Tubacin, an HDAC6 Selective Inhibitor, Reduces the Replication of the Japanese Encephalitis Virus via the Decrease of Viral RNA Synthesis

Japanese encephalitis virus (JEV), a neurotropic flavivirus, annually causes over 30,000 Japanese Encephalitis (JE) cases in East and Southeast Asia. Histone deacetylases (HDACs) modulate lysine acetylation of histones and non-histone proteins, regulating many processes including inflammation and antiviral immune response. This study investigated antiviral activity of pan- and selective-HDAC inhibitors as host-targeting agents against JEV. Among HDAC inhibitors, selective HDAC6 inhibitors (tubastatin-A (TBSA) and tubacin) concentration-dependently inhibited JEV-induced cytopathic effect and apoptosis, as well as reduced virus yield in human cerebellar medulloblastoma cells. The 50% inhibitory concentration (IC50) values of virus yield was 0.26 μM for tubacin and 1.75 μM for TBSA, respectively. Tubacin (IC50 of 1.52 μM), but not TBSA, meaningfully blocked the production of intracellular infectious virus particles. In time-of-addition assays, the greatest potency of antiviral activity was observed in the mode of pre-treatment with tubacin (IC50 of 1.89 μM) compared to simultaneous (IC50 of 4.88 μM) and post-treatment (IC50 of 2.05 μM) modes. Interestingly, tubacin induced the hyperacetylation of a HDAC6 substrate Hsp90 and reduced the interaction of Hsp90 with JEV NS5 protein. Novobiocin, an Hsp90 inhibitor, diminished the NS5 protein amount and virus replication in JEV-infected cells. Meantime, tubacin suppressed the NS5 expression and antisense RNA genome synthesis in infected cells. Tubacin-induced Hsp90 hyperacetylation was suggested to influence the NS5 activity in JEV replication. Therefore, tubacin had a high potential of a host-targeting agent against JEV, exhibiting preventive and therapeutic activities against JEV infection.

concentration-dependently reduced cytopathic effect of JEV infection at an MOI of 0.1 36 h post infection ( Figure 1). Meanwhile, apoptotic cell fraction of infected cells in the presence or absence of HDAC inhibitors was measured using propidium iodide flow cytometric assay ( Figure 2). JEV infection at an MOI of 0.1 caused the appearance of a sub-G1 (apoptotic) fraction. Tubacin and TBSA significantly diminished cell apoptosis post infection (Figure 2A,B). The inhibitory activity of tubacin and TBSA on JEV-induced cytopathic effect (CPE) was associated with the reduction of apoptosis of infected cells in a concentration-dependent manner. Moreover, supernatant virus yield determined using the plaque assay demonstrated that tubacin and TBSA, markedly inhibited JEV production in human cerebellar medulloblastoma cells ( Figure 3A,B). The 50% inhibitory concentration (IC50) values on virus yield were 0.26 µM for tubacin and 1.75 µM for TBSA, respectively. Notably, tubacin, but not TBSA, meaningfully blocked the intracellular production of infectious virus particles ( Figure 3C,D). The IC50 value of tubacin on intracellular infectious virus titers was 1.52 µM. Among these HDAC inhibitors, tubacin showed the highest antiviral potency against JEV.  (Figure 2A,B). The inhibitory activity of tubacin and TBSA on JEV-induced cytopathic effect (CPE) was associated with the reduction of apoptosis of infected cells in a concentrationdependent manner. Moreover, supernatant virus yield determined using the plaque assay demonstrated that tubacin and TBSA, markedly inhibited JEV production in human cerebellar medulloblastoma cells ( Figure 3A,B). The 50% inhibitory concentration (IC50) values on virus yield were 0.26 μM for tubacin and 1.75 μM for TBSA, respectively. Notably, tubacin, but not TBSA, meaningfully blocked the intracellular production of infectious virus particles ( Figure 3C,D). The IC50 value of tubacin on intracellular infectious virus titers was 1.52 μM. Among these HDAC inhibitors, tubacin showed the highest antiviral potency against JEV.    Virus yield in supernatant from infected cells treated with or without tubacin (A) and TBSA (B) was measured by plaque assay 36 h post infection. In intracellular virion production assay, the infected cells treated with or without tubacin (C) and TBSA (D) were lysed by three freeze-thaw cycles. The titer of intracellular infectious particles was determined by plaque assay. ** p value < 0.01; *** p value < 0.001 compared with untreated infected cells.  Virus yield in supernatant from infected cells treated with or without tubacin (A) and TBSA (B) was measured by plaque assay 36 h post infection. In intracellular virion production assay, the infected cells treated with or without tubacin (C) and TBSA (D) were lysed by three freeze-thaw cycles. The titer of intracellular infectious particles was determined by plaque assay. ** p value < 0.01; *** p value < 0.001 compared with untreated infected cells. Virus yield in supernatant from infected cells treated with or without tubacin (A) and TBSA (B) was measured by plaque assay 36 h post infection. In intracellular virion production assay, the infected cells treated with or without tubacin (C) and TBSA (D) were lysed by three freeze-thaw cycles. The titer of intracellular infectious particles was determined by plaque assay. ** p value < 0.01; *** p value < 0.001 compared with untreated infected cells.

