunit (␤ S ), whereas cDNAs for the long form (␤ L ) have been recently isolated by two different groups (9,10). Transcripts and mature proteins for both ␤ S and ␤ L are found in most cell lines, although there is a large excess of ␤ L protein when compared with ␤ S . 2 Lutfalla et al. (10) have also shown that the short and long forms of the ␤ subunit are generated by alternative splicing of the same gene. Interestingly, the type I IFN-binding proteins are not limited to the cellular components described above, because we and others have found a vaccinia virus-encoded protein capable of binding type I IFNs as well as blocking type I IFN signaling (11,12).
Coexpression of the ␣ subunit with either ␤ L or ␤ S in mouse L-929 cells reconstitutes the high affinity receptor, whereas either form of the ␤ chain expressed independently binds type I IFNs with low affinity. Additional studies using these transfectants show that the IFN␣ response can be reconstituted in mouse cells only when the ␣ chain is coexpressed with the ␤ L subunit rather than ␤ S , indicating that both the ␣ and ␤ L subunits are indispensable for IFN␣ signaling (9). Similarly, only the long form of the ␤ subunit is able to complement the mutant cell line U5A that is defective in the IFN␣ pathway (10).
IFNs, cytokines, and growth factors activate the Jak-Stat pathway; more specifically, type I IFNs activate Jak1, Tyk2, Stat1, Stat2, and Stat3 (for recent reviews, see Refs. 13 and 14). Ligand binding to the cytokine receptor subunits activates tyrosine kinases of the Jak family that are constitutively associated with different cytokine receptors. It has been demonstrated that Jak2 binds to the Box 1 motif present in the membrane proximal region of the erythropoietin, prolactin, growth hormone, and granulocyte macrophage colony-stimulating factor receptors (15)(16)(17)(18)(19)(20)(21). In the case of the ␣ subunit of the type I IFN-R, there is no conserved Box 1 motif, and Tyk2 interacts with a distinct region different from that of Box 1 (22). Furthermore, in those cytokine receptors that activate Jak1, the region that interacts with this kinase has not been clearly defined. The presence of a Box 1 motif in the membrane proximal region of the ␤ L subunit of the type I IFN-R suggested a possible role for this motif in the binding and/or activation of Jak1 by this receptor subunit. We therefore sought to determine the role of the Box 1 and other domains of the ␤ L chain in type I IFN signaling using two strategies: (a) we stably cotransfected mouse L-929 cells with the wild-type ␣ subunit and various truncated forms of the ␤ L chain to study in vivo signaling processes; and (b) we produced GST fusion proteins with deletions of the ␤ L chain to demonstrate interactions in vitro.
Our data indicate that a minimal region of the ␤ L subunit cytoplasmic domain, specifically AAs 265-346, is required to activate Jak1 and induce an antiviral state. Using GST fusion proteins we show that this region contains a Jak1 binding domain (AAs 300 -346) that is different from the Box 1 motif (AAs 287-295). The interaction between the ␤ L chain and Jak1 is direct and does not require binding of IFN␣2 to the receptor, although ligand binding slightly increases the amount of Jak1 associated with the ␤ L chain. These results suggest that the homologous Jak kinases may interact with different motifs within the proximal region of the cytoplasmic domain of cytokine receptors.
Construction and Expression of Different Deletions of the ␤ L Subunit of Type I IFN-R in Mouse L-929 Cells-These constructs were made by PCR using proofreading Vent polymerase and primers with an early termination codon at positions 300, 346, 360, or 417 ( Fig. 1), respectively. The same forward primer (GGAAGATCTGTAAAAGTCAA-GAGAAGACTC) at the beginning of the coding sequence of ␤ L was used for all these constructs. Forward and reverse primers contained a BglII restriction site that was used for subcloning into the BamHI site of the pZipNeoSV(X) retroviral vector (constructs pZ␤ L300 , pZ␤ L346 , pZ␤ L360 , and pZ␤ L417 ). The construct in which the ␤ L chain was truncated at position 462 (pZ␤ L462 ) was made by BamHI restriction endonuclease digestion of the 4A1 clone (9) followed by subcloning into the BamHI site of the pZipNeoSV(X) retroviral vector (Fig. 1). Sequencing of the pZ␤ L462 construct revealed the addition of 13 amino acids (GSYMGH-PAPCKLP) before a termination codon in the pZipNeo vector. PCR mutations were confirmed by sequencing. Expression of the ␣ and ␤ L subunit constructs in mouse L-929 cells was achieved by cotransfection of the different ␤ L constructs and the pR4.IFN␣R␣ construct containing the ␣ subunit (cell lines designated as LpZR; i.e. LpZR␣␤ L360.4 ) or transfection of the ␤ L subunit constructs into the LpR␣ cell line that corresponds to L-929 cells stably transfected with the pR4.IFN␣R␣ construct (cell lines designated as LpRZ; i.e. LpRZ␣␤ L346.4 ). Transfectants were selected in medium containing G418 (500 g/ml) and hygromycin B (500 g/ml), and individual clones were isolated and screened for receptor expression using affinity cross-linking and binding methods (see below).
