Tyrosine Phosphorylation of Protein Kinase D2 Mediates Ligand-inducible Elimination of the Type 1 Interferon Receptor*

Type 1 interferons (including IFNα/β) activate their cell surface receptor to induce the intracellular signal transduction pathways that play an important role in host defenses against infectious agents and tumors. The extent of cellular responses to IFNα is limited by several important mechanisms including the ligand-stimulated and specific serine phosphorylation-dependent degradation of the IFNAR1 chain of Type 1 IFN receptor. Previous studies revealed that acceleration of IFNAR1 degradation upon IFN stimulation requires activities of tyrosine kinase TYK2 and serine/threonine protein kinase D2 (PKD2), whose recruitment to IFNAR1 is also induced by the ligand. Here we report that activation of PKD2 by IFNα (but not its recruitment to the receptor) depends on TYK2 catalytic activity. PKD2 undergoes IFNα-inducible tyrosine phosphorylation on specific phospho-acceptor site (Tyr-438) within the plekstrin homology domain. Activated TYK2 is capable of facilitating this phosphorylation in vitro. Tyrosine phosphorylation of PKD2 is required for IFNα-stimulated activation of this kinase as well as for efficient serine phosphorylation and degradation of IFNAR1 and ensuing restriction of the extent of cellular responses to IFNα.

Type 1 interferons (IFNs) play a critical role in modulating the immune responses against many pathogens and directly mounting the anti-viral defenses (for review, see Refs. [1][2][3][4]. A family of these cytokines (including IFN␤ and diverse species of IFN␣) elicit the responses through the cognate Type 1 IFN receptor on the cell surface (5)(6)(7). Interaction of IFNAR1 2 and IFNAR2c chains of this receptor with the ligands triggers the signal transduction pathway that involves activation of the Janus kinases, JAK (TYK2 and JAK1) and subsequent tyrosine phosphorylation and ensuing activation of signal transducers and activators of transcription (STAT1 and STAT2). Transcriptionally active STAT1/2 further associate with IRF3, translocate to the nucleus and interact with IFN-stimulated response elements (ISRE) to induce de novo expression of the IFN-stimulated genes. Protein products of these mediate immunomodulatory and anti-viral responses as well as inhibit proliferation and survival of cells exposed to Type 1 IFN (for review, see Refs. [1][2][3][4]. To alleviate these detrimental effects of Type 1 IFN, cells evolved to develop the mechanisms that limit the magnitude and duration of their responses to these cytokines. For example, some of the IFN-stimulated genes encode the proteins that may interfere with the recruitment of JAK to IFNAR chains (8). Additional modes of negative regulation that is commonly shared between most of cytokines-induced JAK-STAT pathways include inhibition of JAK activity/stimulation of JAK degradation by SOCS proteins, inhibition of tyrosine phosphorylation by phosphatases, and inhibition of STAT-induced transcription by PIAS (for review, see Refs. 6,9). In addition to these modes of negative regulation, which occur in cells that have already executed the IFN-induced programs of signal transduction and transcriptional activation, a rapid elimination of Type 1 IFN receptors from the cell surface serves as a rapid and important mechanism that limits cell sensitivity to continuous exposure to the ligands.
Elimination of the entire receptor is driven by ubiquitination and subsequent endocytosis and lysosomal degradation of the IFNAR1 chain (6,10). This ubiquitination is facilitated by the SCF ␤Trcp E3 ubiquitin ligase that is recruited to IFNAR1 upon its phosphorylation on specific Ser residues within a defined degron ( 534 DSGNYS) (11,12). Stimulation of this phosphorylation in cells exposed to IFN␣/␤ appears to play a key role in subsequent recruitment of ␤Trcp and stimulation of IFNAR1 ubiquitination and degradation (11,13) in a manner that requires catalytic activity of TYK2 (14,15).
