Interaction of TIA-1/TIAR with West Nile and dengue virus products in infected cells interferes with stress granule formation and processing body assembly

Abstract

The West Nile virus minus-strand 3' terminal stem loop (SL) RNA was previously shown to bind specifically to cellular stress granule (SG) components, T cell intracellular antigen-1 (TIA-1) and the related protein TIAR. In vitro TIAR binding was 10 times more efficient than TIA-1. The 3'(-)SL functions as the promoter for genomic RNA synthesis. Colocalization of TIAR and TIA-1 with the viral replication complex components dsRNA and NS3 was observed in the perinuclear regions of West Nile virus- and dengue virus-infected cells. The kinetics of accumulation of TIAR in the perinuclear region was similar to those of genomic RNA synthesis. In contrast, relocation of TIA-1 to the perinuclear region began only after maximal levels of RNA synthesis had been achieved, except when TIAR was absent. Virus infection did not induce SGs and progressive resistance to SG induction by arsenite developed coincident with TIAR relocation. A progressive decrease in the number of processing bodies was secondarily observed in infected cells. These data suggest that the interaction of TIAR with viral components facilitates flavivirus genome RNA synthesis and inhibits SG formation, which prevents the shutoff of host translation.

Fig. 1.

Colocalization of TIA-1 and TIAR with WNV proteins in infected BHK cells. (A–D) Laser scanning confocal microscopy of mock-infected (A) or WNV-infected (MOI of 5) (B–D) BHK cells stained with anti-WNV (red) and anti-TIAR antibody (green) at the indicated times after infection. (E–G) BHK stained with anti-WNV (red) and anti-TIA-1 antibody (green) at the indicated times after WNV (F and G) or mock (E) infection. Nuclear DNA (blue) was stained with Hoechst 33258 dye. (H) Quantification of the amount of TIAR (Left) or TIA-1 (Right) in the nucleus of mock-infected cells (M) and WNV-infected cells at the indicated times after infection. The relative pixel intensity in the nuclei of 20 cells at each time after infection was measured, and the mean values were plotted. Error bars indicate the standard deviation of the mean.

Fig. 2.

Interaction of TIA-1 and TIAR with WNV replication complex components. BHK cells were infected with WNV at a MOI of 0.1. (A–F) Laser scanning confocal microscopy of mock-infected (A and D) and WNV-infected (B, C, E, and F) cells stained with anti-dsRNA (A–C, red) or anti-NS3 (D–F, red) and then with either anti-TIAR (A, B, D, and E, green) or anti-TIA-1 (C and F, green) antibodies at 36 hpi. Enlarged views are included in the B and C merged panels. (G) Coimmunoprecipitation of WNV proteins by anti-TIA-1 and anti-TIAR antibodies. BHK cells were infected with WNV at a MOI of 5. Immunoprecipitates were visualized by Western blot using anti-WNV (Left) or anti-NS3 antibody (Right). Lane 1, mock-infected lysate; lane 2, WNV-infected lysate; lanes 3 and 5, mock-infected immunoprecipitates; lanes 4 and 6, WNV-infected immunoprecipitates. Arrows indicate the positions of the WNV proteins detected. Positions of molecular weight standards are indicated on the left.

Fig. 3.

WNV infection interferes with SG formation. (A–D) Laser scanning confocal microscopy of mock-infected (A) and WNV-infected (MOI of 5) (B–D) BHK cells treated with 0.5 mM arsenite for 30 min at the indicated times after infection. The cells were fixed, permeabilized, and stained with anti-WNV (A–D, red) and with the SG marker, anti-TIAR (A–D, green). Arrows indicate representative SG. (E) Quantification of the number of SGs in mock-infected (M) and WNV-infected cells at the indicated times after infection. Visible SGs in 70 cells were counted for each experimental condition, and the average number of SG per cell was plotted. Error bars indicate the standard deviation of the mean. (F) Western blot analysis of phospho-eIF2α in BHK cells mock-infected or infected with WNV for 24 h or in infected cells treated with 0.5 mM arsenite for 30 min at the indicated times after infection. Lane 1, mock-infected BHK cells; lane 2, cells treated with 0.5 mM arsenite for 30 min; lane 3, cells infected with WNV (MOI of 5); lanes 4–6, cells infected with WNV and then treated with arsenite for 30 min at the indicated times. Phospho-eIF2α (S51) (Top), total eIF2α (Middle), and G3BP (Lower) were detected with specific antibodies (SI Materials and Methods).

Fig. 4.

WNV infection interferes with PB assembly. (A–F) Laser scanning confocal microscopy of mock-infected (A and D) and WNV-infected (MOI of 5) (B, C, E, and F) cells that were treated (Right) or not treated (Left) with arsenite for 30 min at the indicated times. The cells were stained with anti-WNV (A–F, red) and anti-DCP1a (a PB marker) (A–F, green) antibody. Arrowheads indicate representative PBs. (G) Quantification of the number of PBs in mock-infected (M) and WNV-infected cells treated (Right) or not treated (Left) with arsenite. PBs in 50 cells were counted for each experimental condition at the indicated times after infection, and the average number of PBs per cell was plotted. Error bars indicate the standard deviation of the mean.