Inhibition of protein kinase C promotes dengue virus replication

To investigate possible amino acids on NS5 that might be phosphorylated by human kinases, we analyzed amino acid sequence of dengue serotype 2 (DENV2) NS5 by three phospho-algorithms; Scansite™ [18], NetPhosK 2.0 [19], and KinasePhos 2.0 [20]. In accordance with the previous notion describing NS5 as phosphoprotein, the analyses revealed many potential candidate phosphorylation sites on NS5 for several kinases, when compared to other DENV proteins (Fig. 1a). Protein kinases predicted to phosphorylate NS5 by at least one algorithm included PKC, PKA, CDKs (CDCs), MAPK, ATM, CK II, PKB, and others (Fig. 1a). Our analyses also identified 2 kinases that were previously reported to phosphorylate NS5, i.e. Casein Kinase I (CK I), and protein kinase G (PKG) [10, 17, 21]. Among the candidates, we focused on possible functional interaction between NS5 and PKC (predicted by all 3 algorithms). We found that 4 potential phosphorylation sites, in particular Thr244, Thr302, Ser796, and Ser885, were recognized by at least 2 predicting algorithms on DENV2 NS5. Among the 4 candidates, Thr302 and Ser796 are conserved in all vector-borne flaviviruses (Fig. 1b). Thr244, and Ser885 were not in the other mosquito-borne or tick-borne viruses, but were rather limited within the DENVs (Fig. 1b). While Thr244 resides in the methyltransferase domain of NS5, Thr302 Ser796 and Ser885 reside in the RdRp domain of NS5. These 4 candidate sites are exposed to solvent, thus, are potentially be phosphorylated (Fig. 1b, c).

Since RdRp domain contains several of the possible phosphorylation sites for PKC, we set up a PKC in vitro kinase assay using purified recombinant DENV2 RdRp domain, containing Thr302 Ser796 and Ser885, as a substrate. The kinase assay showed that PKC efficiently phosphorylated RdRp in vitro. The signal was approximately 9 folds higher than that of GST (used as non-relevant control). Under these conditions, PKC kinase assay gave a very low signals for the other DENV proteins, such as capsid, prM, E, NS1, NS2B, NS3, NS4B (Fig. 1d). Interestingly, protein kinase C appeared to produce relatively high phosphorylation signal compared with the other potential NS5 kinases, namely PKA, CK I, and CDK4/cyclinD1 (Fig. 1e). Thus, PKC can efficiently phosphorylated the NS5 RdRp domain in vitro. Of note, although the methyltransferase domain contains only one possible PKC phosphorylation site, Thr244, which is not conserved in all of the vector-borne flaviviruses, it is entirely possible that this site may be phosphorylated by PKC. We have yet to test this possibility.

To investigate the functional interaction between host PKC and DENV NS5 in vivo, we depleted PKCα using PKCα-specific siRNAs and examined the level of NS5 phosphorylation in vivo. We first, tested the efficiency of siRNA-mediated PKC depletion. We found that at 24, 48, and 72 h after PKC-specific siRNA transfection, more than 70 % of the endogenous PKC levels were downregulated (Fig. 2a), while host cell viability was unaffected (data not shown). We selected to examine NS5 phosphorylation at the 24 h after siRNA transfection. We expressed His-FLAG-tagged NS5 in host cells by infecting the cells with recombinant DENV expressing the His-FLAG-NS5 protein (etDENV), then pulled down the NS5 using anti-FLAG antibody. The phosphorylation on the His-FLAG-NS5 was analyzed by anti phospho-Ser and phospho-Thr immunoblottings. We found that under this condition, silencing of the PKC clearly decreased Serine and Threonine phosphorylations of NS5 (Fig. 2b, c).

