The 58-kDa cellular inhibitor of the double stranded RNA-dependent protein kinase requires the tetratricopeptide repeat 6 and DnaJ motifs to stimulate protein synthesis in vivo.

Double stranded RNA-dependent protein kinase (PKR) is a double stranded RNA-activated, interferon-induced serine-threonine kinase that participates in both the antiviral and antiproliferative properties of interferon. We previously found that influenza virus inhibited PKR function by recruiting or activating a cellular inhibitor termed P58IPK. The present study was undertaken to complement our earlier analyses, which demonstrated that P58IPK efficiently inhibited PKR autophosphorylation and activity in vitro. We now report that P58IPK down-regulates PKR and, in turn, stimulates protein synthetic rates inside the cell. Using transfection analysis, we show that P58IPK stimulates translation of secreted embryonic alkaline phosphatase reporter gene mRNA. Furthermore, we found that at least two regions of the P58IPK molecule were required for PKR inhibitory activity in COS-1 cells: (i) the DnaJ similarity region at the carboxyl terminus (amino acids 391-504); and (ii) the tetratricopeptide repeat 6 (TPR6) domain (amino acids 222-255) located in the middle of the P58IPK protein and within the eukaryotic protein synthesis initiation factor 2α homology region. P58IPK variants lacking either one of these regions were unable to stimulate secreted embryonic alkaline phosphatase protein synthetic rates. Consistent with this data is the observation that the ΔTPR6 mutant (the P58IPK variant lacking the TPR6 motif) failed to block PKR activity in vitro. Based on these data and our earlier in vitro functional and PKR-P58IPK binding analyses, a revised model of PKR regulation by P58IPK is presented.

PKR 1 is a cAMP-independent serine-threonine kinase that is induced by interferon treatment (1,2). On activation by double stranded RNA, PKR undergoes autophosphorylation and catalyzes the phosphorylation of the ␣ subunit of eukaryotic protein synthesis initiation factor 2 (eIF-2␣), resulting in inactivation of the latter (3,4). These events lead to dramatic decreases in protein synthetic rates inside the cell. This situation is not favorable to viruses, which must make their proteins to replicate. Much of the focus on PKR has related to the role of the kinase in mediating the inhibitory interferon response to viral infection (5)(6)(7)(8). However, there is accumulating evidence implicating PKR, which is constitutively expressed in eukaryotic cells in the absence of interferon (9), in the regulation of normal cellular processes. These include signal transduction (10,11), cellular differentiation (12), growth and proliferation (13)(14)(15), and gene expression at the transcriptional level (16). Finally, the importance of PKR in regulating cell growth and gene expression is underscored by evidence suggesting PKR to be a tumor suppressor gene (17)(18)(19)(20).
PKR, which is efficiently activated by both cellular (21,22) and viral (6,(23)(24)(25) RNAs, is, in turn, subjected to stringent regulation by both viral and cellular gene products. For example, adenovirus encodes a polymerase III gene product, VAI RNA, which binds to and inactivates PKR (26,27). In addition, vaccinia virus encodes the K3L protein, which also binds to PKR and inhibits its activity (28,29). Influenza virus has evolved an unusual mechanism to down-regulate PKR and thus ensure that viral protein synthesis is not compromised during infection. Influenza virus recruits or activates a cellular protein termed P58 IPK , based on its M r of 58,000 (30,31). P58 IPK is in a complex with its own inhibitor, I-P58 IPK , and is thus normally inactive in uninfected cells. On influenza virus infection, P58 IPK is dissociated from I-P58 IPK and available to complex with PKR, thereby preventing phosphorylation of eIF-2␣ (30 -32). Sequence analysis of P58 IPK cloned from bovine, human, and mouse cells showed P58 IPK to be a novel, highly conserved 504-amino acid protein (33,34). Spanning the P58 IPK molecule are nine tandemly arranged motifs referred to as tetratricopeptide repeats (TPRs). TPRs are 34-amino acid motifs that have been demonstrated to form amphipathic ␣ helices that can direct protein-protein interactions (35). The middle region of P58 IPK contains limited homology to the amino-terminal region of eIF-2 ␣, the natural substrate of PKR. Finally, the carboxyl terminus of P58 IPK shows extensive homology to the conserved J domain of the bacterial DnaJ protein (36). Like PKR, P58 IPK likely regulates cellular gene expression in the absence of virus infection. It was demonstrated that overexpression of P58 IPK in NIH 3T3 cells led to the malignant transformation of these cells, suggesting P58 IPK to be involved in cellular growth control (37).