Our previous studies also revealed the role of protein kinase D2 (PKD2) in ligand-stimulated IFNAR1 phosphorylation, ubiquitination, and degradation (16). Whereas an increase in both recruitment of PKD2 to IFNAR1 and in catalytic activity of PKD2 were observed in cells treated with IFN␣/␤, the mechanisms that govern the ligand-inducible JAK and PKD2-stimulated phosphorylation of IFNAR1 degron remains largely to be understood. Here we report that kinase function of TYK2 is dispensable for basal activity of PKD2 or for induction of its recruitment to IFNAR1. Instead, TYK2 activity plays an important role in stimulation of kinase activity of PKD2 by IFN␣ through phosphorylation of specific tyrosine residue, Tyr-438. The latter mechanism is important for IFNAR1 degradation and for tempering the IFN␣-induced signaling and anti-viral defenses.

EXPERIMENTAL PROCEDURES
Plasmids and Reagents-Vectors for mammalian expression of FLAG-IFNAR1 and bacterial expression of GST-IFNAR1 (12), and HA-tagged TYK2 (a gift from J. Krolewski) (17), as well as the 5ϫISRE-luciferase reporter (a gift from C. Horvath) (18) have been described elsewhere. Vectors for mammalian expression of human GST-tagged PKD2 (19) were kindly provided by V. Malhotra. Silent mutations, as well as replacement of Tyr-438 with tyrosine were generated by site-directed mutagenesis. All resulting mutants were verified by dideoxy sequencing. Lentiviral shRNAs against PKD2 constructed in the backbone of the pLKO.1-puro vector were purchased from Sigma (MISSION shRNA, SHGLY-NM_016457). Control shRNA vector targeted against GFP (20) was a gift from J. W. Harper. Recombinant human IFN␣2 (Roferon) was purchased from Roche Applied Science. Cycloheximide and other chemicals were purchased from Sigma.
Transient transfections of cells using Lipofectamine Plus (Invitrogen) were carried out according to the manufacturer's recommendations. For stable transductions, replication-deficient lentiviral particles encoding shRNA against PKD2 or vector control were prepared via co-transfecting 293T cells with three other helper vectors as described previously (21). Viral supernatants were concentrated by PEG8000 precipitation and used to infect HeLa cells or 2fTGH cells in the presence of Polybrene (3 g/ml; Sigma). Cells were selected and maintained in the presence of puromycin (2 g/ml).
In Vitro Kinase Assays-In vitro phosphorylation of PKD2 (purified from HeLa cells) by HA-tagged TYK2 immunopurified from 293T cells (untreated or treated with IFN␣) or by recombinant Src (purchased from Cell Signaling) was carried out in a total volume of 20 l in 50 mM MOPS (pH 7.4), 10 mM MgCl 2 , 5 mM MnCl 2 , 2 mM DTT, and 0.2 mM ATP at 30°C for 20 min. The samples were resolved by SDS-PAGE and analyzed by immunoblotting with an anti-phosphotyrosine antibody (4G10).
Immunocytochemistry-2fTGH or 11.1 cells were plated on coverslips for 24 h and then treated with IFN␣ (2,000 units/ml) for 5 min. Cells were fixed by 3.7% paraformaldehyde at once and permeabilized by 0.3% Triton X-100. Then cells were incubated with anti-PKD2 and anti-TYK2 (sc-5271) for 1.5 h followed by incubation with Alexa Fluor 488 and 594 (Invitrogen) for 1 h at 25°C. DAPI (Sigma) was used at 1 g/ml for 2 min. The cells were embedded in Prolong Gold Antifade mounting medium (Invitrogen). The confocal images were acquired on a Leica Inverted DMI4000 microscope equipped with a 100ϫ HCX PL APO 1.46 NA oil objective, a Yokogawa CSU-10 spinning disk confocal system, and an ImagEM 16-bit cooled EMCCD camera (Hamamatus). Laser excitation was provided by a 488-nm (Spectra Physics) and a 561-nm laser (Cobolt Jive) controlled through LMM5 (Spectral Applied Research).
Viral Infection-The anti-viral effect of Type 1 IFN was determined in the derivatives of 2fTGH cells that harbor diverse PKD2 status. These cells were pretreated with IFN␣ (5 units/ml) for 1 h prior to infection with vesicular stomatitis virus (Indiana serotype, a gift from R. Harty; propagated in HeLa cells) at a multiplicity of infection of 0.1 for 1 h. After removing the virus inoculums, cells were then fed with fresh medium and incubated for 20 h. Culture supernatant was harvested, and the viral titer was determined in HeLa cells overlaid with methylcellulose as described elsewhere (28) and plaqueforming units (pfu/ml) calculated.

