Structural analysis of receptor engagement and antigenic drift within the BA.2 spike protein

Abstract

The BA.2 sub-lineage of the Omicron (B.1.1.529) severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) variant rapidly supplanted the original BA.1 sub-lineage in early 2022. Both lineages threatened the efficacy of vaccine-elicited antibodies and acquired increased binding to several mammalian ACE2 receptors. Cryoelectron microscopy (cryo-EM) analysis of the BA.2 spike (S) glycoprotein in complex with mouse ACE2 (mACE2) identifies BA.1- and BA.2-mutated residues Q493R, N501Y, and Y505H as complementing non-conserved residues between human and mouse ACE2, rationalizing the enhanced S protein-mACE2 interaction for Omicron variants. Cryo-EM structures of the BA.2 S-human ACE2 complex and of the extensively mutated BA.2 amino-terminal domain (NTD) reveal a dramatic reorganization of the highly antigenic N1 loop into a β-strand, providing an explanation for decreased binding of the BA.2 S protein to antibodies isolated from BA.1-convalescent patients. Our analysis reveals structural mechanisms underlying the antigenic drift in the rapidly evolving Omicron variant landscape.

Keywords: COVID-19; CP: Immunology; CP: Molecular biology; Omicron; SARS-CoV-2; cryoelectron microscopy.

Graphical abstract

Figure 1

Global prevalence and S protein mutations of the Omicron sub-lineages (A) Global prevalence of Omicron sub-lineages BA.1, BA.1.1, and BA.2 from November 2021 to March 2022. Only Omicron sub-lineages surpassing 1% global frequency are displayed. Sequence data were downloaded from the Global Initiative on Sharing All Influenza Data (GISAID) and graphed as weekly prevalence. (B) SARS-CoV-2 spike (S) protein amino acid sequence boxplots for the wild-type (D614G), BA.1/BA.1.1, and BA.2 Omicron sub-lineages. The BA.1.1 lineage is identical to BA.1 with the exception of an additional R346K mutation. NTD, N-terminal domain; RBD, receptor-binding domain; RBM, receptor-binding motif. (C) Cryo-EM-derived atomic model of the BA.2 S glycoprotein. Each S protein protomer is colored in different shades of blue. The locations of modeled amino acid mutations are shown as spheres on one protomer. BA.1-, BA.1.1-, and BA.2-specific mutations are colored in magenta, light magenta, and purple, respectively. Shared mutations across BA.1, BA.1.1, and BA.2 sub-lineages are colored in gray.

Figure 2

Binding affinity and cryo-EM structure of the Omicron BA.2 S protein-human ACE2 complex (A) Surface plasmon resonance experiments measuring dimeric human ACE2 (hACE2) binding to immobilized wild-type (WT), BA.1, and BA.2 RBDs, performed in technical triplicates. Summary data are shown at the top with representative surface plasmon resonance (SPR)-binding curves (colored solid line), and fitted 1:1 binding models (black dashed line) are shown on bottom. (B) As in (A) but measuring WT, BA.1, and BA.2 RBDs binding to immobilized dimeric hACE2, performed in at least technical quadruplicates. (C) As in (A) but measuring WT, BA.1, and BA.2 ectodomains binding to immobilized dimeric hACE2, performed in at least technical duplicates. The WT and BA.1 data in (C) were previously reported. Pairwise statistical significance test was performed using a one-way ANOVA test (^∗p ≤ 0.05; ^∗∗p ≤ 0.01; ^∗∗∗p ≤ 0.001; ^∗∗∗∗p ≤ 0.0001, ns, not significant). (D) Focus-refined cryo-EM density map and fitted atomic model of the BA.2 RBD in complex with hACE2 at 2.8 Å. (E) Aligned atomic models of hACE2 bound to BA.1 and BA.2 RBDs. The BA.1 RBD (PDB: 7T9L) and complexed hACE2 atomic models are shown in magenta and dark blue, respectively. The BA.2 RBD and complexed hACE2 atomic models are shown in purple and light blue, respectively. (F) Atomic model of the BA.1 S protein-hACE2 complex, focused on residue S496. The hydrogen bonding interaction between BA.1 S protein residue S496 and hACE2 residue K353 is indicated by an orange dashed line. (G) As in (F) but for the BA.2 S protein-hACE2 complex, focused on residue G496.

Figure 3

Cryo-EM structure of the Omicron BA.2 S protein-mouse ACE2 complex (A) Cryo-EM density map of BA.2 S protein in complex with mouse ACE2 at 2.5 Å. Mouse ACE2 is shown in green, and protomers of the BA.2 S protein are shown in shades of purple. (B) Focus-refined cryo-EM density map and fitted atomic model of the BA.2 RBD-mouse ACE2 (mACE2) complex at 2.7 Å. (C) Aligned atomic models of the BA.1 and BA.2 RBD-mACE2 complexes. The BA.1 RBD and complexed hACE2 atomic models are shown in magenta and dark green, respectively. The BA.2 RBD and complexed hACE2 atomic models are shown in purple and light green, respectively. (D) Atomic model of the BA.2 RBD-mACE2 complex, focused on residues Y501 and H505. (E) As in (D) but focused on residue R493. (F) Atomic model of the WT RBD-hACE2, focused on residues N501 and Y505. (G) As in (H) but focused on residue Q493. (H) Atomic model of mACE2 from the perspective of a binding RBD. Black labels are mACE2 residues, and gray labels denote the interacting residues in a bound RBD. Gold labels denote the interacting residues in a bound RBD that are mutated in the BA.1 and BA.2 Omicron sub-lineages.

Figure 4

Antigenic shift of the BA.2 S protein (A) Percentage of binding of monoclonal antibodies against the BA.1 and BA.2 S proteins relative to WT as assessed by ELISA, performed in technical triplicates. (B) Antibody epitopes with the side chains of contacted residues within the RBD or NTD shown and colored. BA.1- and BA.2-mutated residues are labeled within the antibody epitopes in magenta and purple, respectively, with shared mutations labeled in gray. (C) Alignment of WT, BA.1, and BA.2 RBDs, with shared, BA.1-specific, and BA.2-specific mutations labeled in gray, magenta, and purple, respectively. (D) Alignment of select patient-derived RBD-directed antibodies on the RBD. CB6, PDB: 7C01; REGN10933/REGN10987, PDB: 6XDG; CV2-75, PDB: 7M31; CR3022, PDB: 6YLA. (E) Side-by side comparison of WT (PDB: 7KRS), BA.1 (PDB: 7TNW), and BA.2 NTDs with a focused view on the structural rearrangement of the 67–79 loop and N1 antigenic loop. (F) Alignment of deposited patient-derived NTD-directed antibody atomic models with PDB IDs listed in Table S2. The labels of antibodies that make intermolecular contacts with the N1 and/or 67–79 loop are colored in orange and/or underlined in green, respectively. The NTD is shown in gray with its N1 loop highlighted in orange and the 67–79 loop in green. (G) Schematic and BA.2/BA.1 antibody-binding ratio for domain-enriched (ectodomain, NTD, or RBD) BA.1-convalescent polyclonal sera. Serum was pooled from 18 BA.1-convalescent patients (16 breakthrough cases and 2 infections in non-vaccinated patients) prior to incubation with either BA.1 ectodomain, NTD, or RBD to enrich domain-specific BA.1-convalescent antibodies. The samples were washed prior to quantification of IgG binding by ELISA and plotting of the BA.2/BA.1 ratio of domain-specific antibody binding. Data are derived from serum from 18 pooled BA.1-convalescent patients, and the ELISA assays were performed in technical duplicates.