Amino acid substitutions in the non-structural proteins 4A or 4B modulate the induction of autophagy in West Nile virus infected cells independently of the activation of the unfolded protein response

Abstract

West Nile virus (WNV) is a neurotropic mosquito-borne flavivirus responsible for outbreaks of meningitis and encephalitis. Whereas the activation of autophagy in cells infected with other flaviviruses is well known, the interaction of WNV with the autophagic pathway still remains unclear and there are reports describing opposite findings obtained even analyzing the same viral strain. To clarify this controversy, we first analyzed the induction of autophagic features in cells infected with a panel of WNV strains. WNV was determined to induce autophagy in a strain dependent manner. We observed that all WNV strains or isolates analyzed, except for the WNV NY99 used, upregulated the autophagic pathway in infected cells. Interestingly, a variant derived from this WNV NY99 isolated from a persistently infected mouse increased LC3 modification and aggregation. Genome sequencing of this variant revealed only two non-synonymous nucleotide substitutions when compared to parental NY99 strain. These nucleotide substitutions introduced one amino acid replacement in NS4A and other in NS4B. Using genetically engineered viruses we showed that introduction of only one of these replacements was sufficient to upregulate the autophagic pathway. Thus, in this work we have shown that naturally occurring point mutations in the viral non-structural proteins NS4A and NS4B confer WNV with the ability to induce the hallmarks of autophagy such as LC3 modification and aggregation. Even more, the differences on the induction of an autophagic response observed among WNV variants in infected cells did not correlate with alterations on the activation of the unfolded protein response (UPR), suggesting an uncoupling of UPR and autophagy during flavivirus infection. The findings here reported could help to improve the knowledge of the cellular processes involved on flavivirus-host cell interactions and contribute to the design of effective strategies to combat these pathogens.

Keywords: LC3; West Nile virus (WNV); autophagy; host cells; replication; unfolded protein response.

FIGURE 1

Differences on the induction of LC3 modification and aggregation in cells infected with diverse variants of West Nile virus (WNV). (A) Growth curves of WNVs in Vero cells. Cells were infected (MOI of 0.5 PFU/cell) with WNVs (B13, B956, ArD27875, Egypt101, and NY99) or USUV and supernatant virus yield was determined at different times p.i. by standard plaque assay on Vero cells. (B) Visualization of autophagosome formation by LC3 aggregation in cells infected with the viruses displayed in panel (A). Vero cells were transfected with a plasmid encoding GFP-LC3 and 24 h post-transfection were infected with WNV or USUV (MOI of 5 PFU/cell). Cells were fixed and processed for immunofluorescence using a monoclonal antibody against dsRNA and secondary antibodies AF-594 labeled 24 h p.i. Scale bars: 10 μm. (C) Quantification of the number of LC3 aggregates per cell. The number of fluorescent aggregates on the cytoplasm of cells infected in (B) was determined. Dashed line indicates the mean number of GFP puncta aggregates found in uninfected cells. Statistically significant differences between infected and uninfected cells are denoted by one asterisk (*) for P < 0.05 or two asterisks (**) for P < 0.005. n.s. denotes not statistically significant differences. (D) Monitoring LC3 modification following infection by WNV or USUV. Vero cells infected with different WNVs or USUV (MOI of 0.5 PFU/cell) were lysed at 1 or 24 h p.i. and subjected to western blot analysis using an antibody against LC3 to detect non-lipidated LC3 (LC3-I) and LC3 conjugated to phosphatidylethanolamine (LC3-II). An anti-β-actin antibody was also used as control for protein loading.