Using in vitro assays with purified components, we previously showed that native or recombinant P58 IPK inhibited both the autophosphorylation of PKR and the phosphorylation of eIF-2␣ by an already activated kinase (31,34). We have not, however, directly demonstrated that P58 IPK functions in vivo, inside the cell, to block PKR activity and stimulate protein synthetic rates. We therefore developed an in vivo cotransfection assay using cDNAs encoding P58 IPK and an alkaline phos-phatase reporter gene. This allowed us both to analyze P58 IPK inhibitory activity and to precisely map P58 IPK functional domains. We can now report that PKR inhibition and the resultant translational stimulatory activity appear to require two distinct regions of the P58 IPK molecule: the TPR6 domain located in the central region (amino acids 222-255) and the DnaJ similarity region at the COOH terminus (amino acids 391-504). Based on these data, a model of PKR regulation by P58 IPK will be presented.
The glutathione S-transferase (GST) fusion constructs for the in vitro studies were made as follows. Full-length, wild type P58 IPK was cloned into pGEX2T plasmid (Pharmacia Biotech Inc.) as described previously (34). The GST-⌬TPR6 construct was made by digesting the ⌬TPR6 plasmid (described above) with HindIII and EcoRI restriction enzymes and subcloning this 0.9-kb internal fragment into the GST-P58 IPK backbone.
Transfection Procedures and Secreted Embryonic Alkaline Phosphatase (SEAP) Assays-COS-1 cells were transfected using the DEAEdextran/chloroquine method (38). Monolayers of COS-1 cells were washed once with serum-free Dulbecco's modified Eagle's medium. DNA was added to the cells in serum-free Dulbecco's modified Eagle's medium in the presence of 250 g/ml DEAE-dextran. After incubation for 2 h, chloroquine was added to a final concentration of 80 M. After another 2-h incubation, the transfection mixture was removed, and the cells were "shocked" with a solution of 20% glycerol in HEPES-buffered saline for 2 min. The cells were then washed twice with Hank's balanced salt solution and then incubated in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum. To measure SEAP activity, culture medium was removed from the cells at 40 h after transfection. The cells were then washed and incubated in prewarmed fresh media for 30 min. The media were then collected from the transfected cells, heated to 65°C for 5 min, and then assayed for SEAP activity as previously published (39).
Analysis of Protein Synthesis by 35 S Pulse Labeling and Immunoprecipitation-Transfected cells were labeled for 30 min with [ 35 S]methionine (605 Ci/ml) in methionine-free Dulbecco's modified Eagle's medium. The labeled cells were washed twice with ice-cold Hank's balanced salt solution and lysed in Triton lysis buffer (10 mM Tris-HCl, pH 7.5, 50 mM KCl, 1 mM dithiothreitol, 2 mM MgCl 2 , 100 g/ml aprotinin, 1 mM phenylmethylsulfonyl fluoride, and 1% Triton X-100). The clarified extract was then subjected to immunoprecipitation with protein A-agarose that had been prereacted with a polyclonal antibody against placental alkaline phosphatase (Dako). The immunoprecipitates were washed four times with high salt buffer I (20 mM Tris, pH 7.5, 50 mM KCl, 0.4 M NaCl, 1% Triton X-100, 1 mM EDTA, 10 g/ml aprotinin, 1 mM dithiothreitol, 0.2 mM phenylmethylsulfonyl fluoride, and 20% glycerol) and three times with high salt buffer II (10 mM Tris, pH 7.5, 100 mM KCl, 0.1 mM EDTA, 10 g/ml aprotinin, and 20% glycerol). Bound, radiolabeled proteins were separated on a SDS-9% polyacrylamide gel and visualized using autoradiography. The SEAP protein signal was quantified using PhosphorImager analysis (Molecular Dynamics).
Western Blot (Immunoblot) Analysis-Transfected cells were washed twice with ice-cold Hank's balanced salt solution and lysed in Triton lysis buffer. After SDS-polyacrylamide gel electrophoresis, polypeptides were transferred to nitrocellulose (40) and detected with P58 IPK monoclonal antibodies (37) using the ECL chemiluminescence system (Amersham Corp.).