RESULTS
IFN␣/␤-stimulated and TYK2-dependent phosphorylation on Ser-535 was shown to play a key role in ligand-inducible ubiquitination and degradation of IFNAR1 (11,12,15). Furthermore, our recent studies revealed a key role for PKD2 in this signaling (16). Although PKD2 was expressed in either human fibrosarcoma 2fTGH cells or their isogenic TYK2-null 11.1 derivatives, the activation of PKD2 by IFN␣ was not observed in 11.1 cells (16). Previous work by Pellegrini group established that these cells fail to retain IFNAR1 on cell surface due to constitutive endocytosis of IFNAR1 (29). Reconstitution of these cells with wild-type TYK2 restored the sensitivity of resulting cells (WT-5) to Type 1 IFN. When isogenic cell line (KR-2) was engineered using catalytically incompetent TYK2, it exhibited an attenuated JAK-STAT signaling (15,30) and could only moderately respond to IFN␤ (31). Importantly, phosphorylation of IFNAR1 on Ser-535 in these cells was not robustly stimulated by IFN␣ treatment (Fig. 1A) (14,15).
We sought to delineate the mechanisms by which PKD2 can be coerced to contribute to Ser-535 phosphorylation of IFNAR1 by IFN␣ and to determine the role of TYK2 in these mechanisms. Our hypothesis was that IFN␣-induced activity of JAK/TYK2 may stimulate either recruitment of PKD2 to IFNAR1 or PKD2 kinase activity or both. Consistent with our previous results (16), co-immunoprecipitation analysis determined that PKD2 can interact with IFNAR1 and that this interaction is stimulated by IFN␣ treatment in 293T cells (Fig. 1B).
Similar results were observed in WT-5 cells that express wild-type TYK2 and were capable of responding to IFN␣ by activating STAT1 and phosphorylating IFNAR1 on Ser-535 (Fig. 1C). A greater level of PKD2 recovered in IFNAR1 immunoprecipitates was not reflective of overall levels of PKD2 that remained comparable in either treated or untreated cells under these conditions (see Fig. 3A). Importantly, upon adding IFN␣, neither STAT1 tyrosine phosphorylation nor Ser-535 phosphorylation of IFNAR1 was observed in KR-2 cells that express kinase-deficient TYK2. Yet, these KR-2 cells exhibited a ligandstimulated increase in IFNAR1-PKD2 binding similar that seen in WT-5 cells (Fig. 1C). This result suggests that IFN␣-stimulated recruitment of PKD2 to IFNAR1 does not depend on catalytic activity of TYK2.
We next sought to determine whether TYK2 activity is involved in regulating the IFN␣-induced increase in PKD2 kinase activity. To this end, we expressed GST-tagged PKD2 in either WT-5 or KR-2 cells, treated these cells with IFN␣ for various times, purified PKD2 using affinity beads, and incubated this kinase with bacterially produced GST-IFNAR1 protein as a substrate and ATP in vitro. The resulting kinase activity toward phosphorylation of Ser-535 on GST-IFNAR1 was measured by immunoblotting using phospho-specific antibody as the mode for detection. Basal activity of GST-PKD2 (assessed by its ability to phosphorylate on Ser-535 in vitro) was detected in untreated WT-5 cells (Fig. 2A). Remarkably, whereas treatment of cells with IFN␣ noticeably increased the kinase activity of GST-PKD2 toward Ser-535 phosphorylation in WT-5 cells, this induction was not seen in KR-2 cells (Fig. 2A). Importantly, the status of TYK2 had no bearing on the basal activity of expressed PKD2. These data suggest that TYK2 stimulated by IFN␣ contributes to catalytic activation of PKD2 in response to IFN␣.
