INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Ligand binding to a receptor induces oligomerization of receptor subunits, which results in activation of various signaling pathways. Specificity in ligand receptor systems is achieved at the extracellular level by the specific interaction of a ligand with its distinct receptor components and at the intracellular level by the interaction of the cytoplasmic domain of the receptor subunits with a distinct set of signal transducing proteins. Many cytokine systems share receptor subunits (1-3), and in these situations it is commonly accepted that specificity is determined by the existence of additional ligand-specific receptor chains. For instance, the IL2R¹ has a binding subunit ( chain) and two signaling chains designated as  and c. The IL2Rc chain is also common to the IL4, IL7, and IL9 receptors and functions in association with specific receptor subunits for each of these cytokines (i.e. IL4R, IL7R, and IL9R chains). Therefore, these cytokines have the ability to produce both redundant and distinct biological effects (1-3).

The type I interferon (IFN) family includes 14 subtypes of IFN, as well as one IFN and IFN, all of which bind to the same cell surface receptor designated as IFNR, IFNR, or type I IFN-R (4). The type I IFN-R is composed of at least two subunits: the  chain or IFNAR1 (5-9) and the  subunit or IFNAR2, which has short (_S) and long (_L) forms (10-14). Although expression of the  chain with either _L or _S produces high affinity receptors in murine L-929 cells, only coexpression of  and _L allows activation of the Jak-Stat pathway and induction of an antiviral state in response to stimulation by both huIFN2 and huIFN (13, 14). Interestingly, although both of these human IFNs bind to the same receptor and activate the same components of the Jak-Stat pathway, there are some signaling and biological differences. IFN signaling has two distinctive features: (i) induction of a very strong association of the  and _L subunits of the type I IFN-R (15) and (ii) transcriptional activation of the -R1 gene (16). These signaling differences could be responsible for the disparity in biological effects among the members of the IFN family. For example, IFN is more effective than other type I IFNs in the treatment of multiple sclerosis (17, 18). However, unlike other cytokines, the differences in signaling and biological activities between IFN and IFN do not appear to result from the utilization of different receptor subunits. This has been demonstrated by the finding that mouse L-929 cells stably expressing the human  and _L subunits respond equally well to the induction of an antiviral state by huIFN and huIFN (13), and yet only IFN triggers the aforementioned association of the  and _L chains (15).

In this report we show that IFN2 and IFN require distinct intracytoplasmic regions of the _L chain to elicit an antiviral response. L-929 cells expressing _L truncated at amino acid 417 show a marked decrease in the antiviral response to IFN but not to IFN2. However, no differences in the activation of ISGF3 or SIF factors by IFN2 or IFN were detected. These data suggest that other signaling components, in addition to the Stat pathway, should be activated to obtain an antiviral response. Moreover, IFN seems to activate this unknown pathway through a distinct mechanism that requires the 417-462 region of the _L subunit.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

IFNs, Antibodies, and Antiviral Assays-- Human recombinant IFN2 and IFN were kindly provided by Drs. Paul Trotta (Schering-Plow) and S. Goelz (Biogen). The anti-phosphotyrosine antibody (4G10) was obtained from Upstate Biotechnology Inc. (Lake Placid, NY). Monoclonal antibodies against Jak1 and Stat1 were purchased from Transduction Labs. (Lexington, KY). The anti-Stat1 and anti-Stat2 and anti-Jak1 sera were kindly provided by Drs. A. Larner (Food and Drug Administration, Bethesda, MD) and J. N. Ihle (St. Jude's Children's Hospital, Memphis, TN), respectively. Antiviral assays were performed as described previously (19).

Expression of Different Deletions of the _L Subunit of Type I IFN-R in Mouse L-929 Cells-- These constructs were made by polymerase chain reaction using proofreading Vent polymerase and primers with an early termination codon at positions 346, 417, and 462 (see Fig. 1), respectively (20). Transfectants were grown in medium containing G-418 (500 µg/ml) and hygromycin B (500 µg/ml).

