Chimeric erythropoietin-interferon gamma receptors reveal differences in functional architecture of intracellular domains for signal transduction.

Binding of interferon gamma (IFN-γ) causes oligomerization of the two interferon γ receptor (IFN-γR) subunits, receptor chain 1 (IFN-γR1, the ligand-binding chain) and the second chain of the receptor (IFN-γR2), and causes activation of two Jak kinases (Jak1 and Jak2). In contrast, the erythropoietin receptor (EpoR) requires only one receptor chain and one Jak kinase (Jak2). Chimeras between the EpoR and the IFN-γR1 and IFN-γR2 chains demonstrate that the architecture of the EpoR and the IFN-γR complexes differ significantly. Although IFN-γR1 alone cannot initiate signal transduction, the chimera EpoR/γR1 (extracellular/intracellular) generates slight responses characteristic of IFN-γ in response to Epo and the EpoR/γR1·;EpoR/γR2 heterodimer is a fully functional receptor complex. The results demonstrate that the configuration of the extracellular domains influences the architecture of the intracellular domains.

The interferon ␥ (IFN-␥) 1 receptor complex consists of at least two receptor components, a ligand binding chain and a signal transducing chain, each of which is a member of the class II cytokine receptor family (1,2). Isolation of the two chains of the interferon ␥ receptor (IFN-␥R) has permitted an analysis of the contributions of each to the signal transduction mechanism. The first chain of the receptor (IFN-␥R1) binds ligand (3)(4)(5)(6)(7)(8)(9). The second chain of the receptor (IFN-␥R2) does not bind ligand by itself but is required for signal transduction (3, 10 -16). A large body of experiments has elucidated the involvement of the Jak-Stat pathway in signaling by various cytokines (for reviews, see Refs. [17][18][19][20][21][22]. The Janus kinases (or Jaks) are a family of receptor-associated soluble tyrosine kinases with four known members, Tyk2, Jak1, Jak2 and Jak3. Two of the kinases, Jak1 and Jak2, are required for signal transduction by IFN-␥. Further analyses of the interactions have shown that the IFN-␥R1 chain binds Jak1 (16,23,24) and the intracellular domain of the IFN-␥R2 chain brings Jak2 into the signal transduction complex (16). Upon binding of the ligand, IFN-␥, to the IFN-␥R1 chain, activation of Jak1 and/or Jak2 by reciprocal transphosphorylation causes the phosphorylation of IFN-␥R1 (16,25). Stat1␣, a latent cytoplasmic transcription factor (26), binds to the phosphorylated IFN-␥R1, undergoes tyrosine phosphorylation (27), and forms homodimers that translocate to the nucleus and initiate transcription of IFN-␥ inducible genes (for reviews see Refs. 17 and 21).
As with other cytokine receptors, oligomerization upon ligand binding is the first step in the signaling cascade of IFN-␥. IFN-␥ is a non-covalent symmetrical homodimer (28) that binds to IFN-␥R1 with a stoichiometry of 1:2 (29,30). It is known that a species-specific interaction between the extracellular domains of the IFN-␥R1 and IFN-␥R2 subunits is essential for signaling (10 -12, 31-33). The IFN-␥R2 subunit does not by itself bind the ligand, but can be cross-linked to IFN-␥ when both IFN-␥R1 and IFN-␥R2 chains are present (16). Several lines of evidence (16,34) suggest that the IFN-␥ signaling complex contains two IFN-␥R1 chains, two IFN-␥R2 chains and one IFN-␥ homodimer.
The erythropoietin (Epo) receptor, EpoR, is a member of the class I cytokine receptor subfamily. A single chain encodes both ligand-binding and signal-transducing functions. Epo induces homodimerization of the receptor to initiate signal transduction (for reviews, see Refs. 18, 19, and 35). Jak2 is associated with the cytoplasmic domain of the EpoR and is activated upon ligand-induced dimerization of the receptor (36). Strikingly an Arg 3 Cys mutation in the extracellular domain of EpoR results in ligand independent dimerization/oligomerization and constitutive, ligand-independent activation of Jak2 and mitogenesis (37,38).
