Protein Kinase R (PKR) Interacts with and Activates Mitogen-activated Protein Kinase Kinase 6 (MKK6) in Response to Double-stranded RNA Stimulation*

The double-stranded RNA (dsRNA)-activated protein kinase R (PKR) has been invoked in different signaling pathways. In cells pre-exposed to the PKR inhibitor 2-aminopurine or in PKR-null cells, the activation of p38 mitogen-activated protein kinase (MAPK) following dsRNA stimulation is attenuated. We found that the p38 MAPK activator MKK6, but not its close relatives MKK3 or MKK4, exhibited an increased affinity for PKR following the exposure of cells to poly(rI:rC), a dsRNA analog. In vitro kinase assays revealed that MKK6 was efficiently phosphorylated by PKR, and this could be inhibited by 2-aminopurine. Expression of kinase-inactive PKR (K296R) in cells inhibited the poly(IC)-induced phosphorylation of MKK3/6 detected by phosphospecific antiserum but did not affect the poly(IC)-induced gel migration retardation of MKK3. This suggests that poly(IC)-mediated in vivo activation of MKK6, but not MKK3, is through PKR. Consistent with this observation, PKR was capable of activating MKK6 as assessed in a coupled kinase assay containing the components of the p38 MAPK pathway. Our results indicate that the interaction of MKK6 and PKR provides a mechanism for regulating p38 MAPK activation in response to dsRNA stimulation.

Double-stranded RNAs (dsRNA) 1 formed during virus infection can be potent stimulating agents triggering cellular responses through distinct targets and pathways. DsRNA-dependent interferon synthesis as well as NF-B activation may require the binding of dsRNA with cellular membrane component(s) that includes Toll-like receptor 3 (TLR-3) (1). However, other dsRNA sensor cellular components have also been implicated in the activation of cellular pathways triggered by dsRNA (2)(3)(4).
The dsRNA-dependent protein PKR is a serine/threonine kinase, the expression of which is induced by interferons. This enzyme is activated in response to dsRNA of viral or synthetic origin, most notably the synthetic polyribonucleotide duplex, poly(rI):poly(rC) (pIC), and also by cytokines and cellular stress signals (5). Once activated, PKR functions as one of the mediators of antiviral and antiproliferative activities of interferons (6 -12). PKR has two double-stranded RNA-binding motifs located in the NH 2 -terminal domain (5,13), which bind pIC or viral dsRNA intermediates generated during a viral infection and also permit the recruitment of other dsRNA-binding domain-containing proteins (14,15). The carboxyl-terminal half of PKR harbors the kinase catalytic domain (9). On binding dsRNA, PKR dimerizes and undergoes autophosphorylation at multiple sites (5,16). The direct antiviral activity of PKR is a consequence of the phosphorylation and activation of its major substrate eukaryotic translation initiation factor eIF2␣, thereby inhibiting protein synthesis and impeding virus multiplication (6,10,12).
In addition to playing a role in protein synthesis inhibition, PKR also functions in other biological processes including apoptosis, transcriptional regulation, and cell growth, independent of the phosphorylation of eIF2␣ (5). Some of these mechanisms mediated by PKR in response to dsRNA might be attributed either to protein-protein interaction or to the phosphorylation of substrates other than eIF2␣. Accordingly, PKR has been shown to interact constitutively or in a stimulus-dependent manner with a number of proteins, which may serve as PKR substrates, thus regulating protein translation and transcriptional activity (3,4,14,15,(17)(18)(19)(20)(21)(22)(23)(24)(25)(26)(27)(28). A role for PKR in transcriptional activity in the inflammatory response was suggested by the observation that PKR-null primary fibroblasts and mice had defects in the activation of p38 mitogen-activated protein kinase (MAPK) and downstream proinflammatory gene expression in response to different stimuli (29).
