Regulation of PACT-Mediated Protein Kinase Activation by the OV20.0 Protein of Orf Virus

Cell culture and virus.

The human embryonic kidney cell line (293T) and goat fibroblast cells were cultured in Dulbecco's modified Eagle's medium (DMEM; Gibco BRL, Life Technologies Corporation, Carlsbad, CA) supplemented with 10% fetal calf serum (FCS; HyClone, Logan, UT) and 1% penicillin-streptomycin (Gibco BRL). Cells were maintained at 37°C with 5% CO₂.

Goat fibroblast cells were used for the propagation of recombinant ORFV possessing enhanced green fluorescent protein (EGFP) expression ability (ORFV-GFP) (8). Infection was performed by incubating the indicated multiplicity of infection (MOI) of ORFV with cell monolayers at 80% confluence in infection medium (DMEM without FCS). The infection medium was replaced with 2% FCS DMEM at 1 h postabsorption.

Construction of plasmids.

A series of plasmids harboring various lengths of OV20.0 and PKR were constructed for transient expression in mammalian cells (Table 1). All delineated deletions of OV20.0 and PKR fragments were built by overlap extension PCR (34) with the gene-specific primer sets listed in Table 2. Shortly, two separate rounds of PCRs were performed to generate the upstream or downstream sequences of the deletion region from template containing wild-type genes. Subsequently, products of two separate reactions, with 15-bp overlaps, then served as the template for the next run of PCR using the outermost primer set to generate the final expected fragments.

For expressing OV20.0 with deletions of different regions, DNA fragments of full-length OV20.0L or with the deletion of residues 1 to 79, 80 to 129, 130 to 184, and 80 to 184 were separately amplified from pcDNA3.1-OV20.0-eGFP (8) by PCR using specific primer sets containing recognition sequences of HindIII and NotI on forward and reverse primers, respectively (sequences of the primers are listed in Table 2). All of these insert fragments were digested by corresponding restriction enzymes and ligated with the linearized p3x FLAG-CMV-14 vector (Sigma-Aldrich).

For the expression of different truncated forms of PKR, inserts deleted of the RNA binding domain (ΔRBD1), ΔRBD2, or Δkinase D were created from pcDNA3.1-PKR-HA plasmid (8) by the described overlap extension PCR with degenerated primers containing BamHI and NotI recognition sequences. The amplicons then were digested with corresponding restriction enzymes, followed by ligation with pcDNA3.1-HA plasmid linearized by BamHI and NotI enzymes.

Transfection and immunoprecipitation.

In general, human 293T cells were seeded in 6-well plates at a density of 1.5 × 10⁶ cells/well. The next day, 5 μg of plasmid was used for transfection by Lipofectamine 2000 (Invitrogen), guided by the manufacturer's instructions.

For conducting immunoprecipitation (IP), constructs expressing various truncated forms of OV20.0, PACT (26), or PKR were transfected to cells seeded 1 day before the experiment. After 24 h of transfection, cells were harvested, resuspended in lysis buffer (50 mM Tris HCl, pH 7.4, with 150 mM NaCl, 1 mM EDTA and 1% Triton X-100), and incubated on ice for 10 min, followed by centrifugation at 10,000 × g. Crude cell extract was transferred to a new tube and mixed with anti-FLAG M2 affinity gel (Sigma-Aldrich) at 4°C for 4 h according to the manufacturer's instructions. The gel then was collected by centrifugation at 800 × g and followed by three washes with wash buffer (50 mM Tris-HCl, pH 7.4, with 150 mM NaCl). Lastly, proteins associated with anti-FLAG M2 affinity gel were eluted by SDS sample dye and resolved by SDS-PAGE, followed by Western blotting.

Double-stranded RNA (poly I·C) pulldown assay.

The generation of synthetic dsRNA followed the procedures described previously (8). Briefly, poly(C)-coated agarose beads (Sigma-Aldrich) were incubated with poly(I) (Sigma-Aldrich) in buffer containing 50 mM Tris (pH 7.0) and 150 mM NaCl at 4°C overnight with gentle rotation. Annealed poly(I·C) agarose beads were collected by centrifugation at 800 × g, followed by several washes in buffer containing 50 mM Tris (pH 7.0)–150 mM NaCl, and then were resuspended in the binding buffer A (50 mM Tris, pH 7.5, 150 mM NaCl, 1 mM EDTA, and 1% Nonidet P-40).

