Structural basis for broad sarbecovirus neutralization by a human monoclonal antibody

Abstract

The recent emergence of SARS-CoV-2 variants of concern (VOC) and the recurrent spillovers of coronaviruses in the human population highlight the need for broadly neutralizing antibodies that are not affected by the ongoing antigenic drift and that can prevent or treat future zoonotic infections. Here, we describe a human monoclonal antibody (mAb), designated S2X259, recognizing a highly conserved cryptic receptor-binding domain (RBD) epitope and cross-reacting with spikes from all sarbecovirus clades. S2X259 broadly neutralizes spike-mediated entry of SARS-CoV-2 including the B.1.1.7, B.1.351, P.1 and B.1.427/B.1.429 VOC, as well as a wide spectrum of human and zoonotic sarbecoviruses through inhibition of ACE2 binding to the RBD. Furthermore, deep-mutational scanning and in vitro escape selection experiments demonstrate that S2X259 possesses a remarkably high barrier to the emergence of resistance mutants. We show that prophylactic administration of S2X259 protects Syrian hamsters against challenges with the prototypic SARS-CoV-2 and the B.1.351 variant, suggesting this mAb is a promising candidate for the prevention and treatment of emergent VOC and zoonotic infections. Our data unveil a key antigenic site targeted by broadly-neutralizing antibodies and will guide the design of pan-sarbecovirus vaccines.

Extended Data Fig. 1.. S2X259 Fab binding to recombinant sarbecovirus RBDs, prefusion SARS-CoV-2 S ectodomain trimer and RBD variants.

a, S or RBD antigens were captured on the sensor chip surface and binding to S2X259 Fab at 11, 33, 100, and 300 nM was monitored successively, in single-cycle kinetics format, by surface plasmon resonance. All data have been fit to a 1:1 binding model and the equilibrium dissociation constant (K_D) is reported. For the S-binding data, we report an apparent K_D (K_D,app) since kinetics are affected by conformational dynamics between open and closed RBD states. The colouring scheme matches the phylogenetic tree in Figure 1a. b, Biolayer interferometry binding analysis of the S2X259 Fab to wildtype or VOC SARS-CoV-2 biotinylated RBDs immobilized at the surface of SA biosensors.

Extended Data Fig. 2.. Neutralization tests with control mAbs.

S2E12 and S230 neutralizing activity against VSV pseudotypes harbouring the S glycoprotein of SARS-CoV- and SARS-CoV-2 related strains and VOC.

Extended Data Fig. 3.. CryoEM data processing and validation of S2X259-bound SARS-CoV-2 S datasets.

a-b, Representative electron micrographs (a) and class averages (b) of SARS-CoV-2 S in complex with the S2X259 Fab. Scale bar of the micrograph: 500 Å. Scale bar of the class averages: 100 Å. c, Gold-standard Fourier shell correlation curves for the S trimer bound to three S2X259 Fabs (solid black line) and the locally refined reconstruction of the RBD/S2X259 variable domains (dashed black line). The 0.143 cut-off is indicated by a horizontal dashed grey line. d-e, Local resolution map for the open S trimer bound to three S2X259 (d) and the locally refined reconstruction of the RBD/S2X259 variable domains (e). f, CryoEM data processing flow-chart.

Extended Data Fig. 4.. The S2X259 angle of approach for binding to the SARS-CoV-2 RBD allows to circumvent the SARS-CoV N357 glycan present in all sarbecovirus RBDs except SARS-CoV-2.

Ribbon diagram showing a superimposition of the S2X259-bound and S2A4-bound (PDB 7JVA) SARS-CoV-2 RBD. The SARS-CoV glycan at position N357 was modelled based on the S230-bound SARS-CoV S structure (PDB 6NB6) and is predicted to sterically hinder S2A4 binding (red star) but not S2X259. The mAb light and heavy chains are coloured magenta and purple (S2X259) or light and dark green (S2A4). N-linked glycans are rendered as blue spheres.