Immunoblotting-- Cells were treated with different concentrations of the indicated IFNs for 15 min, rapidly centrifuged at 2000 × g for 30 s in an Eppendorf microfuge, and subsequently solubilized in lysis buffer (20 mM Tris, pH 7.5, 50 mM NaCl, 10 mM sodium pyrophosphate, 20 mM NaF, 1 mM EDTA, 1 mM MgCl₂, 1 mM dithiothreitol, 0.5% Triton X-100, 10 µg/ml leupeptin, 10 µg/ml aprotinin, 100 mM phenylmethylsulfonyl fluoride, 200 µM sodium orthovanadate). Immunoprecipitation and immunoblotting were performed as described previously (7).

Radioiodination of Type I IFNs, Competitive Displacements, and Affinity Cross-linking-- Radioiodination of IFN2 and competitive displacement assays were performed as described previously (6).

EMSA-- Whole cell extracts were prepared as described by Ghislain and Fish (21) and analyzed by EMSA using end labeled ISRE oligonucleotides to detect ISGF3.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

We hypothesized that the differences observed between IFN and IFN signaling occur within the same receptor complex and not by an additional ligand-specific subunit as in other cytokine systems (1-3). Such a model assumes that a particular subtype of the IFN family will use regions of the receptor complex that are not used by another subtype. Thus, we studied the ability of huIFN2 and huIFN to induce an antiviral state in mouse L-929 cells stably coexpressing the wild type  subunit with a _L chain truncated at amino acid 462, 417, or 346, respectively (Fig. 1). All cell lines established were able to respond to huIFN2 as well as to muIFN (positive control), indicating that the transfected receptor and the IFN signaling pathway are functional. However, cells expressing the _L chain truncated proximal to amino acid 462 (Table I, _L417.7, _L417.9, _L346.2, and _L346.4 cells) required significantly higher amounts of huIFN to obtain 50% protection against encephalomyocarditis virus.

View larger version (33K): [in this window] [in a new window]   Fig. 1.   Schematic representation of the human  and L chain constructs expressed in mouse L-929 cells. The different interactions of the cytoplasmic domain of the L subunit are indicated. Constructs were made by polymerase chain reaction using primers with stop codons at the indicated positions (20). The binding sites for the Tyk2 and Jak1 kinases are shown (7, 20, 33). Transfectants were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and selection drugs (500 µg/ml of hygromycin and 500 µg/ml of geneticin).

View this table: [in this window] [in a new window]   Table I Induction of an antiviral state by huIFN2 and huIFN in L-929 cells expressing truncations of the L subunit and wild type  chain Cytopathic effect assay was performed using a concentration of encephalomyocarditis virus stock that produced 100% cytopathic effect. The IFN concentrations (units/ml) shown represent the amount of IFN required to inhibit cytopathic effect by 50%.

To demonstrate that the differences in the response to IFN2 and IFN were not due to alterations in IFN binding, we tested the ability of unlabeled IFN2 and IFN to displace radioiodinated IFN2 from the human receptor subunits in the _L417 and _L462 cell lines. If the defect in the induction of an antiviral state by IFN was a consequence of impaired binding of this IFN to the receptor, then unlabeled huIFN should be less effective as a competitor for binding of ¹²⁵I-IFN2 to the receptor than unlabeled IFN2. Fig. 2 shows that in _L417.7 and _L462.2 cells, unlabeled IFN was even more effective in displacing ¹²⁵I-IFN2 from the receptor than equivalent concentrations of unlabeled IFN2. Thus, the lack of induction of an antiviral state by IFN in _L417 cells is not due to impaired recognition of the receptor by this IFN but rather by deletion of the 417-462 region of _L, indicating that this region may contain a specific IFN response element.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Expression of L chain truncated at amino acid 417 does not affect binding of huIFN to the receptor. Increasing concentrations of unlabeled huIFN2 or huIFN (kindly provided by Drs. P. Trotta and S. Goelz, respectively) were used to displace binding of radioiodinated IFN2 to the receptor expressed in L417.7 and L462.2 cells. The Kd values in L417.7 cells were 33 and 5 pM for IFN2 and IFN, respectively. In L462.2 cells the Kd values were 48 and 40 pM for IFN2 and IFN, respectively. Affinities were calculated using the computer program LIGAND (34). Radioiodinated huIFN2 was selected for these experiments, because it labels to a higher specific activity (96 µg/µCi) than huIFN and can induce an antiviral state in L417 and L462 cells.