Production of GST Fusion Proteins-The GST␣ 465-557 encoding amino acids 465-557 of the ␣ subunit of the type I IFN-R has been described previously (3). The constructs for the cytoplasmic domain of the ␤ S and ␤ L forms of the ␤ subunit were made by PCR amplification using the same forward primer (AAATGGATTGGTTATATATGC) for the common region of the cytoplasmic domain and a specific reverse primer for ␤ S and the SP6 primer for ␤ L . The PCR products were digested with BamHI and subcloned into pGEX-KG vector to generate GST␤ S spanning the whole cytoplasmic domain of ␤ S (amino acids 265-331) (8) and GST␤ L265-462 corresponding to residues 265-462 of ␤ L (9). The GST␤ L265-515 construct corresponding to the whole cytoplasmic domain of ␤ L was generated by SalI/HindIII digestion of the PCR product and subcloned into the XhoI/HindIII restriction sites of pGEX-KG. The GST␤ L265-462 construct was digested with SacI, StuI/SmaI, and NcoI to generate GST␤ L265-375 , GST␤ L265-346 , and GST␤ L265-299 , respectively ( Fig. 1). GST␤ L300 -515 was produced by deletion of an NcoI restriction fragment (from the NcoI site in the pGEX-KG cloning site to the NcoI site at position 959 of the ␤ L chain) from the GST␤ L265-515 construct. The GST fusion proteins were produced in BL-21 cells (Novagen) and purified by affinity chromatography on glutathione-Sepharose (Pharmacia). 5-10 g of the indicated GST fusion proteins were used for precipitations.
Production of Recombinant Jak1 Proteins Using in Vitro Translation Systems-Jak1 cDNA was subcloned into the pGEM vector under the control of the T7 promoter, and [ 35 S]methionine-labeled protein was produced using rabbit reticulocyte (Novagen) or wheat germ (Promega) in vitro translation kits.
Radioiodination of Type I IFNs, Competitive Displacements, and Affinity Cross-Linking-Radioiodination of IFN␣2, competitive displacement assays, and affinity cross-linking procedures were performed as described previously (2).
Electrophoretic Mobility Shift Assay (EMSA)-Whole cell extracts were prepared as described previously (24) and analyzed by EMSA using end-labeled IFN-stimulated response element and m67SIE oligonucleotides to detect ISGF3 and c-sis-inducible factor complexes (24).

