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

TIA-1 and TIAR Colocalize with WNV and DV2 Proteins in Infected Baby Hamster Kidney (BHK) Cells.

A previous study showed that in C57/BL/6 TIAR^−/− MEFs, the replication of WNV was reduced by 6- to 8-fold compared with control MEFs, whereas WNV grew to the same levels in TIA-1^−/− MEFs and control MEFs (13). This was consistent with the 10-fold higher binding efficiency of TIAR for the WNV3′(−)SL as compared with TIA-1 (13). To determine whether a flavivirus infection altered the distribution of these cell proteins, BHK cells were infected with WNV at MOI of 5 and at 3, 6, 12, and 24 hours after infection (hpi), the cells were fixed, permeabilized, incubated with anti-TIA-1 or anti-TIAR and then anti-WNV antibody, and visualized by confocal microscopy as described in Materials and Methods and supporting information (SI) Materials and Methods. As expected, TIAR (Fig. 1A, green) and TIA-1 (Fig. 1E, green) were evenly distributed throughout mock-infected cells (20). Under these infection conditions, WNV proteins were detected in ≈90% of the cells. WNV proteins were faintly detected by 6 hpi (Fig. 1B, red). Although at 6 hpi, TIAR was still found in both the cytoplasm and nucleus, this protein was also concentrated in a few bright foci in the cytoplasm (Fig. 1B, green). Merged images obtained at 6 hpi showed colocalization of TIAR foci, with foci of viral protein (Fig. 1B, merge). Foci of TIA-1 were not observed at this time (data not shown). By 12 hpi, the majority of TIAR was in the cytoplasm (Fig. 1C, green) and the amount of WNV protein had increased and was concentrated in the perinuclear region (Fig. 1C, red). Again, colocalization of TIAR but not TIA-1 (data not shown) with the viral proteins was observed mainly in perinuclear foci (Fig. 1C, merge). By 24 hpi, the intensity of WNV protein staining had increased significantly (Fig. 1 D and F, red), and TIAR strongly colocalized in the perinuclear region with the WNV proteins (Fig. 1D, merge). In contrast, only a few bright cytoplasmic foci of TIA-1 colocalized with the WNV proteins by 24 hpi (Fig. 1F, merge; a pattern similar to that seen at 12 hpi with TIAR). Strong colocalization of TIA-1 with the WNV proteins was observed by 36 hpi (Fig. 1G, merge). Quantification of the amount of TIAR or TIA-1 in the nucleus confirmed that the distribution of both TIA-1 and TIAR was altered with time after WNV infection but the redistribution of TIAR occurred more rapidly (Fig. 1H).

It was previously reported that the expression of TIA-1 was three times higher in TIAR^−/− than in wild-type MEFs (13). To determine whether the time course of TIA-1 redistribution was altered when TIAR was not present, TIAR^−/− MEFs were infected with WNV (MOI of 5), and the locations of TIA-1 and WNV proteins were analyzed at 12 and 24 hpi. TIA-1 was equally distributed in both the cytoplasm and nucleus of mock-infected TIAR^−/− MEFs (SI Fig. 5A, green), but these cells were significantly larger than BHK cells. The time course of appearance, staining intensity, and distribution of the WNV proteins in these cells (SI Fig. 5 B and C, red) were similar to those observed with infected BHK cells. At 12 hpi, a few bright foci of TIA-1 that colocalized with WNV proteins were observed (SI Fig. 5B, merge), similar to what was observed by 12 hpi for TIAR in BHK cells. By 24 hpi, a significant increase in TIA-1 in the perinuclear region and strong colocalization with viral proteins was observed (SI Fig. 5C). At all times, a significant portion of TIA-1 remained in the nucleus. These results indicate that, in the absence of TIAR, colocalization of TIA-1 with the viral proteins occurred more quickly, with a time course similar to that of TIAR colocalization in cells expressing both proteins.

To determine whether TIA-1/TIAR colocalization with viral proteins was also characteristic of other flavivirus infections, BHK cells were infected with DV2, (MOI of 0.1), a divergent flavivirus from a different serocomplex than WNV, and analyzed at 12, 24, 36, and 72 hpi as described in Materials and Methods. Because of the lower MOI, at 12 and 24 hpi, DV2 proteins were detected in only ≈5% of the cells (data not shown). By 36 hpi, clusters of infected cells with bright virus-specific cytoplasmic foci were observed (≈40% of the cells) (SI Fig. 6 B and D, red). At 72 hpi, in ≈60% of the cells, DV2 proteins were detected with a distribution and staining intensity (SI Fig. 6 C and E, red) similar to that observed in WNV-infected cells at 24 hpi (Fig. 1 D and F, red). Strong colocalization of both TIAR (SI Fig. 6B and C, merge) and TIA-1 (SI Fig. 6 D and E, merge) with DV2 proteins was detected at 36 and 72 hpi. This colocalization was primarily in the perinuclear region but was also observed in foci extending into the cytoplasm of the infected cells (SI Fig. 6 B–E, merge). In the uninfected cells visible in the middle panels of SI Fig. 6 B–E, TIA-1 and TIAR were equally distributed in the cytoplasm and nucleus. These results suggest that the colocalization of TIA-1/TIAR with viral proteins is a general feature of flavivirus infections.

