African green monkey kidney (Vero) cells (ATCC CCL-81) were grown in Dulbecco’s Modified Eagle’s (DMEM; Sigma Aldrich) medium containing 10% heat inactivated fetal bovine serum (FBS) (Sera Care Life Sciences, Inc.). 293T cells (ATCC CRL-11268) and Madin-Darby Bovine Kidney (MDBK) cells (ATCC CCL22) were grown in minimum essential medium (MEM; Sigma Aldrich) containing 5% FBS. Wild type BAdV-3 was used to infect MDBK cells at an MOI of 5.

Anti-pVIIIa recognizes a protein of 24 kDa in BAdV-3 infected cells (Ayalew et al., 2014). Anti-DBP recognizes a protein of 48 kDa in BAdV-3 infected cells (Zhou et al., 2001). Anti-DDX3, anti-eIF4G, anti-eIF4E, anti-eIF3, anti-PABP and fluorescence conjugated goat anti-mouse IgG-FITC (Santa Cruz Biotechnology, Inc., USA), Cy-3 conjugated goat anti-rabbit antibody (Jackson Immuno Research), Alexa Flour 680 goat anti-rabbit IgG antibody (Molecular Probes) or IRDye 800 conjugated goat anti-mouse IgG (Li-COR biosciences) and anti-HA and anti-β-actin MAb (Sigma Aldrich) were purchased.

Plasmids were constructed (Supplementary File) using standard procedures.

The recombinant GST-fusion proteins were expressed in Escherichia coli BL21 cells as described (Kulshreshtha and Tikoo, 2008). Glutathione S-transferase (GST), GST.pVIII, and GST.DDX3 fusion proteins were purified using Glutathione sepharose beads (GE Healthcare) as per the instructions of the manufacturers. The purified proteins were dialyzed using Slide-A-Lyzer dialysis cassette (Thermo Scientific). The concentrations of the proteins were measured by Bradford assay (Bio Rad) using Ultrospec^® 3000 spectrophotometer (Pharmacia Biotech).

Individual plasmid DNAs (0.8 μg) were used to synthesize [³⁵S] (Perkin Elmer) labeled human adenovirus (HAdV)-5 pVIII (pC.pVIIIhav) and porcine adenovirus (PAdV)-3 pVIII (pC.pVIIIpav) proteins or unlabeled DDX3 (pcHA.DX3) in vitro by utilizing a TNT T7 Coupled Reticulocyte Lysate System (Promega). Equal amounts of proteins (unlabeled DDX3 and labeled HAdV-5 pVIII or unlabeled DDX3 and labeled PAdV-3 pVIII) were mixed together and incubated for 4–6 h at 4°C. The individual mixture was immuno-precipitated either with anti-DDX3 serum or rabbit pre-immune sera coupled to protein A sepharose beads. The bound proteins were separated by 10% sodium dodecyl sulphate (SDS)-polyacrylamide gel electrophoresis (PAGE), fixed in de-stain solution for 30 min and dried. Subsequently, the gel was exposed to a phosphor screen (Kodak) and visualized on a Molecular Imager FX using Quantity One software (Bio-Rad).

Plasmid pHA.DX3 DNA (0.8 μg) was used to synthesize radio-labeled DDX3 protein in vitro by utilizing a TNT T7 Coupled Reticulocyte Lysate System (Promega) in the presence of 30 μCi of [³⁵S]-methionine (Perkin Elmer). Prior to use, glutathione sepharose (GST) beads were washed three times with GST binding buffer (0.54 M NaCl, 2.7 mM KCl, 10.15 mM Na2HPO4, 1.75 mM KH2PO4, 0.01 M MgCl2, 1 μg/ml each Aprotinin and Leupeptin, 1% Triton X-100, 1 mM PMSF and DNAse I). Purified GST, GST.pVIII, or GST.DDX3 fusion protein (15 μg each) was incubated individually with 20 μl of glutathione sepharose beads (GE Health Care) plus 10 μl of in vitro synthesized indicated proteins at +4°C on a nutator. After overnight incubation, the beads were washed three times, 10 min each with 0.1 M PBS. The bound proteins were analyzed as described above.