Northern Blot RNA Analysis-Poly(A) ϩ RNA was isolated from the transfected cells by the acid guanidinium thiocyanate-phenol-chloroform extraction method (41). The poly(A) ϩ RNA was denatured, electrophoresed in a 1% agarose gel containing 0.2 M formaldehyde, and transferred to a nylon membrane (Hybond N, Amersham). The blot was hybridized to a 32 P-labeled 1.6-kb HindIII-XhoI fragment containing the SEAP gene. Afterward, the same blot was stripped and reprobed with a 32 P-labeled 1.7-kb BamHI fragment containing the human actin gene. Both the radiolabeled SEAP and actin RNA signals were quantified using PhosphorImager analysis.
In Vitro Assay for P58 IPK Activity-Both WT P58 IPK and the ⌬TPR6 variant were expressed in Escherichia coli and purified as GST fusion proteins as described previously (34). To test for kinase inhibitory activity, purified GST fusion proteins were preincubated with purified PKR (4) for 10 min at 30°C in 10 mM Tris-HCl, pH 7.5, 50 mM KCl, 2 mM MgCl 2 , 2 mM MnCl 2 , 2 M ATP, 5 g/ml aprotinin, 1 mM dithiothreitol, and 500 g/ml bovine serum albumin. Subsequently, an activator (0.1 g/ml poly(rI):poly(rC)) was added in the presence of 5 Ci [␥-32 P]ATP and incubated for an additional 10 min. Finally, exogenous substrate (10 g of calf thymus histone IIA) was added with an additional 10 Ci [␥-32 P]ATP, and the mixture was allowed to incubate for another 20 min. The reaction was stopped by the addition of 2 ϫ disruption buffer (180 mM Tris-HCl, pH 6.8, 4.5% SDS, 23% (v/v) glycerol, 17 mM EDTA, 20 g RNase A, and 3 M 2-mercaptoethanol) and boiling. The samples were then separated by 14% SDS-polyacrylamide gel electrophoresis and visualized by autoradiography.

Development of an in Vivo P58 IPK Functional
Assay-A cotransfection assay was developed to measure the ability of P58 IPK to inhibit endogenous PKR function, thereby stimulating mRNA translation inside the cell. A similar assay was frequently and successfully used with other PKR inhibitors, including adenovirus VAI RNA and the reovirus 3 protein FIG. 1. Schematic of the wild type and mutant P58 IPK constructs. Wild type P58 IPK contains nine tandemly arranged 34-amino acid repeats termed TPR domains. P58 IPK amino acids 207-303 are homologous to the amino-terminal region of eIF-2␣ (amino acids 26 -120). The DnaJ homology region resides at the COOH terminus (amino acids 392-463). The Ser-241 mutant contains a single amino acid mutation from a serine to alanine at position 241. A summary of the in vivo translational stimulatory activity of the various P58 IPK variants is indicated at the right. (42)(43)(44)(45)(46). This in vivo assay involved cotransfecting into COS-1 cells a constant amount of SEAP reporter gene cDNA together with increasing amounts of P58 IPK cDNA. SEAP enzymatic activity was then measured as described under "Materials and Methods." Although this assay does not directly measure PKR enzymatic activity, this protocol provides a quantitative measure of mRNA translational stimulation, which occurs as a result of PKR inhibition. To test the efficacy of our assay, we first measured the SEAP activity of cells cotransfected with plasmids encoding the adenovirus VAI RNA and SEAP and compared it with the activity obtained when cells were transfected with SEAP cDNA alone. VAI RNA stimulated SEAP activity approximately 4-fold (Fig. 2). Similarly, after cotransfection with a cDNA encoding P58 IPK , SEAP enzymatic levels increased approximately 3.5-fold (Fig. 2). This level of reporter gene stimulation was comparable to that previously observed for reovirus 3, another protein inhibitor of PKR activity (46). As a negative control, we cotransfected human actin cDNA along with SEAP cDNA into COS-1 cells. In contrast to P58 IPK or VAI RNA, the cotransfection of actin DNA did not result in stimulation of SEAP reporter activity. Instead, SEAP activity actually declined relative to the SEAP-alone control, probably as a result of enhanced PKR activation that occurs in transfected cells (47). To further demonstrate that the observed stimulatory activity was specific for P58 IPK , a titration of P58 IPK cDNA transfected into the cells was attempted. Both SEAP activity measurements and Western analyses showed there was an increase in SEAP reporter activity, which correlated with increased P58 IPK expression (data not shown). This was not the case for the actin control, in which increasing amounts of actin expression did not result in an enhancement of SEAP enzymatic activity (data not shown).