Expression and Characterization of L-929 Transfectants Ex-
pressing Truncation of the ␤ L Subunit-To determine the biological relevance of disrupting the interactions of the ␤ L subunit with the IFN␣ signaling machinery, we made constructs in which the ␤ L chain was truncated at amino acids 462, 417, 360, 346, or 300, respectively (Fig. 1). These ␤ L chain constructs were coexpressed with the wild-type ␣ subunit in mouse L-929 cells, and stable transfectants were selected with G418 and hygromycin B. We characterized the cell surface expression of the different human type I IFN-R chains by cross-linking 125 Ilabeled IFN␣2 to the receptor, followed by immunoprecipitation with specific antibodies against the ␣, ␤ L , or ␤ S subunits. Fig. 2 shows that the anti-␣ subunit antibody detects the ␣ subunit and a high molecular weight complex (lanes 1, 4, 7, 10, 13, and 16) that includes the ␣ and ␤ L chains (9,25). The anti-␤ L subunit antibody immunoprecipitates the ␤ L chain (Fig. 2, ‫,ء‬ lanes 3, 6, 9, 12, 15, and 18) and the high molecular weight complex containing the ␣ and ␤ L chains (arrow) and also coprecipitates the ␣ subunit as described previously (25). The signal corresponding to the ␣ subunit is similar in most transfectants, suggesting that relatively equivalent amounts of high affinity receptors are expressed in the different transfectants. The differences in the intensity of the bands observed in immunoprecipitates performed with the anti-␤ L serum are likely due to the low efficiency of cross-linking of IFN␣2 to this subunit (9) and to the fact that different epitopes recognized by this serum may not be equally exposed in the different mutants. As expected, a specific rabbit serum raised against the carboxyl-terminal region of the ␤ S chain failed to react with any of the complexes (lanes 2, 5, 8, 14, and 17). In human U-266 cells (positive control), ␤ S cross-linked to IFN␣2 was observed in long exposures of the autoradiograms after immunoprecipitation with the anti-␤ S sera, confirming that the ␤ S chain is expressed at much lower levels than ␤ L in this cell line (compare lanes 11 and 12). The anti-␤ L serum detects all truncated forms of the ␤ L chain, except the most proximal at amino acid 300. However, the expression of this truncated form of ␤ L can be detected when cell lysates obtained after affinity crosslinking are directly resolved in an 8% SDS-PAGE (Fig. 2B). Binding of radioiodinated IFN␣2 to all transfectant cell lines was specifically blocked by the IFNaR␤1 monoclonal antibody that recognizes the extracellular domain shared by both forms of the ␤ subunit (Fig. 2B, lane 1 and data not shown). These results indicate that these truncated forms of the ␤ L chain are expressed correctly on the cell surface.
No significant difference in the K d for high affinity receptors was observed among the various cell lines (Table I). The K d for high and low affinity receptors ranged from 22-91 pM and 0.5-5 nM, respectively. Most cell lines expressed between 2,400 and 4,600 high affinity sites/cell, except for LpRZ␣␤ L346.4 , which expressed slightly lower numbers (780 sites/cell). However, it has been previously demonstrated that stable transfectants expressing low numbers of high affinity receptors display a complete antiviral effect in response to human IFN␣2 and human IFN␤ (9).
The First 82 Amino Acids of the Cytoplasmic Domain Are Sufficient to Elicit an Antiviral Response-We first tested the ability of human IFN␣2 to induce an antiviral state in the different transfectants. As can be seen in Table II, all of the transfectants were able to mount an antiviral response to human IFN␣2, except for L-929 cells expressing human ␤ L chain truncated at amino acid 300. The antiviral effect induced by human IFN␣2 in transfectants expressing ␤ L chain truncated at residues 462, 417, 360, or 346 was comparable to that induced by mouse IFN␣␤. Therefore, the lack of response to human IFN␣2 was not due to a defect in the signaling pathway, because these cells did respond to murine IFN␣␤. Similar results were reproduced with a second independent clone (LpZR␣␤ L300.1 ) (data not shown). These results clearly indicate that the first 82 amino acids (residues 265-346) of the cytoplasmic domain of the ␤ L subunit correspond to the minimum cytoplasmic receptor required to elicit an antiviral response.
Activation of the Jak-Stat Pathway in L-929 Transfectants Expressing Truncated Forms of the Human ␤ L Subunit-We next explored the activation (tyrosine phosphorylation) of Jak1 and Tyk2 in cells expressing ␤ L chain truncated at amino acids 417, 346, or 300, respectively. Because the ␤ L subunit associates with Jak1 (see below), we used a GST fusion protein encoding the cytoplasmic domain of this chain to precipitate Jak1. Fig. 3 (center panel) shows that tyrosine phosphorylation of Jak1 is observed in cells expressing ␤ L subunit truncated at residue 417 or 346 but not when the ␤ L chain was truncated at amino acid 300 (␣␤ L300 ). Stripping and reblotting of the filter with an anti-Jak1 antibody demonstrated that similar amounts of Jak1 were precipitated by GST␤ L in all cell lines (Fig. 3, lower panel). Similar results were obtained when precipitations were performed with anti-Jak1 antibodies (data not shown). A tyrosine-phosphorylated protein with slower electrophoretic mobility than Jak1 was precipitated in some experiments by GST␤ L , but not by anti-Jak1 antibodies, in ␣␤ L346 cells (Fig. 3). The nature of this protein remains to be elucidated. We also studied tyrosine phosphorylation of Tyk2 in the different transfectants in response to IFN␣2. Activation of Tyk2 was consistently detected in cells expressing ␤ L chain truncated at amino acids 417 and 346, and in some experiments very low levels of Tyk2 phosphorylation were also observed in ␣␤ L300 cells (data not shown). Murine IFN␣␤ was able to induce Jak kinase phosphorylation in all these stable transfectants (data not shown).
