Fig 1.
Generation of a stable BSR cell line expressing anti-HA-Franken bodies (HA-FB) and HA-tagged BTV virus.
(A) Left: Cartoon showing that BSR cells were transfected with a plasmid expressing GFP-labelled HA-FB, and stably-transfected cells were selected using puromycin. Right: HA-tagged BTV virus was generated using reverse genetics by inserting four copies of the HA tag into the amino acid sequence between positions 218 and 219 of the VP2 tip domain, which allows binding of the HA-FB. (B) Comparison of 10 genomic RNA segments (S1-S10) between recombinant BTVVP2-HA and wild-type virus by agarose gel electrophoresis. (C) Protein expression of VP2 in infected BSR cells at 24h pi analysed by western blot of recombinant BTVVP2-HA vs. wild-type virus. (D) Comparison of plaque size between recombinant BTVVP2-HA and wild-type virus. (E) Cytoplasmic VP2 labelled by anti-VP2 antibody (red, bottom row) can be specifically recognized by GFP-labelled HA-FB (green, bottom row) in BTVVP2-HA infected BSR cells expressing HA-FB at 18h pi compared to the mock infected cells (top row). Colocalization appears yellow on merged images and nuclei stained with Hoechst are shown in blue. Scale bar = 10µm.
Fig 2.
Colocalization of BTV outermost VP2 protein with selected markers of the exocytic secretion pathway by confocal imaging.
Protein markers of the exocytic secretion pathway were labelled by specific antibodies shown in red. Cytoplasmic VP2 was visualized by GFP-labelled HA-FB in BTVVP2-HA infected BSR cells expressing HA-FB at 18h pi, shown in green. Colocalization of BTV and cellular markers appear yellow on merged images. Nuclei were stained with Hoechst and shown in blue in the merged images. Scale bar = 10µm. (B) Bar chart representing results of colocalization quantification analysed using the JACoP plugin of ImageJ using Manders’ coefficients. A Manders’ overlap coefficients threshold above 0.5 was used to indicate a significant degree of colocalization.
Fig 3.
Co-migration of BTV proteins VP2 and NS3 with intracellular vesicle marker proteins by ultracentrifugation in an isopycnic iodixanol density gradient.
(A) The supernatant of a whole cell lysate from infected BSR cells was loaded onto a linear 12-48% (w/v) Optiprep density gradient to allow separation of intracellular compartments and organelles. Both VP2 and NS3 were abundant in fractions 9 to 13 with a calculated range of density from 1.12 to 1.16 g/cm3. NS3 in particularly was not detected in other fractions. MVB-related proteins TSG101 and HSP90, autophagy related protein LC3B, ERAD-related protein EDEM1 and lysosomal membrane glycoprotein LAMP1 were also detected in the same fractions, except for Calnexin, an ER resident protein that is absent from intracellular vesicles. (B) Stacked line charts showing the quantification of relative abundance of each protein in different fractions by densitometry. Peak fractions 9-13 were highlighted in the light grey box.
Fig 4.
NS3 is vital for viral maturation and intracellular trafficking through its direct interaction with outer capsid proteins VP2 and VP5.
(A) Co-IP analysis of the interactions between wild-type or mutant NS3 and VP2 from the cell lysates of BSR cells infected with wild-type or N150A and KKE196-198/AAA NS3 mutant viruses using an anti-NS3 antibody. GAPDH was used as control for equal protein loading. (B) Bar chart showing the normalized VP2 intensity measured by densitometry in Co-IP samples (right). Two-way ANOVA test *p < 0.05. (C) Confocal fluorescence microscopy of BSR cells infected with two NS3 mutant viruses showed Golgi retention of the N150A mutant NS3 and the KKE196-198/AAA mutant NS3. (D) Bar chart showing the percentage of colocalization between NS3 and Golgi (right). Two-way ANOVA test *p < 0.05. (E) Confocal microscopy showed that N150A mutant NS3 and the KKE196-198/AAA mutant NS3 no longer colocalized with VP2 and VP5 compared to the wild-type NS3. Both mutations in the NS3 resulted in a loss of VP2 and VP5 plasma membrane trafficking compared to wild-type NS3 (bottom row). Colocalization appears yellow on merged images. Scale bar = 10µm.
Fig 5.
Isolation and characterization of EVs released from BTV infected mammalian. (A) BSR cells, (B) sheep PT cells and (C) insect KC cells, or (D) NS3-transfected BSR cells.