P58 IPK Stimulates the Rate of SEAP Protein Synthesis in Cotransfected COS Cells-The previous analysis measured only the enzymatic activity of SEAP secreted into the culture media during a 30-min pulse. To determine whether the increase in SEAP enzymatic activity was due directly to enhanced SEAP mRNA translation, cotransfected cells were pulse-labeled with [ 35 S]methionine. Cytoplasmic extracts were prepared and subjected to immunoprecipitation using a SEAPspecific polyclonal antibody as described under "Materials and Methods." When increasing amounts of P58 IPK cDNA were cotransfected into the cells with a constant amount of SEAP cDNA, there was a corresponding increase in SEAP protein synthesis (Fig. 3A, lanes 3-5). This increase in the rate of SEAP protein synthesis (3-6-fold as determined by PhosphorImager analysis) paralleled the up to 3-4-fold increase in SEAP enzymatic activity (Fig. 3A, bottom). To determine whether the increased rate of SEAP protein synthesis was due to increased SEAP mRNA levels, Northern blot analysis of poly(A) ϩ RNA from the P58 IPK -and SEAP-cotransfected cells was performed (Fig. 3B). SEAP RNA levels were quantitated using Phospho-rImager analysis and normalized according to control actin mRNA levels. Cotransfection of WT P58 IPK cDNA with SEAP cDNA resulted not in increased but, rather, decreased levels of SEAP mRNA (Fig. 3B, compare lanes 1 and 2 or 4 and 5). These data demonstrate that cotransfection with P58 IPK cDNA enhanced the translation of SEAP mRNA and caused minor de-  (lanes 3, 6, and 9), 4 g (lanes 4, 7, and 10), or 6 g (lanes 5, 8, and 11) of the P58 IPK cDNA construct indicated at the top. Lane 1, immunoprecipitation from cells transfected with empty vector alone. The total cDNA content of all the transfected samples was supplemented with the pcDNA1/Neo vector DNA to bring the total to 7 g of cDNA for all transfections. The migration of the SEAP protein is indicated at the right. SEAP enzymatic activity (relative to the 100% activity of cells transfected with SEAP alone) was concurrently measured and is indicated at the bottom. B, Northern blot analysis of RNA extracted from cells cotransfected with SEAP and WT or mutant P58 IPK cDNAs (as indicated at the top). Approximately 2 g of Poly(A) ϩ RNA was electrophoresed on a 1% formaldehyde agarose gel, which was subsequently blotted onto nylon filters. The blots were first probed with a 32 P-labeled 1.6-kb HindIII-XhoI fragment containing the SEAP gene, stripped, and then reprobed with radiolabeled actin DNA. The amounts of SEAP and actin mRNAs present in each of the samples were quantified using PhosphorImager analysis. To correct for loading variation, the SEAP mRNA levels were normalized relative to actin mRNA levels. A value of 100 was arbitrarily assigned to levels of SEAP mRNA present in cells transfected with SEAP cDNA alone. creases in SEAP mRNA levels.
The P58 IPK Central Domain and TPR6 Motif Are Critical for PKR Inhibitory Activity and Enhanced mRNA Translation in Vivo-Previous data from our laboratory, using in vitro assays to map the functional domains of P58 IPK , implicated P58 IPK amino acids 168 -277 as critical for function (34). To directly test the importance of the central region (which contains the eIF-2␣ homology region) for PKR inhibitory activity and mRNA translational stimulation in vivo, we tested the ⌬N1 mutant (deleted for amino acids 188 -301; see Fig. 1) in the cotransfection assay. The ⌬N1 variant lacks not only the central domain of P58 IPK , but also TPR motifs 5-7. The ⌬N1 mutant was first tested in our in vitro assay and, as expected, was demonstrated to be approximately 30-fold less active than the WT P58 IPK protein (data not shown). Similarly, the ⌬N1 variant, over a range of concentrations, was unable to stimulate SEAP protein synthetic rates (Fig. 3A, compare lanes 6 -8 with lanes 3-5) or enzymatic activity (Figs. 3A, bottom of lanes 6 -8, and 4A). Western blot analysis confirmed that the ⌬N1 protein was expressed at approximately the same level as WT P58 IPK (Fig.  4B, compare lanes 2 and 3). Consistent with the pulse-labeling experiments, Northern blot studies revealed only minor decreases in SEAP mRNA levels between cells transfected with the ⌬N1 mutant and wild type P58 IPK (Fig. 3B, compare lanes  5 and 7). We conclude that the P58 IPK central domain, spanning amino acids 188 -301, is important for inhibitory function both in vitro and in vivo.
