Role of p38 Protein Kinase in the Ligand-independent Ubiquitination and Down-regulation of the IFNAR1 Chain of Type I Interferon Receptor*

Phosphorylation-dependent ubiquitination and degradation of the IFNAR1 chain of type I interferon (IFN) receptor is a robust and specific mechanism that limits the magnitude and duration of IFNα/β signaling. Besides the ligand-inducible IFNAR1 degradation, the existence of an “inside-out” signaling that accelerates IFNAR1 turnover in the cells undergoing the endoplasmic reticulum (ER) stress and activated unfolded protein responses has been recently described. The latter pathway does not require either presence of ligands (IFNα/β) or catalytic activity of Janus kinases (JAK). Instead, this pathway relies on activation of the PKR-like ER kinase (PERK) and ensuing specific priming phosphorylation of IFNAR1. Here, we describe studies that identify the stress activated p38 protein kinase as an important regulator of IFNAR1 that acts downstream of PERK. Results of the experiments using pharmacologic p38 kinase inhibitors, RNA interference approach, and cells from p38α knock-out mice suggest that p38 kinase activity is required for priming phosphorylation of IFNAR1 in cells undergoing unfolded protein response. We further demonstrate an important role of p38 kinase in the ligand-independent stimulation of IFNAR1 ubiquitination and degradation and ensuing attenuation of IFNα/β signaling and anti-viral defenses. We discuss the distinct importance of p38 kinase in regulating the overall responses to type I IFN in cells that have been already exposed to IFNα/β versus those cells that are yet to encounter these cytokines.

Ubiquitination-mediated down-regulation of signaling receptors plays a key role in restricting the timing of cellular responses to specific ligands (1). This mechanism is especially important for limiting the extent and duration of signaling pathways triggered by those ligands that negatively affect cell proliferation and survival. Such an effect was demonstrated for a number of cytokines, including those that belong to type I interferons (including IFN␣ and IFN␤) known to elicit anti-tumorigenic effects and to mount the antiviral defensive mechanisms (2)(3)(4).
Effects of IFN␣/␤ within cells are attributed to the induction of the interferon-stimulated genes mediated by signal transducers and activators of transcription (STAT1 and STAT2) proteins (5,6). The latter become transcriptionally competent upon their phosphorylation on specific tyrosine residues mediated by JAK (TYK2 and JAK1) that are associated with type I IFN receptor chains (IFNAR1 and IFNAR2c, 2 respectively). JAK themselves become activated as a result of cross-phosphorylation that occurs upon binding of a ligand (e.g. IFN␣ or IFN␤) to the extracellular domains of IFNAR1 and IFNAR2c (reviewed in Refs. [5][6][7][8]. Activation of JAK, particularly of TYK2, is also implicated in activation of the ligand-inducible pathway that leads to the type I IFN receptor down-regulation (9,10). The latter is driven by the endocytosis of the IFNAR1 chain stimulated by a chain-and site-specific ubiquitination of IFNAR1 (11). Ubiquitination of IFNAR1 is catalyzed by the ␤-Trcp E3 ubiquitin ligase (12). This ligase can be recruited to IFNAR1 upon its phosphorylation on specific Ser residues such as Ser-535 in human IFNAR1 or analogous Ser-526 in murine IFNAR1 (12,13). Such phosphorylation is stimulated upon IFN␣/␤ treatment (10,13) in a manner that depends on kinase activity of TYK2 (9,10) and activation of the serine/threonine protein kinase D2 (PKD2) (14).