We next investigated whether PKD2 and TYK2 interact in cells. Immunoprecipitation reactions carried out using anti-TYK2 antibody revealed the presence of PKD2 in a complex with TYK2 in the untreated cells. Treatment of cells with IFN␣ rapidly increased this interaction within 2-5 min followed by subsequent decrease in this interaction at later time points (Fig.  2B). We further sought to determine localization of endogenous PKD2 and TYK2 proteins using immunocytochemistry. We first characterized suitability of antibodies for these analyses. Signals detected using an antibody against TYK2 were readily seen in 2fTGH cells but not in TYK2-negative 11.1 cells (Fig.  2C). The pattern of staining in 2fTGH cells was similar to that reported by Lukashova et al. in monocytoid MonoMac-1 cells (32). In addition, immunochemical detection of PKD2 was much more efficient in 2fTGH cells that received control shRNA compared with those that were transduced with shRNA against PKD2 (Fig. 2D). We further used these antibodies against TYK2 and PKD2 for co-localization studies in 2fTGH cells. These analyses showed that PKD2 and TYK2 co-localize in untreated cells and that this co-localization is noticeably increased in cells treated with IFN␣ (Fig. 2E). In all, these data suggest that PKD2 is capable of interacting with TYK2 and that this interaction can be transiently stimulated by the ligand.
Two nonexclusive mechanisms have been proposed for activation of a better studied PKD2-related kinase PKD1 by numerous stimuli. One is a phorbol ester-induced phosphorylation of the activation loop of PKD on Ser-744/Ser-748 (33,34) attributed to function of protein kinases C (PKC) (33,35). Another is a tyrosine phosphorylation of Tyr-463 that relieves an autoinhibitory effect of the plekstrin homology domain; this phosphorylation could be stimulated by Src (36). Intriguingly, previous reports demonstrated that PKD2 can undergo tyrosine phosphorylation in cells (37). We next assessed phosphorylation status of endogenous PKD2 in WT-5 and KR-2 cells. Phosphorylation of Ser-710 (analog of Ser-748 within PKD1) was indeed stimulated upon treating either WT-5 or KR-2 cells with IFN␣ (Fig. 3A). This result suggests that, similar to an increase in PKD2-IFNAR1 binding (Fig. 1), phosphorylation of Ser-710 can be stimulated by IFN␣ independently of TYK2 kinase activity.
As TYK2 activity contributes to the ligand-induced catalytic activation of PKD2 ( Fig. 2A) and phosphorylation of IFNAR1 on Ser-535 (15), it is plausible that IFN␣-induced phosphorylation on Ser-710 is not sufficient for activation of PKD2 by this  stimulus and that an additional signal is required. Given that TYK2 is a tyrosine kinase we next assessed tyrosine phosphorylation of PKD2. Indeed a ligand-inducible phosphorylation on Tyr residues observed on immunoprecipitated endogenous PKD2 in WT-5 (Fig. 3A) or HeLa (Fig. 3B) but not in KR-2 cells (Fig. 3A). The kinetics of this reaction in TYK2-competent cells revealed a peak at 2-5 min (Fig. 3, A and B) that resembled the kinetics of the PKD2-TYK2 interaction shown in Fig. 2B. In all, these results are consistent with a hypothesis that IFN␣-induced catalytic activity of TYK2 directly contributes to Tyr phosphorylation of PKD2.
To test this possibility, we turned to kinase assays in vitro. In these experiments, the HA-tagged TYK2 was expressed in and immunopurified from 293T cells, which were treated or not with IFN␣ prior to harvesting. After that, HA-TYK2 was incubated with PKD2 (purified from serum-starved and untreated HeLa cells) in the presence of ATP. The reaction was analyzed by immunoblotting using an anti-phosphotyrosine antibody. In this assay, incubation of PKD2 with ATP in the absence of added tyrosine kinase did not produce any phosphotyrosine signals consistent with the fact that PKD2 is a Ser/Thr protein kinase (Fig. 3C, lane 1). Conversely, the tyrosine phosphoryla-FIGURE 2. IFN␣ stimulates kinase activity of PKD2 and its interaction with TYK2. A, activity of GST-PKD2 expressed in cells harboring wild-type or kinase-dead TYK2, and left untreated or treated with IFN␣, was analyzed by in vitro Ser-535 phosphorylation of GST-IFNAR1 as assessed by immunoblotting using a Ser(P)-535-specific antibody. Levels of substrate and kinase were also analyzed by immunoblotting. B, interaction between endogenous TYK2 and PKD2 in HeLa cells treated as indicated was analyzed by co-immunoprecipitation (IP) and immunoblotting. WCL, whole cell lysate. IgG, isotype antibody control. C, localization of endogenous TYK2 in 2fTGH cells or isogenic 11.1 cells was detected using primary anti-TYK2 antibody (1 o Ab, where indicated) and secondary antibodies conjugated with Alexa Fluor 488. Nuclei were counterstained using DAPI. D, localization of endogenous PKD2 in 2fTGH cells transduced with the indicated shRNA was detected using primary anti-PDK2 antibody (1 o Ab, where indicated) and secondary antibodies conjugated with Alexa Fluor 594. Nuclei were counterstained using DAPI. E, co-localization of TYK2 and PKD2 in 2fTGH cells treated with IFN␣ (2000 units/ml for 5 min) or not was determined by confocal microscopy analysis using indicated antibodies. tion of PKD2 was easily detected when recombinant Src protein was used as positive control (Fig. 3C, lane 6). In the presence of activated HA-TYK2, a noticeable tyrosine phosphorylation signal on PKD2 was also observed in an ATP-dependent manner (compare lanes 2 and 5). A lesser extent of in Tyr phosphorylation of PKD2 was seen when HA-TYK2 was purified from untreated cells (Fig. 3C, lanes 4 and 5), indicating that ligandstimulated TYK2 is capable of directly phosphorylating PKD2 on Tyr residues.