To confirm the effect of PKC depletion on NS5 phosphorylation, we inhibited PKC activity of host cells using bisindolylmaleimide I (BisI), a specific chemical inhibitor of conventional and novel PKCs. We found that various concentrations of BisI, (up to 1 μM of BisI) were not toxic to the cells for at least up to 72 h (Fig. 3a). As early as 24 h after treatment, 0.1 and 1 μM of BisI clearly suppressed the activity of PKC, as indicated by a complete loss of the phosphorylation of PEA15, an established PKC substrate (Fig. 3b, c). Importantly, FLAG immunoprecipitation of the tagged-NS5 followed by phospho-Ser or phospho-Thr immunoblotings showed that Serine and Threonine phosphorylations were decreased, when the cells were treated with non-toxic dose (1 μM) of BisI (Fig. 3d, e).

To examine the biological effects of PKC blockage on the viral replication, we quantitated the viral copy number in the host cells pre-treated with BisI using RT-qPCR. BisI, which blocked the PKC activities, significantly increased the viral copy of DENV2 strain 16681 in human hepatocellular carcinoma HepG2 cells, with the highest copy number of the virus attained by the 0.1 and 1 μM BisI treatments (Fig. 4a). On the contrary, we detected a very lowering of dengue virus copy number in HepG2 cells pretreated with 0.1 nM of a PKC activator Phorbol 12-myristate 13-acetate (PMA) (Fig. 4b). We found that, at this dose of PMA, activity of PCK was induced indicated by an upregulation of phospho-PEA15 (Fig. 4c).

To further investigate whether the PKC inhibitor altered DENV entry into the host cells, we assessed the viral copy number in the cells at early time points, i.e. at 2, 5 and 8 h after viral infection. During these time points, the viral replication was not yet fully operated. Therefore, the viral copy number would represent the number of virus entering the host cells after the infections [22]. We found that the amounts of virus in BisI-treated cells were comparable to those in the vehicle-treated cells, during these early time points (Fig. 4d). Therefore, it is likely that the activity of PKC was not involved with DENV entry into the host cells, and inhibition of PKC activity does not interfere with this process.

In contrast to the results from early time points, viral copy numbers at the later time points were augmented by at least 10 folds (after 12 h of BisI-treatment), when compared to that of the vehicle-treated cells. The viral copy numbers continued to be higher at every time point examined until 72 h post infection (Fig. 4e).

These results indicated that inhibition of the PKC activity promotes viral replication, showing in an increase of the viral copy number, whereas stimulation of the PKC activity may be involved in suppression of the DENV replication in the host cells.

Promotion of the viral replication by PKC blockage was also confirmed using siRNA-mediated silencing of PKCα. This also resulted in a significant increase of dengue viral copy number and the amounts of viral NS5 and NS3 proteins in host cells (Fig. 5a, b). Although not statistically significant, we detected an increase of secreted virus into the culture medium of BisI-treated cells as well (Fig. 5c).

The above data showed that depletion of PKC significantly enhanced the dengue viral replication in the host cells. To further assess whether PKC depletion also had an effect on the host cells, we monitored growth of the infected HepG2 cells in the presence or absence of PKC depletion. To this end, HepG2 cells, which either pretreated with non-targeting (control) siRNA or PKC-specific siRNA, were infected with DENV2 16681. We found that, viability of the non-targeting siRNA-treated control cells without viral infection consistently increased over 3 days (Fig. 5d, dashed black line). Infection of dengue virus, or PKC siRNA transfection alone, significantly reduced HepG2 cell viability by 72 h after the PKC knockdown (Fig. 5d, dashed red line and solid black line, respectively). Importantly, depletion of PKC by siRNA together with DENV-infection accelerated suppression of HepG2 cell viability to 48 h after transfection. The loss of host cells became more clearly evident at the 72 h time point (Fig. 5d, solid red line). Inhibition of PKC by BisI also caused a marked reduction of host cell viability at 48 h after DENV infection (Fig. 5e). Altogether, these data indicated that inhibition of PKC led to an enhancement of the DENV growth and suppression of host cell viability.

