Identification of a Domain in the β Subunit of the Type I Interferon (IFN) Receptor That Exhibits a Negative Regulatory Effect in the Growth Inhibitory Action of Type I IFNs*

Expression of human α and long form of the β (βL) subunits of type I interferon receptor (IFN-R) in mouse cells is sufficient to activate the Jak-Stat pathway and to elicit an antiviral state in response to human IFNα2 and IFNβ. We demonstrate herein, however, that these cells respond to the antiproliferative effects of murine IFNαβ but not human type I IFNs. These results suggest that an unknown species-specific component is required for the antiproliferative effect of human type I IFNs. The absence of this component can be complemented by expressing the human βL chain truncated at amino acid 346. Thus, the distal region of βL appears to function as a negative regulator of the growth inhibitory effects of type I IFNs. Further studies looking for possible targets of the βL regulatory domain demonstrated that this region associates with a tyrosine phosphatase. These results suggest that a protein associated with the negative regulatory domain of βL, likely a tyrosine phosphatase, plays a role in regulating the growth inhibitory effects of human type I IFNs.

Expression of human ␣ and long form of the ␤ (␤ L ) subunits of type I interferon receptor (IFN-R) in mouse cells is sufficient to activate the Jak-Stat pathway and to elicit an antiviral state in response to human IFN␣2 and IFN␤. We demonstrate herein, however, that these cells respond to the antiproliferative effects of murine IFN␣␤ but not human type I IFNs. These results suggest that an unknown species-specific component is required for the antiproliferative effect of human type I IFNs. The absence of this component can be complemented by expressing the human ␤ L chain truncated at amino acid 346. Thus, the distal region of ␤ L appears to function as a negative regulator of the growth inhibitory effects of type I IFNs. Further studies looking for possible targets of the ␤ L regulatory domain demonstrated that this region associates with a tyrosine phosphatase. These results suggest that a protein associated with the negative regulatory domain of ␤ L , likely a tyrosine phosphatase, plays a role in regulating the growth inhibitory effects of human type I IFNs.
The most prominent effects of type I interferons (IFN) 1 are the antiviral and antiproliferative actions (1). These effects are mediated through binding to the type I interferon receptor (IFN-R or IFN␣R), which is composed of two subunits termed ␣, or IFNAR1, and ␤, or IFNAR2 (2-10). The genes encoding the different subunits of the type I IFN-R are clustered in the q22.1 region of human chromosome 21 (6,7,(11)(12)(13)(14)(15)(16). This region also harbors an orphan class II cytokine receptor, the CRFB-4 gene, which is encoded on human chromosome 21 between the genes for the ␤ chain of the IFN␣R and the ␤ subunit of the IFN␥R (10,17). Expression of the human ␣ and long form of the ␤ chain (␤ L ) subunits in mouse L-929 cells fully reconstitutes the activation of the Jak-Stat pathway and the induction of an antiviral state in response to HuIFN␣2 and HuIFN␤ (9). Furthermore, only the first 82 amino acids of the cytoplasmic domain of the ␤ L chain are required to activate the Jak-Stat pathway and induce the antiviral effect in response to IFN␣2 (18).
The ␣ and ␤ subunits of the type I IFN-R associate with protein tyrosine kinases of the Jak family (4,8,18). The ␣ subunit interacts with Tyk2 (4,19,20) while the cytoplasmic domain of the ␤ L contains a docking site for Jak1 (18). Binding of type I IFNs to their receptor triggers rapid tyrosine phosphorylation of Tyk2 and Jak1 kinases, type I IFN-R subunits (21)(22)(23)(24)(25), and Stat factors (reviewed in Refs. 26 -28). Regulation of tyrosine kinase activity is mediated in most cytokine systems by protein tyrosine phosphatases (PTPs). For example, SHP1 (also named SHP, SHPTP1, HCP; PTP1C, Ref. 29), a predominantly hematopoietic tyrosine phosphatase that regulates the activity of the erythropoietin and IL-3 systems (30 -35), has also been implicated in IFN␣ signaling in hematopoietic cells (36,37). However, the role of SHP1 in other cell types is not clear since this PTP is mainly expressed in hematopoietic cells, whereas the IFN system functions in almost all, if not all, cell lineages.