Soluble tyrosine kinases including Abl and Src were shown to phosphorylate a related PKD1 on Tyr-463 (36). An increase in this phosphorylation in response to various PKD inducers could be detected using a phospho-specific antibody (38). A homologous tyrosine residue, Tyr-438, was found on PKD2 (Fig. 4A); this residue was proposed to play a role in modulating its activity (39). We next aimed to determine the role of this site in IFN␣-induced PKD2 activation and IFNAR1 phosphorylation. We generated the GST-PKD2 Y438F mutant and expressed it in and purified from 293T cells to use in an in vitro kinase assay as a substrate. Compared with analogous wild-type GST-PKD2, the mutant lacking Tyr-438 revealed a lesser extent of overall tyrosine phosphorylation upon its incubation with recombinant truncated GST-TYK2 833-1187 kinase (Fig. 4B,  lanes 3 and 4). Given that the autophosphorylation activity of TYK2 was not inhibited by the presence of GST-PKD2 Y438F mutant (Fig. 4B, lane 4), this result indicate a possibility that Tyr-438 may represent one of the phospho-acceptor sites used by TYK2 for PKD2 phosphorylation. Consistent with this possibility, overall tyrosine phosphorylation of GST-PKD2 Y438F expressed in 293T cells treated with IFN␣ was noticeably less pronounced than that observed in GST-PKD2 WT protein (Fig.  4C) despite the fact that Tyr-438 mutation did not affect the recruitment of PKD to IFNAR1 (Fig. 4D). Importantly, phosphorylation of these two proteins on Ser-710 was comparable (Fig. 4C), indicating that mutation of Tyr-438 does not affect either basal or IFN␣-induced phosphorylation of the PKD2 activation loop. Together, these results suggest that ligand stimulates phosphorylation of PKD2 on Tyr-438 and that activated TYK2 is directly capable of this phosphorylation.
Given the similarities between key Tyr residue-encompassing sequences within the plekstrin homology domains of PKD1 and PKD2 (Fig. 4A), we further used an antibody developed against Tyr-463 within PKD1 for PKD2 analysis. This antibody was used efficiently for analyses of PKD1 phosphorylation in cells treated with various stimuli (38). Immunoblotting analysis of GST-PKD2 species expressed in 293T cells and purified by affinity beads revealed a robust IFN␣-inducible signal detected on wild-type PKD2 but not on Y438F mutant (Fig. 4E). This result suggests that anti-Tyr(P)-463 antibody can be used for the specific detection of Tyr-438 phosphorylation within PKD2. Subsequent analysis of endogenous proteins also revealed that there is an increase in recognition of PKD2 by this antibody when PKD2 is immunopurified from IFN␣-treated cells (Fig. 4F). These results strongly suggest that IFN␣ stimulates phosphorylation of Tyr-438 within PKD2.
