SARS-CoV-2 mRNA vaccination induces functionally diverse antibodies to NTD, RBD, and S2

Abstract

In this study we profiled vaccine-induced polyclonal antibodies as well as plasmablast-derived mAbs from individuals who received SARS-CoV-2 spike mRNA vaccine. Polyclonal antibody responses in vaccinees were robust and comparable to or exceeded those seen after natural infection. However, the ratio of binding to neutralizing antibodies after vaccination was greater than that after natural infection and, at the monoclonal level, we found that the majority of vaccine-induced antibodies did not have neutralizing activity. We also found a co-dominance of mAbs targeting the NTD and RBD of SARS-CoV-2 spike and an original antigenic-sin like backboost to spikes of seasonal human coronaviruses OC43 and HKU1. Neutralizing activity of NTD mAbs but not RBD mAbs against a clinical viral isolate carrying E484K as well as extensive changes in the NTD was abolished, suggesting that a proportion of vaccine-induced RBD binding antibodies may provide substantial protection against viral variants carrying single E484K RBD mutations.

Keywords: NTD; RBD; SARS-CoV-2; mAbs; mRNA vaccination; plasmablasts; spike.

Graphical abstract

Figure 1

Antibody responses in individuals vaccinated with mRNA-based SARS-CoV-2 vaccines (A–C) Antibody responses of convalescent individuals and vaccinees to full-length spike protein (A) and RBD (B) as measured by ELISA and neutralizing activity of the sera of the same individuals in a microneutralization assay against authentic SARS-CoV-2 (C). Convalescent individuals were grouped based on their initial antibody response (measured in a CLIA laboratory) to spike protein into +, ++, and +++. (D) Ratios between binding and neutralizing antibody levels in vaccinees and convalescent individuals. Higher ratios indicate a bias toward non-neutralizing antibodies. (E–H) show antibody responses against α-coronavirus 229E and NL63 and β-coronavirus OC43 and HKU1 spike proteins over time. Bars represent the geometric mean, error bars represent the 95% confidence intervals.

Figure S1

Comparison of binding to neutralizing titer ratios between naturally infected and vaccinated individuals (A) and full-length spike to RBD ratios (B), related to Figure 1 Statistical analysis was performed in GraphPad Prism using a one-way ANOVA with correction for multiple comparisons, significance was defined as p < 0.05.

Figure S2

Gating strategy for sorting plasmablasts from total PBMCs isolated 1 week after second immunization, related to Figure 2 and 3

Figure 2

Characterization of mAbs derived from vaccine plasmablasts (A–D) Binding of plasmablasts derived from three vaccinees (V3, V5, and V6) against full-length spike (A), RBD (B), NTD (C), and S2 (D). (E) The percentages of the respective antibodies per subject. (F and G) Neutralizing activity of the mAbs against authentic SARS-CoV-2 (F) and the proportion of neutralizing antibodies per subject is shown in (G). (H and I) Reactivity of mAbs to spike protein of human β-coronaviruses OC43 and HKU1. MBC, minimal binding concentration. All experiments except data shown in (H) and (I) were performed in duplicates and the mean of the duplicates is shown with standard deviation. For (H) and (I), a representative dataset from a singlet ELISA run is shown.

Figure 3

Characterization of bulk sorted plasmablasts via single-cell RNA sequencing (A) Uniform manifold approximation and projection (UMAP) of scRNA-seq from bulk plasmablast with recovered BCR sequences (purple) or unrecovered (gray). (B) UMAP overlay of percent of cellular population expressing MZB1, PRDM1, and XPB1. Hexbin equals 80 individual cells. (C) UMAP overlay of BCR sequences with confirmed spike binding activity. (D) Proportional composition of heavy chains genes in the spike binding sequences broken down by sample. (E) Comparison of nucleotide-level mutation frequency in immunoglobulin heavy chain variable (IGHV) genes between plasmablasts clonally related to spike-binding mAbs from SARS-CoV-2 vaccinees, plasmablasts sorted from PBMCs 1 week after seasonal influenza vaccination and found in vaccine-responding B cell clones, and naive B cells found in blood of an influenza vaccinee (left); and between plasmablasts from SARS-CoV-2 vaccinees found to be clonally related to spike-binding mAbs that were, respectively, cross-reactive and non-cross-reactive to human β-coronaviruses spike proteins (right). p values were generated using a two-sided Kruskal-Wallis test with Dunn’s post-test (left) or a Mann-Whitney U test (right).

Figure 4

Mapping of the amino acid substitutions and deletions onto the structure of the SARS-CoV-2 spike glycoprotein (A) Mutations of the three major variants of concern B.1.1.7, B.1.315, and P.1. (B) These mutations mapped onto the structure of the spike glycoprotein (model generated by superposition of PDB: 6M0J and 7C2L) (Chi et al., 2020; Lan et al., 2020). One RBD in the up conformation (red) is bound with ACE2 receptor (pink). The NTD is colored blue and the various amino acid substitutions are shown as yellow spheres. One spike protomer is shown in bold colors while the other two are colored white. (C) Competition between ACE2 and neutralizing RBD targeting mAbs PVI.V3-9 and PVI.V6-4 for binding to RBD. (D) BLI-measured binding affinities of the RBD mutants to ACE2, as well as the calculated fold change compared to wild type, are shown in the table on the right.

Figure S3

Representative Bio-Layer Interferometry binding isotherms from two independent experiments, related to Figure 4 The raw data are shown in pink and the Langmuir 1:1 kinetics fit is shown in black.

Figure S4

Binding of SARS-CoV-2 variant RBDs to ACE2, related to Figure 4 (A) ELISA curves of the RBD variants binding to human ACE2. Shown are the binding curves calculated with nonlinear regression to the arithmetic mean values from eight replicates ± SEM. The calculated steady-state K_D values ± SEM from end-point ELISA measurements and the fold-change in comparison to wild type RBD are reported in (B).

Figure 5

Binding and neutralization of SARS-CoV-2 variants (A–C) Binding of serum samples from convalescent individuals (A), vaccinees (B), and vaccine-derived mAbs (C) to a panel of RBD mutants is shown. The red line in (A) indicates the average reduction. Dotted lines in (A) and (B) indicate 100%, the line with smaller dots in (C) indicated reactivity of the anti-his coating control. For vaccinees, late samples (V1 = d89, V2 = d102, V3 = d47, V4 = d48, V5 = 49, and V6 = 48) were assayed. (D) The spike mutations of virus isolate PV14252 mapped on a structure of the SARS-CoV-2 spike protein with ACE2 (model generated by superposition of PDB: 6M0J and 7C2L) (Chi et al., 2020; Lan et al., 2020). (E and F) The inhibitory effect of vaccine serum and vaccine-derived neutralizing antibodies on both wild-type SARS-CoV-2 and PV14252. (G) Neutralizing activity of the plasmablast-derived neutralizing antibodies against wild-type, B.1.1.7, and B.1.351 virus isolates. Of note, these comparative assays were always performed side by side but sets were run by different operators and on a different Vero cell clone as the neutralization assays shown in Figure 2.