ATG9A regulates the dissociation of recycling endosomes from microtubules to form liquid influenza A virus inclusions
Fig 6
ATG9A depletion arrests viral inclusions on microtubules.
(A–D) Cells (GFP-Rab11a WTlow, magenta) were treated with siRNA non-targeting (siNT) or targeting ATG9A (siATG9A) for 48 h. Upon this period, cells were infected or mock-infected with PR8 virus for 8 h, at an MOI of 3, and simultaneously treated with 200 nM Sir-Tubulin dye to stain the microtubules (green) in live cells. Cells were imaged for 10 min (2 s/frame) under time-lapse conditions at 8 h postinfection. White boxes show viral inclusions/Rab11a. Individual frames with single moving particles highlighted with yellow arrows are shown in the small panels. Bar = 10 μm. Images from selected infected cells were extracted from S11 and S12 Videos. Images from mock-infected cells were extracted from S13 and S14 Videos. For each case, a linescan was drawn as indicated to assess the dynamics of Rab11a and tubulin. The fluorescence intensity of Rab11a endosomes or viral inclusions (magenta) and tubulin (green) at indicated times was plotted against the distance (in μm). Representative analysis was performed using images from (A-D). (E, F) Cells (GFP-Rab11a WTlow) were treated as explained above (in A–D). At 8 h postinfection, cells were treated with DMSO or 10 μg/mL of nocodazole for 2 h. Cells were imaged at 10 h postinfection. White boxes show viral inclusions/Rab11a. Bar = 10 μm. (G) Scheme illustrates how viral inclusion/Rab11a endosome deviation from a reference position (in X and Y direction) was tracked by live cell imaging. The formula used to quantify the mean squared displacement (MSD, μm2) is also shown. (H) Each viral inclusion/Rab11a endosome in a cell was tracked using the TrackMate plugin (FIJI, NIH) and displacement was quantified as explained in (G). Data were plotted as the MSD (μm2) per treatment. The red dot indicates the median in the boxplots. Statistical analysis was done by a Kruskal–Wallis test (***p < 0.001). (I) Colocalization between microtubules (tubulin) and viral inclusions (Rab11a) in live cells was determined as the Manders’ Overlap Coefficient tM1 (thresholded; explained in Methods section). Only infected conditions are shown. Given that very small Rab11a endosomes were scattered throughout the cytosol in mock-infected controls, we could not obtain reliable correlation coefficients. This result is, however, corroborated by quantifying colocalization between microtubules (tubulin) and viral inclusions using a fluorescent virus (PA-mNeonGreen PR8), as shown in S7 Fig. Between 6 and 10 cells per condition were analyzed. Statistical analysis was done by a Student t test (***p < 0.001). Experiments were performed twice. All the values of individual and pooled experiments are provided in S1 Data File.