A Spike-destructing human antibody effectively neutralizes Omicron-included SARS-CoV-2 variants with therapeutic efficacy

Abstract

Neutralizing antibodies (nAbs) are important assets to fight COVID-19, but most existing nAbs lose the activities against Omicron subvariants. Here, we report a human monoclonal antibody (Ab08) isolated from a convalescent patient infected with the prototype strain (Wuhan-Hu-1). Ab08 binds to the receptor-binding domain (RBD) with pico-molar affinity (230 pM), effectively neutralizes SARS-CoV-2 and variants of concern (VOCs) including Alpha, Beta, Gamma, Mu, Omicron BA.1 and BA.2, and to a lesser extent for Delta and Omicron BA.4/BA.5 which bear the L452R mutation. Of medical importance, Ab08 shows therapeutic efficacy in SARS-CoV-2-infected hACE2 mice. X-ray crystallography of the Ab08-RBD complex reveals an antibody footprint largely in the β-strand core and away from the ACE2-binding motif. Negative staining electron-microscopy suggests a neutralizing mechanism through which Ab08 destructs the Spike trimer. Together, our work identifies a nAb with therapeutic potential for COVID-19.

Fig 1. Ab08 neutralizes SARS-CoV-2 Omicron-included variants and binds to RBD with pico-molar affinities.

(A-B) Neutralization assay of Ab08 using pseudoviruses harboring the S proteins from SARS-CoV-2 Wuhan-Hu-1 and Omicron BA.1 (A) or D614G and Omicron subvariants as indicated (B). Data are plotted as mean ± SEM (SEM: standard error of mean; n = 3 biological replicates). Numbers in brackets indicate IC₅₀ values in μg/mL. (C) Neutralization activity of Ab08 against authentic SARS-CoV-2. A Plaque Reduction Neutralization Test (PRNT) was applied. Data are plotted as mean ± SEM (n = 3, biological replicates). (D) Binding affinities of Ab08 to SARS-CoV-2 prototype and variant RBDs using bilayer interferometry (BLI). Biotinylated RBDs were immobilized on a Streptavidin sensor and Ab08 (scFv) was used as analytes at indicated concentrations (nM). Solid lines represent experimental data and dashed lines represent fitted curve.

Fig 2. Ab08 shows therapeutic efficacy in hACE2 mice infected with live SARS-CoV-2.

(A) hACE2 mice intranasally infected with 10⁵ PFU of SARS-CoV-2 were divided into two groups: the Ab08 group was intraperitoneally administered with 10 mg/kg Ab08 (n = 5) and the PBS group was injected with PBS (n = 5) at 4-hour, 28-hour and 52-hour post infection (red arrow). All mice were sacrificed on day 4 post infection. The mouse images are from Openclipart. (B) Weight change of mice was monitored daily (n = 5; Unpaired Student’s t test: *p<0.05; **p<0.01; symbols denote mean ± SEM.) (C) Percent survival was determined for the Ab08 and PBS groups. (n = 5; Kaplan-Meier survival analysis with Log-rank test: * p<0.05). (D) SARS-CoV-2 viral RNA loading in lungs of the Ab08 and PBS groups on day 4 post infection. Viral RNA copies were analyzed by real-time qPCR. (Unpaired Student’s t test: * p<0.05). (E) Representative H&E staining showing pathological changes in the mouse lungs from different groups (Top: Non-infected; Middle: PBS; Bottom: Ab08). The PBS group exhibited severe pneumonia with blocked terminal bronchioles, fibroplasia, and organization (black arrow), and peribronchial and perivascular infiltration (yellow arrow). Scale bars indicate 200 μm.

Fig 3. Structural characterization informs mechanism for the relatively broad activity of Ab08.

(A) The overall structure of the Ab08 (scFv) (cartoon) in complex with the high-chair-shaped RBD (grey surface). The Ab08 epitope is colored green. RBD-interacting CDRs are color-coded as indicated. (B) Interactions between RBD (grey) and Ab08. CDR residues are color-coded as in A. Ab08 residues are labeled black and RBD residues are labeled grey. A prime symbol indicates residues from the light chain. Dash lines indicate distances within 3.6 Å. (C) The β-sheet interaction between Ab08 and RBD. Backbone- and side-chain-mediated H-bonds are indicated by black and orange dash lines, respectively. (D, E) Ab08 is compatible with ACE2-binding. (D) Structural alignment of ACE2-RBD (PDB ID 6M0J) and Ab08-RBD (this work) reveals no clashes between ACE2 (yellow) and Ab08 (red and blue). (E) Ab08 and ACE2 simultaneously bind RBD. A sensor immobilized with RBD was soaked in 100 nM of ACE2 before being further soaked in ACE2-containing buffer with (black) or without (red) 100 nM of Ab08 for BLI signal recording. As a control, the Ab08-RBD interaction was monitored in the absence of ACE2 (blue). (F) The distribution of the Omicron mutations. Of the 15 mutations, only one (Q493R) occurs at the epitope but the relative long Cα distance (8.4 Å) between RBD Gln493 and Ab08 Leu55 suggests neglectable impact of Q493R on Ab08-RBD interactions.

Fig 4. Ab08 neutralizes SARS-CoV-2 by destructing S trimer.

(A-B) Structural alignment of Ab08 to the S trimer structure (PDB ID 7WLY) reveals severe clashes with the N-terminal domain (NTD) in cases of both the “down”-RBD (A) and the “up”-RBD conformation (B). The epitope of Ab08 is colored green. The three protomers of S and the RBD of the protomer A are color-coded as indicated in (i). In (ii), the RBD from protomer A is shown as surface representation, and Ab08 (magenta) and the NTD from protomer C (orange) are shown as ribbon representations. The two glycans are shown as cyan sticks. In (iii), the disruption of the interactions between NTD N165-linked glycans (Cyan) and RBD by Ab08 is highlighted. (C-E) 2D class averages of Negative-staining particles of S-2P showing typical S trimer features (C) were lost upon incubation with Ab08 (D), but not with an RBM-targeting control mAb (CB6, E) [22]. Raw images are in S7 Fig. In E, the 2D classes show binding of two CB6 molecules per S trimer and CB6 may induce more “up”-RBD population, a phenomenon that has been reported for other RBM-targeting antibodies [44].