Role of the Intracellular Domain of the Human Type I Interferon Receptor 2 Chain (IFNAR2c) in Interferon Signaling

A human cell line (U5A) lacking the type I interferon (IFN) receptor chain 2 (IFNAR2c) was used to determine the role of the IFNAR2c cytoplasmic domain in regulating IFN-dependent STAT activation, interferon-stimulated gene factor 3 (ISGF3) and c-sis-inducible factor (SIF) complex formation, gene expression, and antiproliferative effects. A panel of U5A cells expressing truncation mutants of IFNAR2c on their cell surface were generated for study. Janus kinase (JAK) activation was detected in all mutant cell lines; however, STAT1 and STAT2 activation was observed only in U5A cells expressing full-length IFNAR2c and IFNAR2c truncated at residue 462 (R2.462). IFNAR2c mutants truncated at residues 417 (R2.417) and 346 (R2.346) or IFNAR2c mutant lacking tyrosine residues in its cytoplasmic domain (R2.Y-F) render the receptor inactive. A similar pattern was observed for IFN-inducible STAT activation, STAT complex formation, and STAT-DNA binding. Consistent with these data, IFN-inducible gene expression was ablated in U5A, R2.Y-F, R2.417, and R2.346 cell lines. The implications are that tyrosine phosphorylation and the 462–417 region of IFNAR2c are independently obligatory for receptor activation. In addition, the distal 53 amino acids of the intracellular domain of IFNAR2c are not required for IFN-receptor mediated STAT activation, ISFG3 or SIF complex formation, induction of gene expression, and inhibition of thymidine incorporation. These data demonstrate for the first time that both tyrosine phosphorylation and a specific domain of IFNAR2c are required in human cells for IFN-dependent coupling of JAK activation to STAT phosphorylation, gene induction, and antiproliferative effects. In addition, human and murine cells appear to require different regions of the cytoplasmic domain of IFNAR2c for regulation of IFN responses.

Type I interferons (IFNs), 1 IFNs ␣, ␤, and are required for the induction of antiviral responses in a variety of animal species (1). Type I IFNs also elicit important antiproliferative effects in a number of cell lines and play a major role in mediating immunomodulatory activity (2). Cellular responses to type I IFNs require the interaction of type I IFNs with their cognate receptor, which is composed of two receptor subunits, IFNAR1 and IFNAR2c (also designated ␣ and ␤L, respectively). Once activated, the type I IFN receptor initiates signaling events, which culminate in the induction of a broad spectrum of IFNresponsive genes (2,3). One of the major signaling events coupled to receptor stimulation is the activation of signal transducers and activators of transcription (STATs). IFN-activated STAT transcription complexes include heterodimers of STAT1 and STAT2 along with a DNA-binding protein of 48 kDa present in the cell cytoplasm (3). This IFN-stimulated gene factor 3 (ISGF3) binds to IFN-sensitive response elements present in the promoter regions of IFN-inducible genes and initiates gene expression (2,3). The mechanism by which activation of the type I IFN receptor leads to STAT activation and gene expression is unclear. It is known that early stages of signaling require IFN-induced receptor heterodimerization of both receptor chains (2,3). However, the mechanism by which STATs and other regulatory proteins interact with the human type I IFN receptor is unclear, despite some emerging evidence of receptor interactive domains.
