T. Takizawa, Department of Biochemistry, Institute for Developmental Research, Aichi Human Service Center, Kasugai, Aichi 480-0392, Japan. Fax: + 81 568 88 0829, Tel.: + 81 568 88 0829, E-mail: email@example.com
Double-stranded RNA-activated protein kinase (PKR), a serine/threonine kinase, is activated in virus-infected cells and acts as an antiviral machinery of type I interferons. PKR controls several stress response pathways induced by double-stranded RNA, tumor necrosis factor-α or lipopolysaccharide, which result in the activation of stress-activated protein kinase/c-Jun NH2-terminal kinase and p38 of the mitogen-activated protein kinase family. Here we showed a novel interaction between PKR and apoptosis signal-regulating kinase 1 (ASK1), one of the members of the mitogen-activated protein kinase kinase kinase family, which is activated in response to a variety of apoptosis-inducing stimuli. PKR and ASK1 showed predominant cytoplasmic localization in COS-1 cells transfected with both cDNAs, and coimmunoprecipitated from the cell extracts. A dominant negative mutant of PKR (PKR-KR) inhibited both the apoptosis and p38 activation induced by ASK1 in vivo. Consistently, PKR-KR inhibited the autophosphorylation of ASK1 in vitro, and exposure to poly(I)–poly(C) increased the phosphorylation of ASK1 in vivo. These results indicate the existence of a link between PKR and ASK1, which modifies downstream MAPK.
The interferon-inducible, double-stranded RNA (dsRNA)-activated protein kinase (PKR) is a serine/threonine kinase ubiquitously expressed in mammalian cells [1,2]. PKR is activated by a variety of dsRNA molecules generated during viral infection . Upon its activation, PKR autophosphorylates and then phosphorylates eukaryotic translational initiation factor 2 (eIF-2α) , thereby inhibiting cell growth or viral replication [5,6]. Thus PKR mediates the antiviral and antiproliferative actions of type I interferons . On the other hand, catalytically inactive mutants of PKR transform NIH-3T3 cells [7,8], while overexpression of wild-type PKR induces apoptosis of HeLa cells [9,10]. PKR appears to up-regulate expression of the apoptotic receptor Fas induced by viral infection [11,12]. Moreover, mouse embryonic fibroblasts deleted of the PKR gene have been shown to resist apoptosis in response to dsRNA, tumor necrosis factor-α (TNF-α) or lipopolysaccharide (LPS) . PKR has been shown to play some role in the activation of p38 mitogen activated protein kinases (MAPKs) and the stress-activated protein kinase (SAPK)/c-Jun amino-terminal kinases (JNKs) that are strongly activated in response to TNF-α, dsRNA or LPS . However, the precise pathway linking PKR and the MAPK family remains to be elucidated.
Apoptosis signal-regulating kinase 1 (ASK1) is a MAPK kinase kinase (MAPKKK) that acts upstream of JNK and p38 MAPKs [14,15]. ASK1 phosphorylates SEK1/MKK4 or MKK3/MKK6, one of the members of the MAPK kinase family, which in turn activates JNK or p38 MAPK, respectively . A wide variety of stress-related stimuli activate ASK1, including serum withdrawal, TNF-α, reactive oxygen species, microtubule-interfering agents, genotoxic stress, and possibly Fas ligand . Overexpression of the wild type or constitutively active form of ASK1 induces cell death through signals involving the mitochondrial cell death pathway . ASK1 binds proteins associated with death receptors as TNF-receptor-associated proteins (TRAFs) or Daxx, which also results in MAPK activation [18,19]. In the present study, we show that PKR interacts with ASK1 and modifies the ASK1 signaling pathway both in vivo and in vitro. These results suggest that PKR acts as a signal transducer by interacting with MAPKKKs, which then modifies the downstream MAPK cascade.