Preventive and Therapeutic Activities of Tubacin against JEV Infection
To ascertain antiviral mechanism(s) of tubacin, the mode of inhibitory action by tubacin was examined using attachment inhibition and time-of-addition assays (Figures 4 and 5; Figures S2 and S3). In attachment inhibition assays, the TE671 cell monolayer was pre-incubated at 4 • C for 10 min, and then reacted with JEV SRIPs (50 TCID50) or virions (50 pfu) plus tubacin (0, 0.1, 5, 10, and 20 µM) at 4 • C for allowing attachment alone. After one hour of incubation, cell monolayer was washed with PBS; residual infectivity of SRIPs and virions was determined using immunofluorescence microscopy and plaque assay, respectively. Real-time fluorescence imaging of SRIP-infected cells indicated that the green fluorescence intensity of SRIP-driven EGFP reporter was very similar between tubacin-treated and mock-treated groups ( Figure 4). In addition, the plaque assay for residual infectivity of JEV virions indicated that tubacin had no significant inhibitory effect on residual infectivity compared to controls in the attachment assay ( Figure S2). The result of viral attachment assay indicated tubacin did not directly interfere on JEV attachment at early stage of viral replication. Antiviral mechanism(s) of tubacin against JEV was further evaluated using time-of-addition assays with JEV SRIPs and virions, including (1) pre-treatment (one hour prior to infection), (2) simultaneous treatment (at the same time as infection), and (3) post treatment (one hour post infection) ( Figure 5 and Figure S3). The greatest degree of antiviral activity was observed in the mode of pre-treatment with tubacin compared to simultaneous-and post-treatment modes. According to the green fluorescence intensity of SRIP-driven EGFP reporter, IC50 value of tubacin was 1.89 µM in a pre-treatment assay, 4.88 µM in a simultaneous-treatment test, and 2.05 µM in a post-treatment experiment, respectively ( Figure 5). Interestingly, post-treatment with tubacin was also very effective in inhibiting the late stage of JEV replication. Therefore, the results indicated that tubacin exhibited preventive and therapeutic activities against JEV infection, implying tubacin as a host-targeting agent to affect the involvement of cellular factors in JEV replication.

