Fig 1.
Identification of LAH31 as a novel cross-group protective LAH mAb.
(A) LAH mAbs were tested for binding activity against H1 (pink) and H3 (blue) subtypes using ELISA. Full strain names of individual HA are detailed in S1 Table. LAH3, LAH5, V15-5, and 41-5D06 are published LAH clones. The assay was conducted in duplicates and the representative data from two independent experiments are represented as mean ± SD. (B) The binding region of LAH mAbs was investigated using ELISA with the 30-residue overlapping LAH peptides derived from A/Hong Kong/1/68 (H3N2). The amino acid number of the corresponding region is defined above each column in H3 numbering. The binding activity against the whole LAH peptide (amino acids: 76–130) was also assessed. Binding of individual mAb to each antigen is color-coded according to the AUC of the antibody dilution curve as indicated in the color legend. The assay was conducted in duplicates and representative data from two independent experiments are represented. (C) Variable gene usage of V segments in heavy and light chains and CDR-H3 sequence of the individual LAH mAbs. CDR region is defined by IMGT.
Fig 2.
Cross-group protection provided by LAH31.
(A-C) C57BL/6 mice were treated with (A and B) indicated dose of LAH31 IgG2c or PBS or (C) 5 mg/kg of LAH31 Fc variants (IgG2c-WT, -KA, -LALA, or LALA-PG) or PBS. After 3 h of antibody injection, mice were infected with (A and C) X31 (H3N2) or (B) A/Guangdong-Maonan/SWL1536/2019 (H1N1) viruses. Mice presenting a body weight loss higher than 25% of the initial weight were euthanized. The combined data from two independent experiments (n = 10 for each group) are represented. Each plot indicates the data at the indicated time points. Values represent the mean ± SD. The P values were determined using a two-way ANOVA (body weight loss) or log-rank (survival curve) tests. For multi-group comparisons of survival curves, statistical significance was adjusted by Bonferroni correction. (A) **p < 0.0017 and ***p < 0.00017, compared with the PBS group. (B) ****p < 0.0001, compared with the PBS group. (C) †p < 0.05 and ††††p < 0.0001 for weight loss; ***p < 0.0001 for survival, LAH31 IgG2c-WT compared with LAH31 IgG2c-KA, -LALA or -LALA-PG.
Fig 3.
LAH31 epitope is concealed in prefusion state but exposed and stabilized upon the postfusion conversion.
(A) Binding region of LAH31 was investigated using ELISA with 15-residue overlapping LAH peptides. Amino acid sequence of individual peptides is indicated below each bar. Sequence corresponding to LAH31 binding is colored in red. Peptides sharing core sequences, yet undetectable by LAH31, are partially colored in white. The assay was conducted in duplicates and representative data from two independent experiments are represented as mean ± SD. (B) The core region of LAH31 epitope is highlighted in the schematic model of the trimeric form of the prefusion HA (PDB: 1ha0) or postfusion HA2 (PDB: 1qu1). Each HA protomer or LAH31 epitope region are colored in black, grey, and white or in yellow-red, orange, and yellow, respectively. (C) The crystal structure of LAH31 Fab complexed with LAH epitope is shown in the left (green, LAH31 heavy chain; blue, LAH31 light chain; pink, epitope LAH peptide). The sequence of the corresponding peptide is the same as (A). Superimposition of LAH31 complex on the prefusion HA or postfusion HA2 illustrated in (B) is shown in the right. (D) LAH31 Fab binding activity against the native prefusion HA or postfusion HA2, measured using biolayer interferometry. (E and F) In vitro ADCC activity of LAH31. (E) X31 virus-infected MDCK cells or (F) HA-expressing EL4 cells were incubated with serially diluted mAbs (LAH31, FI6, irrelevant control antibody) in the presence of mouse FcγRIV-expressing reporter cells. The assay was conducted in duplicates and representative data from two independent experiments are represented. Values represent mean ± SD fold induction of luminescence relative to the no IgG control.
Fig 4.
Crystallographic and computational analyses reveal the key foot prints for the cross-group recognition of LAH31.
(A) Detailed recognition mode of the CDR-K1 and CDR-K3 (left), CDR-H3 (middle) and CDR-H1, and FR-H2 and CDR-H2 (right) by LAH31. Contact residues are defined by location within 4 Å between LAH31 and epitope peptide (S4 Table). Polar interactions are indicated by dot lines. Residues are colored as Fig 3 and numbered in Kabat and H3 numbering for LAH31 and the peptide, respectively. (B) Binding breadth of LAH31 was tested through ELISA using 18 HA subtypes, including both pre- and post-pandemic strains for H1. Full strain names are detailed in S1 Table. Binding of LAH31 is color-coded as Fig 1B. (C) Amino acid sequence alignment of the epitope region in the HAs tested in (B). Residues same as H3 sequence are colored in white and highlighted in blue, and those different from H3 but conserved in both H1 strains are highlighted in yellow. Contact residues on H3 and corresponding sequences of other subtypes are boxed in red. The sequence is numbered above in H3 numbering. (D and E) In silico alanine scanning of the epitope region in H3 (grey) simulated from experimental structure (D) and in H1 (magenta) and H8 (yellow) simulated from model structures (E). Changes in free energy in LAH31 binding were calculated by replacing each residue with Ala and are represented as bar graphs. The amino acid sequences are indicated below each bar. (F) Detailed recognition mode against the 106th residue of the epitope peptide in the H3 experimental structure (left) and in H1 (middle) or H8 (right) model structures. (G) Numbers of hydrogen bonds were calculated from LAH31 complexed with epitope peptides of H3 (grey), H1 (magenta) and H8 (yellow) model structures. Value in the 10 model is plotted. The P values were determined using the two-tailed Mann–Whitney U test. Not significant (ns), **p < 0.01. (E-G) Model structures (200) of LAH31-peptide complex were produced in silico and (E and F) the highest model or (G) 10 highest models were chosen for representation.
Fig 5.
LAH31 acquires cross-group specificity primarily through light chain mutations.
(A) Four types of LAH31 mAbs were created for bearing mutated heavy and light chains (HL), germline heavy and mutated light chains (gHL), mutated heavy and germline light chains (HgL), and germline heavy and light chains (gHgL). The binding of these LAH31 mAbs were assessed using ELISA with indicated HA subtypes. (B) Alignment of LAH31 CDR-K1 or CDR-K3 and its germline sequence. Contact residues are colored in pink and underlined. Dots indicate identical residues. CDR region is defined by IMGT, and residue positions are noted above the sequence according to Kabat numbering. (C) Contribution of individual light chain mutation was analyzed using ELISA against HAs of H3 X31 or H1 A/Narita/1/2009. Dilution curves of germline-reverted mAbs are indicated in dashed lines. The assay was conducted in duplicates and representative data from two independent experiments are represented. (D) Detailed recognition mode of the S94/S95 or R92 in CDR-K3 of LAH31 against H1 model epitope peptide. Peptide residues are indicated in H3 numbering.