This ligand-inducible pathway mediates IFNAR1 ubiquitination and degradation in cells that have already encountered IFN␣/␤. Given that activated JAK signals both forward to mediate the functions of IFN␣/␤ (via STAT) and toward IFNAR1 elimination (via PKD2), the JAK-and PKD2-dependent IFNAR1 elimination merely serves to limit the extent of already ongoing IFN␣/␤ signaling. Intriguingly, an existence of a basal pathway that does not require either ligands or JAK activity has been also reported (9). This pathway that relies on Ser-535 phosphorylation by constitutively active casein kinase 1␣ (CK1␣) serves to decrease the basal levels of IFNAR1 and to limit the sensitivity of cells to the future encounters with IFN␣/␤ (15). * This work was supported, in whole or in part, by National Institutes of CK1␣ is a constitutively active kinase, yet its ability to phosphorylate diverse substrates can be further augmented via priming phosphorylation of an adjacent proximal Ser/Thr residues (16). IFNAR1 as a CK1 substrate also abides by this rule: phosphorylation of the degron of IFNAR1 by CK1␣ is robustly increased upon phosphorylation of a conserved priming site (Ser-532 in human IFNAR1, Ser-523 in mouse IFNAR1) (17). Intriguingly, the extent of priming phosphorylation (and, accordingly, of the ligand-independent phosphorylation of IFNAR1 degron that is followed by IFNAR1 ubiquitination and down-regulation) can be increased in cells exposed to stress inducers that cause unfolded protein response (UPR). Among such UPR inducers are pharmacologic agents that target the endoplasmic reticulum (ER) and viruses such as vesicular stomatitis virus (VSV) (17,18).
UPR-stimulated priming phosphorylation, ensuing downregulation of IFNAR1 and attenuation of IFN␣/␤ signaling was dependent on activation of PKR-like ER kinase (PERK, (17,18)). PERK is known to phosphorylate Ser-51 on the eIF2␣ translational regulator leading to a decrease in the overall rate of protein synthesis (reviewed in Ref. 19). Accordingly, eIF2␣ phosphorylation appeared to parallel both degron and priming phosphorylation of IFNAR1 (17,18). Yet, we were unable to detect any phosphorylation of IFNAR1 by PERK (17), indicating that it is another kinase downstream of PERK that mediates phosphorylation of the priming site of IFNAR1 in response to the UPR inducers.
Here, we report the results of pharmacologic and genetic analyses in mouse and human cells, suggesting that stress activated p38 protein kinase is a major regulator of the priming phosphorylation of IFNAR1. Activation of p38 kinase is also required for stimulation of IFNAR1 ubiquitination and degradation and ensuing attenuation of IFN␣/␤ signaling. Furthermore, genetic ablation of p38 kinase augments the antiviral defenses indicating that the modulators of this kinase may be considered for potential use in treatment of viral infections.

EXPERIMENTAL PROCEDURES
Plasmids and Reagents-Vector for bacterial expression of GST-IFNAR1 was described previously (9). shRNA constructs for knockdown of p38␣ or control shRNA against GFP were purchased from Sigma. Inhibitors of p38 kinase (SB203580) and JNK (SP600125) were purchased from EMD Biosciences. PI3K inhibitor LY 294002 was purchased from LC Laboratories. p38 inhibitor VX-702 was purchased from ChemTek. Thapsigargin (TG) and cycloheximide were purchased from Sigma. Human IFN␣2 was purchased from Bio-Sidius S.A., and murine IFN␤ was purchased from PBL.
Cell Culture, Treatment, and Viral Infection-Human HeLa and 2fTGH cells were obtained from ATCC. Mouse embryo fibroblasts from p38 Ϫ/Ϫ mice and their wild type counterparts were kindly provided by Angelo Nebreda (Centro Nacional de Investigaciones Oncológicas, Madrid, Spain). All cell lines were maintained in Dulbecco's modified Eagle's medium supplemented with 10% (v/v) fetal bovine serum (Hyclone) and various selection antibiotics where indicated. To generate human cells, in which p38␣ kinase expression is suppressed through the RNAi approach, 2fTGH cells or Hela cells were transduced by lentiviral particles encoding shRNA against GFP or p38␣ and selected in 2 g/ml of puromycin for 2 weeks. 2fTGHderivative 11.1-Tyk2-null cells reconstituted with catalytically inactive Tyk2 (KR2 cells) were a generous gift of S. Pellegrini (Pasteur Institute, Paris, France). VSV (Indiana serotype, a gift from R. Harty, University of Pennsylvania, Philadelphia, PA) was propagated in HeLa cells. For infections, the cells were inoculated with a multiplicity of infection 0.01 of VSV for 1 h, washed, and incubated with fresh medium for 14 h for observing phosphorylation of IFNAR1. Cells were incubated for 18 h post-infection for assays that evaluated IFNAR1 down-regulation or the extent of type I IFN signaling.
