Comprehensive mapping of binding hot spots of SARS-CoV-2 RBD-specific neutralizing antibodies for tracking immune escape variants

Abstract

Background: The receptor-binding domain (RBD) variants of SARS-CoV-2 could impair antibody-mediated neutralization of the virus by host immunity; thus, prospective surveillance of antibody escape mutants and understanding the evolution of RBD are urgently needed.

Methods: Using the single B cell cloning technology, we isolated and characterized 93 RBD-specific antibodies from the memory B cells of four COVID-19 convalescent individuals in the early stage of the pandemic. Then, global RBD alanine scanning with a panel of 19 selected neutralizing antibodies (NAbs), including several broadly reactive NAbs, was performed. Furthermore, we assessed the impact of single natural mutation or co-mutations of concern at key positions of RBD on the neutralization escape and ACE2 binding function by recombinant proteins and pseudoviruses.

Results: Thirty-three amino acid positions within four independent antigenic sites (1 to 4) of RBD were identified as valuable indicators of antigenic changes in the RBD. The comprehensive escape mutation map not only confirms the widely circulating strains carrying important immune escape RBD mutations such as K417N, E484K, and L452R, but also facilitates the discovery of new immune escape-enabling mutations such as F486L, N450K, F490S, and R346S. Of note, these escape mutations could not affect the ACE2 binding affinity of RBD, among which L452R even enhanced binding. Furthermore, we showed that RBD co-mutations K417N, E484K, and N501Y present in B.1.351 appear more resistant to NAbs and human convalescent plasma from the early stage of the pandemic, possibly due to an additive effect. Conversely, double mutations E484Q and L452R present in B.1.617.1 variant show partial antibody evasion with no evidence for an additive effect.

Conclusions: Our study provides a global view of the determinants for neutralizing antibody recognition, antigenic conservation, and RBD conformation. The in-depth escape maps may have value for prospective surveillance of SARS-CoV-2 immune escape variants. Special attention should be paid to the accumulation of co-mutations at distinct major antigenic sites. Finally, the new broadly reactive NAbs described here represent new potential opportunities for the prevention and treatment of COVID-19.

Keywords: Escape variants; Neutralizing antibodies; RBD antigenic sits; SARS-CoV-2.

Fig. 1

Characterization of SARS-CoV-2 natural infection-induced RBD-specific mAbs. a Antibody binding activity (EC₅₀) with native and denatured SARS-CoV-2 RBD as well as native RBD of SARS-CoV was measured by ELISA. Upper panel, four ranges per donor; lower panel, percentage of mAbs from each donor with the indicated EC₅₀ range. N.B., non-binding activity. b SARS-CoV-2 pseudovirus neutralization potency (IC₅₀) (left). Percentage of antibodies with indicated neutralization potencies (right). Results are derived from a single experiment performed in triplicate. c ACE2 blocking activity (IC₅₀) (left). Percentage of antibodies with indicated receptor-blocking potencies (right). For b and c, N.N., non-neutralization or non-blocking activity. The data represent one representative experiment of two independent experiments. d Distribution of heavy-chain variable (VH) germline genes of RBD-specific NAbs

Fig. 2

Determination and characterization of the antigenic sites on RBD. a Mapping of binding sites of a panel of RBD-specific NAbs by global alanine scanning. Thirty-three amino acid positions were identified in four antigenic sites (1–4) on RBD as main targets for RBD-specific NAbs. Degree of binding reduction was defined as percentage by OD₄₅₀ of each mutant relative to OD₄₅₀ of RBD wildtype and is represented as a heatmap from white (reduction) to blue (no impact). The data are representative of at least two independent experiments. b Location of four distinct antigenic sites on the RBD region (PDB ID: 6M0J). The color-coding scheme is described as follows: site 1 (orange), site 2 (green), site 3 (slate blue), site 4 (red); ACE2 (wheat). Top and down are shown from different angles. The key hot spots targeted by NAbs are shown. c Key residues for VH3-53/3-66 dominant NAb recognition. Binding fold change was calculated as follows: EC₅₀ of RBD mutant/EC₅₀ of RBD wildtype. d Antibodies are grouped according to neutralization potency and colored by the usage frequency of each antigenic site. Each antibody was tested for competition with a panel of known antibodies and assigned to an antigenic site based on the competition profile. e Percentage of SARS-CoV-2 and SARS-CoV crossing reactive antibodies targeting each antigenic site. The data represent one representative experiment of two independent experiments