It has been reported that p38 MAPK can be activated by dsRNA and viruses through distinct but unknown mechanisms (29,46,47). The TLR3-mediated activation of MAPKs in response to dsRNA stimulation is normal in cells lacking the TLR adaptor MYD88 (1). Also, we have shown previously that PKR contributes to p38 MAPK pathway activation (29). Here we investigated the intermediate components linking PKR to p38 activation. PKR was found to specifically interact with and activate MKK6 in a dsRNA-dependent manner, thus facilitating the activation of p38 MAPK pathway.

MATERIALS AND METHODS
Cell Culture and Transfections-HT1080 and 293T cells were maintained in 10% fetal bovine serum Dulbecco's modified Eagle's medium medium. THP-1 cells were maintained in 10% fetal bovine serum RPMI medium supplemented with 2 mM glutamine. Bone marrow macrophages were isolated from isogenic wild-type or PKR Ϫ/Ϫ mice (C57/BL6 pure background) (2,29). RAW264.7 cells expressing either PKR-K296R (RAW-PKRK296R cells) or an empty vector (RAW-neo/bla cells) were generated in two rounds of transfections. In the first round, RAW264.7 cells were transfected with the plasmid pRcCMV-PKRK296R and selected both for the resistance to G418 and for the expression of human PKR-K296R by reverse transcriptase-PCR. As increased levels of the human PKR-K296R were not detected by immunoblotting with anti-PKR polyclonal antibody, a positive clone identified by reverse transcriptase-PCR in the first round was transfected with ptV5K296R plasmid, and the selection of blasticidin-resistant clones expressing PKR-K296R was further assessed through immunoblotting by using anti-V5 antibody. The selected clones were grown in 10% fetal bovine serum Dulbecco's modified Eagle's medium in the presence of 400 g/ml G418 and 1 g/ml blasticidin. All cell lines were grown at 37°C in 5% CO 2 . Near confluent 293T or HT1080 cells growing in 10-cm plates were split 1:15 the day before transfection by the calcium phosphate method (48). Transient transfections were performed using 5 g of each plasmid.
Cell Lysis and Immunoblotting-Before lysis, cells were washed twice in cold phosphate-buffered saline. After treatments and/or transfections, cell extracts were prepared by extraction with lysis buffer (50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 50 mM NaF, 10 mM ␤-glycerophosphate, 0.1 mM EDTA, 10% glycerol, 1% Triton X-100) supplemented with protease/phosphatase inhibitors (1 mM phenylmethylsulfonyl fluoride, 1 mM sodium orthovanadate, 2 g/ml each of leupeptin, aprotinin, and pepstatin) followed by incubation on ice for 20 min. Lysates were clarified by centrifugation at 14,000 ϫ g for 20 min, and the supernatant was collected. Proteins were quantified by Bradford assays. Thirty micrograms of cell lysates were fractionated onto 10% SDS-PAGE gel, transferred simultaneously to two PVDF filters (0.45 m), and probed with the antibodies as indicated in the figures. Back filters were used to assess total protein loading. Briefly, before blocking in TBS-T 5% nonfat dry milk for 1 h, the membranes were washed in TBS-T for 5 min at room temperature. After three washes/5 min each in TBS-T, phosphospecific antibodies were diluted in TBS-T 5% bovine serum albumin followed by 12-16 h of incubation at 4°C under rotation. Other antibodies were diluted into TBS-T 5% nonfat dry milk and incubated at room temperature for 1 h. After being washed as above, membranes were incubated with the appropriate horseradish peroxidase-conjugated secondary antibody for 1 h. Proteins were detected by using enhanced chemiluminescence (ECL) according to the manufacturer's recommendations (Amersham Biosciences).
Immunoprecipitation Assays-Cell extracts were prepared as described above. Immunoprecipitations were performed using 500 g of cell extracts obtained from experiments. For PKR pull-down assays, 2 l from 1:10 prediluted monoclonal antibody 71/10 were used. Complexes were rotated at 4°C for 2 h, and antibody-antigen complexes were captured by protein G-Sepharose overnight incubation. Beads were washed five times with ice-cold lysis buffer before immunobloting procedures.
