Respiratory syncytial virus ribonucleoproteins hijack microtubule Rab11 dependent transport for intracellular trafficking

Abstract

Respiratory syncytial virus (RSV) is the primary cause of severe respiratory infection in infants worldwide. Replication of RSV genomic RNA occurs in cytoplasmic inclusions generating viral ribonucleoprotein complexes (vRNPs). vRNPs then reach assembly and budding sites at the plasma membrane. However, mechanisms ensuring vRNPs transportation are unknown. We generated a recombinant RSV harboring fluorescent RNPs allowing us to visualize moving vRNPs in living infected cells and developed an automated imaging pipeline to characterize the movements of vRNPs at a high throughput. Automatic tracking of vRNPs revealed that around 10% of the RNPs exhibit fast and directed motion compatible with transport along the microtubules. Visualization of vRNPs moving along labeled microtubules and restriction of their movements by microtubule depolymerization further support microtubules involvement in vRNPs trafficking. Approximately 30% of vRNPs colocalize with Rab11a protein, a marker of the endosome recycling (ER) pathway and we observed vRNPs and Rab11-labeled vesicles moving together. Transient inhibition of Rab11a expression significantly reduces vRNPs movements demonstrating Rab11 involvement in RNPs trafficking. Finally, Rab11a is specifically immunoprecipitated with vRNPs in infected cells suggesting an interaction between Rab11 and the vRNPs. Altogether, our results strongly suggest that RSV RNPs move on microtubules by hijacking the ER pathway.

Fig 1. Localization of GFP-N and wild type N, P and F proteins in RSV-GFP-N infected cells.

HEp-2 cells were infected with RSV-GFP-N. At 24h p.i. cells were stained with antibodies against N, P or F (red) and Hoechst 33342 (merge). The GFP-N protein was visualized through its spontaneous green fluorescence. IBs are indicated with a star, RNPs are indicated with white arrow heads and viral filaments are indicated with white arrows. Representative images from 2 independent experiments are shown. Images stacks (3 z-steps) were processed as maximum projections and visualized after gaussian filter fixed at 0.5. Scale bar 5 μm.

Fig 2. Characterization of RNP intracellular motions in RSV-GFP-N infected cells.

Live images of HEp-2 or A549 cells infected with RSV-GFP-N for 18 to 20h are analyzed using Imaris software as described in methods section. a) Dynamic behavior of RSV RNPs in HEp-2 cells. Images stacks (3 z-steps) were processed as maximum projections and visualized after gaussian filter fixed at 0.5, the large view is a time projection of 12 consecutive frames (0.21 s between frames). Scale bar 10 μm. The zoomed time series show a fast-moving RNP pointed by white arrow heads, scale bar 2 μm. Representative images from 12 movies from 2 independent experiments are shown. b) Schematic illustration of the motion characterization parameters: instant speed (ratio between the minimal distance between two positions and time interval between these positions), track length (sum of the minimum distance between 2 consecutives positions for the whole track), track displacement (minimal distance between the first and the last position of the particle). c) Smoothed instant speeds of fast-moving particles plotted versus time from the first detection (each color represents one particle, representative examples). d, e) Cumulative distribution of track max speed and track displacement from 6 individual HEp-2 cells from one experiment. f, g and h) Each point is the median of track maximum speed, track velocity and track displacement from one individual cells. ns: No statistical difference between 4 independent experiments on HEp-2 and A549 cells using Brown-Forsythe and Welch’s ANOVA test.

Fig 3. RNPs are decorating the MT network.

A549 cells were infected with RSV for 20h. N (in red) and α-Tubulin (in green) are revealed by immunostaining. High-resolution images were generated using a confocal microscope with Airyscan detector. Images were visualized after gaussian filter fixed at 0.5. A representative image of a whole cell out of 5 cells in 3 experiments is shown. RNPs decorating the MT network are shown on the zoom area. Scale bars 5μm and 1 μm.

Fig 4. Fast long range RNP motions are dependent on microtubules network.

a) Time series of live images of RNPs (in green) moving along MT stained by fluorescent docetaxel-647 (in red) in RSV-GFP-N infected A549 cells. Images stacks (2 z-steps) were processed as maximum projections and visualized after gaussian filter fixed at 0.5. Yellow arrows point positions of a moving RNP. The last image shows a time projection (T Pro). Scale bar 2μm. b to f) RSV-GFP-N infected HEp-2 cells were treated at 17 h.p.i. with nocodazole (10μM, NZ) or DMSO (Ctrl) for 1 h before live-imaging and track analysis. b, c) Centered projection of the tracks of RNPs in a mock (c) or a nocodazole (d) treated cell analyzed over 60 s. Representative images. d, e, f) Each data point represents the median of track maximum speed, track velocity and track displacement from one individual cell. **** p< 0.0001 using t test with Welch’s correction. Data are from 22 cells (DMSO) and 23 cells (NZ) from 3 independent experiments.

Fig 5. Colocalization of RNPs and Rab11a in RSV infected cells.

a) A549-HA-Rab11a cells infected with RSV for 18 h. Rab11a (green) and RNPs (Red) were detected by immunostaining of HA and N. Images were visualized after gaussian filter fixed at 0.5. Scale bar 5 μm. The boxed areas enclose Rab11a and N spots shown magnified. b) Percentage of N spots colocalizing with HA-Rab11a or EEA1 positive spots calculated using Icy Software. Violin plots show the median (center line) and the first and third quartiles (upper and lower hinges). Results are from 70 cells from 3 independent experiments. **** p<0.0001 using Brown-Forsythe and Welch’s ANOVA test followed by Dunn’s test for multiple comparisons.

Fig 6. Rab11 is involved in fast and directed RNP movements.

a) Time series of live images of A549 cells transiently expressing mCherry-Rab11a (red) and infected for 18h with RSV-GFP-N (in green). Images were visualized after gaussian filter fixed at 0.5. Positions of a RNP moving together with a mCherry-Rab11a positive object are pointed out by yellow (RNP) and orange (Rab11a) arrow heads. The last image shows a time projection (T Pro). Scale bar 2μm. b to f) A549 cells were transiently transfected with siRNA targeting Rab11a or non targeting ones (NT) for 48h then infected with RSV-GFP-N for 18 h before live-imaging and track analysis. b, c) Centered projection of the tracks of RNPs in a NT (c) or Rab11a (d) KO cells analyzed during 60 s. Representative images. d, e and f) Each data point represents the median of track maximum speed, track velocity and track displacement from one individual cell. ** p< 0.01, * p<0.05 using t test with Welch’s correction. Data are from 19 cells (NT) and 18 cells (Rab11a) from 3 independent experiments.

Fig 7. Co-immunoprecipitation of Rab11a and RSV RNPs in infected cells.

(a) A549 stably expressing HA-Rab11a were infected with RSV-GFP, RSV-GFP-N or RSV-L-GFP as indicated for 16h, lysed and incubated with beads against GFP. Cell Lysates and bound fractions were analyzed by western-blot using rabbit polyclonal anti-N, anti-P and anti-HA antibodies. One representative experiment out of 3 (RSV-GFP-N) or 2 (RSV-L-GFP) is shown. (b) A549 WT or stably expressing HA-Rab11a as indicated were infected with RSV for 16h, lysed and incubated with beads against HA. Western blot analysis of the co-immunoprecipitated proteins is shown. One representative experiment out of 2.