Extended Data Fig. 5.. S2X259 allows combination with site I and IV-targeting mAbs.

a, View of site I-targeting S2E12 (pink), site II-targeting S2X259 (magenta), and site IV-targeting S309 (purple) mAb binding to SARS-CoV-2 RBD (light blue). b-c, Competition binding assays for S2X259 vs site I-targeting S2E12 (b) and site IV-targeting S309 (c) mAbs on SARS-CoV-2 RBD as measured by biolayer interferometry. One independent experiment out of two is shown.

Extended Data Fig. 6.. S2X259 has a high barrier for the emergence of resistance mutants.

a, FACS gates used in DMS to select escape variants. Top row: yeast controls expressing unmutated SARS-CoV-2 RBD labelled at relevant S2X259 concentrations for setting of selection gates. Bottom row: fraction of cells in SARS-CoV-2 mutant libraries falling into the antibody-escape bin. b, Correlation in site-level (top, sum of escape fractions for mutations at a site) and mutation-level (bottom) escape between independently generated and assayed RBD mutant libraries. c, Line plot of escape mutants along all positions in the SARS-CoV-2 RBD (left). Pink lines indicate sites that escape S2X259 binding illustrated at the mutation-level in logoplots (right). In logoplots, the height of a letter scales with its escape fraction. Letters are coloured according to their deleterious consequences for ACE2 binding (middle) or RBD expression (right) as determined in prior deep mutational scanning experiments. d, Plaque assay on Vero cells with no antibody (left) or S2X259 (right) in the overlay to isolate escape mutants (red arrow). Data are representative of three independent experiments. e, S2X259 in vitro neutralizing activity against SARS-CoV-2 S VSV pseudotyped mutants. For each mutant the fold change of the IC50 geometric mean vs SARS-CoV-2 S WT is reported. *Q506K displayed a 10-fold reduction in viral entry in comparison to the other mutants. Results from two independent experiments are reported.

Extended Data Fig. 7.. S2X259 epitope conservation across sarbecovirus clades.

Protein sequence alignment of representative sarbecovirus RBDs with strictly conserved residues shown as dots. Overall conservation is represented as a bar plot. Positions are based on SARS-CoV-2 RBD. Residues determined to be important for S2X259 binding are denoted in dark green. Substitutions at positions D405 and G504 are indicated in pink and blue/orange, respectively. Additional residues representing extended epitope, are denotated grey. Different clades within the sarbecovirus subgenus are overlayed in grey (clade 1a), red (clade 1b), green (clade 2), and light blue (clade 3).

Extended Data Fig. 8.. S2X259 Fc-mediated activation of FcγRIIa and FcγRIIIa in vitro.

a-b, NFAT-driven luciferase signal induced in Jurkat cells stably expressing FcγRIIa H131 (a) variant or FcγRIIIa V158 (b) variant by S2X259 binding to full-length wild-type SARS-CoV-2 S on ExpiCHO target cells. c-d, NFAT-driven luciferase signal induced in Jurkat cells stably expressing FcγRIIa H131 (c) or FcγRIIIa V158 (d) variants by S2X259 binding to uncleavable full-length pre-fusion stabilized SARS-CoV-2 S (unable to release the S1 subunit) transiently expressed in ExpiCHO cells. SE12, S2M11, S309, S309-GRLR mAbs are included as controls.

Extended Data Fig. 9.. Prophylactic administration of S2X259 mAb protects hamsters against prototypic (Wuhan-1 related) SARS-CoV-2 challenge.

a-b, Viral RNA loads (a) and replicating virus titres (b) in the lungs of Syrian hamsters 4 days post-intranasal infection with prototypic SARS-CoV-2. Results for one independent experiment are shown. c-d, Viral RNA loads and replicating virus titers in the lung 4 days post infection plotted as a function of serum mAb concentrations before infection (day 0). Irrelevant mAb n=3; S2X259 4 mg/kg n=4.