Recombinant Protein Preparation and Purification-Recombinant human PKR was prepared as His-tagged PKR fusion protein (23). Briefly, Escherichia coli BL21(DE3) cells were transformed with human PKR-containing pET15b. Cells were grown in Luria broth culture at 37°C to an A 600 0.6 -1.0, treated with 2 mM isopropyl-1-thio-␤-D-galactopyranoside for an additional 3 h, and centrifuged at 5,000 ϫ g for 20 min at 4°C. Pellets were resuspended in ice-cold lysis buffer (50 mM NaCl, 50 mM Tris-HCl, pH 8, 0.1% Nonidet P-40), sonicated, and centrifuged at 20,000 ϫ g for 20 min at 4°C. Purified histidine-tagged PKR was obtained through passing supernatant over a Ni 2ϩ metal affinity column (His-Bind, Novagen) according to the instructions of the manufacturer. B56␣, eIF2␣, and MKK6 were also expressed as six histidine-tagged proteins and purified by affinity chromatography using His-Bind metal-chelating resin according to the instructions of the manufacturer (Novagen). GST-ATF2-(1-109) was purified as described (50).
Protein Kinase Assays-Immunoprecipitates from cell extracts or purified recombinant proteins were used for the in vitro kinase reactions in kinase reaction buffer containing 10 mM HEPES, pH 7.4, 1 mM dithiothreitol, 5 mM MgCl 2 , 20 M ATP, 5 Ci of [␥-32 P]ATP (6 mCi/mmol). Whenever indicated, poly(IC) was added to a final concentration of 1 g/ml, and kinase reactions were incubated at 30°C for 25 min. The phosphorylation of proteins was examined after fractionation in SDS-PAGE by autoradiogram analysis. To examine the ability of PKR to activate MKK6, a coupled kinase assay was carried out by incubating 0.2 g of His-MKK6 with PKR immunocomplex at 30°C for 15 min in kinase reaction buffer containing 100 M ATP followed by an incubation at 30°C for 7 min with 0.2 g of GST-p38 MAPK (Stratagene). Subsequently, the activated complex was incubated with 2 g of GST-ATF2 and 0.3 Ci of [␥-32 P]ATP (6 mCi/mmol) for an additional 7 min at 30°C. Reactions were fractionated onto 10% SDS-PAGE gel and transferred to PVDF membranes. After transfer, the gel was Coomassie Blue-stained to assess GST-ATF2 loading. The phosphorylation of ATF2 was analyzed by autoradiogram analysis. Subsequently, the membrane was immunoprobed with polyclonal anti-PKR, anti-MKK6, and anti-p38 antibodies to assess loading control.

RESULTS
A PKR-dependent Pathway Mediates the Effect of Doublestranded RNA on p38 MAPK Activation-The contribution of PKR to p38 MAPK activation elicited by dsRNA was examined in bone marrow macrophages derived from PKR-null mice. Comparative immunoblotting analyses of cell extracts reveal that the phosphorylation of p38 MAPK induced by pIC at later time points is attenuated in cells lacking PKR (Fig. 1A). The nucleoside analog 2-AP is known to inhibit the activity and PKR-dependent phosphorylation of eIF2␣ (51). To determine whether the kinase activity of PKR is required specifically for dsRNA-induced p38 MAPK activation in vivo, we treated human promonocytic cell line THP-1 (in which LPS activates p38 MAPK) with pIC or LPS in the presence or absence of 2-AP. The activation of both p38 MAPK and eIF2␣ phosphorylation was severely compromised by 2-AP co-treatment (Fig. 2, left panel). Interestingly, in the presence of 2-AP, eIF2␣ phosphorylation returns to basal levels, whereas p38 MAPK does not. This inhibition was specific to dsRNA stimulation because at the same concentration, 2-AP did not inhibit p38 MAPK phosphorylation in response to an LPS/ TLR4-triggered signal (Fig. 2, right panel). Taken together, these results suggest that pIC triggers both a primary signaling event mediated by TLR-3 that may be PKR-independent and a secondary signal mediated after uptake into the cells that is dependent on PKR. In accord with this, 2-AP completely blocked the dsRNA-triggered phosphorylation of eIF2␣ but not p38 MAPK.