Four pellets were isolated from the cell culture supernatant by differential centrifugation at 300xg (P1), 2000xg (P2) and 10,000xg (P3) followed by PEG-10,000 precipitation (P4), containing detached cells (P1), dead cells (P2), large extracellaule vesicles (LEVs) (P3) and small extracellular vesicles (SEVs) (P4), respectively, and resuspended in equal volume of PBS buffer. The abundance of viral capsid proteins VP2, VP5 and VP7, and nonstructural proteins NS3 and NS2 in P1-P4 from all three cell lines infected with BTV (A-C), or the abundance of NS3 in P1-P4 from NS3-transfected BSR cells were analyzed by western blot. (E) Bar chart showing the ratio of NS3 abundance in SEVs compared to that in LEVs from (A-D) measured by densitometry.
Fig 6.
Comparison between LEVs and SEVs of their different origins and compositions, and the inhibitory effect of different inhibitors on the release of virus in EVs by western blot.
The viral proteins VP2 and NS3, and a series of subcellular organelle protein markers were analyzed against the LEVs and SEVs isolated from mock infected or BTV infected BSR (A) and PT (B) cells. Whole cell lysate (WCL) was included as a positive control. (C) MTT cytotoxicity assay measures cytotoxic effect on BSR cells at the concentrations of GW4869 and EerI used. Two-way ANOVA test *p < 0.05. (D) The expression of sphingomyelin phosphodiesterase 2 (SMPD2), which is inhibited by GW4869 treatment, and (E) ubiquitinated proteins on EerI treatment were analyzed by western blot. GAPDH was used as control for equal protein loading. (F) compared to untreated BSR cells (left), the treatment of 10 µM of GW4869, a potent exosome inhibitor, failed to inhibit the release of SEVs (P4) from BTV infected BSR cells (middle). BTV infected BSR cells treated with EerI, which blocks the process of moving misfolded proteins from the ER to cytoplasm for ubiquitin-mediated protein degradation, resulted in decreased abundance of intracellular NS3 and NS3 in SEVs (P4) (right).
Fig 7.
Cryo-ET visualization of BTV infected BSR cells.
(A) At early time points (12h and 16h pi) lamellae of subcellular regions showed that many virus particles were observed within small single membrane vesicles (SMVs), typically containing only a single particle (top row, 1st and 2nd columns), while large SMVs contain multiple virus particles (bottom row, 1st and 2nd columns). In contrast, at 24h pi many virus particles were observed within double membrane vesicles (DMVs), containing single (top row, 3rd and 4th columns) or multiple virus particles (bottom row, 3rd and 4th columns). (B) Panels showing unmilled areas of the cell periphery with singular or multiple virus particles in SMVs being released from the cell. Scale bar = 100nm.
Fig 8.
Cryo-ET FIB-milled lamellas show the presence of viral particles in EDEMosomes/DMVs in BTV infected BSR cells at 24h pi.
The left panel shows viral particles making contact with the ER. The right panel shows snapshots of the different stages of viral particles being engulfed in EDEMosomes/DMVs. Scale bar = 100nm.
Fig 9.
Viral particles in SEVs induce efficient infection.
(A) Confocal imaging showed that both LEVs and SEVs, labelled with PKH26 red fluorescent dye, were internalized by BSR cells, and were colocalized with viral proteins VP2 and NS3 (shown in green and purple, respectively) in EV uptake experiments. Colocalization appears yellow on merged images and nuclei stained with Hoechst are shown in blue. Scale bar = 5µm. (B) Titre comparison of equal volume of LEVs and SEVs diluted in PBS and free supernatant virus collected from BTV infected BSR, PT and KC cells at 24h pi. (C) Comparison of the specific infectivity of LEVs, SEVs and free supernatant virus calculated as genome copies-to-PFU ratio in BSR, PT and KC cells. Two-way ANOVA test *p < 0.05, **p < 0.01.
Fig 10.
Transmission electron microscopy (TEM) on ultrathin sections of BTV infected BSR and PT cells at 24h pi.
(A) most virus particles were released by non-lytic viral budding in BTV infected BSR cells, although virus particles were also observed to be released in SEVs (B) or LEVs (C) from BTV infected BSR cells but were not predominant. (D) SEVs released from BTV infected PT cells contain only a few virus particles. (E) LEVs released from BTV infected PT cells contain not only multiple virus particles but also other intracellular materials. (F) immunogold labelling with NS3 antibody showed enrichment of NS3 in SEVs released from infected PT cells. Scale bar = 200nm.
Fig 11.
Cartoon showing the hypothesis of the origin of SEVs.
Viral NS3 protein triggers ER remodelling for the formation of NS3-enriched ER-derived double-membrane vesicles (DMVs)/EDEMosomes. SEVs could originate from the fusion between MVBs and these DMVs/EDEMosomes to promote more efficient cell-to-cell virus transmission in addition to the other pathways of virus egress previously described.