To more precisely map P58 IPK functional domains, we generated three smaller TPR deletion mutants: ⌬TPR5, ⌬TPR6, and ⌬TPR7 (see Fig. 1). Similar to ⌬N1, the ⌬TPR6 mutant protein was unable to stimulate SEAP enzymatic activity compared with WT P58 IPK in cotransfected cells. Indeed, SEAP levels were reduced in ⌬TPR6and SEAP-cotransfected cells compared with cells transfected by SEAP cDNA alone (Fig. 4A). Western blot analysis confirmed efficient synthesis of the ⌬TPR6 variant (Fig. 4B, compare lanes 5 and 6). In addition, Northern blot analysis confirmed that the inactivity of this mutant was at the level of translation, since only minor decreases in SEAP mRNA levels were observed when compared with the WT (52 versus 77%, respectively, of mRNA compared with levels in cells transfected by SEAP cDNA alone; Fig. 3B,  compare lanes 5 and 6). In contrast to ⌬TPR6, the ⌬TPR5 and ⌬TPR7 mutants induced enhanced translational stimulatory activity compared with WT P58 IPK (Fig. 5A). This enhanced activity was not due to enhanced variant expression, as Western blot analysis revealed that ⌬TPR5 and ⌬TPR7 were expressed at lower levels than the WT P58 IPK (Fig. 5B). It is tempting to speculate that this stimulation by the two mutants is due to lack of binding of negative regulatory factors, which normally interact with these TPR domains. In support of this notion, we recently identified a novel gene product that was found to interact with the TPR7 domain using a yeast twohybrid screen.  1 and 4). Migration of P58 IPK -related proteins is indicated on the sides.

FIG. 5. Analysis of the function of P58 IPK variants that lack TPR motifs 5 and 7 and the DnaJ similarity region.
A, SEAP enzymatic activity in cotransfected cells. In this experiment, activity is described as the percentage of activity found in cells cotransfected with WT P58 IPK and SEAP cDNAs. The results shown here were derived from three separate transfection experiments, with each performed and sampled in duplicate. SEAP cDNA was cotransfected with 2 g of WT or mutant P58 IPK cDNAs. B, Western blot analysis of WT and mutant P58 IPK protein expression. Analysis was performed as described in the legend to Fig. 4A. Migration of P58 IPK related proteins is indicated on the sides.
The DnaJ Homology Region Is Required for P58 IPK Activity in Vivo but not in Vitro-The carboxyl terminus of P58 IPK shows considerable homology to the conserved J domain of the E. coli DnaJ heat shock family of proteins (36,48). Previous in vitro assays determined that the DnaJ homology region at the COOH terminus was dispensable for P58 IPK inhibitory activity in vitro (34). However, because eukaryotic DnaJ homologs have been suggested to participate in protein-protein interactions (36), we next determined whether the DnaJ similarity region contributed to the PKR inhibitory function inside the cell. The P58 IPK -8-3 mutant (which lacks the DnaJ region and amino acids 391-504; Fig. 1) failed to stimulate SEAP protein synthetic rates in a pulse-labeling analysis (Fig. 3A, lanes 9 -11) or in a SEAP enzymatic activity assay (Figs. 3A, bottom of lanes  9 -11, and 5A). Western blots revealed efficient synthesis of the 8-3 variant (Fig. 5B, lane 7), and Northern blot analysis failed to show significant differences in SEAP mRNA levels between cells cotransfected with SEAP cDNA and wild type P58 IPK cDNA or 8-3 mutant DNA (Fig. 3B, compare lanes 2 and 3). These data demonstrate that, like the TPR6 domain, the DnaJ similarity region is required for PKR inhibitory activity and the resultant translational stimulation in vivo.