We also performed EMSA using probes for the IFN-stimulated response element and the m67SIE elements. These probes allowed us to study the activation of Stat1, Stat2, and Stat3 factors in response to human IFN␣2 in L-929 transfectants carrying deletions of the cytoplasmic domain of the ␤ L chain. Fig. 4A (upper panel) shows that ISGF3 activation is induced by human IFN␣2 (hu␣), human IFN␤ (hu␤), and murine IFN␣␤ (mu␣␤) in L-929 transfectants expressing ␤ L subunit truncated at residues 462 and 346 (lanes 2-4 and 6 -8, respectively), but not when this chain is truncated at amino acid 300 (Fig. 4B, lane 2). However, ISGF3 is induced when ␣␤ L300 cells are treated with mouse IFN␣␤ (Fig. 4C, lanes 2-4),

FIG. 2. Characterization of L-929 transfectants expressing wild-type ␣ subunit and truncations of ␤ L .
Affinity cross-linking was performed as described previously (2). A, cells were lysed, and lysates were immunoprecipitated with the specific rabbit sera ␣ 511-557 (lanes 1, 4, 7, 10, 13, and 16), ␤ S (lanes 2, 5, 8, 11, 14, and 17), and ␤ L265-515 (lanes 3, 6, 9, 12, 15, and 18) against the ␣, ␤ S , or ␤ L subunits, respectively. The migration of the different truncations of ␤ L ‫)ء(‬ and the migration of the high molecular weight complexes (arrowhead) formed by an association of the ␣ and ␤ L subunits are indicated. B, L-929 cells stably transfected with the ␣ and ␤ L300 constructs were cross-linked in the presence or absence of the IFNaR␤1 mAb (50 g/ml) that blocks the binding of all type I IFNs to the receptor. Cell lysates were directly analyzed by SDS-PAGE. The K d and the number of sites/cell were calculated using the computer program Ligand. In all transfectants, the two-site model was statistically more significant than the one-site model (p Ͻ 0.05). Competitive displacement of 125 I-labeled IFN␣2 by unlabeled IFN␣2 was performed as described previously (7).
indicating that the Stat pathway is functional in these cells. The ISGF3 complex is supershifted by an anti-mouse Stat2 serum (Fig. 4C, lane 3).
EMSA experiments performed with the m67SIE element present in the c-fos promoter showed that full activation of Stat1 and Stat3 also requires amino acids 265-346 of the cytoplasmic domain of the ␤ L chain (Fig. 4, A and B, lower panels, compare the induction of the different c-sis-inducible factor complexes in ␣␤ L300 and ␣␤ L346 cells). Low levels of IFN␣2-induced activation of Stat1 and Stat3 were observed in ␣␤ L300.2 cells that are resistant to the antiviral effect of IFN␣2 (Fig. 4B, lower panel, lane 2). The c-sis-inducible factor complexes containing Stat1 and Stat3 were induced by murine IFN␣␤ and supershifted by anti-Stat1 and anti-Stat3 serum (Fig. 4C, lanes 7 and 9).
