Identification of SARS-CoV-2 spike mutations that attenuate monoclonal and serum antibody neutralization

Abstract

Neutralizing antibodies against the SARS-CoV-2 spike (S) protein are a goal of COVID-19 vaccines and have received emergency use authorization as therapeutics. However, viral escape mutants could compromise efficacy. To define immune-selected mutations in the S protein, we exposed a VSV-eGFP-SARS-CoV-2-S chimeric virus, in which the VSV glycoprotein is replaced with the S protein, to 19 neutralizing monoclonal antibodies (mAbs) against the receptor-binding domain (RBD) and generated 50 different escape mutants. Each mAb had a unique resistance profile, although many shared residues within an epitope of the RBD. Some variants (e.g., S477N) were resistant to neutralization by multiple mAbs, whereas others (e.g., E484K) escaped neutralization by convalescent sera. Additionally, sequential selection identified mutants that escape neutralization by antibody cocktails. Comparing these antibody-mediated mutations with sequence variation in circulating SARS-CoV-2 revealed substitutions that may attenuate neutralizing immune responses in some humans and thus warrant further investigation.

Keywords: ACE2 receptor decoys; COVID-19 vaccines; SARS-CoV-2 escape mutants; convalescent sera; coronaviruses; monoclonal antibodies; receptor-binding domain.

Graphical abstract

Figure 1

VSV-SARS-CoV-2 escape mutant isolation (A) Outline of escape mutant selection experiment. 2B04 and a control anti-influenza virus mAb were tested for neutralizing activity against VSV-SARS-CoV-2. The concentration of 2B04 added in the overlay completely inhibited viral infection (middle panel). Data are representative of two independent experiments. Plaque assays were performed to isolate the VSV-SARS-CoV-2 escape mutant on Vero E6 TMPRSS2 cells (red arrow indicated). Plaque assays with 2B04 in the overlay (bottom plaque in the right panel); plaque assays without Ab in the overlay (top plaque in the right panel). Data are representative of three independent experiments. (B) Schematic of S gene, which underwent Sanger sequencing to identify mutations (left panel). For validation, each VSV-SARS-CoV-2 mutant was tested in plaque assays with or without 2B04 in the overlay on Vero cells (right panel). Representative images of two independent experiments are shown. (C) Neutralizing mAbs. ^a The order of immunogens used to immunize the mice, as described in STAR methods. ^b Neutralization of SARS-CoV-2 by each mAb was assessed by focus-reduction neutralization test. The half-maximal effective concentration (EC₅₀ value) was determined by nonlinear regression. Results are the geometric mean from three to four independent experiments. ^c mAb was identified as mouse IgG1 and expressed as human IgG1.

Figure 2

Mapping of escape mutations The surface model of RBD (from PDB 6M0J) is depicted, and contact residues of the SARS-CoV-2 RBD-hACE2 interfaces are colored in brown. Amino acids whose substitution confers resistance to each mAb in plaque assays are indicated for 2B04 (green), 2H04 (lemon), 1B07 (blue), SARS2-01 (yellow), SARS2-02 (teal), SARS2-07 (tangerine), SARS2-16 (violet), SARS2-19 (red), SARS2-32 (fuschia), and SARS2-38 (magenta). See Figures S1 and S2.

Figure 3

Map of cross-neutralizing activity of VSV-SARS-CoV-2 mutants and neutralization potency of hACE2 decoy receptors against each VSV-SARS-CoV-2 mutant (A) Neutralization of VSV-SARS-CoV-2 mutants was evaluated by plaque assays. Degree of resistance was defined as percentage by expressing the number of plaques formed by each mutant in the presence versus absence of antibody and is represented as a heatmap from white (low degree of resistance) to red (high degree of resistance). Representative images of two independent experiments are shown in Figure S3. (B) Neutralization assay of VSV-SARS-CoV-2 mutants in the presence of hACE2-Fc. Virus was incubated with mACE2 or hACE2 at concentrations ranging from 9 ng/mL to 20 μg/mL for 1 h a 37°C, and cells were scored for infection at 7.5 h post-inoculation by automated microscopy. IC₅₀ values were calculated for each virus-hACE2 combination from three independent experiments (^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001, ^∗∗∗∗p < 0.0001; one-way ANOVA with Dunnett’s post-test; error bars indicate SEM). (C) Representative neutralization curves of wild-type and F486S mutant VSV-SARS-CoV-2 with hACE2-Fc and mACE2-Fc. Error bars represent the SEM. Data are representative of three independent experiments. Neutralization curves are provided in Figure S4.