Materials and methods
Plasmids and DNA transfection
Human PKR cDNA was kindly provided by A. Hovanessian (Institut Pasteur, France). A mutant PKR cDNA carrying a point mutation of K to R at position 296 (PKR-KR) was constructed as described . cDNAs for human ASK1 and dominant negative mutant of ASK1 carrying a point mutation of K to M at position 709 (ASK-KM) were kindly provided by H. Ichijo (Laboratory of Cell Signaling, Tokyo Medical and Dental University). The plasmid encoding PKR or PKR-KR fused to enhanced green fluorescent protein (EGFP) (Clontech Laboratories, Inc., Palo Alto, CA, USA) was described previously . Human embryonic kidney 293 (HEK293), COS-1 and NIH-3T3 cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum, and maintained under 5% CO2 at 37 °C. Cells were transfected with 2 µg of plasmid DNA using 6 µL of Lipofectamine-plus and 4 µL of Lipofectamine (GibcoBRL) according to the manufacturer's instructions. To establish permanent transfectants, HEK293 cells were diluted about 10-fold and replated in medium containing 750 µg·mL−1 of G418 two days after transfection, and then drug-resistant colonies were isolated. For microscopic observation, cells were seeded on a cover glass and DNAs were transfected as described above. For in vivo labeling of ASK1 or PKR, cells were washed with phosphate-free DMEM, and then incubated in phosphate-free DMEM in the presence of [32P]-orthophosphate (400 µCi·mL−1) (ICN Biomedicals, CA, USA) for the indicated period as described in the figure legend. The phosphorylation reaction was resolved by SDS/PAGE and visualized by autoradiography. Autoradiogram was scanned and densitometric analysis was performed with Kodak digital science 1d software (Eastman Kodak).
The day after transfection, cells were fixed with 4% paraformaldehyde containing 0.2% Triton X-100 in NaCl/Pi for 30 min and washed with NaCl/Pi. Subsequently, cells were incubated with anti-HA mAb (12CA5, Boehringer Mannheim, Germany) or anti-PKR polyclonal antibody (N-18, Santa Cruz, CA, USA) at a dilution of 200 or 100, respectively, for 60 min. Cells were then stained with anti-(mouse IgG) conjugated with fluorescein isothiocyanate (FITC) (MBL, Nagoya, Japan) or anti-(rabbit IgG) conjugated with tetaramethyl rhodamine isothiocyanate (RITC) (Jackson Immunoresearch Laboratory, West Grove, PA, USA) at a dilution of 200 for 60 min, and observed under a fluorescence microscope at a magnification of 270 (Olympus BX-60, Tokyo, Japan). For the expression of EGFP, cells were fixed with 4% paraformaldehyde, and observed under a fluorescence microscope as described above.
In vitro kinase assay
Cell extracts were prepared with PKR buffer I (20 mm Tris/HCl pH 7.6, 50 mm KCl, 400 mm NaCl, 1 mm EDTA, 5 mm 2-mercaptoethanol, 1% Triton X-100, 0.2 mm phenylmethane sulfonyl fluoride, 100 U·mL−1 aprotinin, and 20% glycerol) and cleared by centrifugation. Then ASK1 in the cell extract was precipitated by incubation with anti-HA mAb (5 µL) for 1 h at 4 °C followed by 50 µL of a 1 : 1 slurry of protein–G Sepharose 4FF (Pharmacia, Piscataway, NJ, USA). The immune complex on the beads was washed four times with PKR buffer I and then once with PKR buffer III (20 mm Tris/HCl pH 7.6, 80 mm KCl, 5 mmβ-mercaptoethanol, 2 mm MgCl2, 2 mm MnCl2, and 20% glycerol). The beads were then resuspended in PKR buffer III containing 2 µm[γ-32P]ATP (5 µCi) (ICN Biomedicals) in the presence or absence of poly(I)–poly(C) at the concentration indicated in the legends for 15 min at 30 °C. The phosphorylation reaction was resolved by SDS/PAGE and visualized by autoradiography. PKR activity was measured as described .
Immunoprecipitation and Immunoblot analyses
Cell extracts were prepared with PKR buffer I and incubated with anti-PKR mAb (2 µL) or anti-HA antibody (5 µL) for 1 h at 4 °C followed by 50 µL of a 1 : 1 slurry of protein-G Sepharose 4FF for another 1 h at 4 °C. The immune complex on the beads was washed four times with PKR buffer I. The beads were then boiled in Laemmli's sample buffer  and resolved by SDS/PAGE. Proteins were transferred onto nitrocellulose filters (Bio-Rad Laboratory, Hercules, CA, USA) and were incubated with polyclonal anti-PKR Ig or anti-HA Ig followed by anti-(rabbit IgG) or anti-(mouse IgG) conjugated with peroxidase. Signals were visualized using an enhanced chemiluminescence (ECL) detection system (Amersham, Boston, MA, USA). Protein was measured by Bradford reagent (Bio-Rad Laboratory).
PKR interacts with ASK1
To investigate whether PKR and ASK1 interact with each other, the localization of PKR and ASK1 was first examined by indirect immunofluorescence. As the wild type of PKR is hardly expressed at all by transfection due to translational inhibition , we used the kinase negative mutant of PKR (PKR-KR), which shows the same localization pattern as the wild type as previously reported . The signal for PKR-KR was predominantly localized in the cytoplasm, whereas that for ASK1 distributed diffusely with relatively intense staining at the periphery of the cells (Fig. 1A, a and b). When PKR-KR and ASK1 were cotransfected into COS-1, both proteins showed predominant cytoplasmic localization (Fig. 1A, c and d), indicating colocalization of PKR and ASK1.