In Vitro Kinase Assay-Lysates from untreated or VSV-infected cells (4 g) were cleared of CK1␣ by immunodepletion as described previously (15). Bacterially expressed and purified GST-IFNAR1 (2 g) used as a substrate was incubated with these lysates in kinase buffer (25 mM Tris HCl, pH 7.4, 10 mM MgCl 2 , 1 mM NaF, 1 mM NaVO 3 ) and ATP (1 mM). The reactions were carried out at 30°C for 30 min with shaking at 600 rpm on the tabletop incubator. The products were analyzed by immunoblotting using phosphospecific antibody and antibodies against GST as described previously (15).

RESULTS
We aimed to characterize a kinase activity that is activated in cells undergoing UPR and is responsible for the priming phosphorylation of IFNAR1. PERK-dependent activation of diverse kinases, including Jun N-terminal kinase (JNK), p38 kinase, and phosphoinositide 3-kinase (PI3K)-Akt was demonstrated in cells undergoing UPR (22,23). As UPR inducers, we used a pharmacologic agent that inhibits ER Ca 2ϩ ATPase TG or infection with VSV. In the latter case, we used a low dose of virus (MOI of 0.01) and timed the harvesting of infected cell to a point where viral proteins are already synthesized and UPR is already activated (ϳ14 h post-infection as judged by expression of VSV-M and phosphorylation of Ser-51 of eIF2␣), and we are capable of detecting phosphorylation of IFNAR1 just before a dramatic down-regulation in the levels of this protein (ϳ17-18 h post-infection, see below).
We initially used human KR2 cells that harbor catalytically inactive form of TYK2 (24, 25); these cells were shown incapa-p38 Kinase Regulates IFNAR1 Stability and Signaling ble of inducing IFNAR1 degron phosphorylation and stimulating IFNAR1 ubiquitination and down-regulation in response to IFN␣ (10) or IFN␤. 3 Consistent with previous reports (17,18) infection with VSV robustly stimulated phosphorylation of both the IFNAR1 degron (Ser-535) and priming site (Ser-532). We then examined the effect of pharmacologic inhibitors of several UPR-induced kinases including PI3K inhibitor LY294002, JNK inhibitor SP600125, and p38 inhibitor VX702. Pre-treatment of KR2 cells with the latter two agents after VSV infection yet prior to cell harvesting noticeably decreased phosphorylation of IFNAR1 without affecting either expression of viral protein VSV-M or specific phosphorylation of eIF2␣ (Fig.  1A). These results suggest that activation of stress activated protein kinases downstream of the UPR pathway may contribute to phosphorylation of IFNAR1.
The effects of p38 inhibitor VX702 were more robust (Fig.  1A); furthermore, the JNK inhibitor SP600125 was reported to also attenuate p38 kinase activity (26). Given that, we hypothesized that p38 kinase might be a regulator of IFNAR1 phosphorylation. Either VX702 or another inhibitor of p38 kinase SB203580 prevented induction of priming phosphorylation of IFNAR1 on S532 in response to either VSV infection or TG treatment in HeLa cells (Fig. 1B). We switched to HeLa cells or KR2-isogenic wild type fibrosarcoma 2fTGH cells because they were shown to respond to IFN␣ by stimulated phosphorylation of IFNAR1 degron (13,17) and, therefore, enabled us to compare the effects of UPR with those of IFN␣. Unlike VSV or TG, treatment with IFN␣ did not induce phosphorylation of eIF2␣ (Fig. 1B). Consistent with previous reports (11,13,17), IFN␣ was not efficient in stimulating the priming phosphorylation on Ser-532, yet it robustly induced IFNAR1 degron phosphorylation on Ser-535. Neither VX702 nor SB203580 prevented Ser-535 phosphorylation by IFN␣, whereas being efficient in decreasing this phosphorylation in cells that received VSV or TG (Fig. 1B). These results further implicate p38 kinase activity in IFNAR1 phosphorylation induced by the stimuli that trigger the ligand-independent pathway.