In this study we used chimeric EpoR, IFN-␥R1, and IFN-␥R2 constructs to investigate the differences between the architecture of Epo and IFN-␥ receptor complexes and shed light on the requirement for one or two receptor-associated tyrosine kinases and the necessity for one or two distinct transmembrane chains for effective signal transduction. Cells and Media-CHO-B7 cells represent the Chinese hamster ovary cell line (CHO-K1) containing a transfected human HLA-B7 gene (12). The 16-9 hamster ϫ human somatic hybrid cell line is a CHO-K1 derivative containing a translocation of the long arm of human chromosome 6 and the human HLA-B7 gene (13). These cells were main-tained in Ham's F-12 medium (Life Technologies, Inc.) containing 10% heat-inactivated fetal bovine serum (Sigma). Transfections were carried out with the DOTAP transfection reagent (Boehringer Mannheim) according to the manufacturer's protocol and the transfected cells were maintained in F-12 medium containing 450 g/ml Geneticin (antibiotic G418). Unless otherwise noted, experiments were performed with cloned cells expressing the various receptor subunits.

Reagents, Restriction Endonucleases, and Other
Construction of Chimeric Receptors-The EpoR expression plasmid was made by cloning the EcoRI-AflIII fragment of the human EpoR cDNA p18R (39) into the EcoRI and EcoRV sites of the eukaryotic expression vector pcDNA3 (Invitrogen). The construction of plasmids expressing Hu-IFN-␥R1 and Hu-IFN-␥R2 chains from cDNA under the control of cytomegalovirus promoter has been previously described (5,12,16,40). For ease of construction of the various chimeric receptors, the polymerase chain reaction (PCR) was employed to incorporate a unique NheI site at the 3Ј end of the extracellular domain (EC) and at the 5Ј end of the transmembrane-intracellular domains (IC) of the receptors. The primers were designed to code for the three amino acids Trp, Leu, and Ala, which are commonly found in the transmembrane domain of several proteins, encompassing the NheI site. The extracellular portions of EpoR, Hu-IFN-␥R1, and Hu-IFN-␥R2, containing an NheI site (designated EpoR EC /NheI, ␥R1 EC /NheI, and ␥R2 EC /NheI) were generated by PCR from the respective cDNAs as templates with the use of the T7 primer (5Ј-TAATACGACTCACTATA-3Ј) and the internal primers 5Ј-GCCGCTAGCCAGGGGTCCAGGTCGCTAGGCG-3Ј (corresponding to nucleotides 1874 -1893 of p18R EpoR cDNA; Ref. 39 . The intracellular portions of the various receptors with the unique NheI site at the 5Ј end of the transmembrane domain (designated EpoR IC /NheI, ␥R1 IC /NheI, and ␥R2 IC /NheI were generated by PCR on corresponding cDNA templates with the use of the SP6 primer (5Ј-ATTTAGGTGA-CACTATA-3Ј) and the internal primers 5Ј-GTGGCTAGCGACGCTC-TCCCTCATCCTCG-3Ј (corresponding to nucleotides 1902-1921 of plasmid p18R), 5Ј-GTGGCTAGCGATTCCAGTTGTTGCTGCTTTAC-3Ј (corresponding to nucleotides 792-814 of the Hu-IFN-␥R1 cDNA), and 5Ј-GTGGCTAGCGATCTCCGTGGGAACATTT-3Ј (corresponding to nucleotides 1398 -1416 of the Hu-IFN-␥R2 cDNA). The NheI site in each primer is underlined. The PCR products encoding the extracellular domains were incubated with T4 DNA polymerase and dNTPs to generate blunt ends; then the PCR fragments, which contained the vector multiple cloning sites, were subsequently digested with the restriction endonucleases EcoRI (EpoR EC /NheI and ␥R2 EC /NheI) or BamHI (␥R1 EC /NheI), and cloned into the EcoRV and EcoRI/BamHI sites of