FIGURE 2

Characterization of the differences between WNV NY99 and B13 variant. (A) Time course analysis of LC3 modification following infection by WNV NY99 or B13. Vero cells infected with WNV NY99 or B13 (MOI of 0.5 PFU/cell) were lysed at 1, 6, 12, 24, or 48 h p.i. and subjected to western blot analysis using an antibody against LC3 to detect non-lipidated LC3 (LC3-I) and LC3 conjugated to phosphatidylethanolamine (LC3-II). An anti-β-actin antibody was also used as control for protein loading. Mock-infected cells were analyzed in parallel to show the levels of LC3-II in uninfected cells. (B) Analysis of LC3 modification following infection by WNV NY99 or B13 including protease inhibitors (E-64d and pepstatin A). Vero cells were infected with WNV NY99 or B13 (MOI of 5 PFU/cell) and harvested at 24 h p.i. Protease inhibitors E-64d and pepstatin A (10 μg/ml each) were added to culture medium 4 h before harvesting the cells. Mock-infected cells were analyzed in parallel to show the effect of inhibitors in uninfected cells. Western blot analysis was performed as described for panel A. (C) Effect of pharmacological inhibition on the infection of WNV NY99 and B13. Cells infected with the different viruses (MOI of 0.5 PFU/cell) were treated with 2.5 mM 3-MA and the virus yield were determined by plaque assay at 24 h p.i. USUV was included as a positive control. Statistically significant differences between each condition and control cells are denoted by one asterisk (*) for P < 0.05.

FIGURE 3

Mutations on WNV NS4A or NS4B can modulate the induction of LC3 modification and aggregation in cells infected with WNV. (A) Growth curves of recombinant WNV recovered from infectious cDNA clone pFLWNV (WT) or its derivatives carrying amino acid substitutions NS4A V67I, NS4B I240M or NS4A V67I + NS4B I240M. Vero cells were infected (MOI of 0.5 PFU/cell) with the viruses and the supernatant virus yield was determined at different times p.i. (B) Visualization of autophagosome formation by LC3 aggregation in cells infected with the viruses displayed in panel (A). Vero cells were transfected with a plasmid encoding GFP-LC3 and 24 h post-transfection were infected (MOI of 5 PFU/cell), fixed and processed for immunofluorescence analysis as described in the legend for Figure 1. Scale bars: 10 μm. (C) The number of fluorescent aggregates on the cytoplasm of cells infected in (B) was determined. Dashed line indicates the mean number of GFP puncta aggregates found in uninfected cells. (D) Analysis of LC3 modification following infection with recombinant WNVs. Vero cells were infected (MOI of 0.5 PFU/cell) and lysed at 1 or 24 h p.i. Western blot analysis was performed as described in the legend of Figure 1. (E) Quantification of the fold change on LC3-II/actin from 1 to 24 h p.i. Results are product of three independent western blots similar to that displayed in panel (D). Statistically significant differences between infected and uninfected cells, or cells infected with WT and mutant viruses, are denoted by two asterisks (**) for P < 0.005. n.s. denotes not statistically significant differences.

FIGURE 4

Infection with WNV variants that differ in induction of autophagy results in similar UPR features. (A) Activation of the unfolded protein response (UPR) following infection with WNV or USUV assayed by detection of spliced Xbp-1 mRNA. RNA was extracted from Vero cells infected with WNV B956, ArD27875, Egypt101, NY99, B13, recombinant WNV recovered from infectious cDNA clone pFLWNV (WT) or its derivatives carrying amino acid substitutions NS4A V67I, NS4B I240M or NS4A V67I + NS4B I240M, and the presence of unspliced or spliced Xbp-1 mRNA was determined by RT-PCR. Cells treated with tunicamycin, or infected with USUV are included as a positive control of the activation of UPR. GAPDH mRNA was also amplified by RT-PCR as a control. (B) Effects of pharmacological modulation of UPR with salubrinal on WNV infection. Cells infected (MOI of 0.5 PFU/cell) with the different viruses described in (A) were treated with 5 μM salubrinal or the same amount of drug vehicle (DMSO, control) and the virus yield was determined by plaque assay at 24 h p.i. (C) Analysis of cellular viability upon treatment with salubrinal. Vero cells were treated with 5 μM salubrinal or drug vehicle (DMSO, control) for 24 h and the viability was estimated by determination of cellular ATP levels by a luminescence assay. Statistically significant differences between each condition and control cells are denoted by two asterisks (**) for P < 0.005.