Fig 1.
VSV G cryo-EM structures in its pre- and post-conformations.
(A) Upper panel: Schematic diagram of VSV G ectodomain (in sky blue) with the position of the two glycosylation sites (N163 and N320) indicated in orange. The transmembrane domain (TMD) and intraviral domain (ID), which are not resolved in our structures, are represented in shades of grey. Lower panel: Domain organization of VSV G ectodomain. The trimerization domain (TrD) is shown in red, the pleckstrin homology domain (PHD) in orange, and the fusion domain (FD) in yellow. Domains are connected by five segments (named R1 to R5) that refold during conformational change: R1 and R4 are in cyan, R2 and R3 are in green, R5 and the C-terminal domain (CTD) are shown in light and deep purple, respectively. (B-C) Cryo-EM density map of full-length VSV G solubilized in LMNG and incubated at pH 8.0 (B) and at pH 5.5 (C). The G ectodomain is depicted in sky blue, the LMNG micelle in grey, and the CTD in magenta. The part of the fusion domain inserted in the micelle is colored in yellow. The starting residues of the N-linked glycosylation chains are depicted in orange, with the positions of N163 and N320 indicated by labels 1 and 2, respectively. (D-E) Close-up view of the fusion loops insertion into the micelle at pH 8.0 (D) and pH 5.5 (E). The outline of the micelle is indicated by the red dotted line. The hydrophilic residues located in the hydrophilic shell are depicted in green. The hydrophobic residues in the fusion loops are in orange.
Fig 2.
VSV G C-terminal domain (CTD) rearrangement.
(A) Ribbon representation of VSV G trimer in its post-fusion conformation, fitted and traced up to residue 426 in the cryo-EM map obtained at pH 5.5. The same color code as in Fig 1A is used. (B) Surface hydrophobicity representation of the interaction between segment 410-426 of chain A (right) and neighboring FD (left) in the post-fusion conformation. The hydrophobicity scale is indicated. (C) Hydrophobic patch on the CTD-FD interface in the post-fusion conformation; Left panel: Overall view of VSV G trimer showing FD from protomer A (in yellow) and FD from protomer B (in green) with the CTD from protomer A in magenta. Residues displayed as sticks contribute to the stabilizing of the CTD-FD interface in the post-fusion state. Right panel: close-up on the hydrophobic patch formed at the level of the CTD (involving L423 and F425) interacting with I82 from the FD of protomer A (in yellow), and L105 and P107 from the FD of protomer B (in green). These residues stabilize the CTD-FD interface in the post-fusion conformation. Putative hydrophobic interactions are indicated by red dashed lines. (D) Ribbon representation of VSV G trimer in its pre-fusion conformation, fitted and traced up to residue 427 in the cryo-EM map obtained at pH 8.0. The same color code as in Fig 1A is used. (E) Close-up view of VSV G CTD in the pre-fusion conformation. The inset shows the folding of the CTD and a detailed view of the environment surrounding I82, L423 and F425 in VSV G pre-fusion conformation. Putative hydrophobic interactions are indicated by red dashed lines.
Fig 3.
mAb 8G5F11 interacts with the pre- and the post-fusion conformations of VSV G.
(A) Coomassie-stained SDS-PAGE analysis (under non-reducing conditions) of interaction experiments between VSV G and mAb 8G5F11 bound to protein A coated magnetic bead, incubated at various pHs. (B) BLI sensograms showing the binding kinetics of VSV Gect to mAb 8G5F11 at pH 8.0. Experimental curves (black) were fitted (red) using a 1:1 binding model. The calculated KD is indicated on the panel. (C) Coomassie-stained SDS-PAGE analysis (under reducing conditions) of interaction experiment between VSV G and Fab 8G5F11 bound to strep-Tactin coated magnetic beads, incubated at various pH values. (D) BLI sensograms showing the binding kinetics of VSV Gect to 8G5F11 Fab at pH 8.0. Experimental curves were fitted using a 1:1 binding model. The calculated KD is indicated on the panel. (E) Neutralization curve of VSVe-GFP by mAb (dark blue) and Fab (light blue). VSV-eGFP was preincubated with increasing concentrations of mAb or Fab. At 5 hours p.i., the percentage of infected cells determined by counting the number of cells expressing eGFP using flow cytometry. This was used to calculate the infectious viral titer. Data depict the mean with standard error from experiments performed in triplicate. Average IC50 value are indicated.
Fig 4.