Madin-Darby Bovine Kidney cells were infected with BAdV-3 at an MOI of 5. Vero cells (6 × 10⁵/well) in one well of six well plate were co-transfected with 2 μg of each plasmid (pHA.DX3 and pEY.pVIII or pHA.DX3 and pEYFPN1) DNA. At 36 h post infection or 48 h post transfection, the cells were lysed, immunoprecipitated with indicated protein specific antibodies, separated by 10% SDS-PAGE, transferred to PVDF membrane and probed by Western blot using indicated protein specific antibodies as described (Poulin et al., 1998). Similarly, DDX3 knockdown 293T cells were transfected with 2 μg of pHA.DX3 or pHA.pVIII or pHA.DX3 and pc.pVIII. 48-h post transfection, the cells were lysed and precipitated using anti-HA affinity matrix (Sigma) and Western blot was performed as indicated above.

Monolayers (60–70% confluent) of Vero cells (5 × 10⁴/well) in two well chamber slides were incubated with Opti-MEM I reduced serum media (Gibco). After 2 h of incubation, the cells were transfected with indicated plasmid DNA(s) (1–2 μg/plasmid DNA/well) using Lipofectamine 2000 (Invitrogen). After 36 h post transfection, the cells were fixed with 3.7% paraformaldehyde, permeabilized with 0.05% triton X-100 and blocked with 2% goat serum. Subsequently, the cells were incubated with protein specific antibody in 0.1 M PBS for 1 h at room temperature, followed by specific secondary antibody for 1 h. Finally, the slides were mounted with mounting medium containing DAPI (Vectashield) and visualized under Zeiss LSM 5 laser scanning con-focal microscope.

Monolayers (90% confluent) of MDBK cells (1 × 10⁵cells/well) grown on two well chamber slides were mock infected or infected with wild type BAdV-3 at an MOI of 5. At 24 h post infection, the cells were fixed with 3.7% paraformaldehyde, immunostained and visualized as described above.

Firefly luciferase (FLuc) mRNAs were synthesized in vitro in the absence (uncapped) or presence (capped) of 40 mM Ribo m7GpppG cap analog (Promega) using T7-RiboMax RNA production system. Ten microliters of prewashed GST-beads were loaded with 750 ng of either GST.pVIII or GST proteins and incubated for 2 h at +4°C with 15 μl of rabbit reticulocyte lysate. Subsequently, the mixtures were centrifuged and the supernatants were used for the translation of the synthesized capped and uncapped mRNAs. Finally, luciferase assay (Promega) was performed in a luminometer as per the company’s procedure.

To examine the effect of pVIII on the translation of IRES or Cap dependent translation, a bicistronic luciferase assay was performed in 293T cells using a plasmid DNA expressing a single transcriptional unit in which translation of renilla luciferase is cap-dependent while translation of FLuc is cap-independent (IRES dependent). The cells grown in six well plates were co transfected with 2 μg/well of a bicistronic reporter plasmid pcDNA3-RLuc-POLIRES-FLuc and 4 μg/well of either plasmid pEY.pVIII DNA or plasmid pEYFP N1 DNA. At 48 h post transfection, FLuc and Renilla reniformis luciferase (RLuc) activities were measured in a luminometer by using a dual luciferase assay kit (Promega) as per the company’s procedure. Expression of EYFP was used to normalize the transfection efficiency.