TIA-1 and TIAR Interact with Sites of WNV and DV2 RNA Replication.

In previous in vitro RNA–protein interaction assays, the TIA-1/TIAR proteins were shown to bind specifically to the WNV3′(−)SL RNA (13). The minus-strand RNA of flaviviruses is present in infected cells only in RNA replication complexes (14, 15). An antibody to dsRNA that does not detect either cellular ribosomal RNA or tRNA was previously used to detect flavivirus replication complexes in infected cells (21). To test whether TIA-1/TIAR colocalized with viral dsRNA in infected cells, BHK cells infected with WNV (MOI of 0.1) or DV2 (MOI of 0.1) were fixed and incubated with anti-dsRNA and anti-TIA-1 or anti-TIAR antibody. A lower MOI for WNV and later time points were used to better visualize the replication complexes. Perinuclear foci stained with anti-dsRNA antibody were observed in WNV-infected cells at 36 hpi (Fig. 2 B and C, red) but not in mock-infected cells (Fig. 2A). In WNV-infected cells, strong colocalization of TIA-1/TIAR with the dsRNA was observed in bright perinuclear foci (Fig. 2 B and C, merge). Similar dsRNA perinuclear foci and similar TIA-1/TIAR colocalization with these foci were observed in DV2-infected cells at 72 hpi (SI Fig. 7 B and C), indicating that TIA-1/TIAR strongly colocalized with flavivirus RNA complexes in infected cells.

Previous immunofluorescence studies using antibodies specific to dsRNA as well as to individual flavivirus NS proteins suggested that NS3 and NS5, as well as NS1, NS2a, and NS4a, associate with ds viral RNA in flavivirus replication complexes located in the perinuclear region of infected cells (22, 23). Broader areas of colocalization were observed with the anti-WNV (Fig. 1) than with the anti-dsRNA antibody, suggesting that TIA-1/TIAR might also interact with a viral nonstructural protein. Coimmunoprecipitation experiments were done by using lysates prepared from mock-infected or WNV-infected BHK cells at 24 hpi and anti-TIAR or anti-TIA-1 antibody, as described in Materials and Methods. Cell lysates, as well as immunoprecipitates, were then analyzed by SDS/PAGE and transferred to nitrocellulose membranes for Western blot analysis using anti-WNV (Fig. 2G Left). Four WNV proteins, presumed to be NS5, NS3, E, and NS1, were detected in lysates from WNV-infected cells (Fig. 2G Left, lane 2) but not in lysates or immunoprecipitates from mock-infected cells (Fig. 2G Left, lanes 1, 3, and 5). In immunoprecipitates from WNV-infected cells, both anti-TIAR (Fig. 2G Left, lane 4) and anti-TIA-1 (Fig. 2G Left, lane 6) antibody coimmunoprecipitated a viral protein with the expected size of NS3. To verify that this was NS3, samples precipitated with anti-TIA-1 or anti-TIAR antibodies were analyzed by Western blot analysis using an anti-WNV NS3 monoclonal antibody. NS3 was detected in lysates from WNV-infected cells after immunoprecipitation with both anti-TIAR (Fig. 2G Right, lane 4) or anti-TIA-1 (Fig. 2G Right, lane 6) antibody.

Anti-WNV NS3 monoclonal antibody was next used to study the colocalization of TIA-1/TIAR with NS3 in BHK cells infected with WNV or DV2 at 36 or 72 hpi, respectively. NS3 was previously shown to be broadly distributed around the nucleus and to colocalize with dsRNA foci in flavivirus-infected cells (21). In the merged images, both TIA-1 and TIAR strongly colocalized with NS3 in WNV-infected cells (Fig. 2 E and F, merge). Bright focal areas, as well as more diffuse areas of colocalization, were observed. Similar results were observed in DV2-infected cells at 72 hpi (SI Fig. 7 E and F).

WNV and DV2 Interfere with SG Formation in Infected BHK Cells.