We next used a genetic approach to complement these pharmacologic analyses. Various tissues are known to express four related yet distinct isoforms of p38 (␣, ␤, ␥, and ␦) of which p38␣ and p38␤ are sensitive to SB203580 (27), and p38␣ protein is a ubiquitously expressed isoform in all tissues (28). We used shRNA against p38␣ kinase, which decreased the levels of endogenous protein by 60 -90% in either HeLa or 2fTGH cells (Figs. 2, A and B, 3A, and 4B). Knockdown of p38␣ kinase in 2fTGH cells did not affect VSV-or TG-induced phosphorylation of eIF2␣ ( Fig. 2A). Yet, shRNA against p38␣ noticeably decreased the priming phosphorylation of IFNAR1 on Ser-532 in cells that received either TG or VSV. Importantly, whereas a similar effect of p38␣ kinase knockdown was detected on the IFNAR1 degron phosphorylation on Ser-535, this phosphorylation was not affected in cells treated with IFN␣ ( Fig. 2A). Similar results were obtained in HeLa cells (Fig. 2B). These results implicate p38␣ in phosphorylation of IFNAR1 that occurs through the ligand-independent pathway.  Analyses of phosphorylation and levels of IFNAR1, p38 kinase, and eIF2␣ were carried out as described in the legend to Fig. 1. B, analyses of IFNAR1 phosphorylation in HeLa cells were carried out as outlined in A. C, MEFs obtained from p38␣ Ϫ/Ϫ or wild type mice were infected with VSV (MOI of 0.01 for 14 h) or treated with TG (1 M) or mouse IFN␤ (5000 international units/ml) for 30 min. Whole cell lysates (WCL) were used for immunoprecipitating IFNAR1. Analyses of phosphorylation and levels of IFNAR1, p38␣ kinase, and eIF2␣ were carried out as described in A. D, whole cell lysates were obtained from wild type or p38␣ knock-out MEFs that were either infected with VSV (MOI of 0.01 for 14 h) or not. These lysates were used as a source of kinase activity to phosphorylate on the recombinant GST-IFNAR1 (1 g) in an in vitro kinase reaction as described under "Experimental Procedures." Phosphorylation of GST-IFNAR1 on Ser-532 and levels of this protein were analyzed by immunoblotting using indicated antibodies.
We sought to corroborate these conclusions in mouse cells. Induction of the phosphorylation of the priming site (Ser-523 in mouse IFNAR1) was observed in mouse embryo fibroblasts (MEFs) derived from the wild type animals but not from p38␣ knock-out mice (Fig. 2C). Activation of PERK evident by eIF2␣ phosphorylation was not affected in these fibroblasts suggesting that the role of p38␣ in priming phosphorylation occurs downstream of PERK. When lysates from these MEFs were used as a source of kinase activity to phosphorylate bacterially expressed GST-IFNAR1 protein on Ser-532 in vitro, we detected a marked increase in activity of a priming kinase in wild type MEFs infected with VSV. This increase was not evident in MEFs from p38␣ knock-out mice (Fig. 2D). Furthermore, whereas phosphorylation of IFNAR1 degron (Ser-526 in mouse IFNAR1) in response to UPR inducers followed the same pattern, status of p38␣ had no bearing on the effects of mouse IFN␤ (Fig. 2B). This result indicates that p38␣ is required for the ligand-independent pathway, yet it is dispensable for the ligand-induced signaling toward phosphorylation of IFNAR1 degron.