The His-tagged NS5 RdRp protein was purified from bacteria expression following Kamkaew et al. [32]. Dengue Capsid, prM, E, NS1, NS2B, NS3 and NS4B coding DNA were amplified by PCR from pD2-IC plasmid containing DENV2 genome [33]. The PCR products were inserted into pGEX-5X-3 plasmid (GE Healthcare Bio-Sciences). E.coli BL21 (DE3) was used to express dengue proteins according to standard protocol recommended by the manufacturer. The assay use 1 μg DENV proteins as substrate in a reaction with 75 ng of purified human PKC (Sigma). Cdk4/cyclinD1, PKA, CK I kinases were purchased from Calbiochem, Sigma and Promega, respectively. The kinase buffers and the reactions were prepared and performed as recommended by the manufacturer’s protocols. After the kinase reactions, the substrates were run on a 12 % SDS-PAGE gel. Expected bands of substrates were visualized by Coomassie blue staining and were excised from the gel to measure the ^γ-32P incorporation. The ^γ-32P was detected by scintillation counter (Perkin Elmer).

The human liver hepatocellular carcinoma cell line HepG2 was a kind gift from Dr. Limjindaporn T, Faculty of Medicine, Siriraj Hospital, Mahidol University. The Baby hamster kidney cell line (BHK-21) was a kind gift from Dr. Leardkamolkarn V, Faculty of Sciences, Mahidol University. They were cultured in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10 % heat-inactivated fetal bovine serum (FBS), 100 U/ml penicillin G and 100 μg/ml streptomycin (Invitrogen). Cells were grown at 37 °C with 5 % CO₂ in a humidified incubator.

Construction of the recombinant DENV harboring the His-FLAG-tagged NS5 DENV (etDENV) was described previously [34]. Briefly, the cDNA from DENV serotype 2 strain 16681 in the pD2-IC plasmid [33] was used as the template for the insertion of the epitope tag. The TAP tag comprised of an octahistidine (His8) tag, a tri-glycine spacer, and FLAG sequence (DYKDDDDK) was inserted into the NS5 gene in pD2-IC via a standard QuikChange procedure. The positive clone containing the tag was verified by DNA sequencing. Wild-type and tagged DENV RNAs were synthesized according to a previous report [33]. Assay for viral kinetics was performed as follows: BHK-21 cells were infected with the viruses, viral titers from day 1 to day 6 post infection were determined using foci formation assay (FFA). We found that the recombinant DENV etDENV produced a comparable kinetic to that of wild type DENV 16681 (Additional file 1: Figure S1) up to 6 days. Once prepared, the viral titration was performed as follows. Supernatant of transfected infected cells was collected for FFA at 7 days post transfection. For FFA, the supernatant was incubated with BHK-21 cells in 96 well plates for 2 h, and subsequently 200 μl of 1.5 % carboxymethylcellulose (CMC) and 2 % FBS in DMEM was added to cover the cells, and the cells were incubated at 37 °C for 3 days. The infected BHK-21 cells were then washed with PBS 3 times and fixed with 3.7 % formaldehyde for 10 min at room temperature. The cells were permeabilized by 1 % Triton X-100 in PBS for 10 min and washed with PBS 5 times. The cells then were incubated with 1:1,000 in PBS anti-DENV E protein (EMD Milipore). Thereafter, the cells were washed with PBS 5 times before incubation with goat anti-mouse conjugated horseradish peroxidase (HRP) antibody at a 1:1,000 dilution for 45 min at 37 °C. The foci were visualized through SigmaFAST™ DAP (Sigma) and counted under microscopy to determine viral titter.

In ever experiment, wild type DENV2 strain 16681 [35] or His-FLAG-tagged DENV was added into host cells at a multiplicity of infection (MOI) of 1, with an exception of the experiment with PMA treatment, in which virus with MOI of 5 was used. Cells were incubated at 37 °C for 90 min to allow incorporation of virus in to cells before being washed twice with PBS to remove excess virus. After that, DMEM medium supplemented with 2 % heat-inactivated FBS was added into the cells. Cells were further incubated at 37 °C for for time points indicated in the figures.

