Fig 1.
High densities of HA on the IAV surface can contribute to host cell attachment and the binding of neutralizing antibodies.
(A) HAs on the viral surface bind to multiple sialylated receptors (depicted in the lower part of the image), increasing receptor-binding avidity. High densities of HA also present opportunities for bivalent attachment of antibodies (upper part of the image). (B) Factors that may influence the ability of antibodies to bind bivalently are the density of HA on the virion surface and the degree to which they can tilt and rotate. (C) Structural model of S139/1 (PDB ID 4GMS, aligned to full-length IgG1 from 1HZH) and C05 (4FQR) bound to the HA head domain (3LZG). Distances indicate the height of the HA ectodomain (~15nm) and typical nearest-neighbor spacing within the viral membrane (~10nm). In all panels, grey portions of antibodies represent the Fc region; colored portions (red for C05, blue-green for S139/1) represent Fabs.
Fig 2.
S139/1 and C05 IgGs are enhanced by avidity to different extents.
(A) The effective off-rate of an IgG antibody (IgG koff) is modeled by the combination of two rates: the off-rate for each of the Fab arms (Fab koff), and the cross-linking rate (kx) at which the second Fab binds when the first Fab is bound. (B) Schematic and images from the assay to measure dissociation kinetics. IAV immobilized onto a flowchamber surface is imaged after equilibrated fluorescent antibody is rapidly washed away. (C) Normalized Fab and IgG dissociation curves for both S139/1 and C05 against the HK68 and WSN strains. Data are combined from three biological replicates at two antibody concentrations each and normalized to the fluorescence intensity at t = 0 for each sample. Time between acquisition is set to 5, 20, or 600 seconds depending on relative rate of dissociation. (D) Kinetic parameters for S139/1 and C05 Fab/IgG against WSN33 and HK68. The dissociation rates for each antibody/strain combination are determined from the initial rates of fluorescence loss. The crosslinking rate kx is fit for each pair by simulating the effective IgG koff for a given Fab koff for a range of kx values. The measured off-rate of S139/1 Fab against WSN33 (0.026 s-1) is likely an underestimate due to low initial signal in this sample and fast dissociation. An estimated kx value and dissociation rate data from Lee et al. are shown in grey as a comparioson.
Fig 3.
S139/1 and C05 avidity persists at ten-fold reductions in HA density.
(A) Schematic and representative image of the fluorescent HA decoy used to reduce the HA density on the viral surface. In the panel to the left, HA is visualized using an scFv derived from FI6v3. The decoy (SEP-HA; shown in green in the panel to the right) contains the cytoplasmic tail and transmembrane domain of WSN33 HA, while the head domain is replaced by super-ecliptic pHluorin. (B) Steady-state binding of the antibodies to the WT and SEP-HA versions of the virus measured by fluorescence intensity. Quantification is performed on large filamentous virions to ensure high signal-to-noise. Measured intensities are normalized to the mean of the WT condition for each pair. P-values are determined by multiple unpaired t-tests for three replicates; significance is indicated by a p-value below 0.05. (C) Schematic of the experimental approach to decrease HA density using synthetic nanoparticles. C-terminally biotinylated HK68 HA ectodomain is bound to streptavidin-coated spheres and diluted with varying concentrations of biotinylated BSA to titrate the relative HA density. (D) Steady-state binding of Fab and IgG antibodies as a function of relative HA density. Relative HA density is determined by normalizing mean HA signal intensity to the 100% HA condition in each replicate. Three separate preparations of beads are measured for each condition. The R2 value of the linear regression for each antibody is shown correspondingly. (E) Comparison of S139/1 and C05 IgG binding at lower HA densities. Top: both HA and the antibody fluorescence intensities are also normalized to the mean 100% HA condition for each antibody for all three replicates. Bottom: antibody occupancy as a function of HA density. For each point, %IgG from the top plot is divided by the corresponding %HA.
Fig 4.
Antibodies prefer different HA orientations for efficient crosslinking.
(A) Model to predict preferred antibody binding geometry. Structures of HA-Fab complexes are subjected to transformations reflecting the degrees of freedom of one Fab relative to the other. The distance between the base of the two bound HAs and the angle between them is determined for each sampled conformation. Plots in the lower right show angle and spacing distributions predicted for S139/1 and C05. Structural models above the plots are built by aligning HA-Fab structures (PDB IDs 4GMS and 4FQR) with human IgG1 (PDB ID 1HZH). (B) Schematic illustrating the proposed effect upon binding of an FISW84 Fab to HA (PDB IDs 6HJP, 6HJR). (C) Images of WSN33 bound by S139/1, + /- FISW84 Fab at 30nM. (D) Steady-state binding for the indicated antibodies. Analysis is performed on filamentous virions to ensure high signal-to-noise, and the median for each population is normalized to that of the corresponding wild-type condition. Three biological replicates are performed for each condition and compared by multiple unpaired t-tests; significance is indicated by a p-value below 0.05.