Preventive and Therapeutic Activities of Tubacin against JEV Infection
To ascertain antiviral mechanism(s) of tubacin, the mode of inhibitory action by tubacin was examined using attachment inhibition and time-of-addition assays (Figures 4 and 5; Figures S2 and  S3). In attachment inhibition assays, the TE671 cell monolayer was pre-incubated at 4 °C for 10 min, and then reacted with JEV SRIPs (50 TCID50) or virions (50 pfu) plus tubacin (0, 0.1, 5, 10, and 20 μM) at 4 °C for allowing attachment alone. After one hour of incubation, cell monolayer was washed with PBS; residual infectivity of SRIPs and virions was determined using immunofluorescence microscopy and plaque assay, respectively. Real-time fluorescence imaging of SRIP-infected cells indicated that the green fluorescence intensity of SRIP-driven EGFP reporter was very similar between tubacintreated and mock-treated groups ( Figure 4). In addition, the plaque assay for residual infectivity of JEV virions indicated that tubacin had no significant inhibitory effect on residual infectivity compared to controls in the attachment assay ( Figure S2). The result of viral attachment assay indicated tubacin did not directly interfere on JEV attachment at early stage of viral replication.
Antiviral mechanism(s) of tubacin against JEV was further evaluated using time-of-addition assays with JEV SRIPs and virions, including (1) pre-treatment (one hour prior to infection), (2) simultaneous treatment (at the same time as infection), and (3) post treatment (one hour post infection) ( Figure 5 and Figure S3). The greatest degree of antiviral activity was observed in the mode of pre-treatment with tubacin compared to simultaneous-and post-treatment modes. According to the green fluorescence intensity of SRIP-driven EGFP reporter, IC50 value of tubacin was 1.89 μM in a pretreatment assay, 4.88 μM in a simultaneous-treatment test, and 2.05 μM in a post-treatment experiment, respectively ( Figure 5). Interestingly, post-treatment with tubacin was also very effective in inhibiting the late stage of JEV replication. Therefore, the results indicated that tubacin exhibited preventive and therapeutic activities against JEV infection, implying tubacin as a host-targeting agent to affect the involvement of cellular factors in JEV replication.  of SRIP-driven EGFP reporter in infected cells was quantified using Image J, and then relative intensity was normalized by the total of cells (bottom). * p value < 0.05; ** p value < 0.01; *** p value < 0.001 compared with untreated infected cells. Scale bar = 50 μm.

Tubacin-Induced Hsp90 Hyperacetylation Was Associated with the Reduction of NS5 RNA Polymerase Activity
HDAC6, a cytoplasmic deacetylase, deacetylates several substrates, such as tubulin, Hsp90, βcatenin, and cortactin [24]. Since Hsp90 is suggested as the key chaperone universally required for the homeostasis of viral replication complexes, particular viral RNA-dependent RNA polymerase [25], the protein-protein interaction between host Hsp90 and JEV NS5 RNA polymerase was analyzed with coimmunoprecipitation assays ( Figure 6A). Western blotting analysis of co-immunoprecipitates with anti-Hsp90 antibodies demonstrated that JEV NS5 protein in infected cell lysate was co-immunoprecipitated by anti-Hsp90 antibodies ( Figure 6A, lane 4 vs. lane 2), revealing that the Hsp90-NS5 interaction might be involved in the JEV replication complex. Hsp90 hyperacetylation via inactivation of HDAC6 causes the loss of chaperone for its binding with Hsp90 client proteins [26,27]. Thus, the interaction of Hsp90 with JEV NS5 as well as Hsp90 acetylation levels in mock and infected cells treated with or without tubacin were also examined using immunoprecipitation assays ( Figure 6B). Western blotting analysis indicated that Hsp90 was immunoprecipitated in all samples, and Hsp90 hyperacetylation, with various amounts of acetyl group, was detected in tubacin-treated cells ( Figure 6B, lanes 3-6), but mock and infected cells ( Figure 6B, lanes 1 and 2). By contrast, the quantity of un-acetylated hsp90 protein decreased post tubacin-treatment ( Figure 6B, lanes 3-6). Importantly, the binding ability of Hsp90 to JEV NS5 concentration-dependently decreased post tubacin treatment ( Figure 6B, lanes 2-4). . Green fluorescence intensity of SRIP-driven EGFP reporter in infected cells was quantified using Image J, and then relative intensity was normalized by the total of cells (bottom). * p value < 0.05; ** p value < 0.01; *** p value < 0.001 compared with untreated infected cells. Scale bar = 50 µm.