A proposed STAT2 binding site on IFNAR1 has been suggested which includes two phosphorylated tyrosines, Tyr 466 and Tyr 481 , in which an SH2 domain within STAT2 mediates the binding of STAT2 to these sites. For one of these sites, critical residues include not only Tyr 466 but also valine ϩ1 and serineϩ5 carboxyl-terminal to tyrosine 466 (4). STAT binding sites on IFNAR2c have also been proposed from in vitro results using glutathione S-transferase-IFNAR2 244 -462 "pull-down" experiments, where both STAT1 and STAT2 have been shown to pre-associate with IFNAR2c, in a manner independent of receptor phosphorylation and dimerization (5). This pre-association entails the binding of STAT2 in the absence of STAT1. More recent studies, using mouse L929 cells, have mapped this constitutive binding site for STAT2 to amino acids 404 -462 of IFNAR2c (6). In addition to this constitutive site on IFNAR2c, tyrosine phosphorylation of the proximal tyrosines (tyrosines 269, 306, 316, 318, and 336) of IFNAR2c is required, however, it is insufficient by itself, for efficient STAT2 activation (6). Therefore, stimulation of mouse cells expressing human IFNAR2c containing only the proximal tyrosines of IFNAR2c or the constitutive docking site with human IFN␣2 results in STAT tyrosine phosphorylation, ISGF3 formation, but no antiviral response. Thus, in this case, efficient STAT2 activation requires both constitutive and phosphotyrosine-dependent binding sites on the receptor. To further define the role of IFNAR2c in IFN signaling, we have chosen to express mutated forms of IFNAR2c having specific modifications in its intracellular domain, in a human cell line, U5A, which lacks IFNAR2c (7). In this way, one can directly measure the effects of such mutants on a variety of IFN-inducible responses, in a human cell that contains complementary Janus kinases and STAT proteins in a background devoid of heterologous receptor chains.
Selection of U5A Cell Lines Stably Expressing IFNAR2c Truncation Mutants-U5A cells (1 ϫ 10 6 cells/well) were transfected with the corresponding plasmid using Superfectin (11). Plasmids containing a neomycin selection marker, IFNAR2c truncations, and the full intracellular tyrosine to phenylalanine substitutions were constructed as described previously (6,12). Multiple stable cell lines for each IFNAR2c mutant were selected in media containing G-418 (1.0 mg/ml). After selection, individual clones were picked and expanded; an integration of each IFNAR2c mutant DNA was determined by polymerase chain reaction using primer sets spanning introns. Positive clones were further expanded and tested for their ability to bind type I IFN.
Ligand Binding Assay-A phosphorylated form of IFN␣2 was used, and ligand binding assays were performed as described previously (10). Ligands were phosphorylated (specific activities of 60 -62 Ci/g) as described previously (10). Binding data were analyzed according to Scatchard (13). Nonspecific binding was determined in the presence of a 100-fold excess of unlabeled IFN.
Electrophoretic Mobility Shift Assay (EMSA)-Gel shift assays were performed using 32 P-labeled double-stranded oligonucleotides representing the human 2-5A oligoadenylate synthetase IFN-sensitive response element and the m67SIE element present in the c-fos promoter. Cell were stimulated with IFN␣2 (1000 units/10 6 cells) or IFN␤1b (1000 units/10 6 cells) for 15 min, and cell pellets collected and processed for EMSA as described previously (14). Reaction mixtures were separated by electrophoresis through a 6% polyacrylamide gel and analyzed by autoradiography (14).

RESULTS
U5A is a human lung fibrosarcoma cell line that cells lacks IFNAR2c but expresses IFNAR1. Sensitivity to type I IFNs can be restored in U5A cells upon transfection with a plasmid encoding full-length IFNAR2c (7). Therefore, U5A cells provide a human cell line in which mutant forms of IFNAR2c can be used to determine the role of this receptor chain in type I IFN signaling in a human cell background. Using this approach, intracellular truncation mutants of IFNAR2c were stably expressed in U5A cells and multiple cell lines of each clone were analyzed with similar results. Initially, integration of cDNA encoding IFNAR2c mutants was demonstrated in transfected U5A cells by polymerase chain reaction (data not shown). For clones of interest, receptor number and binding affinities were then directly determined (Table I). High affinity binding of type I IFNs to the receptor requires both IFNAR1 and IFNAR2c (17)(18)(19). In this study, all cell lines examined, except U5A, bound IFN␣2 with high affinity (200 -500 pM) ( Table I). All cell lines reported here expressed mutant IFNAR2c receptor chains at levels equal to or greater than HT1080 cells, demonstrating that truncations or replacement of all tyrosine residues within the cytoplasmic domain of IFNAR2c, do not affect ligand binding.
