Convergent and clonotype-enriched mutations in the light chain drive affinity maturation of a public antibody

Abstract

Public antibodies that recognize conserved epitopes are critical for vaccine development, and identifying somatic hypermutations (SHMs) that enhance antigen affinity in these public responses is key to guiding vaccine design for better protection. We propose that affinity-enhancing SHMs are selectively enriched in public antibody clonotypes, surpassing the background frequency seen in antibodies carrying the same V genes, but with different epitope specificities. Employing a human IGHV4-59/IGKV3-20 public antibody as a model, we compare SHM signatures in antibodies also using these V genes, but recognizing other epitopes. Critically, this comparison identified clonotype-enriched mutations in the light chain. Our analyses also show that these SHMs, in combination, enhance binding to a previously uncharacterized viral epitope, with antibody responses to it increasing after multiple vaccinations. Our findings offer a framework for identifying affinity-enhancing SHMs in public antibodies based on convergence and clonotype-enrichment and can help guide vaccine design aimed to elicit public antibodies.

Keywords: affinity maturation; cryo-EM; public antibodies.

Figure 1.. Binding, neutralization, and in vivo protection activity of M15.

A) Schematic of binding, in vitro neutralization, and in vivo protection experiments performed using M15, created with BioRender. (B-C) Binding activity of M15 to (B) spike proteins from human coronaviruses, including the XBB.1.5 and JN.1 variants of SARS-CoV-2, and (C) subunits of WA1/2020 spike protein, represented as the minimum binding concentration (μg/mL) measured by ELISA. The dotted line represents the limit of detection (LOD), set at the starting dilution of 30 μg/mL. All values with a minimum binding concentration of > 30 μg/mL, were set to 60 μg/mL for graphing purposes. (D) In vitro neutralization capacity of M15 against WA1/2020, XBB.1.5, and JN.1 variants, represented as half-maximal inhibitory concentrations (IC50). All values with an IC50 of > 30 μg/mL, were set to 60 μg/mL for graphing purposes (E-G) Survival curves of hACE2-K18 mice prophylactically treated with 10 mg/kg (intraperitoneal) of M15 antibody prior to challenge with a 3xLD50 dose of (E) WA1/2020, (F) XBB.1.5, or (G) JN.1. CR9114, an isotype-matched influenza virus anti-hemagglutinin mAb, was used as a negative control mAb.

Figure 2.. Convergent and specific SHM in M15-like antibodies.

(A) Schematic workflow for identifying SHMs that are convergent and specific to M15-like antibodies, created with BioRender. Black traces refer to convergent SHMs. The number of sequences illustrated per group is not representative of the actual numbers. Since the non-M15-like sequences (SARS-CoV-2 non-S2, influenza virus, and malaria) share the same germline V gene (IGH4–59 and IGKV3–20) but have different D and J genes, the lengths of the D and J gene segments vary in the cartoon. (B) Frequencies of heavy (left) and light (right) chain SHM in M15-like clonal lineages (126 sequences) and negative-control malaria (heavy chain: 10 sequences, light chain: 16 sequences), influenza virus (heavy chain: 134 sequences, light chain: 441 sequences), and non-S2 (heavy chain: 208 sequences, light chain: 648 sequences) mAbs from CoV-AbDab using the same germline V genes. Mutations present in at least three individuals are shown. (C) The bar plots represent the frequencies of heavy (left) and light (right) chain SHM in M15-like singleton sequences from Cohort 1, Cohort 2, and CoV-AbDab (21 sequences not previously included in panel B), as well as negative-control antibodies against malaria, influenza virus, and non-S2 epitopes from CoV-AbDab, all using the same germline V genes as above. Underneath the X-axis, a heatmap illustrates the presence or absence of the convergent SHM observed in the original M15 sequence.

Figure 3.. Influence of convergent and clonotype-enriched SHMs on binding affinity of M15.

(A) Schematic workflow for assessing the role of convergent and clonotype-enriched mutations in affinity maturation of M15 and co-occurrence of these SHMs in M15-like antibodies, created with BioRender. (B) Biolayer interferometry (BLI) is used to assess the binding of M15; single reverted mutants G51A, T57I, and S32I; triple mutant with all three mutations reverted; and the unmutated common ancestor (UCA) with S2. Rmax-apparent represents the baseline-corrected maximum binding signal reached during the association step of the assay, measured at different concentrations of S2, ranging from 166.7–16,666.7 nM. The lines represent the best-fit one-site specific binding curves. The BLI assay was performed in duplicates for each antibody (only one of the replicates shown). (C) Binding affinity values of M15, the single mutants, the triple mutant, and the UCA, computed as the dissociation constant, KD. KD values were obtained from the best-fit one-site specific binding in GraphPad Prism independently for the duplicates. (D) Biclustered heatmap representing the occurrence of convergent and clonotype-enriched SHMs in M15-like antibodies from Cohort 1 (M15), Cohort 2 (M15_C2A-F), and CoV-AbDab (M15A-Q). Biclustering was performed using k-means clustering in the ‘pheatmap’ package in R.

Figure 4.. Cryo-EM structure of M15 in complex with S2.

(A) Density map of the trimeric S2-M15 Fab complex, highlighting S2 (light blue), M15 heavy chain (magenta), M15 light chain (pink). Glycan densities are displayed in royal blue. (B) Ribbon representation of S2 protomer-M15 Fab complex structure. (C) S2-M15 interface overview (region designated within black box in panel B. S2 Helices (ribbons) are labeled to designate relative locations of heptad repeat 1 (HR1), central helix (CH), and upstream helix (UH). S2 residues are colored according to whether they form contacts with correspondingly colored heavy chain (magenta) or light chain (pink). S2 residues contacting both heavy and light chains are colored coral. Convergent and clonotype-enriched mutations are colored yellow. (D) Close-up view of M15 heavy chain (HC; left) and light chain (LC; right) binding interfaces with S2. S2 residues are colored, as in panel C, according to whether they form contacts with correspondingly colored heavy chain (magenta) or light chain (pink). Convergent and clonotype-enriched mutations are colored yellow.

Figure 5.. Prevalence of sera antibodies targeting the S2 central interface epitope in SARS-CoV-2-infected and vaccinated individuals.

(A) Schematic workflow of the protocol employed for the serum competition ELISA, created with BioRender. (B) Proportion of responders among the samples tested in a serum competition ELISA between M15 (biotinylated) and human sera from pre-pandemic samples (n = 24) or SARS-CoV-2 convalescent samples (n = 25), two doses of SARS-CoV-2 vaccination (n = 24), or four doses of SARS-CoV-2 vaccination (n = 12). Responders are identified as individuals showing competition ED50 values greater than 1. (C) The ratio of competition ED50 to anti-spike serum AUC values are depicted across individuals from the convalescent, vaccination (2X), and vaccination (4X) exposure groups. ED50/AUC values less than 10⁻²⁰ were set to a detection limit of 10⁻²⁰. Comparisons between exposure groups were performed using the Kruskal-Wallis test followed by Dunn’s multiple correction. All serum competition ELISAs and serum ELISAs were performed in duplicates, and the ED50 values and AUCs were calculated as geometric means of the replicates.