Bioinformatics tools used for screening of the possible PKC sites, Scansite™ 2 [18], NetPhosK 2.0 [19], and KinasePhos 2.0 [20], were used for predicting phosphorylation sites in Dengue NS5 protein. A high stringency level was set for Scansite™, whereas a 0.9 threshold (high) was used for NetphosK, and a 100 % specificity was used for KinasePhos.

To identify the conservation among flaviviruses, DENV1 (P27909), DENV2 (AHG23127), DENV3 (P27915), DENV4(P09816), West Nile virus (P06935), Japanese Encephalitis virus (P27395), yellow fever virus (AAC35930), St. Louis encephalitis virus (P09732), Omsk hemorrhagic fever virus (AAR98531), Zika virus (W8R1T1), Gadgets Gully virus (A0EKU2), Kama virus (W5VRZ2) and tick-borne encephalitis virus (Q01299) NS5 sequences were compared by BioEdit 7.2.5. The phospho-conservative site was shown in crystal structure of NS5 (PDB ID: 4V0R).

To assess cytotoxicity of Bisindolylmaleimide I (BisI) Hydrochloride (Cell signaling) to HepG2 cells, the cells were seeded in 96-well plates at a density of 10⁴ cells/well in DMEM medium supplemented with 10 % heat-inactivated FBS for 24 h before BisI treatment. BisI at concentrations of 0.001 μM, 0.01 μM, 0.1 μM, and 1 μM in DMEM medium supplemented with 10 % heat-inactivated FBS were added onto the cells. DMSO at a final concentration of 0.5 % served as a vehicle control. At 24, 48, and 72 h post-treatment, cell viabilities were assessed by a MTT assay, following the manufacturer protocol. To assess cytotoxicity of siRNA to HepG2 cells, the cells were seeded in a 96-well plate at a density of 10⁴ cells/well in DMEM medium supplemented with 10 % heat-inactivated FBS for 24 h before being transfected with PKCα siRNA at 24, 48, and 72 h post-transfection, MTT was added as described above.

HepG2 cells (5.5 X 10⁵ cells) were seeded into culture dishes for 24 h before being treated with various concentrations of BisI, i.e. 0.001 μM, 0.01 μM, 0.1 μM, and 1 μM. BisI was added into the cells 30 min before viral infection. BisI was kept constant in the medium during infection with DENV2 at MOI of 1. This was followed by an incubation for 90 min to allow internalization of virus into cells. After the excess virus was removed, and the cells were washed twice by PBS, BisI in DMEM supplemented with 2 % FBS was added into cells and further incubated for the desired times, before the cells were harvested for detection of intracellular viral RNA by RT-qPCR.

PKCα siRNA (Santa Cruz biotechnologies, SC-29449) or Silencer® Negative control siRNA (Thermo, AM4611) were transfected into HepG2 cells using Lipofectamine® RNAiMAX reagent (Invitrogen) following the manufacturer’s protocol. HepG2 cells (5.5 × 10⁵ cells) were seeded for 24 h before siRNA complexes were added. Twenty nM of siRNA were effective for suppressing PKCα expression. At 24 h post-transfection, DENV2 at MOI of 1 was added into the HepG2 cells and the culture was further incubated for 90 min to allow internalization of virus into the cells. Then, the excess virus was removed, and the cells were washed twice with PBS. Finally, DMEM supplement with 2 % FBS was added into the cells, and the cells were further incubated for another 24 h. Cells were collected for detection of intracellular viral RNA by RT-qPCR and viral proteins by immunoblots.