IFN-induced receptor dimerization likely leads to conformationally distinct receptor complexes, dependent on the type I IFN subtype (10,20). Such IFN-inducible activation of IFNAR1 and IFNAR2 is independent of the phosphorylation state of the receptor subunits (21). Therefore, type I IFN receptor-binding interactions appear to be directly dependent on unique interactions of type I IFNs with the two receptor chains (22). It is assumed that IFN-induced receptor assembly leads to the specific activation of the Janus kinases, TYK2 and JAK1, and the subsequent phosphorylation of both receptor chains. The assembly of both receptor chains initiates activation of these kinases in an IFN-dependent manner.
Earlier work has demonstrated a specific association between IFNAR1 and TYK2 and IFNAR2c and JAK1 (18,(23)(24)(25)(26). Accordingly, to confirm that the various truncations or mutations to IFNAR2 had no effect on TYK2-IFNAR1 function, we examined the extent of IFN-inducible TYK2 activation in trans- fectants expressing variant IFNAR2c constructs. Our data indicate that IFNAR1 function, at least in the context of IFN-inducible TYK2 activation, is unaffected in transfectants expressing the mutant IFNAR2c constructs, but IFNAR1-associated TYK2 activation is ablated in the U5A cells (Fig. 1A). Similarly, JAK1 activation was also observed in all mutant cell lines expressing IFNAR2c mutants. However, JAK1 activation was not observed in IFN-stimulated U5A cells (Fig. 1B). Therefore, Janus kinase activation (TYK2 and JAK1) is dependent on receptor dimerization, and does not require the intracellular region distal to residue 346 in IFNAR2c or tyrosine phosphorylation of IFNAR2c. In order to determine the contribution of distinct regions of the intracellular portion of IFNAR2c to STAT recruitment and activation, we examined IFN-inducible STAT1 and STAT2 activation in each of the transfectants expressing mutant IFNAR2c. U5A cells expressing IFNAR2c truncation mutants were stimulated for 15 min with either IFN␣2 (1000 units/10 6 cells) or IFN␤1b (1000 units/10 6 cells) and the level of STAT1 or STAT2 activation determined by measuring the extent of STAT1 or STAT2 tyrosine phosphorylation. In contrast to U5A cells in which IFN-inducible STAT activation does not occur, STAT activation could be completely rescued by expressing full-length IFNAR2c in the U5A cells (Fig. 2). Removal of the distal 53 residues (R2.462) from the cytoplasmic region of IFNAR2c (R2c) had no effect on STAT1 or STAT2 activation (Fig. 2). However, further truncation to residue 417 (R2.417) or residue 346 (R2.346) resulted in a complete loss of STAT1 and STAT2 activation, as measured by IFN-dependent tyrosine phosphorylation (Fig. 2). Furthermore, substitution of all tyrosines present in the cytoplasmic domain of IFNAR2c (R2.Y-F) with phenylalanine residues, also resulted in a complete loss of STAT1 and STAT2 activation (Fig. 2). These results suggest an obligatory role for intracellular tyrosine residues and the 417-462 region of IFNAR2c in activation of STAT1 and STAT2. Furthermore, the distal 53 residues of IFNAR2c, which contain a potentially phosphorylatable tyrosine residue at position 512, are apparently not required for STAT activation.

FIG. 1. Activation of TYK2 and JAK1 kinases in U5A cells expressing IFNAR2c mutants.