To identify the dsRNA-binding domain of OV20.0, cells transfected with various OV20.0 plasmids were harvested and lysed in radioimmunoprecipitation assay (RIPA) buffer (50 mM Tris, pH 7.9, 150 mM NaCl, 0.1% SDS, 1.0% NP-40, 0.5% sodium deoxycholate) on ice followed by centrifugation at 10,000 × g. Supernatants (whole-cell lysate) were mixed with annealed poly(I·C) beads at 4°C. After 1 h of incubation, the beads was collected by centrifugation at 800 × g, followed by washing twice with buffer C (10 mM Tris-HCl, pH 7.8, 6 mM MgCl₂, 80 mM KCl, 2 mM dithiothreitol, 250 mM sucrose, and 0.1 mM EDTA). Finally, the poly(I·C) beads were collected by centrifugation and resuspended in SDS sample dye.

Immunofluorescence assay (IFA).

Human 293T cells were individually transfected with plasmids expressing EGFP fused with full-length OV20.0 or deletion of residues 1 to 79 (Δ1-79) or 130 to 184 (Δ130-184) (8). At 24 h posttransfection, cells were fixed by 4% paraformaldehyde for 30 min, followed by permeabilization with 0.1% Triton X-100 for another 10 min. Subsequently, cells were incubated with anti-PKR antibody (Abcam) at a dilution of 1:1,000 for 1 h at room temperature. After six washes, goat anti-rabbit IgG (1:2,000-fold diluted) (Alexa Fluor 594; Invitrogen) was used as the secondary antibody. After 1 h of incubation, nuclei were stained with 4,6-diamidino-2-phenylindole (DAPI) for 10 min, followed by confocal microscopy. Between each step, cells were washed at least three times with PBS containing 1% FCS.

Western blot analysis.

Protein samples were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) and then transferred to a polyvinylidene difluoride (PVDF) membrane, followed by immune staining procedures described previously (8). In particular, the antibodies (and corresponding dilutions) were the following: FLAG antibody (1:5,000) (F7425; Sigma-Aldrich), PKR antibody (1: 2,000) (ab32052; Abcam), PKR-p antibody (1:1,000) (ab32036; Abcam), β-actin antibody (1:5,000) (GTX26276; Gene Tex), EGFP antibody (1:5,000) (YH80005; Yao-Hong Biotechnology, Taiwan), hemagglutinin (HA) antibody (1:5,000) (YH80008; Yao-Hong Biotechnology, Taiwan), PACT antibody (ab117704; Abcam), and OV20.0 antibody from mice immunized with purified OV20.0 recombinant protein (1:1,000) (8). Goat anti-rabbit or anti-mouse IgG-conjugated horseradish peroxidase (HRP) (1:10,000-fold diluted) (Jackson Laboratory) served as the secondary antibody. After reacting with secondary antibodies for 1 h, the signal was detected by an enzyme-linked chemiluminescence system (ECL; Amersham, GE Healthcare) and revealed by ImageQuant LAS 4000 (GE Healthcare). Washes with PBS-T (PBS with 0.1% Tween) were done between each step.

Sequence analysis.

Sequences of OV20.0 (ABY41265.1), PACT (AAC25672.1), and PKR (XP_011531289.1) proteins were obtained from GenBank. Multiple-sequence alignment was conducted by the ClustalW2 online service (http://www.ebi.ac.uk/Tools/msa/clustalw2/). Fully conserved residues are labeled with an asterisk. Strong similar properties, with a scoring of >0.5 in the Gonnet PAM 250 matrix among groups, are indicated by a colon. A score of <0.5, indicating weak similarity among groups, is labeled with a dot.

The dsRNA binding region of OV20.0.