Figure 4

Neutralization potency of human serum against each VSV-SARS-CoV-2 mutant (A) Neutralization potency of four human sera against VSV-SARS-CoV-2 mutants. IC₅₀ values were calculated from three independent experiments. Neutralization potency is represented as a rainbow color map from red (most potent with low IC₅₀) to violet (less potent with high IC₅₀). LOD indicates limit of detection (1:80). (B) Representative neutralization curves of wild-type, S477N, and E484A mutant with four different human sera. Error bars represent SEM. Data are representative of three independent experiments. (C) Neutralization potency of additional 16 human sera against VSV-SARS-CoV-2 mutants. IC₅₀ values were calculated from one independent experiment each. Neutralization potency is represented as a rainbow color map from red (most potent with low IC₅₀) to violet (less potent with high IC₅₀). Neutralization curves are provided in Figure S5. (D) Serum samples from 18 individuals were collected at different time points post-onset of COVID-19 symptoms and screened using two ELISA assays (Euroimmun or Epitope). The serum identifier numbers in the first column correspond to those of Figures 4 and S5. IgG index values were calculated by dividing the optical density (O.D.) of the serum sample by a reference O.D. control, and ratios were interpreted using the following criteria as recommended by the manufacturer: negative (−) < 0.8, indeterminate (+/−) 0.8–1.1, and positive (+) ≥ 1.1.

Figure 5

Mapping of additional VSV-SARS-CoV-2 escape mutants The surface model of RBD (from PDB 6M0J) is depicted, and contact residues of the SARS-CoV-2 RBD-hACE2 interfaces are colored in brown. Amino acids whose substitution confers resistance to each mAb in the plaque assays are indicated for SARS2-21 (lime), SARS2-22 (green), SARS2-23 (blue), SARS2-31 (yellow), SARS2-34 (cyan), SARS2-55 (orange), SARS2-58 (magenta), SARS2-66 (red), and SARS2-71 (pink). See Figures S6 and S7.

Figure 6

Position and frequency of RBD amino acid substitutions in SARS-CoV-2 (A) RBD amino acid substitutions in currently circulating SARS-CoV-2 viruses isolated from humans. For each site of escape, we counted the sequences in GISAID with an amino acid change (323,183 total sequences at the time of the analysis). Variant circulating frequency is represented as a rainbow color map from red (less circulating with low frequency) to violet (most circulating with high frequency). A black cell indicates that the variant has not yet been isolated from a patient. A rainbow cell with cross indicates that the variant has been isolated from a patient but does not appear in those 50 escape mutants. (B) Location of natural sequence variation within the RBD. The RBD is modeled as a surface representation. Variant frequency is rainbow colored, as in (A). Black coloration indicates that variation at that residue has not yet been isolated.

Figure 7

Sequential selection of 2B04 and 2H04 escape mutants (A) Plaque assays were performed to isolate the VSV-SARS-CoV-2-S wild-type, E484A, E484K, and F486S escape mutant on Vero E6 TMPRSS2 cells in the present of the indicated mAb in the overlay. Representative images of three independent experiments are shown. (B) The surface model of RBD (from PDB 6M0J) is depicted, and contact residues of the SARS-CoV-2 RBD-hACE2 interfaces are colored in brown. 2B04 escape mutants including E484A, E484K, and F486S are indicated in green. Amino acids whose substitution confers resistance to 2H04 in the plaque assays are indicated in lemon. (C) Wild-type and sequentially identified double mutants were tested for neutralizing activity using a plaque assay with the indicated mAb in the overlay. mAb concentrations added were the same as those used to select the escape mutants. Representative images of two independent experiments are shown.