A, cells were either untreated with IFN as a negative control (1) or stimulated (2,3) with IFN␤1b (1000 units/ 10 6 cells) for 15 min, after which cell lysates were prepared and subjected to immunoprecipation using either a nonspecific (2) or TYK2-specific antibody (3) as described under "Materials and Methods." Immunoprecipitates were then resolved by SDS-PAGE and subjected to immunoblotting using an anti-phosphotyrosine-specific antibody (p-tyr). The resultant membrane was then stripped and reprobed with a TYK2-specific antibody. B, cells were stimulated with IFN␤1b (1000 units/10 6 cells) for 15 min, cell lysates were prepared, and JAK1 was immunoprecipitated as described under "Materials and Methods," then resolved by SDS-PAGE and subjected to immunoblotting. Detection of activated JAK1 was performed using an "activation-dependent" duel phosphospecific antibody as present in the promoter region of the c-fos gene (6,28). IFN-dependent formation of both ISGF3 (Fig. 3A) and SIF (Fig. 3B) complexes were determined for all of the IFNAR2c truncation mutants. Formation of ISGF3 and SIF complexes were observed in response to IFN␣2 or IFN␤1b stimulation in HT1080, R2c, and R2.462 cell lines but not in U5A, R2.Y-F, and cells expressing the R2.417 and R2.346 IFNAR2c truncation mutants (Fig. 3, A and  B). These results are consistent with observations demonstrating a loss of STAT1 and STAT2 phosphorylation-activation in the R2.Y-F and R2.417 and R2.346 IFNAR2c mutants (Fig. 2).
In a recent report, it has been suggested that IFN-inducible STAT complex formation and DNA binding does not necessarily correlate with IFN-inducible transcriptional activation (29). Apparently, following STAT-DNA binding, there is an obligatory event, which is p38-dependent, which is required for transactivation and transcription (29). Accordingly, we undertook experiments to determine whether any of the mutations introduced into IFNAR2c in the transfectants affected IFN-inducible gene induction. Specifically, IFN-inducible gene expression for known IFN-responsive genes was examined by TaqMan ® analysis (15) and RNase protection assays (16). As shown in Table II, IFN-inducible ␤R1, ISG 54, and ISG 6 -16 gene expression was observed in the HT1080, R2c, and R2.462 cell lines but not in U5A cells expressing the R2.417, R2.346, and R2.Y-F IFNAR2c mutants. A similar pattern of gene expression was observed using RNase protection assays (data not shown).
In all cases for which gene expression could be measured, both IFN␣2 and IFN␤1b were capable of inducing gene expression, although some variation in gene expression levels was observed depending on whether IFN␣2 or IFN␤1b was used. As expected, differential expression of ␤R1 and ISG-54 was observed in HT1080 cells and to some extent in the R2c and R2.462 cell lines. Consistent with the gel shift data, IFN␣-and IFN␤-dependent gene expression was absent in cells expressing IFNAR2c truncation mutants R2.417 or R2.346 and IFNAR2c lacking intracellular tyrosine residues (R2.Y-F).
IFN-dependent gene expression leads to a number of important cellular responses such as control of cell growth. Therefore, we examined the effects of the various IFNAR2c mutants on IFN-dependent growth inhibitory effects as measured by short term [ 3 H]thymidine incorporation. Consistent with STAT activation and gene expression studies, IFN-inducible growth inhibition was not observed in the U5A, R2.417, R2.346, and R2.Y-F transfectant cell lines (Fig. 4). However, IFN stimulation of U5A cells expressing IFNAR2c (R2c) or the R2.462 mutant did result in a strong inhibition of [ 3 H]thymidine incorporation (Fig. 4). A similar pattern of antiproliferative effects was observed over a 4 -5-day period (data not shown).   (6,12). Such cells simultaneously express a heterologous combination of both mouse and human type I IFN receptor chains in which the specific response to human IFNs is generally dependent on the absolute species specificity of type I IFNs.