In Vitro Analysis of PKR Inhibitory Function-To provide a direct link between our in vivo cotransfection assay and the ability of P58 IPK to regulate PKR, we examined the PKR inhibitory activity of the key ⌬TPR6 mutant using an in vitro kinase inhibition assay. Purified GST-P58 IPK or GST-⌬TPR6 fusion proteins were preincubated with purified PKR for 10 min at 30°C, after which the poly(rI):poly(rC) activator and histone substrate was added, as described under "Materials and Methods." As shown in Fig. 6, neither the vector control (GST, lanes 2 and 3) nor the GST-⌬TPR6 fusion construct (lanes 6 and 7) was able to reduce the amount of PKR-mediated histone phosphorylation. In contrast, GST-P58 IPK (Fig. 6, lanes 4 and 5) efficiently blocked PKR phosphorylation activity. At the higher concentration of GST-P58 IPK , the percentage of inhibition of PKR activity was 71% as determined by Phosphor-Imager analysis. These data confirmed our cotransfection analyses demonstrating the inability of the ⌬TPR6 P58 IPK mutant to inhibit PKR. DISCUSSION In the present study, we developed a cotransfection assay to analyze P58 IPK function inside the cell. As was the case in similar assays with viral-encoded PKR inhibitors, such as the reovirus 3 protein (44 -46) and the adenovirus VAI RNA (42,43,49), P58 IPK down-regulated endogenous PKR activity, leading to the stimulation of reporter gene mRNA translation. Although cotransfection with wild type P58 IPK and variants caused minor decreases in SEAP mRNA levels (possibly due to competition for transcription factors; Ref. 44), the predominant effect of P58 IPK action was at the level of mRNA translation. This was perhaps most dramatically demonstrated by the cotransfections with WT P58 IPK and SEAP, in which the levels of SEAP mRNA were reduced approximately 2-fold, whereas the levels of SEAP mRNA translation were concurrently stimulated up to 3-6-fold (Fig. 3). Unexpectedly, we found that both the central region, containing the TPR6 motif, and the COOH terminus, encompassing the DnaJ similarity region, were required for P58 IPK activity in vivo (summarized in Fig. 1). In accordance with this functional data, we found that P58 IPK lacking the TPR6 motif failed to interact with PKR using the yeast two-hybrid system to analyze in vivo binding (50). In addition, the ⌬TPR6 variant failed to inhibit PKR phosphorylation activity in vitro (Fig. 6). In contrast, the COOH region of P58 IPK is not required for P58 IPK -PKR interactions (32,50), nor is the DnaJ region required for P58 IPK function in vitro (34). This region shows homology as high as 55% identity and 80% similarity to select DnaJ proteins that play an essential role in the chaperone function of the hsp70-like DnaK protein of E. coli (36,48,51).
Based on our in vitro and in vivo functional and binding data, we present what has become an increasingly complex model of PKR regulation by P58 IPK (Fig. 7). As we have previously shown, P58 IPK likely exists in an inactive complex with its own inhibitor, I-P58 IPK , in uninfected cells. After influenza virus infection, P58 IPK dissociates from I-P58 IPK , allowing P58 IPK to FIG. 6. In vitro analysis of P58 IPK function. Equimolar amounts (0.6 and 1.2 pm) of the P58 IPK wild type and ⌬TPR6 variant were tested for their ability to inhibit PKR-mediated histone phosphorylation as described under "Materials and Methods." As a negative control, we tested the PKR inhibitory activity of material that bound to and eluted from glutathione-agarose beads exposed to E. coli extracts that expressed the GST fusion vector alone. The histone bands were subjected to PhosphorImager analysis for quantitation. The degree of histone phosphorylation in lane PKR when no GST fusion protein had been added was arbitrarily set at 100%. Relative to this PKR control, the GST vector and the ⌬TPR6 mutant showed little or no inhibition, whereas WT P58 IPK inhibited PKR-mediated histone phosphorylation by more than 70% at the 1.2-pm concentration.

FIG. 7.