The monolayers of MDBK cells in six well plates were mock infected or infected with BAdV-3 at an MOI of 5 or monolayers of Vero cells grown in six well plates were transfected with indicated concentrations of plasmid pEY.pVIII or pEYFPN1. At different times post transfection, the cells were starved in media without methionine for 2 h before labeling with 100 μCi of [³⁵S] methionine for 10 min. The radiolabelled cells were collected and lysed with RIPA buffer containing protease inhibitor cocktail. Proteins from radiolabelled cell lysates were separated by 10% SDS-PAGE. The gel was fixed, exposed to a phosphor screen (Kodak) overnight and visualized on a Molecular Imager FX using Quantity One software (Bio-Rad).

To determine whether the reduction in the level of the translation of capped luciferase mRNA in the presence of pVIII is due to the reduction in the level of eIFs, initially 10 μl of Flexi rabbit reticulocyte lysate was incubated with 10 μl of GST bead containing 750 ng of either GST or GST.pVIII fusion protein at +4°C and centrifuged for 10 min. Finally, the supernatant and pellet were collected and analyzed by Western blot using anti-DDX3 or anti-eIF specific antibodies, followed by Alexa Flour 680 goat anti-rabbit IgG (Molecular Probes) or IRDye 800 conjugated goat anti-mouse IgG (Li-COR biosciences) as secondary antibody. The membranes were scanned using Odyssey LiCOR infrared scanning system and the intensity of the bands measured by Odyssey Software v2.1.

Similarly, 50 μl of the cytoplasmic fraction of cells were incubated with 10 μl of GST beads (loaded with 750 ng of either GST or GST.pVIII fusion protein) at +4°C and centrifuged for 10 min. Finally, the samples were analyzed by Western blot as described above.

The indicated cells grown on six well plates were infected with five MOI of BAdV-3. 293T cells grown on six well plates were transfected either with 2 μg/well of plasmid pEYFPN1 or pEY.pVIII DNAs. At 36 h post infection or transfection, the cells were lysed with RIPA buffer containing protease inhibitor cocktail and centrifuged at 10,000 rpm. The supernatants were incubated with 50 μl of 7-methyl GTP sepharose beads on a nutator overnight at +4°C. The beads were then washed three times 5 min each and boiled with protein loading dye with 10% beta mercaptoethanol. Proteins were separated by 10%SDS-PAGE gel, transferred to nitrocellulose and probed in Western blot using anti-DDX3 or anti-eIF specific antibody followed by Alexa Flour 680 goat anti-rabbit IgG (Molecular Probes) or IRDye 800 conjugated goat anti-mouse IgG (Li-COR biosciences) as secondary antibody. The expression of pVIII in BAdV-3 infected cells or EY.pVIII in transfected cells was checked by Western blot using anti-pVIII serum.

Confluent MDBK cells grown in six well plates were starved with media without phosphate for 1 h before radiolabelling with 30 μci of [³²P] UTP. After labeling for 1 h, the cytoplasmic RNA was extracted using SurePrep^TM nuclear or cytoplasmic RNA purification kit (Fisher Bioreagents). Finally, poly A⁺ RNA was purified using Oligotex^® mRNA mini kit (Qiagen). The purity of the RNA was further checked by digestion with either RNase or DNase enzymes followed by gel electrophoresis. Subsequently, RNA EMSA was carried out using a modified protocol by Yakhnin et al. (2012). Radiolabelled (10,000 cpm) RNA probe of oligo dT purified polyA⁺ RNA was incubated with 500 ng of GST alone, or GST fusion proteins GST.pVIII or GST.100K (BAdV-3 100K; used as positive control) at room temperature. After incubation for 30 min, the reaction mixture was separated by 4% acrylamide native gel for 4–6 h at 150 V before the gel was dried, exposed onto a phosphor screen (Kodak) and scanned on a Molecular Imager FX using Quantity One software (Bio-Rad).