The ␤ L Chain Specifically Associates with Jak1-To test for a possible interaction between the ␤ subunit and Jak1, we produced GST fusion proteins encoding the entire cytoplasmic domain of the ␤ L and ␤ S chains. These GST fusion proteins were used to precipitate lysates from U-266 cells treated with IFN␣2 or left untreated. Precipitates were resolved by SDS-PAGE and then subjected to immunoblotting with anti-Jak1 antibodies. Fig. 5A shows that Jak1 is precipitated with the ␤ L265-515 fusion protein and the control anti-Jak1 antibody before and after IFN␣2 treatment (lanes 7 and 12). GST␤ S (lane 2), along with antibodies against Stat1, Stat2, and Tyk2, failed to detect Jak1. To define the role of the Box 1 motif in the interaction between Jak1 and the ␤ L chain we produced two GST fusion proteins: (a) one containing only the first 34 amino acids (␤ L265-299 ) of the cytoplasmic domain, which include the Box 1 motif (AAs 287-295); and (b) a second consisting of all the cytoplasmic residues except the first 34 amino acids (AAs 300 -515), thus deleting the Box 1 region. Interestingly, the ␤ L265-299 construct precipitated Jak1 at levels that were barely above background (Fig. 5A, lane 4), whereas the 300 -515 aminoterminal deletion, lacking the entire Box 1 domain, bound Jak1 roughly at the same level (80%) as the full-length protein (Fig.  5A, lanes 7 and 8). This result indicates that the binding of Jak1 to the ␤ L chain requires a region carboxyl-terminal to the Box 1 motif rather than Box 1 itself. To further define the Jak1 binding site, we produced GST fusion proteins encoding truncated forms of the cytoplasmic domain. GST fusion proteins encoding the ␤ L chain truncated at residues 462, 375, or 346 all bound Jak1 at comparable levels (Fig. 5A, lanes 3, 5, and 6), suggesting that the primary Jak1 binding site resides between residues 300 -346.
To determine whether the interaction between Jak1 and the ␤ L chain could be detected in vivo, lysates obtained from U-266 cells treated with IFN␣2 or left untreated were immunoprecipitated with antibodies against the different receptor subunits and Jak1. Fig. 5B shows that an association between the  1-4 and 5-8, respectively). Cells were treated with 10,000 units/ml murine IFN␣␤ (lanes 2 and 6), human IFN␤ (lanes 3 and 7), and human IFN␣2 (lanes 4 and 8) or left untreated (lanes 1 and 5) for 20 min at 37°C. EMSA was performed using IFN-stimulated response element and m67SIE probes (upper panel and lower panels, respectively) (24). B, whole cell extracts were obtained from mouse L-929 cells stably expressing wild-type ␣ subunit and ␤ L chain truncated at amino acids 300 or 346 (lanes 1 and 2 and  lanes 3 and 4, respectively). Cells were treated with human IFN␣2 as described in A. EMSA was performed using IFN-stimulated response element and m67SIE probes (upper panel and lower panels, respectively). C, ␣␤ L300 cells were treated with murine IFN␣␤ (ϩ) or left untreated (Ϫ). Supershifts of the ISGF3 complex were performed with an anti-Stat2 serum (Santa Cruz Laboratories; lane 3) and with anti-Stat1 and anti-Stat3 sera for the c-sis-inducible factor complexes (lanes 7 and 9). Normal rabbit serum was used as a negative control (lanes 4 and 8). The migration of the different complexes is indicated. a Cytopathic effect assay was performed using a 1:25,000 dilution of encephalomyocarditis virus stock that produced 100% cytophatic effect in 24 h. The data shown represent the amount of the respective IFN that was able to inhibit cytopathic effect by 50% (23). The S.D. in these assays was always below 20%.
␤ L chain and Jak1 can be detected in lysates of control and IFN-treated U-266 cells. Interestingly, coprecipitation of Jak1 with the ␣ subunit was only detected after binding of IFN␣2 (Fig. 5B) or IFN␤ (data not shown) to the receptor, suggesting that ligand binding alters interactions among the components of the receptor complex. The ␤ L chain was not detected in anti-Jak1 immunoprecipitates after reprobing the blot a second time with an anti-␤ L serum. This is most likely due to the lower sensitivity of this serum, compared with that of the Jak1 mAb, in Western blots (data not shown).
Finally, to determine whether the interaction between Jak1 and the ␤ L chain required additional proteins or phosphorylation, we performed experiments using an in vitro translation system. The Jak1 protein was produced using a wheat germ in vitro translation kit, labeled with [ 35 S]methionine, and precipitated with the appropriate GST fusion proteins or control antibodies. Fig. 5C shows that ␤ L265-515 , but not GST␣ or GST␤ S , precipitates Jak1 produced in the cell-free system. Because the wheat germ system should not contain adaptor proteins and no tyrosine phosphorylation of the GST fusion proteins or Jak1 was detected by either anti-phosphotyrosine immunoblotting or in vitro kinase assays with [␥-32 P]ATP (data not shown), these results strongly suggest a direct interaction between Jak1 and the ␤ L subunit of the type I IFN-R.