Model for the regulation of PKR activity and mRNA translation by the cellular P58 IPK inhibitor. P58 IPK is normally in an inactive complex with its own regulator, referred to as I-P58 IPK . Subsequent to influenza virus infection, I-P58 IPK dissociates from P58 IPK , allowing P58 IPK to interact with and inhibit PKR. Both the TPR6 and DnaJ regions are required for P58 IPK function in vivo. Based on the data presented in the current report, we hypothesize that for P58 IPK -PKR interactions in vivo, protein X must first interact with the DnaJ region of P58 IPK . Protein X then facilitates P58 IPK -PKR interaction either: (i) by targeting P58 IPK to PKR inside the cell, or (ii) by inducing the proper folding of P58 IPK such that the inhibitor can then interact with PKR. Earlier work has found that the TPR6 motif is critical for P58 IPK binding to PKR, and that P58 IPK binds to PKR in a region spanning the protein kinase regulatory (R) and catalytic (C) domains. See text for additional details. interact with PKR. Although we found that P58 IPK can interact with PKR in the absence of other protein or RNA factors in vitro (32,34), the present study suggests that additional factors must interact with P58 IPK and/or PKR to allow for P58 IPK -PKR interactions inside the cell. We hypothesize that such a factor (Fig. 7, X) binds to the P58 IPK DnaJ region, a region already known to promote protein-protein interactions (36). Two alternative pathways can further explain how factor X facilitates interactions between P58 IPK and PKR. In one scenario, we propose that X binds to P58 IPK and targets P58 IPK to the correct cellular compartment and then to the protein kinase. This model would likely require that X interact with both P58 IPK and PKR. Such a model could explain why P58 IPK variants, lacking the DnaJ region, can still interact with and inhibit PKR in vitro. Alternatively, factor X, which might itself be a molecular chaperone (52,53), may bind to the DnaJ region of P58 IPK and alter the conformation or folding of P58 IPK , such that it can now interact with PKR and inhibit the protein kinase. Our earlier studies mapped the P58 IPK -interactive site of PKR to amino acids 244 -296, spanning the regulatory and catalytic borders of PKR but including the ATP binding region of the protein kinase catalytic domain (50). In any case, the end result of either of these two model pathways would be the P58 IPK induced inhibition of PKR activity and resultant stimulation of mRNA translation. The interplay between P58 IPK , PKR, and protein X is reminiscent of other systems in which a TPR protein either interacts with other TPR proteins or else another regulatory protein, such as ␣2, to control its activity (54,55). Moreover, there is even precedence in the literature for the interaction of another TPR protein, Hip, with the molecular chaperone Hsc70 (56).
Before closing, it is worth commenting on the requirement of the TPR6 motif for both P58 IPK binding and function. Curiously, the central TPR motifs, often including TPR6, were found to be critical for the function of other TPR proteins, including Nuc2, CDC-23, CDC-27, and BimA (57-60). Furthermore, single amino acid substitutions in TPR domains 5-7 have been shown to abolish the biological function of the yeast CDC23 gene (58,59). These data suggest a critical function for this region of these TPR proteins despite their diverse regulatory roles. The P58 IPK TPR6 domain contains sequences with limited similarity to a region of eIF-2␣, the natural PKR substrate. This homologous region, however, is not extensive, with a maximum 31% identity with five gaps. Nonetheless, this region of P58 IPK does contain the PKR consensus eIF-2␣ phosphorylation motif ELS (P58 IPK amino acids 239 -241 and eIF-2␣ amino acids 49 -51). Since the eIF-2␣ serine 51 is phosphorylated by PKR (61), these data suggested that P58 IPK may function as a PKR substrate (62). This would require the presence of serine 241 to functionally mimic the eIF-2␣ phosphorylation site. However, we found that a serine to alanine mutation at amino acids 241 (see Fig. 1) did not inactivate P58 IPK function either when P58 IPK was assayed in vitro (34) or in vivo (data not shown), suggesting that P58 IPK probably does not function as a PKR pseudosubstrate. In accordance with these results, we found that P58 IPK likely binds at or near the ATP binding domain on PKR, a region clearly distinct from the substrate binding site. In contrast, the vaccinia virus-encoded PKR inhibitor K3L likely does function as a PKR substrate (28,29,63). This argument is supported by our recent evidence that, unlike P58 IPK , K3L binds to a region of PKR spanning amino acids 367-551, which localizes to the large lobe of the PKR protein kinase, thought to be largely responsible for substrate recognition and binding (50,64).