MKK6 Physically Interacts with PKR in Response to dsRNA Stimulation-It has been reported that the activation of MAPKs by dsRNA does not depend on TLR adaptor proteins MYD88 and/or TIR-containing adapter protein (1,52). Also, it has been shown previously that proinflammatory stimuli, including pIC, can activate the p38 MAPK pathway via a PKRdependent signal in mouse embryo fibroblasts (29). Accordingly, we reasoned that upstream activators of p38 MAPK could be potential targets for PKR following pIC stimulation. To test this, HT1080 cells were transiently transfected with constructs encoding MKK3, MKK4, or MKK6 and then treated with pIC. The phosphorylation of eIF2␣ was induced by pIC treatment and remained largely unaffected by expressing these constructs (not shown). Co-immunoprecipitation assays from these cell extracts (Fig. 3A) showed that MKK6, but not MKK3 or MKK4, had an increased affinity for PKR following dsRNA stimulation. Although JNK activator MKK4 could not be efficiently pulled-down from cell extracts, the result showed that MKK3, the most structural and functional p38 MAPK activator related to MKK6 (33), did not interact with PKR. Thus, pIC induces a specific interaction between PKR and MKK6. The poly(IC)-dependent interaction of PKR and MKK6 also occurred in 293T cells (Fig. 3B). The signal-and time-dependent interaction of PKR and MKK6 was further confirmed by performing the immunoprecipitation of endogenous PKR from pIC-treated 293T cells (Fig. 3C). Although a weak constitutive interaction of PKR and MKK6 was noted in this experiment, an increasing amount of overexpressed MKK6 became associated with endogenous PKR with in- Bone marrow macrophages derived from wild-type or PKR-null (PKR Ϫ/Ϫ ) mice were treated with pIC (100 g/ml) for the times indicated. Cell extracts were resolved by 10% SDS-PAGE gel, transferred to a PVDF membrane, and immunoprobed with the indicated antibodies. Bottom panel, graph representation of densitometry analysis of p38 MAPK phosphorylation from the autoradiogram shown above. Open circles and filled squares indicate wild-type and PKR-null cells, respectively. PKR-ko, PKR knock-out.

FIG. 3. Physical interaction between PKR and MKK6. A, anti-PKR Western blot (WB) analysis of cell extracts immunoprecipitated
with anti-MKK3, anti-MKK4, and anti-MKK6 antibodies. Forty-eight hours after transfection with the indicated constructs, HT1080 cells were left untreated (Ϫ) or treated (ϩ) for 90 min with 2 g/ml pIC in complex with FuGENE transfection reagent. Four hundred micrograms of cell extracts were used to immunoprecipitate specific MKKs as indicated (IP antibody), and interaction with endogenous PKR was examined by immunoblotting analysis with polyclonal anti-PKR antibody. Membranes were reprobed with anti-FLAG or anti-MKK3 antibodies as a loading control for the immunoprecipitation. B, anti-MKK6 Western blot analysis of cell extracts immunoprecipitated with anti-PKR monoclonal antibody. Forty-eight hours after transfection with empty vector (Ϫ) or pCEFL-GST-MKK6 plasmid (ϩ), 293T cells were left untreated or treated with pIC (100 g/ml) for 90 min. Four hundred micrograms of cell extracts were used to immunoprecipitate PKR immunocomplexes, and the physical association of MKK6 with PKR was examined by immunoblotting with polyclonal anti-MKK6 antibody. Immunoblots were subsequently used to assess the immunoprecipitation of PKR. The levels of expression of MKK6 and endogenous PKR were further examined in immunoblotting analysis of the same cell extracts (lower panels). C, following 48 h of mock-transfection (NT) or transfection with pCEFL-GST-tagged MKK6 construct, 293T cells were treated with pIC (100 g/ml) for the times as indicated. Cell extracts were subjected to immunoprecipitation-immunoblotting analysis as described above.