Degron phosphorylation of IFNAR1 is essential for ability of this receptor to recruit the ␤-Trcp E3 ubiquitin ligase that facilitates ubiquitination of IFNAR1 (12). Consistent with this paradigm, both IFN␣ and non-ligand inducers of degron phosphorylation (VSV and TG) stimulated ubiquitination of IFNAR1 in 2fTGH cells that received an irrelevant control shRNA (Fig.  3A). Under these conditions, VSV and TG induced p38 kinase (as evident from specific p38 phosphorylation) much more efficiently than IFN␣. Knockdown of p38␣ noticeably decreased the IFNAR1 ubiquitination induced by either VSV or TG but not by IFN␣ (Fig. 3A). Furthermore, similar studies in mouse cells revealed that MEFs from p38␣ knock-out mice were not capable of responding to VSV or TG by increasing IFNAR1 ubiquitination (Fig. 3B). This defect could not be attributed to hypothetical alterations in the ␤-Trcp-dependent ubiquitination machinery that may have occurred in the absence of p38␣ because mouse IFN␤ robustly induced IFNAR1 ubiquitination in either wild type or p38␣ knock-out cells (Fig. 3B). These results suggest that p38␣ is essential for IFNAR1 ubiquitination simulated within the ligand-independent pathway. We next sought to determine whether p38 kinase plays a role in the regulation of IFNAR1 stability and IFNAR1-mediated signaling. Consistent with previous reports (18), infection of HeLa cells with VSV accelerated the rate of turnover of IFNAR1 assessed via a cycloheximide chase assay (Fig. 4A). A similar effect was observed either in 2fTGH cells that received irrelevant control shRNA against GFP (Fig. 4B) or in wild type MEFs (Fig. 4C). Under these conditions, the acceleration of the rate of proteolytic turnover of human or murine IFNAR1 could be reverted by either inhibition of p38 kinase using treatment with VX702 inhibitor (Fig. 4A) or knockdown of p38␣ using shRNA (Fig. 4B) or genetic ablation of p38␣ in MEFs (Fig. 4C). These data implicate p38 kinase in general (and its ␣-isoform in particular) in the regulation of IFNAR1 degradation in cells exposed to VSV.
Infection with low doses of VSV was shown to induce IFNAR1 phosphorylation and down-regulation in a manner that required PERK activation (18). We carried out a time course for PERK and p38 activation after a pulse infection of HeLa cells. Activation of PERK (assessed by eIF2␣ phosphorylation) preceded activation of p38 kinase as evidenced from its phosphorylation; both events occurred before noticeable decrease in the levels of endogenous IFNAR1 (Fig. 5A). This A, 2fTGH cells stably transduced with control shRNA or shRNA against p38␣ were infected with 0.01 MOI of VSV or treated with 1 M of TG or 5000 international units/ml of IFN␣ as described in the legend to Fig. 2. Cell lysates obtained under denaturing conditions were subjected to IFNAR1 immunoprecipitation and analysis by immunoblotting using the indicated antibodies. Phosphorylation and levels of p38 kinase in the lysates is also shown. B, experiment described in A was carried out on MEFs from wild type or p38␣ knockout mice. Mouse IFN␤ (5000 international units/ml) was used instead of human IFN␣. Analyses of ubiquitination and levels of IFNAR1 and of phosphorylation and levels of p38␣ are shown. FIGURE 4. p38␣ kinase regulates IFNAR1 stability. A, HeLa cells were exposed (or not) to VSV (MOI of 0.01) for 1 h and then incubated for 8 h. After that, cells were treated or not with or not with VX702 (1 M) for 6 h before being treated with cycloheximide (CHX; 50 g/ml) for the indicated times and harvested. Levels of IFNAR1 were analyzed by immunoprecipitation followed by immunoblotting. Levels of ␤-actin in the supernatants of the immunoprecipitation reactions were determined to control equal loading. B, 2fTGH cells that harbor control shRNA or shRNA against p38␣ were infected, treated, and analyzed as described in A. Levels of p38␣ kinase in the whole cell lysates was determined by immunoblotting using anti-p38␣ antibody to determine the efficacy of knockdown. C, MEFs obtained from p38␣ Ϫ/Ϫ mice or wild type mice were processed and analyzed as described in B.