Fig. 3

The mutations that impair the global folding of RBD are evolutionally conserved. a The residues of which alanine substitution impaired RBD antigenic conformation and ACE2 binding ability are shown in a surface representation (PDB ID: 6M0J). Mutants located in the core region are highlighted in green and RBM are in cyan. Yellow sticks indicate disulfide bridges. b Conservation of residues that are essential for RBD folding across clades of sarbecoviruses. Human and animal SARS-related coronaviruses were classified by clades. SARS-CoV-2 genome sequences (n = 364,409) retrieved from GISAID and Genbank on January 19, 2021 (n = 11,839) and human SARS-CoV genome sequences (n = 200) from Genbank were used to annotate variants of the spike glycoprotein. Dashes indicate identity to SARS-CoV-2 consensus residues. Variants found in at least two sequences are parenthesized. c Top predicated destabilizing mutants by structural analysis. The score of MAESTRO and DUET is for predicted impact on stability from the RBD structure (PDB ID: 7C01). Average percent binding versus wildtype RBD for conformation-dependent RBD antibodies are shown

Fig. 4

The impacts of natural mutations at antigenic sites on binding and neutralizing activities of RBD-specific NAbs. a Binding activity of RBD natural mutants with a panel of NAbs was evaluated by ELISA. Degree of binding reduction was defined as percentage by OD₄₅₀ of each mutant relative to OD₄₅₀ of RBD wildtype is represented as a heatmap from white (reduction) to blue (no impact). The data are representative of at least two independent experiments. b Binding of 18 purified mammalian expressed RBD single-point mutants and two co-mutations with a panel of NAbs. Binding fold change was calculated as follows: EC₅₀ of RBD mutant/EC₅₀ of RBD wildtype. N.B., not binding. c Details of the structural interaction between CB6 and RBD (PDB ID: 7C01). d Details of the structural interaction between S309 and RBD (PDB ID: 6WPT). e Details of the structural interaction between CR3022 and RBD (PDB ID:7A5S). For c–e, polar interactions are indicated by yellow dashed lines. Cyan, heavy chain; light blue, light chain; gray, RBD. Key residues on RBD highlighted in orange sticks; key residues on heavy chain highlighted in cyan sticks. f Neutralization of 19 pseudotyped variants by a panel of RBD-specific NAbs. Neutralization fold change was calculated as follows: IC₅₀ of pseudotyped variants/IC₅₀ of the wildtype. IC₅₀ values were calculated from three independent experiments. The data represent one representative experiment of two independent experiments

Fig. 5

The binding affinity between antibody escape mutants and ACE2. a Affinity measurement of purified RBD mutants for binding to immobilized ACE2-Fc by surface plasmon resonance (SPR). The K_on and K_off were determined by BIAcore and the KD were computed as K_off/K_on. Neutralization fold change was calculated as follows: KD value of RBD wildtype/KD value of RBD mutants. b The residues that are important for resistance to antibodies are presented on the interface of the ACE2 and RBD. The position of mutants that enhance ACE2 binding affinity are highlighted in red (affinity fold change ≥ 2.5); comparable to wildtype are in orange (affinity fold change between 2.5 and 0.4); reduce ACE2 binding affinity are in slate blue (affinity fold change ≤ 0.4). The data represent one representative experiment of two independent experiments

Fig. 6

Effect of natural mutations on neutralizing activity of convalescent plasma. a Neutralization potency of nine convalescent plasma against pseudotyped variants. The data are presented as the highest plasma dilution giving a ≥ 50% inhibition of pseudotyped virus infection (NT₅₀). NT₅₀ values were calculated from three independent experiments. b The neutralizing antiserum titer against pseudotyped RBD variants decreases relative to wildtype pseudotyped virus. Neutralization fold change was calculated as follows: NT₅₀ of the wildtype/NT₅₀ of pseudotyped variants. The data represent one representative experiment of two independent experiments