We next compared kinase activities of GST-tagged PKD2 WT and PKD2 Y438F proteins expressed in HeLa cells using an in vitro assay that enables detection of the phosphorylation of IFNAR1 on Ser-535. The basal Ser-535-phosphorylating activity of GST-PKD2 Y438F mutant was similar to that of wild-type enzyme. However, activation of this mutant kinase in response to IFN␣ treatment was visibly impaired (Fig. 5A). This result suggests that phosphorylation of Tyr-438 might be required for stimulation of PKD2 activity by IFN␣.
To test this possibility further, we generated GST-PKD2 expression constructs (depicted as GST-PKD2*) that contained silent mutations, making them insensitive to shPKD2 that we previously used for kinase knockdown (16). For subsequent experiments we chose to use human 2fTGH cells because, similar to HeLa, the role of PKD2 in modulating IFNAR1 phosphorylation and degradation was firmly established; yet, unlike HeLa cells, 2fTGH cells do not harbor human papillomavirus genes and exhibit a robust anti-viral effects in response to IFN␣ (16). 2fTGH cells were stably transduced with control shRNA or shPKD2 and with either empty vector or GST-PKD2* constructs (wild type or Tyr-438 mutant). Consistent with our previous report (16), knockdown of endogenous PKD2 impeded IFN␣-induced phosphorylation of IFNAR1 on Ser-535 (Fig. 5B,  lane 2 versus lane 4). Whereas reexpression of wild-type PKD2 partially restored this phosphorylation, such an effect was not observed when PKD2 was mutated at Tyr-438 (Fig. 5B, lane 6  versus lane 8). These results collectively suggest that ligandinduced phosphorylation of this tyrosine residue within PKD2 is required for IFN␣-stimulated PKD2 activation and ensuing PKD2-mediated phosphorylation of the IFNAR1 degron.
We next sought to investigate the role of these mechanisms in regulating IFNAR1 stability and signaling. Cycloheximide chase was used to determine the rate of IFNAR1 degradation in cells treated with IFN␣. Under these conditions, proteolytic turnover was noticeably delayed in cells that received shRNA against PKD2 (Fig. 6A). These results are in line with our previous report and are consistent with the model wherein PKD2 plays a key role in regulation of the ligand-inducible IFNAR1 degradation (16). Subsequent expression of shPKD2-insensitive wild-type GST-PKD2* in these cells noticeably reverted the FIGURE 3. TYK2 activity is required for IFN␣-induced tyrosine phosphorylation of PKD2. A, phosphorylation and levels of endogenous PKD2 immunopurified (IP) from WT-5 or KR-2 cells treated with IFN␣ as indicated were analyzed by the indicated antibodies. B, tyrosine phosphorylation and levels of endogenous PKD2 immunopurified from HeLa cells treated with IFN␣ as indicated were analyzed by immunoblotting. C, HA-tagged TYK2 was expressed in and immunopurified from 293T cells that were IFN␣-treated (lanes 1-3 and 5) or not (lane 4). TYK2 proteins were then incubated with PKD2 purified from serum-starved and untreated 2fTGH cells (except in lane 2) in the presence of ATP (except in lane 2). The reaction was analyzed by immunoblotting using an anti-phosphotyrosine antibody. The recombinant Src protein was used as positive control. OCTOBER 14, 2011 • VOLUME 286 • NUMBER 41 delay in IFNAR1 turnover. Remarkably, mutation of Tyr-438 rendered this construct incapable of restoring the rates of IFNAR1 degradation (Fig. 6A). These results suggest that tyrosine phosphorylation of PKD2 on Tyr-438 is important for rapid degradation of IFNAR1 in the presence of IFN␣.

Tyr Phosphorylation of PKD2 Regulates IFNAR1 Signaling
Our previous studies demonstrated that PKD2 plays a major role in negative regulation of magnitude and duration of IFN␣ signaling via mediating the ligand-stimulated IFNAR1 degradation (16). The role of PKD2 tyrosine phosphorylation in limiting the extent of the cellular responses to IFN␣ was next assessed using transactivation assay based on expression of the ISRE-driven luciferase in 2fTGH cells. A pulse IFN␣ treatment of 2fTGH cells that received irrelevant control shRNA led to a noticeable activation of ISRE-dependent transcription (Fig.  6B). Consistent with previously reported results (16), knockdown of PKD2 significantly augmented the IFN␣-stimulated luciferase activity. Importantly, expression of shRNA-insensitive wild-type GST-PKD2* (but not of the Tyr-438 mutant) tempered the effects of PKD2 knockdown (Fig. 6B).