We next used a coimmunoprecipitation assay to define the interaction between PKR and ASK1. They were transfected into COS-1 cells, and immunoprecipitated with antibody against either PKR or ASK1. When the immunecomplexes were precipitated with anti-PKR Ig and analyzed by Western blotting with anti-HA Ig, the signal of ASK1 was detected only in the cell extracts transfected with both cDNAs (lane 2 in Fig. 1B). PKR-KR was also coimmunoprecipitated with anti-HA Ig (lane 4 in Fig. 1B). Expression of these proteins was verified by Western blotting (Fig. 1B, right panel). To examine whether endogenous PKR is coimmunoprecipitated with ASK1, we established two HEK 293 cells permanently expressing ASK1 (ASK-4 and ASK-8 cells) and control cells containing empty plasmid (pcDNA). The expressions of ASK1 and endogenous PKR in these cells were confirmed by Western blotting (Fig. 1C, right panel). ASK1 or endogenous PKR was immunoprecipitated with anti-PKR or anti-HA Ig, respectively (Fig. 1C, left panel). All these results indicate that PKR directly interacts with ASK1. Alternatively, the interaction might be bridged by RNA. However, as all these coimmunoprecipitation assays were conducted in the presence of high salt (0.45 m), and immunecomplexes were resistant to RNase treatment (data not shown), direct protein–protein interaction seems to be likely.
Dominant negative mutant of PKR inhibits ASK1 activity
To explore the potential influence of PKR on ASK1 in vivo, the effect of PKR-KR on the ASK1-induced apoptosis was investigated. We used constructs of PKR fused with EGFP to directly visualize cell morphology. We have shown that EGFP-PKR induced apoptosis without poly(I)–poly(C), whereas EGFP-PKR-KR inhibited Fas-induced apoptosis . NIH-3T3 cells were transfected with ASK1 and either pEGFP-PKR-KR or pEGFP. Serum was removed from the medium 24 h after transfection, and the cells were incubated for another 24 h. The cells were then fixed and ASK1 was stained with anti-HA Ig followed by RITC-labeled secondary antibody. ASK1 and EGFP-expressing cells exhibited a round shrunken morphology indicating an induction of apoptosis (Fig. 2B, arrow heads in upper panel), whereas the cells expressing both ASK1 and pEGFP-PKR-KR exhibited a flat spread shape (Fig. 2B, arrow heads in lower panel). EGFP-expressing cells without ASK1 expression were a flat shape as well (Fig. 2B, upper panel). The expression of these proteins was verified by Western blotting (Fig. 2A). The number of cells exhibiting a shrunken morphology was counted in several fields and summarized (Fig. 2C). Transfection of ASK1 and subsequent serum deprivation caused about 60% of cells to die, whereas cotransfection of pEGFP-PKR-KR suppressed the cell death to almost the control level (Fig. 2C). Transfection of either empty vector or PKR-KR alone did not cause significant cell death.
As PKR-KR inhibited the ASK1-induced apoptosis, it seems reasonable to speculate that PKR-KR inhibited ASK1 signaling. As ASK1 has been shown to activate stress-activated MAPKs, the effect of PKR-KR on the activation of p38 by ASK1 was examined. COS-1 cells were transfected with ASK1 and/or PKR-KR, and p38 activation was examined by Western blotting using antibody against the phosphorylated form of p38 (Fig. 3). ASK1 increased p38 phosphorylation 24 h after transfection (lane 6 in Fig. 3) (an average of 5.4-fold increase in the intensity from three independent experiments compared with that of pcDNA at 24 h), whereas cotransfection of PKR-KR inhibited its increase to about 60% level of ASK1 (lane 8 in Fig. 3) (an average of 3.4-fold increase).
Dominant negative mutant of PKR inhibits ASK1 activity in vitro
The above results suggested that PKR modifies ASK1 activity. We therefore examined by means of an in vitro kinase assay whether PKR directly affects ASK1 activity. ASK1 was immunoprecipitated with anti-HA Ig from the extract of HEK293 cells permanently expressing ASK1, and an autophosphorylation reaction was induced in the presence or absence of poly(I)–poly(C). Poly(I)–poly(C), however, did not affect ASK1 activity at all (Fig. 4A). This might be due to the amount of PKR coimmunoprecipitated, which was so small that its effect could not be detected. Therefore, PKR-KR was further transfected into ASK-8 cells and ASK1 activity was examined. An increase in the intensity of the PKR signal was observed in PKR-KR-transfected cells by Western blotting (lanes 2 and 4 in Fig. 4B), while the amount of ASK1 in ASK-8 cells did not change (lanes 3 and 4 in Fig. 4B). Transfection of PKR-KR revealed a 35% decrease (an average from three independent experiments) in the autophosphorylation activity of ASK1 (lanes 7 and 8 in Fig. 4B), suggesting that PKR affects the activity.