Using this approach, two IFN-regulatory regions of IFNAR2c have been reported. These include an IFN␤ response region (IBR) located between residues 417-462 (30) and a distal negative regulatory domain (31). Furthermore, a STAT2 binding site was mapped to the 404 -462 region of IFNAR2c using glutathione S-transferase fusion proteins encoding different regions of the intracellular domain (6). The existence of an IFN␤ response region (IBR) or distal negative regulatory domain was not observed in the current study in human cells due to a complete loss of receptor function in IFNAR2c mutants truncated at residue 417. This is in contrast to results obtained when identical IFNAR2c truncation mutants were stably expressed in mouse L929 cells. In this case, IFN␣2-dependent antiviral effects and detectable STAT2 and STAT1 tyrosine phosphorylation were observed in mouse cells expressing IFNAR2c truncated at residues 417 or 346. Furthermore, a negative regulatory effect on cell growth as measured by [ 3 H]thymidine incorporation was not observed for any of the IFNAR2c mutants analyzed in the present study. Only IFNAR2c and the R2.462 truncation mutant expressed in U5A cells were capable of producing an inhibition of [ 3 H]thymidine incorporation upon stimulation with type I IFN. Therefore, a complete loss of receptor function occurs in U5A cells expressing IFNAR2c mutants R2.417, R2.346, and R2.Y-F as measured by STAT1 and STAT2 activation, ISGF3/SIF complex formation, gene expression (␤R1, ISG 54, ISG 6 -16), and antiproliferative activity. A differential induction of an antiviral state was previously reported for human IFN␤ and IFN␣2 on murine cells expressing the R2.417 or R2.346 IFNAR2c truncation mutant and wild type IFNAR1 (30). However, this observed antiviral activity did not correlated with impaired Janus kinase, STAT1, and STAT2 activation or ISGF3 complex formation. Therefore, the induction of an antiviral state in these cells by human IFNs is likely to require a functional JAK/STAT activation pathway, which is absent in U5A cells expressing similar IFNAR2c truncation mutants.
Our data confirm previous results demonstrating the requirement of the 417-462 region of IFNAR2c for STAT activation and receptor function. However, in human cells, this region is absolutely required for receptor function and cannot be compensated for by additional STAT binding sites on either receptor chain. It is likely that, even though in U5A cells expressing IFNAR2c mutants, TYK2 and JAK1 activation occurs in response to type I IFNs, the inability of R2.417, R2.346, and R2.Y-F to induce STAT phosphorylation is due to the inability of these mutants to bind STATs or correctly present them as substrates for activated JAKs. The lack of any STAT activation in human U5A cells, which express IFNAR1, confirms that IFNAR1 on its own is unable to induce downstream signaling events such as STAT activation. In addition, it is interesting to note that, even though the 417-462 region of IFNAR2c was demonstrated to be critical for STAT1 and STAT2 phosphorylation, there are no tyrosine residues within this region. This confirms that the interaction of STATs with this site is independent of tyrosine phosphorylation. Furthermore, the distal tyrosine at residue 512 (Tyr 512 ) is not present in the R2.462 IFNAR2c truncation mutant, demonstrating that Tyr 512 appears not to be required for IFN-dependent STAT1 and STAT2 activation, transcription complex formation, gene expression, or antiproliferative effects.
Species differences in STATs and JAKs have been documented (25) and may lead to variations in the way mouse and human cells couple IFN-dependent receptor activation to gene expression. Indeed, the differences in receptor function observed using IFNAR2c mutants expressed in either mouse or human cells is unclear but may be partly due to such species differences. However, it is clear from studies using either human or mouse cells that both phosphotyrosine-dependent and -independent sites exist within the intracellular domain of IFNAR2c, which couple downstream signaling to type I IFN receptor activation. Clearly, our data demonstrate for the first time that differences exist in the manner in which the human and murine IFNAR2c influences IFN-dependent STAT activation. It will now be necessary to determine which phosphotyrosine residues in IFNAR2c are critical for IFN signaling and what role they play in regulating differential type I IFN signaling and gene expression in human cells.