Tubacin-Induced Hsp90 Hyperacetylation Was Associated with the Reduction of NS5 RNA Polymerase Activity
HDAC6, a cytoplasmic deacetylase, deacetylates several substrates, such as tubulin, Hsp90, β-catenin, and cortactin [24]. Since Hsp90 is suggested as the key chaperone universally required for the homeostasis of viral replication complexes, particular viral RNA-dependent RNA polymerase [25], the protein-protein interaction between host Hsp90 and JEV NS5 RNA polymerase was analyzed with co-immunoprecipitation assays ( Figure 6A). Western blotting analysis of co-immunoprecipitates with anti-Hsp90 antibodies demonstrated that JEV NS5 protein in infected cell lysate was co-immunoprecipitated by anti-Hsp90 antibodies ( Figure 6A, lane 4 vs. lane 2), revealing that the Hsp90-NS5 interaction might be involved in the JEV replication complex. Hsp90 hyperacetylation via inactivation of HDAC6 causes the loss of chaperone for its binding with Hsp90 client proteins [26,27]. Thus, the interaction of Hsp90 with JEV NS5 as well as Hsp90 acetylation levels in mock and infected cells treated with or without tubacin were also examined using immunoprecipitation assays ( Figure 6B). Western blotting analysis indicated that Hsp90 was immunoprecipitated in all samples, and Hsp90 hyperacetylation, with various amounts of acetyl group, was detected in tubacin-treated cells ( Figure 6B, lanes 3-6), but mock and infected cells ( Figure 6B, lanes 1 and 2). By contrast, the quantity of un-acetylated hsp90 protein decreased post tubacin-treatment ( Figure 6B, lanes 3-6). Importantly, the binding ability of Hsp90 to JEV NS5 concentration-dependently decreased post tubacin treatment ( Figure 6B, lanes 2-4). Novobiocin, an Hsp90 inhibitor, was used to evaluate the role of Hsp90 in the stability of JEV NS5 protein and virus yield ( Figure 6C,D). Novobiocin treatment caused a marked decrease in the amounts of NS5 protein in infected cells ( Figure 6C, lane 3 vs. lane 2), as well as reduced JEV yields in a concentration dependent manner ( Figure 6D). The results indicated that tubacin induced Hsp90 hyperacetylation and reduced the binding activity of Hsp90 to JEV NS5 in infected cells. The Hsp90 inhibitor, novobiocin, affected the amount of NS5 protein and the JEV production in infected cells. The finding implied that tubacin-induced Hsp90 hyperacetylation might alter the binding interaction of Hsp90 with JEV NS5 which is involved in JEV replication. Subsequently, the protein amount of NS5 and the synthesis of antisense RNA genome in JEV infected cells were determined 36 h post treatment with tubacin using immunofluorescent staining and real-time RT-PCR, respectively (Figure 7). Immunofluorescent staining with anti-JEV NS5 antibodies indicated that tubacin at 5 and 10 µM significantly decreased the protein amount of NS5 in JEV-infected cells ( Figure 7A). Meantime, tubacin concentration-dependently reduced NS5-mediated synthesis of JEV antisense RNA genomes in infected cells ( Figure 7B). The overall results illustrated that tubacin significantly repressed the functional activity of JEV NS5 in vitro replication. Novobiocin, an Hsp90 inhibitor, was used to evaluate the role of Hsp90 in the stability of JEV NS5 protein and virus yield ( Figure 6C,D). Novobiocin treatment caused a marked decrease in the amounts of NS5 protein in infected cells ( Figure 6C, lane 3 vs. lane 2), as well as reduced JEV yields in a concentration dependent manner ( Figure 6D). The results indicated that tubacin induced Hsp90 hyperacetylation and reduced the binding activity of Hsp90 to JEV NS5 in infected cells. The Hsp90 inhibitor, novobiocin, affected the amount of NS5 protein and the JEV production in infected cells. The finding implied that tubacin-induced Hsp90 hyperacetylation might alter the binding interaction of Hsp90 with JEV NS5 which is involved in JEV replication. Subsequently, the protein amount of NS5 and the synthesis of antisense RNA genome in JEV infected cells were determined 36 h post treatment with tubacin using immunofluorescent staining and real-time RT-PCR, respectively (Figure 7). Immunofluorescent staining with anti-JEV NS5 antibodies indicated that tubacin at 5 and 10 μM significantly decreased the protein amount of NS5 in JEV-infected cells ( Figure 7A). Meantime, tubacin concentration-dependently reduced NS5-mediated synthesis of JEV antisense RNA genomes in infected cells ( Figure 7B). The overall results illustrated that tubacin significantly repressed the functional activity of JEV NS5 in vitro replication.