To further examine the differences observed in the antiviral response, time course and dose response experiments were performed in _L417.7 and _L462.2 cells to determine the length and intensity of tyrosine phosphorylation for components of the Jak-Stat pathway. Fig. 3A shows that IFN (lanes 3, 5, and 7) induced even higher levels of Jak1 tyrosine phosphorylation than IFN2 (lanes 1, 4, and 6) in _L417.7 cells at all time points studied. Intense phosphorylation could be detected at 10 min and returned to base-line levels by 90 min. Dose response experiments showed that activation of Jak1 was achieved at doses as low as 300 units/ml of IFN2 or IFN (Fig. 3A, lower panel, lanes 2-5). Similar results were obtained in time course and dose response experiments performed with _L462.2 (data not shown). We also studied the activation of Tyk2 kinase in both cell lines. Fig. 3B shows that Tyk2 tyrosine phosphorylation was induced by both IFN2 and IFN in _L417.7 and _L462.2 cells at 10 min (lanes 2, 3, 9, and 10) and subsequently decreased at 30 min (lanes 4, 5, 11, and 12). The level of Tyk2 phosphorylation returned to base-line levels by 90 min in _L462.2 cells (lanes 13 and 14) and significantly decreased in _L417.7 cells (lanes 6 and 7). Tyk2 tyrosine phosphorylation was also detected with concentrations of IFN2 or IFN as low as 300 units/ml (Fig. 3C, lanes 2, 3, 7, and 8). It is worth mentioning that the level of tyrosine phosphorylation detected in _L462 cells were slightly higher than in _L417 cells (Fig. 3C, compare _L417 and _L462 cells).

View larger version (63K): [in this window] [in a new window]   Fig. 3.   Activation of the Jak-Stat pathway in response to huIFN2 and huIFN. A, cells were stimulated with huIFN2 or huIFN (20,000 units/ml) for the indicated periods of time (top panel) or for 10 min with the indicated concentrations of IFN (bottom panel). Cell lysates were immunoprecipitated with an anti-Jak1 serum, resolved in a 8% SDS-polyacrylamide gel electrophoresis, immunoblotted with the anti-phosphotyrosine antibody 4G10, stripped, and reblotted with an anti-Jak1 monoclonal antibody (Transduction Laboratories). B, tyrosine phosphorylation of Tyk2 in response to IFN2 or IFN. Cells were treated as described in A, immunoprecipitated with an anti-Tyk2 antibody, and immunoblotted with the anti-phosphotyrosine antibody 4G10. Stripping and immunoblotting of the same filter with the anti-Tyk2 serum could not be performed due to the high background produced by this antibody when used for immunoblotting as described previously (13). C, experiment similar to that in B in which lower doses of IFNs were used. Cellular lysates were immunoprecipitated with an anti-Tyk2 antibody and immunoblotted with the anti-phosphotyrosine antibody 4G10. The electrophoretic mobility of Tyk2 and a background protein (Bkgd) detected in some experiments are indicated. IP, immunoprecipitation; WB, Western blot; CT, control.

We next studied the activation of the Stat proteins. Fig. 4 shows that at all time points studied there were no differences in the induction of tyrosine phosphorylation of Stat1 by IFN2 or IFN in _L417.7 (panel A) or _L462.2 (panel B) cells. Tyrosine phosphorylation was maximal at 30 min and then progressively declined. Fig. 4 (A and B) also shows that after IFN2 or IFN treatment the anti-Stat1 serum coimmunoprecipitates another tyrosine phosphorylated protein that has an electrophoretic mobility similar to that of Stat2. The lower panels in A and B of Fig. 4 show that equivalent amounts of Stat1 were immunoprecipitated in all conditions. Similar results were obtained when the activation of Stat1 and Stat2 (ISGF3) was studied by EMSA with an ISRE probe. Fig. 4C shows that equivalent levels of the ISGF3 were induced by huIFN2 and huIFN in _L417.7 cells even though these cells are resistant to the induction of an antiviral state by huIFN but remain sensitive to IFN2. Similar results were observed with _L462.2 cells (Fig. 4C, lower panel). Dose response experiments were then performed to examine the ability of varying concentrations of huIFN2 or huIFN to affect activation of Stat proteins. Fig. 4D shows that Stat1 and Stat2 activation are equivalent for human IFN and IFN at concentrations as low as 300 units/ml. The higher levels of tyrosine phosphorylation observed after treatment with 300 units/ml of IFN or IFN in _L462 cells are likely due to the higher amounts of Stat1 protein precipitated in these lanes (Fig. 4D, lower panel, anti-Stat1 immunoblot). The middle panel in Fig. 4D shows an immunoblot with an anti-Stat2 antibody that identifies Stat2 as the protein coprecipitated by the anti-Stat1 antibody. Coprecipitation of Stat2 is proportional to the intensity of tyrosine phosphorylation as expected from a SH2-phosphotyrosine interaction. In summary, the study of activation of the Jak kinases and Stat factors in _L417 and _L462 cells did not reveal significant differences that would account for the impaired response to IFN observed in _L417 cells.