the expression vector pcDNA3 (Invitrogen) to yield the plasmids pEpoR EC , p␥R1 EC and p␥R2 EC . Analogously, the PCR products encoding the intracellular domains of the various receptors were treated with T4 DNA polymerase to generate blunt ends, digested with XbaI restriction endonuclease, and cloned into the EcoRV and XbaI sites of pcDNA3 to yield the plasmids pEpoR IC , p␥R1 IC , and p␥R2 IC . To introduce the Stat1␣ binding site of Hu-IFN-␥R1 into the cytoplasmic domain of EpoR, two-step asymmetric PCR (detailed in Ref. 41) was carried out sequentially on Hu-IFN-␥R1 cDNA and pEpoR IC cDNA templates with vector primers and the internal primer CTTGTCCTTCTGTTTTTATT-TCagagcaagccacatagetggg. The uppercase letters denote sequences corresponding to the Hu-IFN-␥R1 cDNA, and the lowercase letters represent sequences corresponding to the EpoR cDNA. The Hu-IFN-␥R2 chain with the Stat1␣ binding site of Hu-IFN-␥R1 was constructed by restriction enzyme digestion of p␥R2 IC and IFN-␥R1 cDNA with BspEI and AvaI, respectively, followed by ligation. For construction of the chimeric receptors, plasmids encoding the suitable extracellular or intracellular domains were digested with NheI and XbaI restriction endonucleases and ligated together. All constructs were sequenced for verification of the entire nucleotide sequence of the receptor. Sequencing was done in an Applied Biosystems model 373 automated DNA sequencer with dideoxy dye-terminator chemistry.
Electrophoretic Mobility Shift Assays (EMSA)-EMSAs were performed with the 22-base pair sequence containing a Stat1␣ binding site (5Ј-GATCGATTTCCCCGAAATCATG-3Ј) corresponding to the GAS element in the promoter region of the human IRF-1 gene (42). Two oligonucleotides, 5Ј-GATCGATTTCCCCGAAAT-3Ј and 5Ј-CATGATT-TCGGGGAAATC-3Ј, were annealed by incubation for 10 min at 65°C, 10 min at 37°C, and 10 min at 22°C, and labeled with [␣-32 P]dATP and the Klenow fragment of DNA polymerase I by the filling-in reaction (43). Whole cell extracts were prepared as follows (44). Cells were grown to confluence in six-well plates, and harvested by scraping in ice-cold phosphate-buffered saline. Cells from each well were washed with 1.0 ml of cold phosphate-buffered saline, pelleted, and resuspended in 100 l of lysis buffer (10% glycerol, 50 mM Tris⅐HCl, pH 8.0, 0.5% Nonidet P-40, 150 mM NaCl, 0.1 mM EDTA, 1 mM dithiothreitol, 0.2 mM phenylmethylsulfonyl fluoride, 1 mM Na 3 VO 4 , 3 g/ml aprotinin, 1 g/ml pepstatin, and 1 g/ml leupeptin). After 30 min on ice, the extracts were centrifuged for 5 min at full speed in a microcentrifuge and the supernatant was recovered for use in the assay and stored at Ϫ80°C.
EMSA reactions contained 2.5 l of the whole cell extracts, 1 ng 32 P-labeled probe (specific activity approximately 10 9 cpm/g), 24 g/ml bovine serum albumin, 160 g/ml poly(dI⅐dC), 20 mM HEPES, pH 7.9, 1 mM MgCl 2 , 4.0% Ficoll (Pharmacia Biotech Inc.), 40 mM KCl, 0.1 mM EGTA, and 0.5 mM dithiothreitol in a total volume of 12.5 l. For the supershift assay, 1 l of a 1:10 dilution of anti-Stat1␣ antibody was included in the reaction. Competition experiments contained a 100-fold excess of the unlabeled oligonucleotide. Reactions were incubated at 24°C for 20 min. Then 8 l of the reaction mixture was electrophoresed at 400 V for 3-4 h at 4°C on a 5% polyacrylamide (19:1, acrylamide: bisacrylamide) gel. The dried gel was exposed to Kodak XAR-5 film with an intensifying screen for 12 h at Ϫ80°C.