Fig. 1.. Identification of a broadly neutralizing sarbecovirus mAb.

a, Phylogenetic tree of sarbecovirus RBDs constructed via maximum likelihood analysis of amino acid sequences retrieved from GISAID and GenBank. Cross-reactivity within the sarbecovirus subgenus for S2E12, S309, and ADG-2 is included for comparison. b, Flow cytometry analysis of S2X259 cross-reactivity with a panel of 30 S glycoproteins representative of sarbecovirus clades 1a, 1b, 2, and 3 as well as SARS-CoV-2 VOCs. c, S2X259 binding to RBDs representative of the different sarbecovirus clades and SARS-CoV-2 variants as measured by ELISA. d, S2X259-mediated neutralization of SARS-CoV-2-Nluc authentic virus and SARS-CoV-2 S MLV- pseudotyped virus. e-f S2X259-mediated neutralization of VSV pseudotypes harbouring SARS-CoV-2 S from isolates representing the B.1.1.7, B.1.351, P.1 and B.1.429 VOC (e) as well as single RBD mutants (f). g-h, S2X259-mediated neutralization of VSV pseudotypes harbouring SARS-CoV-related (clade 1a, g) or SARS-CoV-2-related (clade 1b, h) S glycoproteins. One independent experiment out of at least two is shown. Error bars indicate standard deviation of duplicates or triplicates.

Fig. 2.. The S2X259 broadly neutralizing sarbecovirus mAb recognizes RBD antigenic site II.

a-b, CryoEM structure of the prefusion SARS-CoV-2 S ectodomain trimer with three S2X259 Fab fragments bound to three open RBDs viewed along two orthogonal orientations. c. The S2X259 binding pose involving contacts with multiple RBD regions. Residues corresponding to prevalent RBD mutations are shown as red spheres. d-e, Close-up views showing selected interactions formed between S2X259 and the SARS-CoV-2 RBD. In panels a-e, each SARS-CoV-2 S protomer is coloured distinctly (cyan, pink and gold) whereas the S2X259 light and heavy chain variable domains are coloured magenta and purple, respectively. N-linked glycans are rendered as blue spheres in panels a-c.

Fig. 3.. S2X259 is resilient to a broad spectrum of escape mutations.

a, Complete mapping of mutations reducing S2X259 binding using yeast-displayed RBD deep mutational scanning (DMS). Mean mutation effect on ACE2 affinity, RBD folding, and contribution to S2X259 binding for substitutions at each position in the S2X259 epitope is reported. * mutations introducing a N-linked glycosylation sites that may not be tolerated in full spike. b, Plaque assays performed to validate the VSV-SARS-CoV-2 G504D mutant in Vero cells in the presence (right) or absence (left) of S2X259 in the overlay. Representative image of two independent experiments is shown. c, Frequency of mutants at positions encompassing S2X259 epitope based on SARS-CoV-2 genome sequences available on GISAID as of April 2021. S2X259 neutralizing activity against selected mutations is reported. d, Zoomed-in view of the S2X259/RBD interface showing that the G504D substitution would disrupt mAb binding due to steric hindrance.

Fig. 4.. S2X259 blocks ACE2 engagement, promotes shedding of the S1 subunit and protects hamsters against B.1.351 SARS-CoV-2 challenge.

a, S2X259 (purple/pink) and ACE2 (dark green) bind partially overlapping binding sites on the SARS-CoV-2 RBD (blue). b, Pre-incubation of serial dilutions of S2X259 with SARS-CoV-2 (red) or the SARS-CoV (black) RBDs prevents binding to immobilized human ACE2 (hACE2) ectodomain in ELISA. c, mAb-mediated S1 subunit shedding from cell-surface expressed SARS-CoV-2 S as determined by flow-cytometry. S2E12 was included as positive control whereas S2M11 was included as negative control. d-e, Quantification of viral RNA loads (d) and replicating virus titres (TCID50) (e) in the lungs of Syrian hamsters 4 days post intranasal challenge with B.1.351 SARS-CoV-2 VOC following prophylactic administration of S2X259 at 1 mg/kg (n=6), 4 mg/kg (n=6), and in combination with S309 (1+1 mg/kg, n=6). Mann–Whitney test was used for statistical analysis of significance. *p < 0.05, **p < 0.01. f, Correlation between concentration of mAbs measured in the serum before infection (day 0) infectious virus (TCID50) in the lung 4 days post infection. Data from one independent experiment are presented.