Discussion
Selective HDAC6 inhibitors (tubacin and TBSA), but not pan-HDAC inhibitors, exhibited potent antiviral efficacy against JEV in this study. Tubacin and TBSA substantially reduced JEV-induced cytopathic effect and apoptosis, and concentration-dependently lessened virus yields in human cerebellar medulloblastoma cells (Figures 1-3). The pan-HDAC inhibitors, such as VPA and arginine butyrate, effectively enhanced the sensitivity of Epstein-Barr virus (EBV)-positive lymphoma cells to ganciclovir [28]. In addition, treatment with pan-HDAC inhibitors (TSA and SAHA) decreased the replication of respiratory syncytial virus (RSV) in vivo [29]. Tubacin, HDAC6-specific inhibitor, dosedependently boosted the release of influenza A virus (IAV) progeny through the increase of acetylated microtubules for the effective movement of viral components to the plasma membrane [30]. TBSA, a selective HDAC6 inhibitor, caused the acetylated α-tubulin accumulation in hepatocyte cells which was associated with the decrease of hepatitis C virus (HCV) RNA genome synthesis in hepatocyte cells [22]. Therefore, HDACs might be involved in the replication process of many viruses, such EBV, IAV, RSV, and HCV. Moreover, virus-specific HDACs could be the targets for development of virus-distinctive antiviral agents. The study proposed HDAC6 as a JEV-specific host target for exploiting antiviral agents.
Both tubacin and TBSA were selective HDAC6 inhibitors; however, tubacin (IC50 of 0.26 μM in virus yield reduction) was a more potent inhibitor of JEV than TBSA (IC50 of 1.75 μM) (Figure 3). Importantly, tubacin, but not TBSA, markedly reduced the intracellular production of JEV infectious particles in human cerebellar medulloblastoma cells ( Figure 3C vs. 3D). The result might be linked with higher HDAC6 selectivity and inhibition potency of tubacin compared to TBSA [31]. Therefore, antiviral activity of tubacin relies on its ability to target a host protein. In a time-of-addition assay,