Antibodies-Rabbit anti-Jak1 antibody was developed against a synthetic peptide (KTLIEKERFYESRCRPVTPSC) corresponding to the end of the second kinase-like domain of murine Jak1. Rabbit anti-Stat1␣ antibody, raised against the carboxyl-terminal region of Stat1␣, was a gift from James Darnell. Rabbit anti-Jak2 antibody (catalogue no. SC-294) and rabbit anti-Stat5 antibody (catalogue no. SC-835) were from Santa Cruz Biotechnology. Monoclonal anti-phosphotyrosine antibody was purchased from Sigma (catalogue no. P3300).

Construction of Chimeric Receptors-
The schematic illustration of the various chimeric receptor molecules that were produced is shown in Fig. 1. In one set of chimeric constructs, the extracellular domain of the EpoR was spliced to the transmembrane domain and the cytoplasmic domain of each of the two IFN-␥R subunits. In the other set of chimeras, the transmembrane and intracellular domain of EpoR was fused to the extracellular domain of IFN-␥R1 and IFN-␥R2.
Class I MHC Antigen Induction-To investigate the role of the intracellular domain of IFN-␥R2 in the signal transduction complex of IFN-␥, we constructed a chimeric receptor chain consisting of the extracellular domain of IFN-␥R2 and the intracellular domain of EpoR. This chimeric construct, ␥R2/ EpoR, and the native IFN-␥R2 subunit were separately transfected into CHO-B7 as well as CHO-16-9 cells. The ability of the transfected chimeric cDNA to transduce a signal upon induction with Hu-IFN-␥ was assayed by measurement of enhanced MHC class I antigen expression in the transfected cells and by activation of Stat1␣. CHO-B7 cells transfected with IFN-␥R2 or ␥R2/EpoR cDNA showed no response to Hu-IFN-␥ as they lack the ligand-binding receptor subunit, Hu-IFN-␥R1 (data not shown). Parental CHO-16-9 cells, which contain human chromosome 6q and express the Hu-IFN-␥R1 subunit, showed no induction of MHC class I antigens in response to Hu-IFN-␥ (Fig. 2, panel A) but when stably transfected with expression vectors encoding Hu-IFN-␥R2 cDNA or ␥R2/EpoR chimera, exhibited enhanced cell surface expression of class I MHC antigens in response to Hu-IFN-␥ (Fig. 2, panels B and C). To assess how effectively the intracellular domain of EpoR could substitute for the intracellular domain of the IFN-␥R2 subunit, we measured the induction of MHC class I antigens as a function of IFN-␥ concentration. As depicted in Fig. 3, there was a slightly lower induction of MHC class I antigens in the cells containing the chimeric ␥R2/EpoR than in the cells containing the native Hu-IFN-␥R2 chain at each concentration of Hu-IFN-␥ used. Nevertheless, the fact that the EpoR intracellular domain can be substituted for the Hu-IFN-␥R2 intracellular domain shows that another sequence that can recruit Jak2 into the signal transduction complex can substitute for the intracellular domain of Hu-IFN-␥R2.