Effect of PKR on the activity of ASK1 in vivo
To examine in vivo the effect of PKR on ASK1 activity, HEK293 cells permanently transfected with wild type or dominant negative mutant of ASK1 were exposed to either poly(I)–poly(C) or H2O2 in the presence of [32P]orthophosphate, and cell lysates were prepared. ASK1 was immunoprecipitated with anti-HA Ig. Immunecomplex was then resolved by SDS/PAGE, and phosphorylation reaction was visualized by autoradiography. Exposure to H2O2 markedly increased the intensity of ASK1 about 2.7-fold (an average of two independent experiments) compared with that without H2O2 (AR in Fig. 5B), indicating H2O2 activates ASK1 as described . Exposure to poly(I)–poly(C) also increased the signal for ASK1 about 2.1-fold (an average of two independent experiments) compared with that without poly(I)–poly(C) (AR in Fig. 5A), suggesting that PKR could activate ASK1. The amount of ASK1 in the immunecomplexes was not changed by these exposures, which was verified by Western blotting (WBs in Fig. 5A and B). On the other hand exposure to poly(I)–poly(C) did not markedly increase the phosphorylation state of ASK-KM, a kinase negative mutant of ASK1 [15,19]. This may indicate that PKR does not directly phosphorylate ASK1 but rather supports autophosphorylation activity of ASK1. All these results suggest that PKR could activate ASK1, although it remains possible that poly(I)–poly(C) directly activates ASK1. However, the latter might be unlikely, as poly(I)–poly(C) did not activate ASK1 in vitro (Fig. 4A).
Besides having the antiviral activity of type I interferons, PKR has been shown to transduce signals such as dsRNA, LPS, platelet-derived growth factor, Fas, and TNF-α[13,26,27]. As most of these signals are capable of inducing apoptosis, PKR seems to transduce apoptotic signals, especially receptor-mediated stimuli. As these stimuli also have been shown to activate protein kinases of the MAPK family , cross talk between PKR and the MAPK cascade has been proposed. However, the pathway linking PKR and the MAPK family remains to be clarified.
In the present study, we showed that PKR colocalized and was coimmunoprecipitated with ASK1 when cotransfected into COS-1 cells. The interaction of PKR with ASK1 does not require kinase activity of PKR, as a kinase negative mutant as well as endogenous PKR were coimmunoprecipitated with ASK1. This seems to be consistent with recent reports that the interaction of PKR with IκB kinase β or the signal transducer and activator of transcription1 does not require kinase activity of PKR [29,30]. However, PKR-KR decreased the autophosphorylation activity of ASK1 in vitro, and inhibited both the activation of p38 MAPK and apoptosis induced by ASK1 in vivo. Moreover, exposure to poly(I)–poly(C) increased ASK1 phosphorylation. All these results indicate that PKR dose not play only a structural role but rather modulates a signaling pathway of ASK1. Therefore the binding of PKR to ASK1 might cause a conformational change in ASK1 for activation, of which is dependent on PKR activity, or an additional factor(s) might be requited to activate ASK1.
It has been shown that the activation p38 by poly(I)–poly(C) or LPS treatment was abrogated in PKR-null fibroblasts, while generally acting stimuli such as osmotic shock or H2O2 did not require PKR to activate MAPKs . By contrast, a variety of stimuli such as TNF-α, IL-1, Fas, ceramide, H2O2, osmotic shock, heat shock, anticancer drugs, protein synthesis inhibitors and so on activate ASK1 . Thus, PKR seems to be a specific transducer of inflammatory stimuli, while ASK1 is a general transducer. Our results indicate that the signaling pathway directed from PKR to ASK1 may define the roles of these kinases. Determining the binding site(s) in the PKR and ASK1 molecules will help to confirm this speculation.
We are grateful to Dr Ara Hovanessian for providing the anti-PKR mAb and PKR cDNA, and Dr Hidenori Ichijo for ASK1 and ASK-KM cDNAs. This work was supported in part by a grant-in-aid for scientific research from the Ministry of Education, Science, Sports, and Culture, Japan.