We further used these functional PKD2 knock-in approaches in 2fTGH cells to assess the antiviral effects of the pulse treatment with a low dose (5 units/ml) of IFN␣. Pretreatment of cells with this dose did not significantly decrease the titer of vesicular stomatitis virus upon subsequently infection with this virus unless PKD2 was knocked down (Fig. 6C). The knockdown phenotype was rescued by expression of shRNA-insensitive wild-type PKD2 but not of Tyr-438 mutant. A similar trend was observed when the dose of IFN␣ was increased to 50 units/ml (Fig. 6C). The extent of the observed differences is probably underestimated due to an incomplete knockdown of endogenous PKD2 (Fig. 5). Nevertheless, these results collectively suggest that Tyr-438 phosphorylation of PKD2 play an important role in restricting the effects of IFN␣ signaling and in limiting its ability to mount an efficient anti-viral state.

DISCUSSION
Continuous responses of cells to Type 1 IFN are disrupted by proteolytic elimination of IFNAR1 (10). These events are facilitated by PKD2-dependent phosphorylation of IFNAR1 (16) followed by the recruitment of SCF ␤Trcp E3 ubiquitin ligase and ensuing IFNAR1 ubiquitination, endocytosis, and postinternalization sorting into the lysosomes (11)(12)(13). Activation of PKD2 by Type 1 IFN was shown to play an important role in these processes (16); in addition, the role of TYK2 catalytic activity in ligand-stimulated IFNAR1 degradation was revealed by Pellegrini group (15). Although we previously proposed that PKD2 may function downstream of TYK2, the detailed mechanisms by which this signaling occurs remained poorly characterized.
Here we describe studies that reveal that TYK2 interacts with PKD2 and that this interaction is noticeably stimulated by IFN␣ (Fig. 2). Furthermore, activated TYK2 is capable of directly phosphorylating PKD2 on tyrosine residues. A specific Tyr-438 within the plekstrin homology domain of PKD2 has been characterized as an important amino acid residue that can serve as a . C, phosphorylation and levels of GST-PKD2 (wild type or Y438F mutant) expressed in 293T cells treated or not with IFN␣ as indicated before harvesting. GST-PKD2 species were purified from the lysates using affinity beads and analyzed by immunoblotting using the indicated antibodies. D, interaction of endogenous IFNAR1 with exogenous GST-tagged PKD2 (wild type or Y438F mutant) expressed in 293T cells determined by immunoprecipitation (IP) followed by immunoblotting using the indicated antibodies. E, specific Tyr-438 phosphorylation of GST-tagged PKD2 species (wild type or Y438F mutant) purified from 293T cells treated or not with IFN␣ for 5 min analyzed by immunoblotting using the indicated antibodies. F, specific Tyr-438 phosphorylation of endogenous PKD2 immunopurified from HeLa cells treated with IFN␣ for the indicated times analyzed by immunoblotting using the indicated antibodies.
phospho-acceptor site for IFN␣-induced and TYK2-facilitated phosphorylation (Fig. 4). The integrity of this Tyr-438 site plays a key role in the ligand-inducible activation of PKD2 and ensuing phosphorylation of IFNAR1 degron (Fig. 5). Furthermore, Tyr-438 appears to be important for PKD2-dependent regulation of the IFNAR1 stability and the restrictions imposed on the extent of IFN␣-induced signaling and anti-viral effects (Fig. 6).
Based on these data, we propose a model wherein the interaction of Type 1 IFN with the cognate receptor chains and ensuing activation of catalytic activity of JAKs may lead to a direct phosphorylation of recruited PKD2 on Tyr-438 by TYK2 (and, most likely, by JAK1 as well). This phosphorylation mediates catalytic activation of PKD2 and results in the phosphorylation of IFNAR1 degron leading to IFNAR1 degradation followed by decreased sensitivity of cells to subsequent effects of IFN␣.