DISCUSSION
The IFN␣ system is similar to many other cytokine systems in that ligand binding activates kinases of the Jak family and transcription factors of the Stat family. To determine the contribution of the ␤ subunit of type I IFN-R in activating the Jak-Stat pathway, we expressed truncated forms of the IFN␣R␤ L chain in mouse L-929 cells. Our results from in vivo experiments demonstrate that a minimum receptor containing only the 82 proximal amino acids of the cytoplasmic domain (AAs 265-346) can activate Tyk2 and Jak1 kinases and Stat1, Stat2, and Stat3 transcription factors and elicit an antiviral response in L929 cells after stimulation with IFN␣2. Because Tyk2 associates with the ␣ subunit (3), it is likely that Jak1 interacts with the other component of the receptor, the ␤ subunit. The proximal region of the ␤ L chain contains a Box 1 motif, which is essential for binding and activation of Jak2 by single-chain cytokine receptors (15-18, 26 -34). Therefore, we wished to determine whether binding of homologous Jak1 is localized to the Box 1 site (AAs 287-295) or other discrete sites in the cytoplasmic region of the ␤ L subunit.
We show that Jak1 from several different sources binds to different GST␤ L chain fusion proteins and that the majority of the binding is localized to a region (AAs 300 -346) carboxylterminal to the Box1 domain, whereas only a minimal interaction was observed for the Box1 domain. This corresponds directly with our in vivo data showing that cells expressing the ␤ L300 construct, which contains the Box1 region (but excludes the major Jak1 binding site), do not activate Jak1 in the presence of IFN␣2. Similarly, previous reports showed that mutation of residues in the Box 1 domain of the interleukin 2 receptor ␤ chain do not significantly effect activation of Jak1 (35). Our data indicate that the primary Jak1-binding region corresponding to residues 300 -346 on the ␤ L chain is required for activation of the Jak-Stat pathway and induction of the antiviral response; however, it is possible that the Box 1 site is also necessary for activation of Jak1. For example, the minimal binding interaction between Jak1 and the fusion protein containing only Box 1 may result from noncontiguous binding of Jak1 at both Box 1 and AAs 300 -346 sites or an extended Jak1-binding site that encompasses a region extending from Box 1 to AA 346 on the ␤ subunit. The data presented herein for Jak1 and previous studies for Jak2 suggest that the homologous Jak kinases interact differently with their specific receptor subunits. For example, Jak2 requires the Box 1 motif for binding and, in addition, a less-conserved region distal to Box 1, designated as Box 2, that seems to be necessary in most cases for full kinase activation (15-18, 26 -34). In contrast, we have observed that for the IFN␣R␤ L chain, Jak1 binding and activation relies on a site carboxyl-terminal to Box 1. Similarly, an equivalent region designated in some cases as Box 2 (i.e. interleukin 2 receptor ␤, gp130, and leukemia inhibitory factor receptor) has been reported to be necessary for Jak1 activation by other cytokines (34 -36). Experiments are currently under way to establish the role of the distal part of the ␤ L chain in IFN␣ signaling. FIG. 5. Jak1 interacts with the ␤ L chain. U-266 cells treated with IFN␣2 or left untreated were used as a source of Jak1. A, cell lysates were precipitated with the various GST fusion proteins (5 g) or indicated control antibodies, resolved by SDS-PAGE, transferred to polyvinylidene difluoride, and blotted with anti-Jak1 monoclonal antibody. B, the ␤ L chain interacts with Jak1 in vivo. Cell lysates obtained from U-266 cells treated with 20,000 units/ml IFN␣2 or left untreated were immunoprecipitated with 5 l of sera directed toward the indicated regions of the receptor or normal rabbit serum (NR), transferred to polyvinylidene difluoride, and immunoblotted with an anti-Jak1 mAb (1:1000 dilution). C, direct interaction between the ␤ L subunit and Jak1. Jak1 was produced in an in vitro translation assay. Proteins were labeled with [ 35 S]methionine and then precipitated with the indicated GST fusion proteins or a Jak1 antibody. Precipitated proteins were resolved by SDS-PAGE, transferred to polyvinylidene difluoride, and exposed to x-ray film overnight. No tyrosine phosphorylation of Jak1 or the GST fusion proteins was detected when Jak1 was produced by this method. It is possible that Jak inhibitors are present in the in vitro transcription/translation reactions or that the low amounts produced are not sufficient for transphosphorylation resulting in kinase activation observed when this kinase is produced in other systems (i.e. baculovirus).