Mouse L-929 cells that coexpress wild-type human ␣ and ␤ subunits respond to the antiviral effects of human type I IFNs, demonstrating the presence of functional human type I IFN-R (9). We therefore decided to test these cells for their ability to respond to the antiproliferative effects of type I IFNs. Human IFN␣ and IFN␤ induced only a minimal antiproliferative response, whereas murine type I IFNs produced a marked inhibition of cell proliferation. These data indicate that (i) induction of the antiproliferative and antiviral responses occurs through partially divergent pathways and (ii) that a novel species-specific signaling component is required, in addition to the ␣ and ␤ L chains, for the growth inhibitory effect. Surprisingly, the antiproliferative response was observed in cells that express the human ␤ L chain truncated at amino acid 346. Thus, the distal part of ␤ L apparently contains a negative regulatory domain that controls the growth inhibitory effects of type I IFNs. Further characterization of this negative regulatory domain revealed that it interacts with a PTP that appears to be distinct from SHP1 and SHP2. The data herein suggest that a novel species-specific component is required for the growth inhibitory effect and that this effect is regulated by a distal region corresponding to amino acids 346 -417 on the ␤ subunit.
Constructs and Expression of the Human Type I IFN-R Subunits in Mouse L-929 Cells-Mouse L-929 cell lines coexpressing different constructs of the human ␣ and ␤ L chains LpZR␣␤ L , LpRZ␣␤ L462 , LpRZ␣␤ L417 , LpRZ␣␤ L346 , and LpZR␣␤ L300 were described previously (9,18). The L-929 transfectants stably coexpressing mutations of tyrosine 466 and truncation at amino acid 511 of the ␣ subunit (designated as ␣Y1F511) with either wild type or truncation 346 of the ␤ L subunit (LpZR␣ Y1F(511) ␤ Lwt and LpZR␣ Y1F(511) ␤ L346 , respectively), as well as coexpressing wild-type ␣ chain and ␤ L subunit carrying a mutation of tyrosine 411 to phenylalanine, are described elsewhere. 2 Cell Proliferation Assays-Cell proliferation was assessed by performing MTT assays (7,40) and cell counts after treatment with the indicated amount of human and mouse IFNs. Briefly, cells were seeded at 6,000 cells/well in 24-well plates in a final volume of 0.6 ml and treated with the indicated concentrations of IFNs. The numbers of cells per well were determined by trypsinization and counting of duplicate wells in a hemocytometer. Experiments were performed at least twice with two independent clones carrying the same mutation.
Immunoblotting-Cells were treated with different concentrations of the indicated IFNs for 15 min, rapidly centrifuged at 2000 ϫ g for 30 s in an Eppendorf microfuge, and subsequently solubilized in lysis buffer. Immunoprecipitation and immunoblotting were performed as described previously (4).
Phosphatase Assays-For protein phosphatase assays, cells expressing wild-type ␣ chain and the ␤ L subunit truncated at the indicated amino acids were treated with or without IFN␣2 for 10 min at 37°C and lysed in lysis buffer as described previously (4). The ␤ L subunit was precipitated using a polyclonal serum raised against a GST fusion protein encoding the entire cytoplasmic domain (␤ L265-515 ) (39), immunoprecipitates were washed three times in cold phosphatase buffer to remove phosphatase inhibitors. The phosphatase activity of the immunocomplexes was determined using pNPP (Sigma) as a substrate. The phosphatase assay was carried out at 37°C for 0.5 h in 50 l of reaction mixture (100 mM HEPES, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1 mM dithiothreitol, 10 mM pNPP). The reaction was terminated by adding 950 l of 1 M NaOH. The reaction product, p-nitrophenolate, was quantified by measuring absorbance at 405 nm.