Initially, proteomic analysis of the nucleoli of BAdV-3 infected and uninfected MDBK cells identified a number of nucleolar proteins, which appeared to be involved in BAdV-3 infection (Patel and Tikoo, unpublished data). Since one of these proteins identified as DDX3 appeared to be important for BAdV-3 replication, we initially determined if DDX3 interacted with any BAdV-3 protein using matchmaker GAL4 Yeast two hybrid assays (Supplementary file). Our results suggested that BAdV-3 pVIII interacted with DDX3 (Supplementary Figure 1). The interactions were confirmed by GST-pull down assay (Figure 1A) and co immunoprecipitation assay using transfected cells (Figures 1B,C). Moreover, using co-immunoprecipitation coupled with Western blot analysis, we determined the interaction of DDX3 with pVIII in BAdV-3 infected MDBK cells. As seen in Figure 1, a 73 kDa DDX3 protein could be co-immunoprecipitated from the lysates of BAdV-3 infected MDBK cells using anti-pVIII serum and probed in Western blot with anti-DDX3 MAb (panel D).

Finally, confocal microscopy analysis revealed that pVIII co-localized with DDX3 around the perinuclear region and in the cytoplasm of the transfected cells (Figure 1G, panels h1–h4) or BAdV-3 infected cells (Figure 1G, panels g1–g4). The perinuclear localization may probably correspond to rough endoplasmic reticulum, the site of mRNA translation. However, DDX3 localized diffusely in the cytoplasm in uninfected MDBK or untransfected Vero cells (Figure 1E, panels a and b). The DDX3 also localized diffusely in the cytoplasm of transfected cells (Figure 1E, panel c). Additionally, pVIII localized both in the cytoplasm and in the nucleus in BAdV-3 infected (Figure 1F, panel d) or pEY.VIII transfected (Figure 1F, panel e) cells. No pVIII was detected in uninfected MDBK cells (Figure 1E, panel f).

To verify if interaction of DDX3 with pVIII is conserved among members of Mastadenovirus genus, GST-pull down assay was performed using purified GST or GST.DDX3 fusion protein (Figure 2A). As seen in Figure 2B, radiolabelled HAdV-5 pVIII protein interacted with GST-DDX3 fusion protein but not with GST. Similarly, radiolabelled PAdV-3 pVIII protein (lanes c, e, and f) interacted with GST-DDX3 fusion protein but not with GST protein alone.

Next, we performed co-immunoprecipitation assay to corroborate the results of GST pull down assay. As seen in Figure 2C, radiolabelled HAdV-5 pVIII co-immunoprecipitated with unlabeled DDX3 using anti-DDX3 serum. Similarly, radiolabelled PAdV-3 pVIII co-immunoprecipitated with unlabeled DDX3 using anti-DDX3 serum. No such proteins could be co-immunoprecipitated using pre immune sera.

To determine if BAdV-3 infection inhibits the synthesis of cellular proteins at late times post infection, mock or BAdV-3 infected MDBK cells were pulsed for short duration with [³⁵S] methionine at different times post infection. Proteins from the lysates of the labeled cells were separated by 10% SDS-PAGE and analyzed by autoradiography. As seen in Figure 3A, compared to mock infected MDBK cells, reduction in cellular protein synthesis is apparent in BAdV-3 infected cells at 18 h post infection, which coincides with the detection of pVIII (a late viral protein) in BAdV-3 infected cells. By 36–48 h post infection, there is significant decrease in cellular protein synthesis, which coincides with steady expression of pVIII protein (Figure 3A).

To determine if expression of pVIII alters the cellular protein synthesis, Vero cells were transfected with indicated plasmid DNA, labeled with [³⁵S] methionine, and analyzed by SDS-PAGE. As seen in Figure 3B, there was noticeable reduction in the synthesis of cellular protein in cells transfected with plasmid pEY.pVIII (express EYFP-pVIII fusion protein; Supplementary file) DNA than in cells transfected with plasmid pEYFP-N1 (express EYFP; Supplementary file) DNA at 24–48 h post transfection. Moreover, the reduction in cellular protein synthesis appeared to depend on the amount of plasmid DNA used for transfection (Figure 3C).

