Fig 1.
Construction of an Antibody Surface Display System in HEK 293T.
(A) Schematic diagram of antibody scFv surface display system on mammalian cells. The VL and VH domain of scFv receptor construct used in anti-CD19 CAR was replaced with a SARS-CoV-2 neutralizing antibody REGN10987. VH, variable region of Ig heavy-chain; VL, variable region of Ig light-chain. (B) Confocal microscopic images of REGN10987 scFv displayed on HEK 293T cells. HEK 293T cells were infected by lentivirus packaged with a plasmid directing surface-expression of REGN10987 scFv. Infected cells (red) were fixed with DAPI nuclear staining followed by detection with biotinylated RBD, and then with streptavidin-FITC (green). Merged staining patterns are shown. Scale bar: 20 μm. (C) RBD bound to REGN10987 scFv-expressing HEK 293T cells at different titers. Concentration of RBD was from 0 nM to 1,000 nM. RBD-FITC signal was measured by flow cytometry. (D) Confocal microscopic images of mutated REGN10987 scFv displaying on HEK 293T cells. Several key amino acids in CDRs were mutated into alanine. Mutated amino acid was showed in S1B Fig. Transfected cells (red) were fixed with DAPI nuclear staining followed by detection with biotinylated RBD, and then with streptavidin-FITC (green). Scale bar: 100 μm. (E) RBD-His added to RBD-FITC competitively bound to REGN10987 scFv-expressing HEK 293T cells. Concentration of RBD-FITC was 10 nM, while different titers of RBD-His was from 0 nM to 10,000 nM. RBD-FITC signal was measured by flow cytometry.
Fig 2.
High-throughput Saturation Mutagenesis Screen.
(A) A novel method for the generation of primers for high-throughput saturation mutagenesis. High-throughput chip-synthesized oligo was used to produce saturation mutagenesis primers of REGN10987 scFv. (B) Flow chart of obtaining REGN10987 scFv saturation mutagenesis library. High-throughput saturation mutagenesis primers were used to construct a full-length REGN10987 scFv lentiviral vector. (C) Timeline for establishment and screening of mutagenesis library. REGN10987 scFv vector-packaged lentivirus (MOI < 0.3) was transduced to HEK 293T cells and a mammalian cell mutant library (P0) was obtained after one round of FACS. After the incubation with SARS-CoV-2 RBD-FITC, a second round of FACS yielded RBD positive and negative groups (P1, N).
Fig 3.
Acquisition of Mutant scFv With a Higher Affinity against Wild-type, Beta and Delta SARS-CoV-2 using a HEK 293T Antibody Surface Display Library.
(A) REGN10987 scFv saturated mutagenesis heatmap of affinity binding to WT-RBD, Beta-RBD and Delta-S1. Fitness scores based on the average Log2 enrichment ratios from several replications of the WT-RBD (P2/P0), Beta-RBD (BP/BP0) and Delta-S1 (DP3/DP0) sorts were plotted from depletion or deleterious (orange) to enriched (blue). Positions on REGN10987 scFv were shown on the horizontal-vertical axis, and amino acid substitutions were indicated on the vertical axis. VL and VH of REGN10987 were shown in light green and yellow, respectively. CDRs of VL were shown in purple and CDRs of VH were shown in pink. Four replications for WT-RBD, three replications for Beta-RBD, two replications for Delta-S1. (B) Magnified views of CDR positions of VL (upper) and VH (lower). Positions on REGN10987 scFv were shown on the horizontal-vertical axis, and amino acid substitutions were indicated on the vertical axis. CDRs of VL were shown in purple and CDRs of VH were shown in pink. (C) Volcano plots of increased and decreased amino acid mutations in binding affinity to WT-RBD (P2/P0), Beta-RBD (BP/BP0) and Delta-S1 (DP3/DP0). Representative mutations were highlighted in red. Data were generated from n = four independent experiments for WT-RBD, three for Beta-RBD, two for Delta-S1. FC: relative fold change (mean frequency (after screening)/mean frequency (before screening)). The p-value was calculated using a two-sided Student’s t-test. (D) Structure of screened mutation sites with high affinity on REGN10987. Structure (PDB: 6XDG) of REGN10987 bound to RBD (cyan). VL and VH of REGN10987 were shown in light green and yellow, respectively. CDRs of VL were shown in purple and CDRs of VH were shown in pink. S97, D54, N57 and L106 were labelled in red.
Fig 4.
Verification of Mutant scFvs using HEK 293T Antibody Surface Display System.
(A) Verification of single amino acid substitution REGN10987 scFv to WT-RBD and Beta-RBD using HEK 293T antibody surface display system (left). Data were normalized by REGN10987 scFv RBD binding fluorescence. Data are mean ± SEM, n = 2 replicates. Unpaired t test was used to analyze differences between groups. * p < 0.05, ** p < 0.01, *** p < 0.001. Right: eleven kinds of mutated combination of different single amino acid substitution. (B) Verification of single amino acid substitution REGN10987 scFv to WT-RBD and Delta-S1 using HEK 293T antibody surface display system (lower). Data were normalized by REGN10987 scFv RBD binding fluorescence. Data are mean ± SEM, n = 2 replicates. Unpaired t test was used to analyze differences between groups. * p < 0.05, ** p < 0.01. Upper: three kinds of mutated combination of different single amino acid substitution. (C) Binding of combined amino acid substitution REGN10987 scFv to WT-RBD (left) and Beta-RBD (right) using HEK 293T antibody surface display system. Data were normalized by REGN10987 scFv RBD binding fluorescence. Data are mean ± SEM, n = 2 replicates. Unpaired t test was used to analyze differences between groups. * p < 0.05, ** p < 0.01, *** p < 0.001. (D) Binding of combined amino acid substitution REGN10987 scFv to WT-RBD (left) and Delta-S1 (right) using HEK 293T antibody surface display system. Data were normalized by REGN10987 scFv RBD binding fluorescence. Data are mean ± SEM, n = 2 replicates. Unpaired t test was used to analyze differences between groups. *** p < 0.001.
Fig 5.
Broad Efficient Neutralization level of Optimized Antibodies.
EC50 values of optimized antibodies prevalent SARS-CoV-2 spike-pseudotyped viruses. The entry of SARS-CoV-2 and its variants (Alpha, Beta, Gamma, Delta, Lambda or Mu) into cell lines expressing hACE2 was blocked with monoclonal antibodies. Unpaired t-test was used to analyze differences between groups. * p < 0.05, ** p < 0.01, *** p < 0.001. All results were shown as mean ± SEM. In each experiment, the infection assay was performed in duplicate wells.