Discussion
Selective HDAC6 inhibitors (tubacin and TBSA), but not pan-HDAC inhibitors, exhibited potent antiviral efficacy against JEV in this study. Tubacin and TBSA substantially reduced JEV-induced cytopathic effect and apoptosis, and concentration-dependently lessened virus yields in human cerebellar medulloblastoma cells (Figures 1-3). The pan-HDAC inhibitors, such as VPA and arginine butyrate, effectively enhanced the sensitivity of Epstein-Barr virus (EBV)-positive lymphoma cells to ganciclovir [28]. In addition, treatment with pan-HDAC inhibitors (TSA and SAHA) decreased the replication of respiratory syncytial virus (RSV) in vivo [29]. Tubacin, HDAC6-specific inhibitor, dose-dependently boosted the release of influenza A virus (IAV) progeny through the increase of acetylated microtubules for the effective movement of viral components to the plasma membrane [30]. TBSA, a selective HDAC6 inhibitor, caused the acetylated α-tubulin accumulation in hepatocyte cells which was associated with the decrease of hepatitis C virus (HCV) RNA genome synthesis in hepatocyte cells [22]. Therefore, HDACs might be involved in the replication process of many viruses, such EBV, IAV, RSV, and HCV. Moreover, virus-specific HDACs could be the targets for development of virus-distinctive antiviral agents. The study proposed HDAC6 as a JEV-specific host target for exploiting antiviral agents.
Both tubacin and TBSA were selective HDAC6 inhibitors; however, tubacin (IC50 of 0.26 µM in virus yield reduction) was a more potent inhibitor of JEV than TBSA (IC50 of 1.75 µM) (Figure 3). Importantly, tubacin, but not TBSA, markedly reduced the intracellular production of JEV infectious particles in human cerebellar medulloblastoma cells ( Figure 3C vs. 3D). The result might be linked with higher HDAC6 selectivity and inhibition potency of tubacin compared to TBSA [31]. Therefore, antiviral activity of tubacin relies on its ability to target a host protein. In a time-of-addition assay, fluorescence intensity of JEV SRIP-driven EGFP reporter and the number of JEV plaques clarified that tubacin pre-treatment showed the highest anti-JEV activity compared to simultaneous treatment and post-treatment ( Figure 5, Figure S3). Notably, tubacin post-treatment also significantly suppressed the JEV replication in concentration-dependent manners. The results explicated that a short-term treatment of tubacin presented the potential of the host-targeting agent with preventive and therapeutic activities against JEV infection. Tubacin might be combined with direct-acting antivirals for preventing and treating JE which accomplishes complementary antiviral actions in a synergistic manner and reduces the possibility of viral resistance during clinical therapy.
Hsp90 hyperacetylation induced by an inhibitor of HDAC6/8 MC1568 has been reported to cause the decrease of IAV replication in lung epithelial cells [32]. In this study, JEV NS5 protein was identified as one of the Hsp90 client proteins ( Figure 6A). Co-immunoprecipitation indicated that tubacin-treatment caused Hsp90 hyperacetylation and decreased the binding ability of Hsp90 to JEV NS5 in infected cells in concentration-dependent manners ( Figure 6B). Tubacin-induced Hsp90 hyperacetylation has been reported to induce the loss of chaperone activity of Hsp90 that resulted in the functional deficiency of Hsp90 client protein glucocorticoid receptor, such ligand binding and transcriptional activation [27]. Hsp90 inhibitor, novobiocin, decreased the protein amounts of NS5 and the virus production in JEV-infected cells ( Figure 6C,D). The results indicated the importance of Hsp90-NS5 interaction for JEV NS5-mediated replication. In addition, tubacin treatment caused the diminution of NS5 protein and antisense RNA genome expression in infected cells (Figure 7). Therefore, our results suggested that tubacin-induced Hsp90 hyperacetylation might influence the NS5 activity in JEV replication, as one of the antiviral mechanisms of tubacin against JEV. Hsp90 has also been demonstrated to interact with viral non-structural (except RNA polymerase) and structural proteins to improve the structural folding and functional activity of those viral proteins [25]. Tubacin-induced Hsp90 hyperacetylation could affect the interaction of Hsp90 with JEV proteins, except NS5, involved in RNA translation and replication complex. Interestingly, microtubule acetylation modulated by HDAC6 inhibitors affected the viral replication through interrupting viral components transportation along microtubules [33]. Therefore, the involvement of tubacin-induced hyperacetylation of the other HDAC6 substrates that assist viral translation and replication could not be excluded in the antiviral mechanism(s) of tubacin against JEV.
In conclusion, HDAC6-selective inhibitors exhibited the potential of antiviral activity against JEV. Particularly, tubacin presented the high-potent inhibition of JEV yield (IC50 of 0.26 µM) and intracellular infectious virion production (IC50 of 1.52 µM). Tubacin was demonstrated as a host-targeting agent with preventive and therapeutic activities against JEV. Tubacin treatment caused the decrease of the Hsp90-NS5 interaction and the reduction of viral proteins and antisense RNA genomes in infected cells. The combination of tubacin and direct-acting antiviral agents provides a novel approach for prophylaxis and treatment against JEV infection.