According to previous research, both OV20.0 and its N-truncated isoform, sh20 (i.e., Δ1-79), are dsRNA binding proteins in vitro. However, the specific residues or domains required for dsRNA binding ability remain uncertain. Sequence alignments based on aligning ORFV OV20.0 with its orthologs from goatpox virus (GPV; a capripoxvirus) and VACV indicate that the dsRNA-binding domain of OV20.0 resides at the C terminus (8). To investigate the minimal region needed for dsRNA binding, several constructs, namely, full-length OV20.0 and various OV20.0 mutants with deletions of residues 1 to 79, 80 to 129, 130 to 184, and 80 to 184, were generated (Fig. 1A). To assess their dsRNA binding ability, lysates of cells transiently expressing each of these OV20.0 proteins were incubated with beads conjugated with annealed poly(I·C), which is a synthetic dsRNA. As shown in the upper portion of Fig. 1B, in addition to full-length OV20.0, only Δ1-79 was able to be pulled down by poly(I·C) beads (Fig. 1B, lanes 1 and 2), while EGFP and the other OV20.0 mutants with various deletions were not detected in the dsRNA-binding fraction. These results indicated that residues 80 to 184 are required for interacting with dsRNA.

Region of OV20.0 responsible for interaction with PKR.

Interaction between OV20.0 and PKR was reported previously. In this study, we investigated in more detail the binding domain that allows interaction between these two proteins. An earlier study has shown that sh20, a shortened isoform of OV20.0 that consists of residues 80 to 184 (Δ1-79), is sufficient to allow PKR binding; hence, in these experiments, full-length OV20.0 and Δ1-79 were used as positive controls, and the PKR binding ability of the other three mutants was determined. Affinity gel conjugated with FLAG antibody was used to pull down the FLAG-tagged OV20.0 proteins, and the interaction of these proteins with endogenous PKR then was revealed by immunoblot analysis; the results are presented in Fig. 2A. Wild-type OV20.0 and Δ1-79 were able to interact with cellular PKR consistently (Fig. 2A, lanes 1 and 2). Importantly, the deletion mutant lacking residues 80 to 129 (Δ80-129), which has been shown earlier not to bind dsRNA, still is able to interact with PKR (Fig. 2A, lane 3). Nevertheless, the other two mutants (Δ130-184 and Δ80-184), which are missing the last 55 amino acids at the C terminus (residues 130 to 184), are unable to efficiently bind PKR and give only a faint or lack a PKR signal (Fig. 2A, lanes 4 and 5, respectively). These results were confirmed by the immunofluorescent staining (Fig. 2B). Full-length OV20.0 and Δ1-79 fused with EGFP were expressed in the 293T cells and found to colocalize with endogenous PKR. However, when the last 55 amino acids of OV20.0 protein were deleted, this mutant version of OV20.0 (Δ130-184) was found to be present mainly in the nucleus, while PKR remained in the cytoplasm. These findings indicate that OV20.0 uses a domain that contains residues 130 to 184 for PKR binding.

To identify the region of PKR interacting with OV20.0, PKR mutants with deletions of each individual functional domain, namely, deletions of RBD1 (ΔRBD1), RBD2 (ΔRBD2), and the kinase domain (Δkinase D), were generated (Fig. 2C). The results of an immunoprecipitation assay showed that there was a strong interaction between wild-type PKR and OV20.0 (Fig. 2D, lane 1); this was used as the positive control, while EGFP acted as a negative control. This approach reduces concerns regarding nonspecific binding in these experiments (Fig. 2D, lane 5). When a similar approach was used with the three deletion mutants, it was found that the absence of the kinase domain (Δkinase D) weakened the association of PKR with OV20.0 (Fig. 2D, lane 4), indicating the contributions of the kinase region to this interaction. However, when ΔRBD1 and ΔRBD2 were used in similar experiments, the binding of PKR to OV20.0 was completely abrogated (Fig. 2D, lanes 2 and 3). This indicates that the interaction between PKR and OV20.0 involves both RBD1 and RBD2. Taking the above-described findings as a whole, the region containing the last 55 amino acids (130 to 184) of OV20.0 protein seems to be critical to the interaction of OV20.0 with PKR via RBD1 and RBD2.

The region of OV20.0 involved in suppression of PKR phosphorylation.