Expression of the Human ␣ and ␤ L Subunits of the Type I IFN-R Is Not Sufficient to Reconstitute the Antiproliferative
Response-Mouse L-929 cells transfected with the human ␣ and ␤ L subunits, LpZR␣␤ L.10 , activate the Jak-Stat pathway and are highly responsive to the antiviral effects of HuIFN␣2 and HuIFN␤ (9). To characterize the antiproliferative effect of type I IFNs in these cells, we first performed MTT proliferation assays using L-929 cells stably transfected with wild-type human ␣ and ␤ L subunits (LpZR␣␤ L.10 ). Fig. 1A shows that high doses of HuIFN␣2 (100,000 units/ml) had little effect on cell proliferation. However, treatment with MuIFN␣␤ or MuIFN␤ at doses between 1,000 and 10,000 units/ml induced a significant antiproliferative effect. To confirm the results observed with MTT assays, we performed similar experiments in which cell numbers were assessed. Fig. 1B shows that treatment of LpZR␣␤ L.10 cells (Fig. 1B, ␣␤ L cells) with MuIFN␣␤ reduced cell proliferation more than 95% over a period of 6 days, whereas HuIFN␣2 produced a minimal response (20%, from 545,000 cells/well to 445,000 cells/well in control and HuIFN␣2-treated cells, respectively). Similar results were ob-served when MuIFN␤ and HuIFN␤ were used (data not shown). These results indicate that the endogenous mouse type I IFN-R expressed in L-929 cells can trigger a complete antiproliferative effect in response to mouse IFNs. Thus, in contrast to the antiviral effect, reconstitution of the human receptor with the ␣ and ␤ L chains in these cells is not sufficient to trigger an antiproliferative effect in response to human type I IFNs, suggesting that an additional human signaling component is required for this effect.
It has been reported that human/rodent somatic cell hybrids carrying human chromosome 21 acquire the ability to respond to the antiviral effect and induction of HLA class I antigens by human type I IFNs (7,(11)(12)(13)(14)(15)(16)(41)(42)(43)(44)(45). This is in part due to the fact that the ␣, ␤ S , and ␤ L subunits of the receptor are encoded by genes on this chromosome. However, it has also been reported that additional IFN signaling components may also reside on the distal part of human chromosome 21q (46,47). Therefore, we tested mouse A9ϩ21 cells, which carry several copies of human chromosome 21, for their ability to respond to the antiproliferative effects of HuIFN␣2 and HuIFN␤. Fig. 1C shows a proliferation assay performed with HuIFN␣2 and MuIFN␣␤. An 85% reduction in the growth occurred for A9ϩ21 cells in response to MuIFN␣␤, whereas only a 40% decrease (from 165,000 cells/well to 100,000 cells/well in controls and HuIFN␣2-treated cells, respectively) in cell proliferation was induced by HuIFN␣2. These data show that binding of human type I IFNs by human ␣ and ␤ L chains expressed in mouse cells (either by transfection in L-929 cells or by incorporation of human chromosome 21 in A9ϩ21 cells) induces only a partial 2 Domanski and Colamonici, manuscript in preparation. growth inhibitory response. Since human/rodent somatic cell hybrids carrying human chromosome 21, including A9ϩ21 cells, have been shown to develop an antiviral state and induce HLA class I antigens in response to human type I IFNs (7,(11)(12)(13)(14)(15)(16)(41)(42)(43)(44)(45), the inability of these cells to respond to the antiproliferative effects of human type I IFNs further suggests that an additional species-specific signaling component is required. Moreover, it is unlikely that this signaling element is encoded by any genes present on human chromosome 21.