Various chimeric receptors between the EpoR and Hu-IFN-␥R1 and Hu-IFN-␥R2 subunits were constructed in order to gain an understanding of the events leading to signal transduction. CHO-16-9 cells were stably transfected with expression vectors coding for EpoR, EpoR/␥R1, EpoR/␥R2, the combination of EpoR/␥R1 and EpoR/␥R2, and ␥R1/EpoR(p91). In response to Epo, the EpoR transfectants showed no response (Fig. 2, panel D). The EpoR/␥R1 transfectants showed a slight enhancement of expression of MHC class I antigens (Fig. 2, panel E), which shows that the intracellular domain of the Hu-IFN-␥R1 chain, by itself, can recruit all the requisite components for signal transduction. At lower concentrations of Epo (less than 100 units/ml), there was little or no increased MHC class I antigen expression in these cells (Fig. 4). The transfectants containing both EpoR/␥R1 and EpoR/␥R2 chains exhibited substantial expression of MHC class I antigens (Fig. 2,  panel F; Fig. 4). Cells transfected with the expression vector coding for EpoR(p91) chimeric cDNA (EpoR with the p91 recruitment site from IFN-␥R1) respond to Epo with enhanced expression of class I MHC antigens, while the ␥R1/EpoR(p91) transfectants were unresponsive (Fig. 2, panels H and G, respectively). Furthermore, the ␥R1/␥R2(p91) receptor chain is unable to transduce a signal upon binding ligand, 2 whereas the cells expressing the EpoR/␥R2(p91) chimeric receptor exhibited enhanced class I MHC antigen expression in response to activation by Epo (Fig. 2, panel I).
Activation of Jak Kinases-IFN-␥ activates Jak1 and Jak2 kinases (46), whereas Epo activates Jak2 (36) during signal transduction. Thus, we tested the ability of the various chimeric receptors to activate Jak1 and Jak2 kinases in response to binding of ligand. Phosphorylation of Jak1 and Jak2 (Fig. 7) was examined by immunoprecipitation of cellular lysates with anti-phosphotyrosine antibodies, followed by a Western blot visualized with specific anti-Jak1 and anti-Jak2 antibodies. Both Jak1 and Jak2 were phosphorylated in response to Hu-IFN-␥ treatment in 16-9 cells expressing parental IFN-␥R2 or chimeric ␥R2/EpoR receptors. Induction with Epo phosphorylated both Jak1 and Jak2 kinase in the cell line expressing both EpoR/␥R1 and EpoR/␥R2 chains. In the cell line expressing only the chimeric EpoR/␥R1 receptor, only Jak1 kinase was phosphorylated in response to Epo. The cell line transfected with the ␥R1/EpoR chimeric receptor did not exhibit phosphorylation of either Jak1 or Jak2 kinase upon IFN-␥ treatment. Relative fluorescence values are based on the mean fluorescence of cell populations (n ϭ 10,000). The data were normalized so that the mean fluorescence intensity was adjusted to 1.0 for cells in the absence of Epo.

FIG. 5. Electrophoretic mobility shift assays of cells expressing chimeric receptors.
Clones of transfected 16-9 cells stably expressing native IFN-␥R2, chimeric IFN-␥R2/EpoR, or chimeric IFN-␥R1/EpoR(p91) receptor subunits were induced with 1,000 units/ml IFN-␥. Whole cell extracts were prepared, incubated with 32 P-labeled GAS probe, and complexes resolved by separation on 5% polyacrylamide gels (16,41) and detected by autoradiography. Competition experiments contained a 100-fold molar excess of unlabeled GAS oligonucleotide. Supershift assays were performed by the addition of 0. receptor. Depending on the ligand, this can take the form of receptor homodimers (Epo, growth hormone), heterodimers (ciliary neurotrophic factor, leukemia inhibitory factor), homotrimers (tumor necrosis factor), and more complex assemblies (reviewed in Ref. 47). In the case of IFN-␥, the oligomerization involving IFN-␥R1 and IFN-␥R2 initiates the signal transduction events: activation of Jak1 and Jak2, phosphorylation of IFN-␥R1 on Tyr-457 (16,25), followed by phosphorylation and activation of Stat1␣ (27). A major function of receptor dimerization is to bring two receptor-associated kinases together for transactivation and phosphorylation of the receptor chains. The cytoplasmic domain of the IFN-␥R2 subunit serves to bring Jak2 kinase into the signal transduction complex (16). This is a crucial event since deletion of the membrane-proximal region of the intracellular domain of the IFN-␥R2 chain, which encompasses the Jak2 association site, completely abrogates its ability to transduce signals in response to IFN-␥ (16), and cells lacking Jak2 do not respond to . This is further supported by the observation that the IFN-␥R2/EpoR chimeric receptor, which recruits Jak2, is almost as effective as the native IFN-␥R2 chain in supporting signal transduction in response to IFN-␥ (Figs. 2, 5, and 7). The IFN-␥R2 subunit is a helper receptor subunit with a Jak2 association site, but no Stat recruitment site; its intracellular domain can be substituted with the cytoplasmic domain of any receptor subunit that can bring a Jak kinase to the IFN-␥ receptor complex to support signal transduction (40).