Cells and Virus
Human cerebellar medulloblastoma TE671 cells (kindly provided by Wen-Kuang Yang, China Medical University, Taiwan) used this study were cultured in minimum essential medium (MEM, GE Healthcare Life Sciences, Pittsburgh, PA, USA), 100 U/mL penicillin-streptomycin, and 5% fetal bovine serum (FBS) (ThermoFisher, Waltham, MA, USA). TE671-CprME stable cell line generated in our laboratory is a packing cell line expressing JEV structure proteins (C, prM, and E) which was described in our previous work [34]. TE671-CprME cells were culture in MEM, 5% FBS, and 500 µg/mL G418 (Sigma, Saint Louis, MO, USA). Baby hamster kidney BHK-21 cells for JEV amplification and plaque assay were also grown in in MEM, 5% FBS and 100 U/mL penicillin-streptomycin. All cells were incubated at 37 • C in an atmosphere containing 5% carbon dioxide.

Quantitative Assays of Virus Yield and Intracellular Viral Titer
For determining anti-JEV activity of TBSA and tubacin on virus yield, JEV yield in a cultured supernatant of infected cells at an MOI of 0.1 in the presence or absence of TBSA and tubacin was measured using a plaque assay. Serial dilution of the supernatant was added onto BHK-21 cell monolayer in 6-well plates, incubated at 37 • C in 5% CO 2 for 1 h, and then overlaid with 2 mL MEM medium containing 1.1% methylcellulose. After a 3-day incubation, the cell monolayer was stained with naphthol blue-black dye; in which viral yields were calculated by the number of viral plaques per mL. For measuring the inhibitory effect of TBSA and tubacin on intracellular viral titer, infected cells were harvested after a 36 h treatment with TBSA and tubacin (0, 1, 5, and 10 µM), and then lysed through three freeze-thaw cycles. The intracellular titer of infectious viruses in lysate was measured by the plaque assay mentioned above.

Virus Attachment Assays with Single Run Infectious Particles (SRIPs) and Virions
JEV SRIPs were produced from CprME-expressing packaging cells transfected with JEV replicon containing EGFP reporter [34]. In brief, TE671 cells grown to 90% confluence were transfected with pFlag-CMV3-CprME using Lipofectamine LTX (Invitrogen, Carlsbad, CA, USA); a stable transfected cell line was established after a 10-day-selection with 500 µg/mL of G418. Next, the packaging cells were transfected with pBR322-JEV-EGFP replicon; JEV-EGFP SRIPs were released in cultured media of transfected packaging cells. For the attachment assay, SRIP was mixed with or without tubacin (10 µM), then immediately added to TE671 cell monolayer at 4 • C to allow viral attachment. After 1-h incubation, cells were washed with cold PBS, then incubated with MEM containing 2% FBS at 37 • C in 5% CO 2 . The green fluorescence of the EGFP reporter derived from the replication of JEV SRIPs was detected by fluorescent microscopy at different time courses (0, 6, 12, 24, 30, and 36 h). In addition, the attachment assay was performed with JEV virions, as described in our prior reports [35]. The TE671 cell monolayer in 6-well plates was incubated with the mixtures of JEV (50 pfu) and tubacin (0, 0.1, 5, 10 and 20 µM) at 4 • C for 1 h, the mixture was then removed, washed with cold PBS, and then overlaid with MEM medium containing 1.1% methylcellulose. After a 3-day incubation at 37 • C in 5% CO 2 , the cell monolayer was stained as described in the plaque assay. Residual plaques were counted; the relative percentage of plaque formation was determined as the ratio of plaque number of each tubacin-treated group to that of mock-treated control.