Deletion of a Negative Regulatory Region of the ␤ L Chain Promotes the Growth Inhibitory Effects of Human Type I IFNs-Since deletion of the distal region of the ␣ chain produces an increase in HLA class I response induced by human IFN␣2 (48), we wished to determine whether the antiproliferative response could be induced by human IFNs in L-929 cells that express ␣ and ␤ L subunits with deletions of various regions of their cytoplasmic domains. Cells that express the ␤ L chain truncated at amino acid 417 are more responsive to MuIFN␣␤ than to HuIFN␣2 ( Fig. 2A, panel A). These results parallel our findings for L-929 or A9ϩ21 cells, which express wild-type receptor subunits (Fig. 1, B and C). By direct contrast, L-929 cells coexpressing both the wild-type ␣ and ␤ L truncated at amino acid 346 (Fig. 2, panels B, ␣␤ L346.2 cells), responded to the growth inhibitory effects of human type I IFNs. Truncation of the cytoplasmic domain of ␤ L at amino acid 300, which removes the Jak1 binding site (18), abolished the antiproliferative response to human type I IFNs but did not affect the antiproliferative response to MuIFN␣␤ (Fig. 2C), demonstrating that the mouse signaling machinery is intact in these cells. We also studied L-929 transfectants expressing wild-type ␤ L and an ␣ chain with a deletion of the negative regulatory domain (truncation at amino acid 511 and tyrosine 466 mutated to phenylalanine; Fig. 2, panel D, ␣ Y1F ␤ L.11 cells) (48). These cells showed a significant response to MuIFN␣␤, whereas human type I IFNs induced only a partial antiproliferative effect, which was similar to that observed in cells expressing wild-type receptors. This result suggests that the negative regulatory region of the ␣ subunit does not control cell proliferation (48). However, L-929 cells coexpressing the same mutations of the ␣ subunit and ␤ L truncated at amino acid 346 (Fig. 2, panel E, ␣ Y1F511 ␤ L346.3 ) were extremely sensitive to the antiproliferative effect of human and mouse type I IFNs. Sim-ilar results were obtained with two independent clones carrying the same mutation (data not shown). These data strongly suggest that a region corresponding to amino acids 346 -417 in the ␤ L chain contains a negative regulatory domain and may be a possible target for mouse regulatory proteins. Thus, removal of this negative regulatory domain appears to complement the absence of an unknown species-specific component required for the antiproliferative effect (see "Discussion").
The ␤ L Subunit of the Type I IFN-R Associates with a Phosphatase-Tyrosine phosphorylation plays a central role in IFN and cytokine signaling. Therefore, to determine if the negative regulatory domain of ␤ L was associated with a PTP, we performed in vitro phosphatase assays after immunoprecipitation with anti-␤ L sera. As a source of ␤ L chain, we used cell lysates obtained from mouse L-929 cells cotransfected with the wildtype ␣ subunit and truncations of the ␤ L chain at amino acids 346, 417, or 462, respectively. Cells were treated with IFN␣2 for 10 min, cell lysates were immunoprecipitated with an antibody that recognizes all truncated forms of the ␤ L chain, and in vitro phosphatase assays were performed on the immunoprecipitates. Fig. 3 shows that significant phosphatase activity is associated with ␤ L462 and ␤ L417 , but not with ␤ L346 after IFN␣2 treatment, indicating that the 346 -417 region of ␤ L associates with a PTP. Since the increase in phosphatase activity associated with the 346 -417 region of ␤ L is observed only after IFN␣ treatment, we could not elucidate whether the ␤ L -associated PTP is recruited to the receptor complex after IFN␣ stimulation or is constitutively associated with the ␤ L chain and activated by IFN␣2 stimulation.
