Design of Beads-on-a-string (BOAS) immunogens

(A) Phylogenic tree of influenza hemagglutinin (HA) subtypes. Subtypes are further categorized as group 1, group 2 or influenza B, and stars indicate subtypes included in the BOAS. (B) Native trimeric HA structure (A/H3/Hong Kong/1968; PDB: 4FNK) is shown on the left, with the HA1 domain in light grey, and HA2 in dark grey. A construct diagram of full-length HA is shown at the top with the locations of head domain segments within HA1. Head domains of H1, H3, and H7 span residues 57-261, and head domains of H2, H3, H4, H9, and B span 37-316 of the HA1 region of HA. (C) Construct diagram of HA BOAS. BOAS are assembled by linking HA head domains (light grey) into a single polypeptide chain where HA head domains are separated by a flexible linker with 4 (GSS) repeats. The C-terminus consists of an HRV-3C protease cleavable His tag and streptavidin-binding-protein (SBP) tag for purification and other affinity tags. (D) BOAS used in this study.

Biochemical Characterization of BOAS

A) SDS-PAGE analysis of 3mer-8mer BOAS under non-reducing conditions. B) SDS-PAGE analysis of 3mer-8mer BOAS under non-reducing conditions following glycan digestion with PNGase-F. C) Size exclusion chromatography traces of 3mer-8mer BOAS on an Superdex 200 column in PBS. D) Negative stain electron microscopy (NS-EM) images of the 3mer. Arrows point to individual HA monomers. Inset image is 2.5X zoomed enlarged image of boxed area. E) Negative stain electron microscopy (NS-EM) images of the 8mer. Arrows point to individual HA monomers. Inset image is 2.5X zoomed enlarged image of boxed area. F) Characterization via ELISA of BOAS components with subtype-specific antibodies. KD values in µM are reported for each antibody for each BOAS immunogen. Antibodies corresponding to each component are as follows: S5V2-29 (interface, all components except B/Mal/04), 5J8 (RBS, H1), 2G1 (RBS, H2), 3E5 (RBS, H3), P2-D9 (H4), H5.3 (RBS, H5), H7.167 (RBS, H7), H1209 (RBS, B).

Murine immunizations with 3mer-8mer BOAS

A) Schematic of immunization regimen with 3mer-8mer BOAS. Mice were primed with 20µg of BOAS, followed by two homologous boosts two weeks apart. B) Immunogenicity of BOAS over time as determined by serum ELISA. Serum reactivity to each immunogen is reported as the reciprocal of the EC50 dilution factor for each group (n=3) and d14, d28, and d42. Data points are an average of n=3 mice and error bars are +/- 1 s.d. C) Serum titers to each BOAS at the final time point, d42. Data points for single time point antigen titers are from individual mice (n=3), bars represent mean titers and error bars are +/- 1 s.d., * = p<0.05 as determined by Kruskal-Wallis test relative to 3mer. D) Serum titers to full-length HA trimers elicited by 3mer-8mer BOAS. Solid bars below each plot indicate a matched sub-type, and striped bars indicate a mis-matched subtype (i.e. not present in the BOAS).

BOAS component epitope homology and serum competition

A) Amino acid conservation of each BOAS projected on the A/H2/Japan/1957 head domain structure (PDB: 2WRE). Variability and conservation scores were determined as the % conservation of each residue at each position for each BOAS composition. Cut-outs are RBS and Interface epitopes with their respective conservation. B) Homology score as determined by the average % conservation of each residue across the entire protein sequence (overall, circle), and the RBS (square) and interface (triangle) epitopes. C) Serum competition to RBS and interface epitopes as determined via competition with a cocktail of RBS-directed antibodies, interface-directed antibodies, or a stem-directed negative control antibody. % competition is determined by the relative decrease in serum binding to each BOAS normalized to an untreated control. Each data point represents a mean from n=3 mice and error bars are +/- 1 s.d. D) Trend overlay of % serum competition with RBS-directed antibodies and homology score of the RBS epitope as a function of BOAS length. E) Trend overlay of % serum competition with interface-directed antibodies and homology score of the interface epitope as a function of BOAS length.