The requirement for two distinct Jak kinases in the IFN-␥ signaling pathway was demonstrated with the use of kinasedeficient cell lines (46,48). Based on our results with the chimeric erythropoietin-interferon ␥ receptors, we propose that this reflects two features characteristic of the IFN-␥ receptor complex: the unique properties of the receptor relative to the positioning of the Jaks, and the idea that Jak1 is relatively ineffective in one or more of the following phosphorylation steps (trans-phosphorylation of itself, phosphorylation of IFN-␥R1, and activation of Stat1␣). The presence of Jak2 facilitates effective phosphorylation of the above steps. In contrast to the growth hormone receptor (49) and the EpoR (37) complexes, when one IFN-␥ homodimer binds two IFN-␥R1 molecules, the two receptor subunits do not interact with one another and are separated by 27 Å (50) at their closest point. Therefore, although the IFN-␥R1 chain possesses both a Jak1 association site and a Stat1␣ recruitment site, alone it is unable to transduce a signal on homodimerization as the two Jak1 kinases are not in physical proximity to permit transphosphorylation (Fig.   FIG. 6. Electrophoretic mobility shift assays of cells expressing chimeric receptors. Clones of transfected 16-9 cells stably expressing EpoR/␥R1, EpoR(p91), or EpoR/␥R2(p91) chimeric receptor subunits, or both EpoR/␥R1 and EpoR/␥R2 chimeric receptors were treated with erythropoietin at 100 units/ml for 15 min at 37°C. Whole cell extracts were made and the electrophoretic mobility shift assay performed. As shown in the figure, induction with Epo causes activation of Stat1␣ in cells expressing EpoR/␥R1, EpoR(p91), and EpoR/␥R2(p91), as well as in those cells expressing both EpoR/␥R1 and EpoR/␥R2. Addition of anti-Stat1␣ antibody to the reaction mixture caused the Stat1␣ complex to be shifted. 8A). Crystallographic analysis of the IFN-␥⅐IFN-␥R1 complex suggests that each monomer of the IFN-␥ homodimer binds one IFN-␥R1 and one IFN-␥R2 subunit (50). Thus the signal-transducing complex of IFN-␥ consists of the IFN-␥ homodimer bound to two IFN-␥R1 and two IFN-␥R2 chains, which recruit Jak1 and Jak2, respectively (16,40); and Jak2 phosphorylates Jak1, following which either kinase phosphorylates Tyr-457 of the IFN-␥R1 chain ( Fig. 8B; see also Refs. 25, 45, and 51). The phosphorylated segment of each IFN-␥R1 chain recruits Stat1␣, which is then phosphorylated by Jak1 or Jak2, then released to dimerize and form the active Stat1␣. In contrast, with the EpoR/␥R1 dimer, two Jak1 kinases are brought sufficiently close together to activate one another (Fig. 8C), albeit inefficiently. In the case of the EpoR/␥R1⅐EpoR/␥R2 dimer, one Jak1 and one Jak2 are in close apposition for Jak2 to phosphorylate Jak1 and initiate efficient downstream signaling events (Fig. 8D). Cells expressing the EpoR/␥R2(p91) chimeric receptor (Fig. 2I) exhibit a stronger biological response than cells expressing both EpoR/␥R1 and EpoR/␥R2 (Fig. 2F) or even the native IFN-␥ receptor (␥R2, Fig. 2B), which supports a modulating role for Jak1 in the IFN-␥R complex. In cells expressing both EpoR/␥R1 and EpoR/␥R2 chains, binding of Epo can induce the formation of three types of receptor dimers: EpoR/␥R1 homodimers, EpoR/␥R2 homodimers, and EpoR/␥R1⅐EpoR/␥R2 heterodimers. The EpoR/␥R1 homodimer is barely active (Figs. 2E and 4), and the EpoR/␥R2 homodimer is inactive. The major functional receptor complex therefore must be the EpoR/ ␥R1⅐EpoR/␥R2 heterodimer (Fig. 8D).