TIA-1/TIAR were previously reported to be required for the formation of SGs in cells after treatment with arsenite (2), and both proteins were shown to colocalize with viral nonstructural proteins and replication complexes in flavivirus-infected cells. To determine whether a flavivirus infection would interfere with cellular SG formation, BHK cells were mock-infected or infected with WNV and exposed to oxidative stress by treatment with 0.5 mM arsenite for 30 min. Infected cells were treated at 6, 12, or 24 hpi. The cells were then fixed and stained with anti-WNV (Fig. 3 A–D, red) and anti-TIAR antibody (Fig. 3 A–D, green). At 6 hpi, when only a small amount of colocalization of TIAR with WNV proteins was observed (Fig. 3B, merge), the number of SGs induced by the arsenite was slightly reduced (Fig. 3B, green; Fig. 3E) as compared with mock-infected cells (Fig. 3A, green; Fig. 3E). At 12 hpi, SG formation further decreased (Fig. 3C, green; Fig. 3E) as colocalization of TIAR and when WNV proteins increased (Fig. 3C, merge). Few, if any, SGs were detected at 24 hpi (Fig. 3D, green; Fig. 3E) in cells showing strong colocalization between TIAR and WNV proteins (Fig. 3D, merge). WNV infection alone did not induce SG formation at any time after infection (Fig. 1 B–D, green). The effect of DV2 infection on SG formation was also investigated. BHK cells infected with DV2 for 36 or 72 h were treated with 0.5 mM arsenite for 30 min and immunostained with anti-DV2 and anti-TIAR (SI Fig. 8). At both times, strong colocalization of TIAR and DV2 proteins was observed. Few SGs were observed in infected cells (SI Fig. 8 B and C, green; SI Fig. 8D), whereas many SGs were observed in uninfected cells (SI Fig. 8A, green).

Because phosphorylation of eIF2α at S51 also regulates SG formation, the effect of WNV and DV2 infections on eIF2α phosphorylation was assayed. BHK cells were infected with either virus, and, at various times after infection, cells were treated for 30 min with 0.5 mM arsenite. Control infected cells were not treated with arsenite. Proteins in cell lysates were analyzed by Western blot using a polyclonal antibody specific for eIF2α phosphorylated at S51 (Fig. 3F). Under these experimental conditions, phosphorylation of eIF2α was not detected in mock-infected cells or 24 h after WNV (Fig. 3F, lane 3) or 36 h after DV2 (SI Fig. 8E, lane 1) infection. High levels of eIF2α phosphorylation were observed in extracts of mock-infected cells treated with arsenite (Fig. 3F, lane 2). Reduced levels of phospho-eIF2α were detected in WNV-infected extracts by 12 hpi and were further decreased by 24 hpi (Fig. 3F, lanes 4–6). In the DV2 extracts, decreased phosphorylation of eIF2α was observed at later times after infection (SI Fig. 8E, lanes 2–5). The large number of uninfected cells present in the DV2-infected cultures are the source of the higher levels of phospho-eIF2α observed until 6 days after infection (dpi). The levels of total eIF2α and of another protein involved in SG formation, G3BP, remained constant under all conditions tested and after infection with both viruses (Fig. 3F and SI Fig. 8E). Although neither WNV nor DV2 infection induced SG formation in BHK cells, cells infected with both flaviviruses became progressively more resistant to SG induction by arsenite with time after infection.

WNV and DV2 Infections Interfere with PB Assembly.

SGs and PBs share some components and are dynamically linked by their role in regulating cytoplasmic mRNA storage and/or degradation (7, 24). To test whether WNV and DV2 infections also interfere with PB assembly, BHK cells mock-infected or infected for various times with either WNV or DV2 were treated or not treated for 30 min with 0.5 mM arsenite. The cells were then fixed and incubated with antibody to the PB marker DCP1a and anti-WNV (Fig. 4) or anti-DV2 (SI Fig. 9). PBs were detected in untreated cells (Fig. 4 A–C, green; SI Fig. 9 A–C, green) as well as in arsenite-treated cells (Fig. 4 D–F, green; SI Fig. 9 D–F, green). In uninfected, WNV-infected, and DV2-infected cells, PBs were evenly distributed throughout the cytoplasm (Fig. 4 and SI Fig. 9). In both uninfected and infected cells treated with arsenite, the number of PBs was decreased (Fig. 4G; SI Fig. 9G), and the remaining PBs were mainly located around the nucleus (Fig. 4 D–F, green; SI Fig. 9 D–F, green). In WNV-infected cells, a noticeable reduction in PB formation was observed by 24 hpi in untreated (Fig. 4B, green; Fig. 4G Left) and arsenite-treated (Fig. 4E, green; Fig. 4G Right) cells as compared with mock-infected cells (Fig. 4 A and D, green; Fig. 4G). A further reduction in PB formation was observed in both arsenite-untreated (Fig. 4C, green; Fig. 4G Left) and treated (Fig. 4F, green; Fig. 4G Right) WNV-infected cells at 36 hpi. In DV2-infected cells, a reduction in PB formation was observed at later times after infection (SI Fig. 9 B, C, E, and F, green; SI Fig. 9G), consistent with the slower replication kinetics of DV2. These results indicate that both WNV and DV2 infections interfere with PB assembly in BHK cells.