View larger version (74K): [in this window] [in a new window]   Fig. 4.   Activation of Stat proteins by IFN and IFN. A and B, time course experiments. Antiphosphotyrosine immunoblotting (upper panels) were performed as described in Fig. 3, but immunoprecipitations were performed with an anti-Stat1 antibody instead. The same filters was stripped and then probed with anti-Stat1 antibody (lower panels) to determine the amounts of Stat1 protein loaded in each lane. C, EMSA to detect ISGF3 induction. Whole cell extracts were obtained from L417.7 and L462.2 transfectants treated with huIFN2 (20,000 units/ml), huIFN (20,000 units/ml), or left untreated for the indicated times at 37 °C as described previously (21). EMSA was performed using an ISRE probe (21) for the detection of the ISGF3 (arrows). D, dose response experiment. Tyrosine phosphorylation of Stat1 and Stat2 was studied after stimulation of L417.7 and L462.2 cells with different concentrations of huIFN2 and huIFN. Cell protein homogenates were immunoprecipitated with anti-Stat1 serum and then sequentially immunoblotted with anti-phosphotyrosine (upper panel), anti-Stat2 (middle panel), and anti-Stat1 (lower panel) antibodies.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Cytokine systems that share subunits usually show both overlapping and distinct biological effects that result from activation of common and discrete receptor chains, respectively (3). In the type I IFN system, however, several lines of evidence indicate that the different subtypes of IFN, IFN, and IFN utilize a common type I IFN-R. For example, all type I IFNs compete for binding to the same receptor (reviewed in Refs. 4 and 22) and specific monoclonal antibodies that recognize the  and chains, block binding, and inhibit biological activity of several type I IFNs (11, 23). Other anti- subunit antibodies precipitate radiolabeled huIFN2, huIFN7, huIFN8, huIFN, and huIFN cross-linked to the  subunit (24). Finally, expression of the human wild type  and _L subunits in mouse L-929 cells gives them the ability to mount an antiviral response after treatment by either huIFN2 or huIFN (13). Our data for the type I IFN system indicate that binding of different ligands to the same receptor chains produces changes that go beyond simple dimerization of receptor subunits. IFN2 and IFN apparently utilize different regions of the intracellular domain of the _L subunit to generate an antiviral state. One possible explanation is that each IFN promotes distinct conformational changes in the receptor that affects signaling through the 417-462 region of _L, similar to the mechanism proposed for the Tar protein (25, 26). This model theoretically increases the number of distinct biological responses that can be promoted by different ligands utilizing a common set of receptor subunits, which may explain how different type I IFNs exert slightly different biological responses (27). The presence of two binding sites on the  subunit of the type I IFN-R receptor (28) may contribute to the diverse biological responses induced by different subtypes of type I IFNs. This is supported by the finding that the splice variant of the  subunit, which lacks the N-terminal binding site, preferentially binds certain type I IFNs (29).

Finally, the differences in induction of an antiviral state by IFN and IFN do not correlate with differences in activation of the Jak-Stat pathway. This is not surprising because mutant U4 cells complemented with kinase deficient Jak1 can induce IFN-activating factor and the interferon-stimulated gene but not an antiviral state in response to IFN (30). Therefore, although activation of these DNA binding complexes is required (31, 32), it is not sufficient to elicit an antiviral effect. Similarly, the IFN response is partially conserved in mutant cells that lack Tyk2 (35, 36). These findings suggest that additional signaling mechanisms should be triggered by IFNs. It is tempting to speculate that the IFN-induced activation of the IFN response element (IBR) region of _L is responsible for the formation of other DNA binding complexes that interact with an element different from the ISRE or IRE. The existence of such elements has been also proposed by others to explain the differences in induction of the IFN-specific gene R-1 (16).