creasing time of exposure of cells to pIC, peaking at 60 min after pIC treatment. Taken together, these results suggest that MKK6 is recruited to PKR to form a complex in response to dsRNA stimulation.
MKK6 Is a Substrate for PKR-The above physical interaction studies suggested that MKK6 phosphorylation might be modulated by PKR. To address this possibility, we determined whether recombinant MKK6 could be efficiently phosphorylated by recombinant PKR in vitro. In the absence of pIC, PKR exhibited constitutive autocatalytic activity, and MKK6, as well as the PKR substrates, eIF2␣ and B56␣, were phosphorylated (Fig. 4A). However, in the presence of pIC, there was an increase in PKR autophosphorylation, which was accompanied by the robust phosphorylation of MKK6, eIF2␣, and B56␣. To determine whether inhibiting PKR activity by 2-AP resulted in subsequent inhibition of MKK6 phosphorylation, we performed an in vitro kinase assay of PKR and MKK6 in the presence of pIC and 2-AP. The pIC-mediated phosphorylation of recombinant MKK6 by PKR was blocked by the PKR inhibitor 2-aminopurine (Fig. 4B). Thus, MKK6 is a substrate for PKR in vitro.
In Vivo Regulation of MKK6 Phosphorylation by PKR-To determine whether PKR activated in vivo can mediate MKK6 phosphorylation, kinase-inactive PKR (PKR-K296R) expressing macrophage-like RAW264.7 cells and bone marrow macrophages derived from PKR-null mice were treated with poly(IC).
In PKR-K296R-expressing cells treated with pIC, no increase in eIF2␣ phosphorylation was observed (Fig. 5A). This indicates that PKR-K296R interferes with the activation of the endogenous PKR in RAW264.7 cells. Comparative immunoblotting analysis of cell extracts revealed that the phosphorylation of MKK3/6 induced by poly(IC) is attenuated in these cells (Fig. 5, B and C). The residual activation of MKK3/6 observed in RAW-PKRK296R and PKR-null cells following pIC stimulation is more likely to be phosphorylated MKK3 since the phospho-specific antibody used for detection does not discriminate between MKK3 or MKK6 when these proteins are activated. This observation is corroborated with the results shown in Fig. 5B (middle panel) in which, following pIC treatment, the migration of MKK3 was retarded to the same extent, in a time-dependent manner in both RAW-neo/bla and RAW- A, anti-phospho eIF2␣ (p-eIF2␣) Western blot (WB) analysis of RAW264.7 cells expressing a catalytically inactive PKR-K296R. RAWneo/bla and RAW-PKRK296R cells were treated with pIC (100 g/ml) for the times indicated, and cell extracts were analyzed by immunoblotting as shown. B, anti-phospho MKK3/6 (p-MKK3/6) and anti-MKK3 (p-MKK3) Western blot analysis of RAW264.7 cells expressing PKR-K296R. Cells were treated with 100 g/ml pIC for the times indicated, and cell extracts were analyzed by immunoblotting as shown. C, antiphospho MKK3/6 and anti-phospho p38 (p-p38) Western blot analysis of bone marrow macrophages derived from wild-type or PKR-null (PKR Ϫ/Ϫ ) mice. Cells were treated with 100 g/ml pIC or 100 ng/ml LPS for the times as indicated. Cell extracts were obtained and resolved by 10% SDS-PAGE gel, transferred to PVDF membranes, and immunoprobed with the indicated antibodies. PKRK296R cells. The activation of p38 by pIC was also attenuated in PKR-null cells but not significantly affected in the mutant PKR-expressing cells (not shown), likely due to efficient MKK3 activation. Moreover, the observation that MKK6 phosphorylation was decreased in PKR-defective cells supports the observations that PKR specifically regulates MKK6 in vivo. These observations are supported by immune complex kinase assays of PKR immunoprecipitated from pIC-treated cells. Cell extracts from HT1080 cells treated with pIC for different time intervals exhibited a significant PKR autophosphorylation at 90 and 180 min (not shown). PKR