Cells.

BHK-21, strain WI2 cells (32) were maintained as described (33). C57/BL/6 and C57/BL/6 TIAR^−/− MEF cell lines were provided by Paul Anderson (Brigham and Woman's Hospital, Boston, MA) and grown in MEM supplemented with 10% FCS, 1× MEM nonessential amino acids (Invitrogen, Carlsbad, CA), 1 mM sodium pyruvate (Invitrogen), 2 mM glutamine (Invitrogen), and 10 μg/ml gentamycin.

Viruses.

A stock of WNV strain Eg101 was prepared in BHK cells (2 × 10⁸ PFU/ml) as described (33). DV2, strain Bangkok D80-100, was provided by Walter Brandt (Walter Reed Army Institute of Research, Washington, DC). The titer of a 10% wt/vol suckling mouse brain homogenate was 2 × 10⁶ PFU/ml. Cells were infected with WNV at a MOI of 5 or 0.1 and with DV2 at a MOI of 0.1. Virus was adsorbed to cells for 1 h at room temperature, the inoculum was removed, and the cells were washed three times with serum-free medium. MEM containing 5% FCS was added, and the infected cells were incubated at 37°C in a CO₂ incubator. To induce stress, cells were treated at various times after infection with 0.5 mM sodium m-arsenite (Sigma–Aldrich St. Louis, MO) in MEM for 30 min. Virus growth experiments were done as described (13).

Western Blotting.

Western blot analysis was carried out as described (33). Membranes were blocked in 1× Tris-buffered saline (TBS) containing 5% nonfat dry milk (NFDM). Primary antibodies were diluted in 1× TBS containing 5% BSA (1:1,000 for anti-phospho-eIF2α, 1:500 for anti-eIF2α, 1:3,000 for anti-G3BP, and 1:500 for anti-WNV and anti-WNV NS3 antibodies), and the secondary antibody was diluted 1:2,000 in 1× TBS containing 5% NFDM.

Coimmunoprecipitation Assays.

BHK cells (2 × 10⁶) were mock-infected or infected with WNV at a MOI of 5. At 24 hpi, cell lysates were prepared by using lysis buffer containing 50 mM sodium phosphate (pH 7.2), 150 mM NaCl, 1% Nonidet P-40, and Complete, Mini, EDTA-free protease inhibitor mixture (Roche, Indianapolis, IN). Cell lysates were incubated on ice for 30 min, sonicated, and then centrifuged at 2,000 × g for 5 min at 4°C. The supernatant (250 μl per ≈100 μg of total protein) was precleared with protein A/G-magnetic beads (New England Biolabs, Beverly, MA) for 1 h at 4°C with rotation. The clarified supernatants were incubated with 1 μg of anti-TIA-1 or anti-TIAR antibody at 4°C for 1 h with rotation, and then beads were added and incubated for 1 h. Beads were collected magnetically, washed seven times with 1 ml of lysis buffer, and proteins were eluted by boiling for 5 min. Proteins were separated by 12% SDS/PAGE, transferred to nitrocellulose membranes, and analyzed by Western blotting using either mouse anti-WNV or anti-NS3 antibody.

Detection of Intracellular and Cellular Viral Proteins by Immunofluorescence.

BHK, C57BL/6, and TIAR^−/− cells (2 × 10⁴) were seeded onto 3-mm coverslips (Fisher Scientific, Pittsburgh, PA) and grown to ≈50% confluency. At various times after infection, cells were fixed in 4% paraformaldehyde (Sigma–Aldrich) in PBS for 10 min at room temperature, permeabilized with 100% chilled methanol for 10 min at −20°C, washed three times in PBS, and incubated with blocking buffer [PBS containing 5% heat inactivated horse serum (Invitrogen)] overnight at 4°C. Primary antibodies were diluted in blocking buffer (1:1,000 for anti-TIAR, anti-TIA-1, and anti-Dcap1; 1:200 for anti-dsRNA; and 1:100 for anti-WNV, anti-DV2, and anti-WNV NS3 antibody) and incubated with cells at 37°C for 1 h. Cells were then washed three times for 10 min with PBS and incubated for 1 h at 37°C with Texas red- and FITC-labeled secondary antibodies diluted 1:300 in blocking buffer containing 0.5 μg/ml Hoechst 33258 dye (Molecular Probes, Bedford, MA) to stain the nuclear DNA. Coverslips were mounted in Prolong Gold antifade reagent (Invitrogen), and the cells were scanned with a confocal microscope LSM 510 (Zeiss, Oberkochen, Germany) using a 100× oil-immersion objective. The images compared for each experimental series were collected by using the same instrument settings, and the images were analyzed by using Zeiss software version 3.5.