Deletion of the Negative Regulatory Domain of the ␤ L Subunit at Amino Acid 346 Results in Strong and Prolonged Tyrosine Phosphorylation of Jak1-We next sought to test if deletion of the 346 -417 region of the ␤ L chain and, consequently, removal of the phosphatase interaction site had an effect on tyrosine phosphorylation. We performed time course and dose response experiments with mouse L-929 cells stably cotransfected with wild-type ␣ chain and ␤ L truncated at amino acids 346 or 417, respectively. Fig. 4A shows that more intense phosphorylation of Jak1 was observed at lower doses of IFN␣2 in cells expressing the ␤ L subunit truncated at amino acid 346, as compared with the ␤ L chain truncated at residue 417. Moreover, deletions distal to amino acid 346, but not at amino acid 417, prolonged Mouse L-929 cells coexpressing different constructs of the ␣ and ␤ L subunits were evaluated for their response to the antiproliferative effect of human and murine IFNs. ␣␤ L417 , ␣␤ L346 , and ␣␤ L300 correspond to clones expressing wild-type ␣ subunit and truncations at residues 417, 346, or 300 of the ␤ L chain, respectively. ␣Y1F␤ L clones express wild-type ␤ chain and ␣ subunit with a deletion of the negative regulatory domain (truncated at amino acid 511 and mutation of Tyr-466 to phenylalanine). ␣Y1F␤ L346 clones express the same mutations of the ␣ subunit (truncation 511 and mutation of Tyr-466 to phenylalanine) and ␤ L truncation at amino acid 346. ␣␤ LY411F cells correspond to L-929 cells expressing wild-type ␣ chain and ␤ L with a mutation of Tyr-411 to phenylalanine. Proliferation was assessed as described in Fig. 1. the period of time that Jak1 was phosphorylated (Fig. 4B). The increase in tyrosine phosphorylation observed was not due to different amounts of immunoprecipitated Jak1 protein since stripping and reblotting of the same membranes with an anti-Jak1 mAb showed a similar amount of Jak1 in all lanes (Fig. 4,  A and B, lower panels). Two independent clones expressing ␤ L truncated at amino acid 346, ␣␤ L346.2 and ␣␤ L346 . 4 , produced equivalent results. The intensity of Tyk2 phosphorylation, however, was unaffected by truncation of ␤ L at residue 346 as revealed by immunoprecipitation experiments with anti-Tyk2 sera followed by immunoblotting with antiphosphotyrosine antibodies (data not shown). Altogether, these data suggest that the 346 -417 region of ␤ L interacts with a phosphatase that regulates Jak1 phosphorylation.
Mutation of Tyrosine 411 in the ␤ L Chain of Type I IFN-R Does Not Alter the Response of Mouse Cells to Human Type I IFNs-It has been previously reported that SH2-containing phosphatases (SHP1 and SHP2) interact with the ␣ subunit of type I IFNR (37,49). Since the negative regulatory region of ␤ L contains only one tyrosine (Tyr-411), which (if phosphorylated) may serve as docking site for SH2-containing tyrosine phos-phatases, we studied the effect of a phenylalanine mutation of tyrosine 411 on cell proliferation. Panel F (Fig. 2) shows that mutation of tyrosine 411 to phenylalanine does not reconstitute the antiproliferative effect of type I IFNs. Similar results were also obtained with different clones carrying the same mutations (data not shown). Therefore, tyrosine 411 is not critical for induction of the negative regulatory effect and is presumably not a docking site for SH2-containing phosphatases. DISCUSSION The type I IFNs have multiple biological actions, including antiviral and antiproliferative effects, which are the most prominent of these cellular responses (1). Binding of human type I IFNs to mouse L-929 cells that coexpress human ␣ and ␤ L chains is sufficient to trigger activation of the Jak-Stat pathway and to produce a full antiviral response, which demonstrates the presence of functional human type I IFN-R subunits (9). We therefore tested these cells for their ability to respond to the antiproliferative actions of human type I IFNs.
Although these cells were able to respond to the full growth inhibitory effects of MuIFN␣␤, only a minimal response was observed for human type I IFNs. Thus, the antiproliferative pathway is intact in these cells but not fully activated via the human receptors. Similar results were also observed for different clones of mouse A9ϩ21 cells that carry several copies of human chromosome 21, which is thought to contain the type I IFN-R cluster (7,(11)(12)(13)(14)(15)(16)(41)(42)(43)(44)(45) and an uncharacterized signaling component (46,47). Altogether, these data indicate that (i) induction of the antiproliferative and antiviral responses occurs through partially divergent pathways and (ii) that a novel species-specific signaling component is required, in addition to the ␣ and ␤ L chains, for the growth inhibitory effect.