Structural basis for the neutralization of VSV by mAb 8G5F11.
(A) Cryo-EM density map of VSV G in complex with Fab 8G5F11 at pH 8.0 shown in side view orientation. One 8G5F11 Fab molecule was observed to bind to each protomer of VSV G. G ectodomain is in sky-blue, the LMNG micelle in grey, the starting residues of the glycosylation chains are in orange. Fusion loops are in yellow and CTD in magenta. Fab variable heavy (VH) and variable light (VL) chains are in green and in cyan respectively. (B) Ribbon diagram of VSV G trimer in the pre-fusion conformation in complex with Fab 8G5F11, viewed from the side (left panel) and from the top (right panel). The VSV G model was fitted in the density and traced up to residue 426. The VH and VL chains of the 8G5F11 Fab were modelized and fitted in the identified density. The same color code as in Fig 1A is used for VSV G. Fab VH chain is in green and VL chain is in cyan. (C) Cryo-EM density map of VSV G in complex with Fab 8G5F11 at pH 5.5, shown in side orientation. The same color code as in Fig 4A is used. (D) Ribbon diagram of VSV G post-fusion trimer in complex with Fab 8G5F11, viewed from the side (left panel) and from the top (right panel). As in the pre-fusion state, one molecule of Fab binds each VSV G protomer of the post fusion trimer. The VSV G model was fitted in the density and traced up to residue 426, while the VH and VL chains of the 8G5F11 Fab were fitted in the identified density. The same code for G as in Fig 1A is used. Fab VH chain is in green and VL chain is in cyan. (E) Interaction between 8G5F11 and VSV G. Key residues are depicted in stick representation and are highlighted in the boxed panel. For clarity only, labels of key residues on VSV G are indicated. (F) Close up on the residues involved in the antibody-antigen interaction at high pH. VSV G PHD is in orange, and the key residues interacting with 8G5F11 are depicted in sticks.
Fig 5.
Residues D241 and D243 are essential for the interaction of 8G5F11 with VSV G and for neutralization.
(A) Schematic description of the binding assays used to assess the capacity of VSV G WT and VSV G alanine mutants to bind CR domains and 8G5F11 mAb or Fab. (B) Surface expression test for VSV G WT and VSV G alanine mutants by measuring their ability to bind the conformational probe (CR3-GST labelled by and ATTO550 fluorophore). The histogram indicates the mean fluorescence intensity of ATTO550 positive cells for each G construct. Three independent experiments were performed. Error bars represent the standard deviation. (C) Antibody recognition assay for VSV G WT and mutants G by measuring their ability to bind 8G5F11 mAb (left panel) or Fab (right panel). Statistically significant differences with WT are indicated by stars (**p < 0.005, ***p < 0.0005). (D) Schematic description of the neutralization assay using VSVΔG/G pseudotypes. (E) Incorporation of VSV G WT and VSV G alanine mutants in VSVΔG viral particles, assessed using a polyclonal anti-VSV G and an anti-VSV M antibodies. (F) Neutralization of VSVΔG/GWT, VSVΔG/GD241A, and VSVΔG/GD243A by mAb (left panel) or Fab 8G5F11 (right panel). VSV pseudotypes were preincubated with increasing concentrations of mAb or Fab. At 16 hours p.i., the percentage of infected cells was determined by counting the number of cells expressing eGFP using flow cytometry. This was used to calculate the infectious viral titer. Data depict the mean with standard error for experiments performed in triplicate. Average IC50 are indicated.
Fig 6.
Inhibition of VSV G mediated cell-cell fusion by 8G5F11.
(A) Schematic description of cell-cell fusion assay. BSR cells were co-transfected with plasmids encoding VSV G and a cytoplasmic fluorescent marker (P-GFP). Twenty-four hours later, cells were incubated in fusion buffer at different pH values in the absence or presence of 10 nM 8G5F11 and syncytium formation was assessed. (B) Cell-cell fusion assay of VSV G with or without 8G5F11 after incubation at the indicated pH values. Nuclei were stained with DAPI. All images are representative examples from three independent experiments. (C) Schematic representation of 8G5F11 mechanism of action. There are no constraints for 8G5F11 to bind its epitope on the glycoprotein in its pre-fusion conformation when G is inserted into the viral membrane. However, the density of spikes on the surface of the viral particle prevents the structural transition due to the size of the antibody and its orientation when bound to the post-fusion state, which induce significant steric hindrance (indicated by arrows).