Disulfide stabilization of human norovirus GI.1 virus-like particles focuses immune response toward blockade epitopes

Figures

• 

  Fig. 1. Dissociation of GI.1 norovirus VLPs leads to exposure of occluded epitopes. a Representative micrograph of negatively stained GI.1 WT VLPs. 2D class averages of small objects are shown in the bottom right corner, and 2D class averages of intact VLPs in the top right corner. Notice the presence of dissociated VP1 dimers and larger oligomers. The shell and P domain are indicated by white arrows. b Representative micrograph of GI.1 WT after incubation with A1227 Fab fragment. 2D class averages of small objects show 2 Fabs bound to each VP1 dimer. Intact particles do not show any A1227 Fab bound. c Representative micrograph of GI.1 WT after incubation with 512 Fab fragment. Class averages show VP1 dimers bound to two Fabs and intact VLPs decorated with a layer of 512 Fabs. d–f Analytical size-exclusion chromatography of d GI.1 WT, e GI.1 WT + A1227 Fab, and f GI.1 WT + 512 Fab. In all cases, the peak at 8.5 mL elution volume corresponds to void volume. g SDS–PAGE analysis of fractions from chromatogram in d. h SDS–PAGE analysis of fractions from chromatogram in e. Fraction 4 contains equal amounts of VP1 and A1227 Fab, indicative of 1:1 complex (1 Fab: 1 VP1), while fraction 5…

• 

  Fig. 2. Structure-based design of interprotomer disulfides between VP1 monomers leads to stabilized particles with increased thermal stability of the shell domain. a Two strategies can be used to form interprotomer contacts between adjacent VP1 protomers: (1) design of disulfides (or other intermolecular interactions) at the strict fivefold and strict threefold symmetry axes (top panel), and (2) design of disulfides (or other interactions) within the icosahedral asymmetric unit (iASU; bottom panel). One example of each strategy is shown on a portion of GI.1 VLP structure (P domain omitted for clarity). b Screening of eight double-point mutants within GI.1 shell domain after oxidation in diamide. Top gel shows the expression levels of VP1 mutants in reducing conditions. Bottom gel shows the same samples in nonreducing conditions. Asterisks indicate mutants that prevent VLP dissociation (intact VLPs fail to enter the separating gel due to their size). Notice the presence of other VP1 oligomers (dimers, tetramers, and larger species). c Preparative-scale purification of diamide-treated mutant N116C–G193C (separation performed on Sephacryl S500 column). d Stabilization of the final product…

• 

  Fig. 3. Cryo-EM reconstruction of GI.1 DS1 VLP provides details of interprotomer disulfide bonds between Cys116 and Cys193. a Overall electron density of GI.1 DS1 VLPs. Atomic model of the shell is colored according to each nonequivalent monomers. Monomers A, B, and C are shown in magenta, orange, and green, respectively. b, c Close-up view of the region around b the strict fivefold and c threefold symmetry axes of the icosahedron. d, e Details of the electron density for one of the interprotomer disulfide bond between d A–A monomers and e B–C monomers.

• 

  Fig. 4. Stabilization of GI.1 VLPs prevents VLPs from disassembling but preserves accessibility to 512 blockade epitope. a Representative micrograph of stabilized GI.1 VLPs (GI.1 DS1) in the absence of antibody. Particles appear to be intact, and no disassembled components are visible. 2D class averaging yields very homogeneous particle classes with an approximate diameter of 40 nm (top right panel) and no sign of dissociated molecule (bottom right panel). b Representative micrograph of stabilized VLPs after incubation with A1227 Fab. 2D class averages show two species: intact particles with no Fab bound (top right panel) and Fab fragments (bottom right panel). c Representative micrograph of stabilized VLPs after incubation with 512 Fab. 2D class averages show two species: intact particles decorated with 512 Fabs (top right panel) and free Fab fragments (bottom right panel). Notice that the sizes of the intact particles increase by ~7 nm, consistent with the presence of a layer of Fabs bound to each particle. d The sample in a was separated by size-exclusion chromatography and c fractions analyzed by SDS–PAGE in reducing and nonreducing conditions. Only one peak is present at the void…

• 

  Fig. 5. Mice immunizations with stabilized GI.1 VLPs produces high titers of blockade antibody responses after two injections without alum. a Balb/c mice were immunized with two micrograms of either GI.1 WT or GI.1 DS1 VLPs intramuscularly at weeks 0, 3, 6, and 9. Blood draws were performed at weeks 3, 5, 8, and 11. To test durability of immune response, final bleed was performed 13 weeks after the last immunization. For each blood draw, HBGA blockade titers were assessed. b Comparison of blockade antibody titers between GI.1 WT (gray circles) and GI.1 DS1 (red circles) in the absence of adjuvant. Disulfide stabilization of norovirus VLPs leads to fast development of blockade titers compared to wild-type VLPs (compare blockade titers after second immunization). c Comparison of blockade titers between GI.1 WT (green circles) and GI.1 DS1 (blue circles) in the presence of adjuvant (alum). No significant differences at any time point are observed, indicative of potentially stabilizing effect of GI.1 WT VLPs by alum. Limit of detection (titer = 80) is indicated with a horizontal dotted line. Geometric mean titers ± geometric SD are shown in scatter dot plot. P values were determined by…

• 

  Fig. 6. Stabilization of GI.1 VLPs focuses immune responses toward blockade epitopes and away from occluded epitopes. a Ratio of blockade titers over total GI.1-reactive IgG titers at week 22. In the absence of alum, immunization with GI.1 DS1 leads to a twofold increase in blockade titers relative to total GI.1-reactive titers, compared to immunization with GI.1 WT. b Ability of serum antibodies at week 22 to compete with blockade antibody 512, relative to total GI.1-reactive titers. For both non-adjuvanted and adjuvanted groups, sera from mice immunized with stabilized VLP can compete with 512 binding more than sera from mice immunized with wild-type VLPs. c Ability of serum antibodies at week 22 to compete with occluded-site antibody A1227, relative to total GI.1-reactive titers. In the presence of alum, sera from mice immunized with wild-type VLPs can compete with A1227 more than sera from immunization with stabilized VLPs. Mean ± SD are shown in each box plot. P values were determined by two-tailed Mann–Whitney tests. *P ≤ 0.05, **P ≤ 0.01.