p38 Kinase Regulates IFNAR1 Stability and Signaling
result suggests that sequential activation of PERK and p38 may promote down-regulation of IFNAR1. Whereas the role of PERK in IFNAR1 down-regulation in VSV-infected cells has been established previously (18), here, we focused on the importance of p38 kinase. Either inhibition of p38 kinase using VX702 or knockdown of p38␣ kinase using RNAi prevented down-regulation of IFNAR1 in VSV-infected human cells (Fig.  5, B and C). In addition, treatment with TG did not efficiently decrease IFNAR1 levels in cells that received shRNA against p38␣ knockdown (Fig. 5D). Furthermore, neither VSV infection nor TG treatment caused a robust down-regulation of murine IFNAR1 in MEFs from p38␣-null mice (Fig. 5, E and F). These data strongly suggest that p38 kinase plays an important role in regulating IFNAR1 levels.
Given that cells from heterozygous Ifnar1 ϩ/Ϫ mice display a partially decreased response to type I IFN, it has been proposed that levels of IFNAR1 play an important role in IFN signaling (29). Indeed, pre-exposure of human or murine cells to PERK inducers such as VSV was shown to attenuate their responses to the future encounter with IFN␣/␤ (18). In line with these data, IFN␣-induced phosphorylation of STAT1 was decreased in VSV-infected cells (Fig. 6A). Under these conditions, treatment of cells with p38 kinase inhibitor VX702 noticeably restored the extent of IFN␣ signaling (Fig. 6A). Similar results were observed when modulation of p38␣ status was achieved using the RNAi approach (Fig. 6B). Furthermore, VSV-mediated inhibition of STAT1 phosphorylation induced by exogenous IFN␤ was seen in wild type MEFs but not in MEFs from p38␣ knock-out mice (Fig. 6C). These results implicate p38 kinase in controlling IFN␣/␤ signaling in infected cells.
We further sought to corroborate the role of p38 kinase in cells that undergo UPR using a non-viral UPR inducer. Whereas pre-treatment of either human or mouse cells with TG noticeably decreased the efficacy of STAT1 activation by type I IFN, this efficacy could be restored (at least in part) by either knockdown or knock-out of p38␣ (Fig. 6, D and E). These results suggest that p38 kinase plays an important role in the regulation of proximal IFN␣/␤ signaling in cells subjected to UPR.
While carrying out the experiments that use p38 inhibitors, we noticed that to achieve a similar viral load (as shown in Fig.  1A), it was important to first infect cells with VSV and then, upon 10 -12 h of incubation, to use VX702 or SB203580. Otherwise, if p38 inhibitors were added earlier (prior to infection or together with infection), they robustly decreased the efficacy of VSV replication (data not shown). Given these data as well as an important role of type I IFN in induction of an anti-viral state, we hypothesized that negative regulation of p38 kinase should stabilize IFNAR1 and increase the resistance of cells to viral infections. We tested this possibility using genetic approaches to avoid confounding issues related to the specificity of phar-  Whole cell lysates were analyzed for STAT1 levels and tyrosine phosphorylation using the indicated antibodies. B, 2fTGH cells stably transduced with shp38␣ or control cells (shCON) were treated and analyzed as described in A. C, MEFs obtained from wild type mice or from p38␣ knock-out mice were exposed to VSV (0.01 MOI) as indicated. Twenty hours following infection, cells were treated with mouse IFN␤ (50 international units/ml for 30 min) and harvested. Analyses of STAT1 phosphorylation and levels are shown. D, analyses of STAT1 levels and phosphorylation in lysates from 2fTGH cells that received indicated shRNA were pretreated or not with TG (1 M, 3 h) and then treated with human IFN␣ (200 international units/ml for 30 min) as indicated. E, analyses of STAT1 levels and phosphorylation in lysates from MEFs pretreated or not with TG (1 M, 3 h) and then treated with mouse IFN␤ (50 international units/ml for 30 min) as indicated.
macologic inhibitors of p38. Knockdown of p38␣ kinase in human 2fTGH cells has dramatically decreased the efficacy of VSV infection as evident from a decreased viral titer and expression of VSV-M protein (Fig. 7A). Similarly, replication of VSV was much less evident in MEFs from p38␣ knock-out mice compared with their wild type counterparts (Fig. 7B). These data strongly suggest that p38 kinase plays an important role in regulation the anti-viral defenses.