To determine the effect of pVIII on mRNA translation, capped or uncapped FLuc mRNAs were synthesized (Figure 4A) and translated in the presence of purified GST or fusion protein; GST.pVIII. (Supplementary file). As seen in Figure 4B, GST.pVIII protein significantly inhibited the in vitro translation of capped, but not of uncapped luciferase mRNA. No such effect could be observed in the presence of GST on the translation of capped or uncapped luciferase mRNAs. No significant difference could be observed using different concentration of GST.pVIII fusion protein (Supplementary Figure 2).

To examine the effect of BAdV-3 pVIII on the translation of capped and uncapped mRNA translation in vivo, a dual luciferase assay was performed. As shown in Figure 4B, in the presence of EY.pVIII (Supplementary file), the cap dependent translation of renilla luciferase mRNA (capped) was significantly reduced compared to IRES dependent translation of FLuc mRNA (uncapped) in 293T cells. In contrast, there was no significant difference in the translation of renilla luciferase (cap-dependent) mRNA or FLuc (IRES dependent) in 293T cells in the presence of EYFP protein (Figure 4B). Moreover, cytoplasmic RLuc-POLIRES-FLuc mRNA levels appeared similar in EY.pVIII or EYFP expressing plasmid transfected 293T cells (Figure 4B).

To determine if the interaction of pVIII and DDX3 also affects the level of eIFs, the pellet and supernatant fractions of the rabbit-reticulo Lysates incubated with glutathione sepharose beads preloaded with purified GST.pVIII or GST were analyzed by Western blot using protein specific antibodies. As seen in Figure 5A, addition of GST.pVIII in reticulo lysate leads to the detection of DDX3, eIF3β, eIF4G, and PABP in the pellet fraction. No detection of DDX3, eIF3, eIF4G, or PABP could be observed in the pellet fraction when purified GST alone was added to the reticulo Lysate (Figure 5A). Similarly, addition of GST.pVIII in cytoplasmic fraction of MDBK cells leads to the detection of DDX3, eIF3β, eIF4G, and PABP in pellet fraction (Figure 5B). No detection of DDX3, eIF3β, eIF4G, and PABP could be observed in the pellet fraction when purified GST alone was added to cytoplasmic fraction of MDBK cells (Figure 5B). Thus, we conclude that pVIII forms immunoprecipitable complex with eIFs.

To determine the effect of BAdV-3 on the association of eIFs with 5′-Cap of mRNAs, a Cap (m5GpppG) binding assay was performed using cleared lysates of BAdV-3 infected cells. The captured components of cap-binding complex were identified by Western blot using protein specific antibodies. As seen in Figure 6A, no significant difference could be observed in the binding of eIF4G and eIF4A to 5′-Cap (m7GpppG) in mock or BAdV-3 infected cells. However, significant decrease in the binding of DDX3, eIF3, eIF4E, and PABP proteins to 5′-Cap (m7GpppG) was observed in BAdV-3 infected cells compared to mock infected cells. However, pVIII was not retained in the m7GTP resins. To confirm for virus infection of cells, Western blot was performed on BAdV-3 infected cell lysates using anti-pVIII serum which detected the expression of pVIII.

Similarly, no significant difference could be observed in the binding of eIF4A to 5′-Cap (m7GpppG) in the cells (Figure 6A) transfected with plasmid pEY.pVIII or plasmid pEYFP.NI (Supplementary file) DNA. However, there was significant decrease in the binding of DDX3, eIF3, eIF4E, eIF4G, and PABP proteins to 5′-Cap (m7GpppG) in the cells transfected with plasmid pEY.pVIII DNA than the cells transfected with plasmid pEYFP.N1 (Supplementary file) DNA. Interestingly, pVIII was not retained in the 7-methyl GTP resins. Moreover, anti-pVIII serum detected the expression of EY.pVIII fusion protein in the transfected cells (Figure 6A). The quantified intensity of the protein bands were plotted to show the relative amount of each protein retained in the m7GTP-Sepharose beads as compared to the total amount. In addition, to determine if BAdV-3 infection degrades eIFs, VIDO R2 cells were mock infected or infected with BAdV-3 and analyzed by Western blot 36 h post infection. As seen in Figure 6B, significant level of PABP, eIF3β, eIF4G, DDX3, and eIF4E could be detected in BAdV-3 infected cells. β-actin was used as a loading control.