The above-described findings identify the binding sites that allow interaction between OV20.0 and PKR. Based on these findings, we next determined whether OV20.0 binding is responsible for repressing dsRNA-induced PKR activation. Five constructs expressing various regions of OV20.0, namely, OV20.0, Δ1-79, Δ80-129, Δ130-184, and Δ80-184, were transfected individually into cells, and this was followed by poly(I·C) treatment in order to stimulate PKR phosphorylation. As indicated in Fig. 3, the expression level of actin and the basal expression level of PKR were similar across each of the transfection groups; furthermore, poly(I·C) treatment, as a positive control (PC; lane 7), indeed stimulated phosphorylation of PKR compared to that of mock treatment (lane 1). Consistent with our previous results, OV20.0 and Δ1-79 completely ablated PKR phosphorylation (to a level that was below the sensitivity of the assay) (Fig. 3, lanes 2 and 3). Compared with the positive control, a much lower phosphorylation level was observed when Δ80-129 was present (Fig. 3, lane 4). Noticeably, the deletion of residues 130 to 184 and 80 to 184 did not have any significant impact on PKR phosphorylation (Fig. 3, lanes 5 and 6). These findings indicate that a lack of the PKR-interacting sites of OV20.0 (residues 130 to 184; Δ130-184) results in a weakening of the ability to antagonize dsRNA-induced PKR phosphorylation.

OV20.0 inhibition of PKR activation is mediated by PACT.

In addition to dsRNA, PKR also can be activated by cellular PACT, which is also a dsRNA binding protein (24). We next explored whether OV20.0 is able to regulate PKR via this alternative PACT-mediated route. Increased PKR phosphorylation was observed in 293T cells that were transiently expressing PACT alone (Fig. 4A, lane 1), while in mock-transfected cells, PKR phosphorylation was below the detection level (Fig. 4A, lane 6). To identify the effect of OV20.0 on PKR activation mediated by PACT, increasing amounts (1 μg, 3 μg, 5 μg, and 10 μg) of OV20.0 plasmid were cotransfected with PACT into the cells and the PKR activation level was monitored. As shown in Fig. 4A (lanes 2 to 5), PACT-mediated PKR activation was strongly suppressed by OV20.0 in a dose-dependent manner; image quantification of the intensities of the bands indicated that the group transfected with 10 μg OV20.0 showed a significant inhibition of PKR activation compared with that of cells expressing none or lower levels of OV20.0 (Fig. 4B). These findings suggest that OV20.0 is able to block PACT-induced PKR activation.

Interaction of OV20.0 with PACT.

We investigated whether OV20.0 is able to inhibit PACT-mediated PKR phosphorylation through protein-protein interaction. A series of immunoprecipitation reactions were set up to identify the association between endogenous PACT and OV20.0 protein; this involved individually expressing FLAG-tagged wild-type OV20.0 and its various mutants in 293T cells. Wild-type OV20.0, but not empty FLAG vector, strongly bound to cellular PACT (Fig. 5A, lanes 1 and 6, respectively). It appears that PACT is able to interact with the N terminus of the truncated form of OV20.0 (Δ1-79) to a much lesser extent than wild-type OV20.0 (Fig. 5, lane 2). In contrast, the other C-terminal deletion mutants of OV20.0 had completely lost their ability to associate with PACT (Fig. 5A, lanes 3 to 5). This observation suggests that amino acids 80 to 184 of OV20.0 are necessary for its binding to endogenous PACT.

Subsequently, the region of PACT interacting with OV20.0 was determined using three PACT mutants deficient in domain I (Δ1), domain II (Δ2), and domain III (Δ3); domains I and II are responsible for dsRNA binding, while domain III contributes to PKR binding and activation (Fig. 5B). HA-tagged OV20.0 and FLAG-tagged PACT were coexpressed in cells, and FLAG-based immunoprecipitation assays were performed. Consistent with this finding, FLAG-tagged wild-type PACT, which acts as a positive control, interacted with full-length OV20.0 (Fig. 5C, lane 1), while the empty FLAG vector did not pull down HA-OV20.0 (Fig. 5C, lane 5). Furthermore, PACT with a deletion of either domain II or III retained the ability to interact with OV20.0 (Fig. 5C, lanes 3 and 4), while PACT that lacked domain I (Δ1) was found to have lost the ability to allow binding between PACT and OV20.0 (Fig. 5C, lane 2). These findings demonstrated that residues 80 to 184 of OV20.0 and domain I of PACT are required in order to establish the interaction between these two proteins.