That Jak1 is relatively ineffective in transphosphorylation is supported by the observation that cells expressing the EpoR/ ␥R1 chimera show a smaller response than the cells expressing both EpoR/␥R1 and EpoR/␥R2 or the EpoR/␥R2(p91) chimeric receptor chains. Thus, even though homodimerization of the EpoR/␥R1 receptor by Epo brings the cytoplasmic domains of the two ␥R1 subunits into close proximity (Fig. 8C), the data of Figs. 2, 4, and 7 indicate that Jak2 is more effective at phosphorylating Jak1 than the latter is at cross-phosphorylating itself. This is consistent with the results of Briscoe et al. (52), who reported that a Jak1 molecule with an inactive kinase domain can replace the normal Jak1 in signal transduction by IFN-␥ and suggested a structural role for Jak1 in the receptor complex.
As noted above, the Jak kinases do not mediate Stat selectivity and are promiscuous in their activity; each of the Jak kinases can substitute for Jak2 in signal transduction by IFN-␥ (40). Selectivity is likely maintained from the extracellular receptor-ligand interaction to the final signal transduction mechanism by other regions of the intracellular domains. For example, Heim et al. (53) suggested that the SH2 recognition domain of Stat1␣ maintains some of the specificity. It remains to be established, however, how Stat1␣ can be activated by many different cytokines and maintain specificity through transcription. Other molecules that interact with Jaks and Stats may contribute to the specificity of the interaction (54). 4 We propose that the multichain cytokine class II receptors have two major chains exemplified by the IFN-␥ receptor complex (Fig. 8B). The ligand binding chain (IFN-␥R1) and the accessory chain (IFN-␥R2; helper receptor) serve as a foundation for the functional IFN-␥R complex (16,40). The geometry of the IFN-␥R1 chain is such that its homodimerization yields a non-functional intracellular receptor complex. The accessory chain completes this function (Fig. 8A). The question arises: why should two separate chains have evolved when one in the correct configuration would suffice? We postulate that the presence of two distinct chains provides for more effective control and fine tuning of responses to ligand. For example, the differences in response of T H 1 and T H 2 cells to IFN-␥ result from the lack of expression of the IFN-␥R2 chain in the T H 1 subset (55-57) and allows exquisite fine tuning of sensitivity to IFN-␥. It is also possible that receptors with multiple chains could recruit additional factors into the complex to generate a wider variety of intracellular signals. This could explain how receptors with multiple subunits could activate a greater number of specific pathways and signals than those with fewer elements in the receptor complex. Our experiments begin to provide an insight into these possibilities. FIG. 8. Schematic representation of receptor complexes. A represents the IFN-␥R1 homodimer bound to IFN-␥. The cytoplasmic domains of the two chains are too far apart to permit transactivation of the two Jak1 kinases. B represents the active heteromeric IFN-␥ receptor complex with two IFN-␥R1 and two IFN-␥R2 subunits per complex. The IFN-␥ homodimer binds to two IFN-␥R1 chains, followed by its interaction with two IFN-␥R2 chains. The associated Jak2 and Jak1 kinases activate one another by transphosphorylation, with subsequent phosphorylation and dimerization of Stat1␣. C depicts the EpoR/␥R1 homodimer, which, unlike the IFN-␥R1 homodimer, permits transactivation of the two Jak1 molecules. D illustrates the structure of the heterodimer of EpoR/␥R1 and EpoR/␥R2, which is the putative active receptor complex.