Total RNA was extracted from DENV-infected cells by Trizol reagent (Invitrogen), according to manufacturer’s protocol. The reverse transcriptase reaction was set up using iScriptTM select cDNA synthesis kit (Bio-Rad) for first-strand cDNA synthesis. One microgram of total RNA was used as a template for all RT reaction. The reactions were performed with random primers. qPCR reactions were carried out using 2X KAPPA SYBR Green (KAPPA Biosystem). Amplification and detection were performed using an Applied Biosystems 7500 Real-time PCR system (ABI, California). qPCR conditions were 95 °C for 3 min, followed by 40 cycles of 95 °C for 3 s and 60 °C for 1 min.

Primers are:

NS5-Forward primer 5’-TCCATACATGCTAAACATGA-3’

NS5-Reverse primer 5’-GGGATTTCCTCCCATGATTCC-3’

β-actin-Forward prime 5’-TCTTCCAGCCTTCCTTCCT-3’

β-actin-Reverse primer 5’-AGCACTGTGTTGGCGTACAG-3’

HepG2 cell lysates were suspended in RIPA buffer solution containing 50 mM Tris-base pH 7.4, 150 mM NaCl, 1 % triton X-100, 0.1 % SDS, 1 % sodium deoxycholate, 1 mM EDTA, 1 mM Na₃VO₄, 1 mM NaF, 1 mM β-glycerophosphate, 0.5 % NP40, 1 mM PMSF and 7× protease inhibitor cocktail. Protein samples were mixed with 6× protein loading dye, boiled and separated using SDS-PAGE. Gel was run for 2 h at 120 V before the proteins were transferred onto PVDF membrane by semi-dry blotting for 30 min at 10 V. After the transfer, nonspecific binding was blocked using 5 % skim milk for 1 h at room temperature. To determine inhibitory activity of PKC by BisI, after 24 post-infection, primary and secondary antibodies were rabbit antibodies against p-PEA15, PEA15, and goat anti rabbit antibody conjugated with horseradish-peroxidase (HRP) (Thermo scientific), respectively. To test whether PKCα siRNA can knock down PKCα after 24 and 48 h post-transfection, expressions of PKCα were detected by using mouse antibody specific to PKCα, and goat antibody specific to GAPDH (loading control) (both were from Santa cruz biotechnologies) as primary antibodies and goat anti-mouse (Novagen) and mouse anti-goat antibodies conjugated with HRP (Santa cruz biotechnologies) as secondary antibodies. To verify expression of NS5, rabbit antibody against NS5 (Thermo scientific) was used, followed by goat anti-rabbit antibody conjugated with HRP as secondary antibodies. The antigen-antibody complexes were detected by using chemiluminesence detection (ECL, Amersham Pharmacia Biotech).

In this test, we performed infection of the recombinant virus on BHK-21 cells, due to the relatively high expression level of viral proteins obtained from this cell line. BHK-21 cells were treated with 1 μM BisI before, at the time of, and after DENV infection as described above. At 24 h post-infection, cells were washed with cold PBS before extraction of total protein using CLB buffer (50 mM Tris HCl PH 7.4, 250 mM NaCl, 5 mM EDTA, 1 mM Na₃VO₄, 1 mM NaF, 1 mM β-glycerophosphate, 0.5 % NP40, 1 mM PMSF and 7X protease inhibitor cocktail). Three mg of cell lysate were used for FLAG-NS5 immunoprecipitation (IP) with 20 μl of anti-FLAG M2 affinity gel (Sigma) in 1 ml reaction according to the manufacturer’s instruction. IP products were separated by SDS-PAGE and transferred onto a PVDF membrane. The membrane was probed with rabbit anti-NS5 (Thermo scientific), mouse anti-phospho-Serine or anti-phospho-Threonine (Sigma) followed by incubation with goat anti-rabbit or goat anti-mouse IgG conjugated with HRP. The antigen-antibody complexes were detected by using chemiluminesence detection (ECL, Amersham Pharmacia Biotech).

All experiments were performed in triplicate and data are presented as the mean ± standard deviation (SD). Statistical comparisons were performed using Student’s t test. p-value ≤0.05 was considered to indicate a statistically significant difference.