immunoprecipitated from these cell extracts was able to strongly phosphorylate recombinant human MKK6 with increasing efficiency (Fig. 6A). This observation is further corroborated with data obtained in a coupled kinase assay in which activated PKR, through MKK6, is capable of triggering the subsequent activation of downstream effectors targets of MKK6 (Fig. 6B). Thus, we suggest that activated PKR is capable of directly regulating MKK6 activity following the exposure of cells to pIC. DISCUSSION The activation of p38 MAPK in response to a variety of stimuli is preferentially mediated by its direct activator MKK6 (30). Immediate upstream activators for MKK6, which include ASK1, TAK1, MLK3, MEKK3, cot/tpl-2, and MTK1, have been already described (35, 40 -44). Here we provide evidence that PKR is a direct upstream activator of MKK6 forming a catalytic complex in response to dsRNA treatment of cells and thus contributing in the regulation of p38 MAPK activation in vivo.
Association of MKK6, but not the relatives p38 kinase kinases MKK3 or MKK4, with endogenous PKR was detected following pIC stimulation (Fig. 3). Our results suggest that MKK6 has an increased affinity for activated PKR. This observation is in accord with previous observations showing that specific PKR-interacting proteins preferentially recognize activated PKR (28). Moreover, the finding that MKK6 and PKR physiologically interact with each other after dsRNA stimulation in cells supports the idea that this association is important for dsRNA signaling. The observation that MKK6 is a substrate for PKR kinase activity in vitro suggests that PKR can also regulate the activation of MKK6 in vivo. Other protein kinases that are directly upstream to MKK6 phosphorylate it on its activation loop at serine and threonine residues (40). However, the phosphorylation of MKK6 by PKR in vitro also occurred at other serine and/or threonine residues since the mutation of the consensus site on MKK6 did not alter phosphorylation by PKR (not shown). Whether these other residues are required to activate the p38 MAPK or other potential downstream targets remains to be determined.
Because p38 MAPK activation in the absence of PKR still occurs but with reduced intensity, PKR could be functioning either as a structural enhancer or in a compensatory mechanism, which in turn contributes to the total kinetic efficiency of the p38 MAPK pathway. Thus, the amplification of p38 activation could be a result of the specific kinase-substrate association of PKR and MKK6. Different mammalian proteins have been shown to function as MAPK pathway scaffolds. For example, kinase suppressor of Ras and Raf kinase inhibitor protein have been implicated in the regulation of Ras/MAPK (53,54). Also, the activation of the JNK pathway can be modulated by JIP1 or MEKK1, which enhances JNK activity by specifically interacting with MKK7 but not with MKK3, MKK4, or MKK6 (55). What makes these proteins fit a scaffold feature is their ability to interact with at least two components of the kinase pathway. Once the binding occurs, they can act as insulators to prevent cross-talk between pathways or enhance the kinetics of that signaling. As we have shown in this study, following pIC stimulation of cells, PKR specifically interacted with MKK6 but not with MKK3 or MKK4. Also, we have recently shown that in 293 cells, PKR can associate with TAK1 in a pIC-dependent manner in an interleukin-1 receptor-associated kinase-independent TLR3-mediated pathway (47). Moreover, the MAPKKK ASK-1 has also been shown to interact with endogenous PKR in 293 cells, suggesting a role in the MAPK cascade (56). Thus, it is possible that PKR might act as a scaffold rather than a central player in the p38 activation by regulating the activity of additional components of the p38 MAPK pathway.