OV20.0 represses PKR activation via its association with PKR rather than with PACT.

Since OV20.0 interacts with both PKR and its activator PACT, the question arises as to which association is critical to the OV20.0-dependent inhibition of PKR activation induced by PACT. The problem associated with complicated interactions was simplified by the use of a PACT mutant (Δ1) that fails to interact with OV20.0 but retains the ability to induce PKR phosphorylation. In such a system, if the inhibition effect of OV20.0 still exists, then the physical interaction between PACT and OV20.0 is not absolutely essential for repressing PACT-mediated PKR activation. Human 293T cells were transfected simultaneously with PACT Δ1 and OV20.0, and this was followed by an assay to detect the level of PKR phosphorylation. The results initially revealed that the overexpression of PACT Δ1 in 293T cells led to PKR phosphorylation compared to the level of expression in untreated cells (Fig. 6, lanes 1 and 2). Noticeably, as the doses of OV20.0 increased, the inhibition effect on PKR phosphorylation was gradually elevated (Fig. 6, lanes 3 to 5), while no effect was observed in the cells transfected with the empty vector (Fig. 6, lanes 6 to 8). Thus, it seems that interaction between OV20.0 and PKR is more critical to the inhibition of PACT-induced PKR phosphorylation than is the binding of OV20.0 to PACT.

OV20.0 interferes with the interaction of PKR with its inducer, PACT.

OV20.0 has the ability to inhibit PACT-induced PKR activation, even after losing its ability to interact with PACT (Fig. 6). A previous report indicated that the existence of domain III alone is not sufficient to allow the association of PACT with PKR, and that there is a requirement for either domain I or domain II in order to stabilize this interaction (28). Hence, it is worth exploring whether OV20.0 blocks PACT-mediated PKR activation by initially occupying PKR, which in turn will prevent the binding of PACT to PKR. By means of immunoprecipitation, the results show that an increase in OV20.0 is associated with a decreased association between PACT and PKR (Fig. 7, lanes 3 to 5). These results suggest the possibility that binding of OV20.0 interferes with the interaction between PACT and PKR, which in turn leads to a repression of PACT-induced PKR activation.

dsRNA facilitates the association of OV20.0 with PACT but not with PKR.

OV20.0 protein interacts with PKR and its activators PACT and dsRNA. Coincidently, the three proteins all have dsRNA binding ability; therefore, whether dsRNA participates in these associations needs further investigation. To define the role of dsRNA, RNase V1 (Life Technologies Corporation), an RNase that specifically targets dsRNA, was included during the immunoprecipitation reaction. As indicated in Fig. 8A, RNase V1 digestion did not influence the binding of endogenous PKR with OV20.0 in transient expression cells (Fig. 8A, upper, lanes 2 and 3). Nevertheless, the association between OV20.0 and PACT was weakened under RNase V1 treatment (Fig. 8A, middle, lanes 2 and 3).

This finding was further confirmed in cells infected with recombinant ORFV (ORFV-GFP) that also were expressing FLAG-tagged OV20.0 (8). In infected goat fibroblast cells, both PACT and PKR are able to be coprecipitated with viral OV20.0 protein (Fig. 8B, upper and middle). Consistent with this finding, RNase V1 treatment significantly reduced the association between OV20.0 and PACT, while the binding of PKR remained unaffected (Fig. 8B, lanes 2 and 3). These results indicate that the presence of dsRNA increases the interaction and affinity between viral OV20.0 and cellular PACT.

Sequence alignment of the RBD regions of OV20.0, RKR, and PACT.

It is worth noting that sequence alignment of the RBDs of OV20.0 (residues 80 to 184), PACT (domain I and II), and PKR (RBD1 and RBD2) contain several consensus residues that are conserved (Fig. 9), especially residues located at the C terminus. These residues, F148, K167, R168, and K171, have been reported to be involved in the dsRNA binding capacity of these proteins (35, 36). Since members of the dsRNA binding protein family are likely to form homodimers with themselves or heterodimers with other members of the family via their mutual RBDs (37, 38), it is possible that OV20.0 would form heterodimers with PKE and PACT proteins.