Design of BOAS nanoparticle (NP) immunogens

A) Schematic of design of BOAS nanoparticles. The eight original 8mer components were split into two 4mers, the first containing H1, H2, H3, and H4 (H1-H2-H3-H4), and the second containing H5, H7, H9, and B (H5-H7-H9-B). Each were expressed with an N-terminal SpyTag, then attached at equal molar ratios to an N-terminally fused SpyCatcher ferritin nanoparticle (SpyCatcher NP). This yields a ferritin nanoparticle with both 4mer BOAS conjugated to its surface (2×4mer BOAS NP). B) Dynamic light scattering (DLS) distributions of SpyCatcher NP and 2×4mer BOAS NP. Curves represent an average of 3 technical replicate measurements and error bars are +/- 1 s.d. C) Size exclusion traces of SpyCatcher NP and 2×4mer BOAS NP on an S400 column. D) Representative NS-EM image of SpyCatcher NP. Arrows indicate individual nanoparticles, and the inset is a 2X zoom of the boxed area. Representative NS-EM image of BOAS NP. Arrows indicate individual nanoparticles, and the inset is a 2X zoom of the boxed area. E) ELISA of BOAS components, SpyCatcher NP, and 2×4mer BOAS NP with structure-dependent subtype-specific antibodies and a negative control stem-directed antibody, MEDI8852. ELISA data points are technical duplicates and error bars are +/- 1 s.d.

Murine Immunizations with BOAS NP

A) Schematic of immunization regimen with 2×4mer BOAS NP and SpyCatcher NP control. Mice were primed with 20µg of NP, followed by two homologous boosts two weeks apart. B) Serum reactivity to each immunizing antigen at d42. ** p<0.01 as determined by Mann-Whitney test. C) Serum reactivity to Control NP scaffold by both groups. ** p<0.01 as determined by Mann-Whitney test. D) Serum reactivity elicited by Control NP group (n=5) at d42 to both individual BOAS components (H1-H2-H3-H4 and H5-H7-H9-B), Control NP, and BOAS NP. n.s. = not significant (p>0.05) as determined by Kruskal-Wallis test. E) Serum reactivity elicited by BOAS NP group (n=5) at d42 to both individual BOAS components (H1-H2-H3-H4 and H5-H7-H9-B), Control NP, and BOAS NP. *** p < 0.001 and n.s. = not significant (p>0.05) as determined by Kruskal-Wallis test. F) Serum reactivity elicited by BOAS NP group (n=5) to individual full-length HA components of BOAS at d42. G) Serum reactivity to each individual full-length HA component of BOAS (H1 green, H2 blue, H3 orange, H4 purple, H5 yellow, H7 pink, H9 maroon, B salmon) by 3mer-8mer BOAS and BOAS NP. Solid bars below each plot indicate a matched sub-type, and striped bars indicate a mis-matched subtype (i.e. not present in the BOAS or NP). * p<0.05 and ** p<0.01 as determined by Kruskal-Wallis test with multiple comparisons. In all plots, data points for single time point antigen titers are from individual mice (n=5 for NP studies, n=3 for 3mer-8mer BOAS studies), bars represent mean titers and error bars are +/- 1 s.d.

Microneutralization titers to matched and mis-matched virus

Microneutralization of matched and mis-matched psuedoviruses: H1N1 (green, top left), H3N2 (orange, top right), H5N1 (yellow, bottom left), and H7N9 viruses (pink, bottom right) with d42 serum. H1N1 and H3N2 and H5N1 viruses match components in all BOAS constructs tested and BOAS NP. The H7 component is present and matched in 6mer-8mer and BOAS NP and mismatched in 3mer-5mer. Dashed lines represent the first serum dilution (1:32), and points at the dashed line indicate IC50s at or below the limit of detection. Serum samples are pooled from n=3 (BOAS) or n=5 (BOAS NP) mice and individual data points are technical duplicates. The mean is denoted by a bar +/- 1 s.d.