The kinase inhibitor 2-AP can block the induced expression of a number of genes in response to virus, double-stranded RNA, and interferons (57). This inhibitor is relatively specific for PKR at the concentrations used (51,58,59) and is thought to compete for ATP at the ATP-binding site of PKR and thereby inhibit its autophosphorylation (51). It is clearly noticed that in the presence of 2-AP, the phosphorylation of FIG. 6. Regulation of MKK6 activation by PKR. A, an autoradiogram of an immunoprecipitation (IP) kinase assay revealing MKK6 phosphorylation by PKR immunoprecipitated from cell extracts of pICtreated HT1080 cells. HT1080 cells were left untreated or treated with 2 g/ml pIC complexed with FuGENE transfection reagent for the times as indicated in the top of panel. PKR was immunoprecipitated from the cell extracts, and kinase reactions were carried out at 30°C for 25 min in kinase reaction buffer in the presence of 1 g of recombinant MKK6 and 5 Ci of [␥-32 P]ATP. Samples were subjected to SDS-PAGE, transferred to PVDF membranes, and exposed to autoradiogram films. The gel was used to assess MKK6 loading by Coomassie Blue staining. The levels of PKR immunocomplexes were analyzed by immunoblotting the membrane with polyclonal anti-PKR antibody. WB, Western blot. B, autoradiogram and Western blot analyses of coupled kinase assay showing the PKR-dependent activation of the MKK6-p38-ATF2 module. PKR was immunoprecipitated with anti-human PKR monoclonal antibody from pRC-CMV-PKR-transfected 293T cells. The immune complex was incubated in the presence (ϩ) or not (Ϫ) of His-MKK6 and GST-p38, and the kinase activity in the complex was measured with the substrate GST-ATF2-(1-109). Samples were subjected to SDS-PAGE and transferred to PVDF membranes. Membrane was first used to monitor the incorporation of [␥-32 P]ATP in GST-ATF2 by autoradiogram analysis. After successive stripping washes, the levels of PKR, MKK6, and p38 were examined through Western blot with specific antibodies as indicated. The gel (after transfer) was Coomassie Bluestained to assess GST-ATF2 loading.
Levels of MKK3 phosphorylation are still detected in PKRnull cells in response to dsRNA (Fig. 5), and it remains to be established which other components can associate with and activate MKK3 in response to dsRNA in the absence of PKR (Fig. 7). The role of TLR3-dependent pathways in PKR-independent signaling is unknown (47), although it has been proposed that PKR plays a role in LPS signaling via physical interaction with the adapter protein MyD88 adaptor-like/TIRcontaining adapter protein (60). However, PKR is not required for LPS-induced interferon-␤ production (61). Clearly, double knock-outs of both PKR and TLR-3 will assist in the elucidation of other mechanisms of dsRNA signaling.
The results we have provided in this work by no means exclude the possibility that the dsRNA-mediated activation of p38 MAPK is also regulated by a mechanism that does not require PKR (Fig. 7). However, our finding that MKK6 associates with PKR following dsRNA stimulation and that disruption of PKR in cells accounts for both reduced MKK6 phosphorylation and reduced p38 MAPK phosphorylation suggests that such an interaction is a cellular response to viral dsRNA replicative intermediates, which determines the outcome of infection. FIG. 7. Schematic illustration of the regulation of dsRNA-activated p38 MAPK mediated by PKR-MKK6 interaction. Once activated by binding to dsRNA, PKR undergoes autophosphorylation followed by interaction with and the activation of substrates such as eIF2␣ and MKK6. PKR has also been shown to interact with MAPKKKs (47,56). It remains to be established which other components can associate with and regulate MKK3-p38 MAPK activation in response to dsRNA in the TLR3-dependent pathway.