DISCUSSION
In this study, we aimed to characterize the PERK-dependent pathway that leads to the ligand-independent priming phosphorylation of IFNAR1 followed by phosphorylation of IFNAR1 degron, ubiquitination, and degradation of IFNAR1. Pharmacologic analyses using inhibitors against protein kinases known to be activated by UPR downstream of PERK implicated p38 kinase in the ligand-independent induction of IFNAR1 phosphorylation (Fig. 1). Activation of p38 kinase by inducers of UPR has been long established in the literature (22,30). Given that p38 kinase can be activated by tumor necrosis factor ␣ (TNF␣) (31,32), identification of TNF receptor associated factor-2 as an interacting protein for IRE1 transmembrane kinase/endoribonuclease (33), a known sensor of UPR (19), initially suggested the role of IRE1 in UPR-induced activation of stress-activated protein kinases (e.g. JNK (33)). However, subsequent analysis using genetically defined cell system firmly placed p38 kinase activation by UPR inducers downstream of PERK (22). In our study, we further observed that modulation of p38␣ expression by knockdown or knock-out approaches had a profound negative effect on IFNAR1 phosphorylation stimulated by UPR inducers but not by IFN␣/␤ (Fig. 2).
Subsequent pharmacologic and genetic analyses revealed p38 kinase as a key regulator of IFNAR1 ubiquitination, degradation, down-regulation, and downstream signaling (Figs. 3-6). As VSV infection presents a convenient and robust yet inexpensive tool for inducing UPR and PERK (34), we have widely used this approach in our studies. Cells infected with VSV indeed displayed phosphorylation of eIF2␣ paralleled by activation of p38 kinase and subsequent down-regulation of IFNAR1 (Fig. 5A). Intriguingly, we found that the VSV load is decreased in human or mouse cells where p38␣ was either knocked out or knocked down (Fig. 7). Thus, the interpretation of inefficient IFNAR1 down-regulation, degradation, and signaling seen in p38-defective cells should include both the abrogation of p38-mediated effects on IFNAR1 itself (also evident from experiments that used p38 pharmacologic inhibitors added after VSV infection has been completed) and resistance of these cells to viral replication. Nevertheless, all presented data are supportive of our conclusions that p38 kinase plays an important role in regulating type I IFN signaling and antiviral defenses. Furthermore, experiments that use TG as a non-viral UPR stimulus clearly demonstrate that UPR-induced IFNAR1 phosphorylation ( Figs. 1 and 2), IFNAR1 ubiquitination (Fig. 3), down-regulation of IFNAR1 (Fig. 5, D and F) and attenuation of IFN␣/␤ signaling (Fig. 6, D and E) require the activity of p38␣. Whereas genetic data obtained in fibroblasts and fibrosarcoma cells highlight the importance of p38␣ form, the importance of other forms (␤, ␥, and ␦) in other cell types cannot be ruled out.
Although present data establish the role of p38 kinase in priming phosphorylation of IFNAR1, it remains to be seen whether this regulation is direct. Our pilot biochemical data suggest that immunopurified p38 kinase is capable of phosphorylating the priming site (Ser-532) within IFNAR1 in an in vitro kinase reaction. 4 However, given a known preference of this p38 kinase for the proline-directed Ser and Thr residues as phospho-acceptor sites ((S/T)P (35)) and the fact that the priming site on IFNAR1 does not conform to these criteria (QTSQ), it is plausible that the direct phosphorylation of Ser-532 in cells might be carried out by another kinase that is activated downstream of p38 kinase. A number of such candidate kinases including Mnk, Msk, and MK2/3 that physically interact with p38 kinase and become activated in a manner that requires p38 catalytic activity has been identified (36). Remarkably, some of these kinases were shown to be induced by IFN␣/␤ (37,38).