It has been reported that elements of the cytoplasmic domains of cytokine receptors have a negative regulatory role in signaling. For example, removal of the docking site for SHP1 in EPO-R results in hypersensitivity to EPO and prolonged phosphorylation of Jak2 (31). Similarly, deletion of the distal region of the ␣ subunit of the IFN␣R results in increased sensitivity to induction of HLA class I antigens by IFN␣2 (48). Correspondingly, removal of the distal region (346 -417) of the cytoplasmic domain of ␤ L resulted in a gain in response to the growth inhibitory effects of IFN␣, indicating that this region encodes a negative regulatory domain. The effect of the negative regulatory domain of ␤ L is specific, as demonstrated by the finding that deletion of a homologous region of the ␣ chain did not have an effect on cell proliferation (Fig. 2, ␣Y1F511␤ L cells). The negative regulatory domain of ␤ L interacts with a PTP as indicated by detection of phosphatase activity associated with the 346 -417 region of ␤ L and the finding that deletion of this region resulted in prolonged phosphorylation of Jak1 in mouse cells. Mutation of the only tyrosine in the negative regulatory domain (Tyr-411) did not have the same effect on the antiproliferative response as deletion of the 346 -417 region, indicating that the putative phosphatase is not docked to ␤ L through an SH2 domain. Moreover, immunoprecipitation with antibodies against the ␣ and ␤ L chains, and pull-down experiments with GST fusion proteins encoding the cytoplasmic domain of the ␣ and ␤ L chains failed to precipitate SHP1 or SHP2 (data not shown). Taken together, these data indicate that the 346 -417 region of ␤L functions as a negative regulator for the antiproliferative effect of IFNs, possibly by recruiting a regulatory phosphatase through an SH2-independent mechanism. Consistently with the finding in mouse cells, immunoprecipitations with anti-␤ L antibodies also revealed phosphatase activity specifically associated with the ␤ L subunit expressed in human cells (data not shown). It should be noted, however, that in human cells no conclusive data have been obtained using A, dose response experiment. Cells coexpressing the ␣ chain and truncation mutants of the ␤ L chain at amino acids 346 (␣␤ L346 ) or 417 (␣␤ L417 ) were treated with the indicated doses of IFN␣2 for 8 min at 37°C. Cells were lysed and immunoprecipitated with anti-Jak1 sera as described under "Materials and Methods." Immunoblotting was first performed with anti-phosphotyrosine antibody (top panel) followed by stripping and reprobing of the filters with an anti-Jak1 monoclonal antibody (bottom panel) to demonstrate that similar amounts of proteins were present in all lanes. B, kinetics of Jak1 phosphorylation. ␣␤ L346 and ␣␤ L417 cells were treated with 20,000 units/ml of IFN␣2 for the indicated periods of time. Cells were lysed, and lysates were immunoprecipitated with an anti-Jak1 sera followed by sequential immunoblotting with anti-phosphotyrosine (top panel) and -Jak1 (bottom panel) antibodies.
GST␤ L fusion proteins.
Since mouse IFNs completely inhibit proliferation of cells expressing human ␣ and ␤ L chains, we can conclude that the mouse receptor couples the signal induced by murine IFNs with the intracellular proteins responsible for antiproliferative pathway in these cells. Consequently, the inability of LpZR␣␤ L and A9ϩ21 cells to respond to the growth inhibitory effect of human IFNs centers the defect at the level of the human receptor. One possibility is that the intracellular domains of the human ␣ and/or ␤ L chains, which are responsible for activation of the antiproliferative pathway, are not homologous to their murine counterparts. However, this possibility is highly unlikely, based on two primary observations. First, deletion of the negative regulatory domain of ␤ L allows a full human IFN␣-induced antiproliferative effect, indicating that the human receptor subunits are capable of interacting with the appropriate mouse signaling proteins. Second, these transfectants respond to the antiviral effect of IFNs. Another possibility is that the missing species-specific component corresponds to a third receptor subunit. In this scenario, the antiproliferative response requires the assembly of a receptor composed of three subunits; in this complex, the third receptor subunit regulates the activity of a protein associated with the negative regulatory domain of ␤ L , presumably a PTP. If, in fact, a phosphatase is involved, the third receptor subunit may delay activation of the PTP or releases the PTP from the complex, resulting in prolonged activation of Jak1 and generation of the antiproliferative response. Thus, deletion of the negative regulatory domain of ␤ L has the same outcome as activating the third receptor subunit, blocking the action of the PTP or an unknown protein that associates with this region.