Time-of-Addition Assay
To examine the inhibitory effect of tubacin by time of addition on JEV replication, pretreatment (1 h prior to JEV infection), simultaneous treatment (at the same time as JEV infection), and post treatment (1 h after JEV entry) experiments were performed. For the pretreatment experiment, the TE671 cells monolayer was pretreated with tubacin (0, 1, 5, and 10 µM) for 1 h, and then infected with 10 TCID50 of JEV SRIPs or 100 p.f.u. of JEV virions at 37 • C in 5% CO 2 . The images of replicon-derived EGFP reporter and virus-induced CPE in JEV SRIP-infected cells were taken by fluorescent and optical microscopies 36 h post infection. The fluorescent intensity of EGFP reporter in SRIP-infected cells was quantified by Image J. In the assays with JEV virions, cell monolayers were overlaid with 2 mL MEM medium containing 1.1% methylcellulose. After 72-h of incubation, viral plaques were counted after staining with naphthol blue-black dye. In the simultaneous treatment experiment, cells were simultaneously treated with tubacin and infected with SRIPs or virions. In the post treatment experiment, tubacin was added into cell monolayers 1 h post infection with SRIPs or virions. The following procedures in both these experiments were performed as the pre-treatment test. The photography of EGFP reporter and CPE in SRIP-infected cells was taken using a microscope; plaque number in virion-infected cells was counted after naphthol blue-black dye staining mentioned above.

Detection of Viral NS5 Expression Using Immnunofluorescence
TE671 cells were infected with JEV at an MOI of 0.1, and simultaneously treated with tubacin (0, 1, 5, and 10 µM). Cells were rinsed once with PBS 36 h post infection, fixed with 4% formaldehyde for 30 min, permeabilized with 0.1% Triton X-100, and blocked with 10% BSA in PBS for 1 h at room temperature. Viral NS5 proteins were detected using rabbit polyclonal anti-JEV-NS5 (GeneTex, Inc., Irvine, CA, USA) and secondary AF546 goat anti-rabbit IgG (ThermoFisher). The image of fluorescent signals in treated infected cells was photographed by fluorescent microscopy.

Quantification of Replicon RNA Expression Using RT-PCR
For examining the inhibitory effect of tubacin on the synthesis of viral antisense RNA genomes, total RNAs of TE671 cells with JEV at an MOI of 0.1 were extracted using PureLink Mini Total RNA Purification Kit (ThermoFisher) 36 h post treatment tubacin, reverse transcripted into cDNA with antisense RNA-specific capture primer (5 -GCAGCAGAAGGAAAGACC GTGAT-3 ), and followed by measuring antisense RNA genomes using SYBR Green Master Mix kit with JEV-specific primer pairs (5 -TCCACTTCCTCAACGCAATG-3 at nucleotide 9724-9743 and 5 -CAGTCGTGCCAGCCATG-3 at nucleotide 9799-9783). The Real-time RT-PCR was performed by 7300 Realtime PCR system (Applied Biosystems, Foster City, CA, USA), and then the corresponding threshold cycle value (C t ) was measured. Relative levels of RNA genomes were normalized by the housekeeping gene GAPDH, described in a prior report [26].

Co-Immunoprecipitation and Western Blotting Assays
Lysate from mock-and JEV-infected cells was incubated with the anti-Hsp90 antibodies (Cell Signaling, Danvers, MA, USA) in a cool room overnight, followed by addition of protein A-Sepharose beads for an additional 4-h. The immunoprecipitate was collected after centrifugation, and analyzed using Western blotting with the anti-JEV NS5 (GeneTex, Inc.), anti-acetyl lysine (Cell Signaling), or anti-Hsp90 antibodies in a cool room overnight. After 4-h of incubation with horseradish peroxidase-conjugated secondary antibodies, the immune-reactive complexes were detected using enhanced chemiluminescence reaction (Amersham Pharmacia Biotech, Piscataway, NJ, USA).

Statistical Analysis
Each data was shown as mean ± standard deviation (S.D.) of three independent experiments. p value in each comparison was determined by Student-t test, in which the comparison was recognized as a statistical significance if p value was lower than 0.05.