Under conditions used in this study, status of p38 kinase did not affect phosphorylation or ubiquitination of IFNAR1 stimulated by IFN␣/␤ (Figs. 1-3) suggesting that functions of p38 might be limited to the ligand-independent pathway. This scenario is very intriguing given that, in a number of influential studies, the inducing effect of IFN␣/␤ on p38 kinases as well as the kinases that function downstream of p38 was clearly demonstrated in diverse cell types (5, 39 -41). Whereas it is plausible that IFN␣ and/or IFN␤ obtained from various sources may possess diverse ability to induce p38 kinase, 4 there might be another explanation for this paradox. Our studies have demonstrated that IFNAR1 degron phosphorylation induced by the ligand does not require priming phosphorylation (17). Furthermore, we recently identified TYK2-dependent activation of PKD2 as a mechanism essential for IFN␣/␤-induced IFNAR1 degron phosphorylation, ubiquitination, and degradation (14). Given that p38 kinase mediates priming phosphorylation-de-4 S. Bhattacharya and S. Y. Fuchs, unpublished data. p38 Kinase Regulates IFNAR1 Stability and Signaling pendent, TYK2-independent IFNAR1 ubiquitination and degradation (Ref. 17 and present results), the activity of p38 might be indeed dispensable for the ligand-inducible pathway.
In discussing the role of p38 kinase in regulating the cellular responses to type I IFN, it is also critical to distinctly evaluate the importance of p38 in degradation of IFNAR1 in cells that have not yet encountered IFN␣/␤ and specific function of p38 within the context of IFN␣/␤ signaling and subsequent transcriptional events. Compelling studies by Platanias and others (42,43) clearly demonstrated that stimulation of p38 activity in cells treated with IFN␣/␤ does not affect the extent of proximal signaling (e.g. STAT tyrosine phosphorylation and DNA binding). Yet, induction of p38 activity by type I IFN contributes to the overall activation on the interferon-stimulated response elements leading to the superinduction of transcription of IFNstimulated genes (43,44). Accordingly, p38 kinase and its downstream kinase effectors play an important role in stimulating the outcome of IFN␣ signaling (reviewed in Ref. 41). Conversely, when activated within the context of UPR, p38 kinase functions to promote down-regulation of IFNAR1 and to decrease the extent of proximal IFN signaling (STAT1 activation) as well as the sensitivity of cells to the future encounters with IFN␣/␤ (this study).
Altogether, available data suggest that the role of p38 kinase in the cell defenses against viruses might be fairly complex. On one hand, presented here data suggest that UPR-induced p38 kinase activation supports viral infections, perhaps in part due to its ability to stimulate the degradation of IFNAR1 prior to activation of Type I IFN receptor. Under conditions used in this study, genetic ablation of p38␣ kinase rendered the cells more resistant to infection with VSV. These results are in line with published reports that inhibitors of p38 kinase negatively affect replication of other viruses, including human cytomegalovirus (45), herpes simplex virus (46), human immunodeficiency virus (47), Japanese encephalitis virus (48), coxsackievirus (49), hepatitis B virus (50), and human coronavirus 229E (51) in diverse cell types. In contrast, activation of p38 kinase in the cells that have been already treated with IFN before they encounter virus should augment the IFN-induced anti-viral state via p38-dependent stimulation of transcription of the IFN-stimulated genes. Indeed, p38 kinase inhibitor treatment attenuated the IFN-stimulated protection of cells against the cytopathogenic effects of encephalomyocarditis virus in leukemic KT1 cells (52). Future detailed delineation of the role of p38 kinase and its downstream effectors in antiviral resistance may help to develop novel agents for treatment of viral infections.