To determine if pVIII interacts directly with eIFs, the cytoplasmic fractions of VIDO R2 cells (Reddy et al., 1999) and DDX3 knocked down VIDO R2 cells (Ayalew and Tikoo, in preparation) were collected, incubated with purified GST or GST.VIII, centrifuged to collect supernatant and pellet fractions and analyzed by Western blot. As seen in Figure 7A, addition of GST.pVIII lead to the detection of DDX3, eIF3β, eIF4G, and PABP in the pellet of DDX3 positive VIDO R2 cell cytoplasmic fractions. No such detection of DDX3, eIF3β, eIF4G, and PABP could be observed in the pellet fraction when purified GST.VIII was added to the cytoplasmic fraction of DDX3 knocked down VIDO R2 cells. Similar results were observed when HA.pVIII, HA.DDX3 or HA.DDX3 plus pVIII expressing plasmid transfected DDX3 knocked down 293T cells were immunoprecipitated using anti-HA affinity matrix. The expression of HA.pVIII did not co-immunoprecipitate the eIFs in DDX3 knockdown 293T cells. In contrast, expression of HA.DDX3 or HA.DDX3 and pVIII co-immunoprecipitated eIFs in DDX3 knockdown 293T cells (Figure 7B).

To determine if the silencing of DDX3 in BAdV-3 permissive VIDO R2 cells affects the cellular protein synthesis, [³⁵S] methionine pulse labeling experiment was performed using mock or BAdV-3 infected DDX3 +Ve VIDO R2 (DDX3 positive) or DDX3kd VIDO R2 (DDX3 knockdown) (Figure 8A) cells. As seen in Figures 8B,C, there was significant difference in the synthesis of cellular proteins between mock infected and BAdV-3 infected VIDO R2s cells at 36–48 h post infection. In contrast, there was less inhibition of synthesis of cellular proteins in DDX3 kd VIDO R2 cells as compared to DDX3 +Ve VIDOR2 cells. This further strengthens the critical importance of DDX3 in suppression of cellular protein synthesis in BAdV-3 infected cells.

To assess the possibility of the interaction of pVIII with RNA, ³²P (UTP) labeled Poly A⁺ containing mRNA was purified from the cytoplasmic fraction of MDBK cells and RNA electrophoretic mobility shift assay was performed using purified GST alone or fusion protein GST.pVIII. Since adenovirus 100K protein interacts with mRNAs (Cuesta et al., 2004), RNA electrophoretic mobility shift assay was also performed using GST.100K as a positive control. As seen in Supplementary Figure 3, GST.pVIII did not interact with the cellular poly A RNA. Similarly, no interaction was observed between the cellular polyA⁺ RNA and GST. As expected (Cuesta et al., 2004), GST.100K interacts with cellular polyA mRNA.

To examine whether cytoplasmic mRNA levels are affected or their stability compromised, cDNAs were synthesized from cytoplasmic mRNAs purified from mock infected cells or BAdV-3 infected cells collected at different time points post infection. Real time PCR (Supplementary file) was performed using primers targeting different species of bovine housekeeping genes (Lisowski et al., 2008). The results indicate that (Figure 9) as compared to mock infected cells, the level of cytoplasmic mRNAs is not altered in BAdV-3 infected cells even at 48 h post infection Moreover, the integrity of cellular mRNAs appears intact during BAdV-3 virus infection. The detection of expression of DNA binding protein in BAdV-3 infected cell lysates by Western blot using anti-DBP serum (Zhou et al., 2001) confirm virus infection.