Given the reports suggesting that phosphorylation of analogous Tyr-463 in the pleckstrin homology domain of PKD1 leads to activation of this related kinase in response to oxidative stress via the Src-Abl pathway (40), it is plausible that analogous Tyr-438 phosphorylation of PKD2 by JAK would achieve a similar outcome. Indeed, TYK2 catalytic activity and PKD2 phosphorylation on Tyr-438 appears to be dispensable for basal activity of PKD2 yet play an important role in its activation by IFN␣ ( Figs. 2A and 5A). However, the exact mechanisms of activation of PKD2 by IFN␣ and of PKD1 by H 2 O 2 might be rather distinct. It was proposed that phosphorylation of Tyr-463 releases the plekstrin homology domain, thus exposing the activation loop of PKD1 for subsequent phosphorylation by PKC (40). A later report indeed suggested that hydrogen peroxide-induced PKD1 activation includes sequential Tyr-463 phosphorylation by c-Abl followed by the loop phosphorylation on Ser-738/Ser-742 by PKC␦ (41). Although IFN␣ treatment also increases phosphorylation of PKD2 on its activation loop (Ser-710), this phosphorylation is dependent neither on TYK2 activity (Fig. 3A) nor on Tyr-438 phosphorylation (Fig. 4C). This distinction is likely to reflect the intrinsic biochemical and biologic differences between PKD2 and PKD1. The latter kinase is poorly recruited to IFNAR1 (16) (Fig. 1B) and exhibits a strong preference for specific phospho-acceptor motif (L/I/V-X-R/K-XX-s/t) (42)(43)(44), that is absent in IFNAR1 degron. Seemingly for that reason PKD1 does not efficiently phosphorylate IFNAR1 on Ser-535 (16). In addition, soluble c-Abl kinase activated by H 2 O 2 might act differently from the receptor-associated JAK; accordingly, treatment of cells with hydrogen peroxide does not efficiently stimulate IFNAR1 phosphorylation in cells. 3 The mechanisms by which the recruitment of PKD2 to IFNAR1 and to TYK2 is stimulated by IFN␣ remain to be understood. Neither this increase in recruitment (Fig. 1C) nor in IFN␣-stimulated phosphorylation of PKD2 on Ser-710 ( Fig.  3A) required TYK2 activity. Although participation of a remaining JAK member (JAK1) cannot be ruled out, the dichotomy between requirements for Tyr phosphorylation and recruitment/Ser-710 phosphorylation of PKD2 strongly suggests that a separate signal transduction pathway must regulate these latter events. Given that a great wealth of published studies implicated various PKC species in PKD1 loop activation (33,(45)(46)(47)(48)(49), it is plausible that this mechanism could be also engaged by the Type 1 IFN pathway. Indeed, extensive work by many groups has demonstrated that IFN␣/␤ are capable of activating diverse PKC species and that this activation plays an important role in mediating the signal transduction induced by these cytokines (50 -55). Future studies will identify specific members of PKC superfamily that may contribute to PKD2 loop phosphorylation in a manner that does not depend on TYK2 kinase function. Additional studies are also warranted to delineate the mechanisms that govern the recruitment of PKD2 into the vicinity of IFNAR1 as well as to determine the role of these mechanisms in regulation of IFNAR1 stability and signaling. As IFN␣/␤ are often used as anti-cancer or anti-viral drugs 3 H. Zheng and S. Y. Fuchs, unpublished data. FIGURE 5. Integrity of Tyr-438 site is required for IFN␣-induced PKD2 activity to phosphorylate IFNAR1 at Ser-535. A, activity of GST-PKD2 (wild type or Y438F mutant) expressed in HeLa cells treated as indicated was analyzed by in vitro Ser-535 phosphorylation of GST-IFNAR1 as assessed by immunoblotting, using a Ser(P)-535-specific antibody. Levels of substrate and kinase were also analyzed by immunoblotting. B, derivatives of 2fTGH cells that stably harbor control shRNA or shRNA against PKD2 were transfected with the indicated plasmid constructs. Cells were treated with IFN␣ for 30 min. Immunoblot analysis of phosphorylation and levels of endogenous IFNAR1 are shown. Levels of actin and endogenous PKD2 (lower panels; short and long exposures) and exogenous GST-PKD2 in whole cell lysates (WCL) are also shown. (56 -58), targeting the mechanisms that desensitize cells to these cytokine is of potential clinical importance.