A structure-function analysis shows SARS-CoV-2 BA.2.86 balances antibody escape and ACE2 affinity

Abstract

BA.2.86, a recently described sublineage of SARS-CoV-2 Omicron, contains many mutations in the spike gene. It appears to have originated from BA.2 and is distinct from the XBB variants responsible for many infections in 2023. The global spread and plethora of mutations in BA.2.86 has caused concern that it may possess greater immune-evasive potential, leading to a new wave of infection. Here, we examine the ability of BA.2.86 to evade the antibody response to infection using a panel of vaccinated or naturally infected sera and find that it shows marginally less immune evasion than XBB.1.5. We locate BA.2.86 in the antigenic landscape of recent variants and look at its ability to escape panels of potent monoclonal antibodies generated against contemporary SARS-CoV-2 infections. We demonstrate, and provide a structural explanation for, increased affinity of BA.2.86 to ACE2, which may increase transmissibility.

Keywords: ACE2 binding; BA.2.65; SARS-CoV-2; antigenic escape; coronavirus; receptor binding; virus evolution; virus structure.

Graphical abstract

Figure 1

Sequence changes in BA.2.86 compared with other Omicron sublineages (A) Sequence alignments of BA.2.86 RBD with Omicron sublineages BA.1, BA.2, BA.4/5, XBB.1.5, EG.5/EG.5.1, XBB.1.5.70/HK.3, JN.1, and JN.4. (B) Surface representation of BA.2.86 mutations shown on BA.2 RBD. (C) XBB.1.5 mutations on BA.2 RBD. Mutations in common are colored in magenta, further mutations in BA.2.86 and XBB.1.5 in cyan, and V483 deletion in BA.2.86 in green. If there are two letters after the residue number in the labels, the first letter indicates the residue type for BA.2 and the second BA.2.86 or XBB.1.5. (D) Phylogenetic tree generated by aligning spike sequences of the SARS-CoV-2 variants. (E) Evolutionary tree of BA.2.86 with spike mutations indicated in red. See also Figure 2 and Table S1.

Figure 2

Pseudoviral neutralization assays of BA.2.86 by vaccine and infected serum samples (A and B) Geometric mean PVNT50 values for the indicated viruses using serum obtained from vaccinated volunteers after 18 months (n = 17), neutralization assays were performed in duplicate and the average titer was taken after a third dose of Pfizer BNT162b2 or Moderna vaccine (A) and 6 months after a fourth bivalent vaccine dose (n = 23) (B). (C–E) Serum from vaccinees suffering breakthrough infections by (C) BA.2 (n = 19), (D) BA.4/5 (n = 10), and (E) a set of samples collected following vaccine breakthrough infections in the last year (n = 19). (F) A composite figure for the geometric means of all serum samples against selected Omicron sublineages. The Wilcoxon matched-pairs signed rank test was used for the analysis and two-tailed p values were calculated. (G) Antigenic map showing BA.2.86 in the context of the positions of previous lineages including several Omicron-related sublineages calculated from pseudovirus neutralization data. The distance between two positions is proportional to the reduction in neutralization titer when one of the corresponding strains is challenged with serum derived by infection by the other (see STAR Methods). We have previously described the method; however, while in previous reports we generated a 3D map here we were able to describe the map in 2D with minimal impact on the target function (starting error function for random positions was 1.25, final errors were 0.038 and 0.039 for 3D and 2D models, respectively). An approximate scale bar is shown, the scale of the map is linear and is the same in all directions. See also STAR Methods and Figure 1.

Figure 3

Pseudoviral neutralization assays using BA.2-specific monoclonal antibodies (A) Heatmap of BA.2.86 IC₅₀ neutralization titers of 25 potent human mAbs made following BA.2 infection. BA.2 neutralization titers are taken from Dijokaite-Guraliuc et al. The likely BA.2.86 mutations leading to loss of activity in BA.2.86 for each mAb are indicated in the final column. (B) BA.2 mAb binding positions (blue spheres) mapped on RBD surface by BLI competition measurements and structure determinations. RBD shown in gray surface representation with BA.2.86 mutation sites colored in magenta. S309 and Omi-42 are also shown for reference. See also Figure S1.

Figure 4

Neutralization curves for XBB.1.5 RBD-specific mAbs XBB-specific mAb isolated from breakthrough infection with recent variants. (A and B) (A) Titration curves for BA.2.86 are compared with BA.2, BA.4/5, XBB.1.5, EG.5, EG.5.1, XBB.1.5.70, HK.3, JN.1, and JN.4. Assays were performed twice in duplicate. Data are presented as mean values ± SEM. IC₅₀ titers are shown as a heatmap with numbers in parentheses indicating heavy chain V genes in (B). (C) Surface representation of RBD with ACE2 footprint colored in green and the sites of mutations L455F and F456L highlighted in red (F456L in EG.5, L455F + F456L in XBB.1.5.70). (D) Heatmap of IC₅₀ neutralization titers of mAbs developed for clinical use. (E) Binding pose of S309 (sotrovimab) and its interactions with K356. (F) Measurement of the affinity of ACE2 with BA.2.86, XBB.1.5, and Beta RBDs by surface plasmon resonance. Titration curves for ACE2 that flowed over the indicated immobilized RBDs are shown together with the calculated K_D values. (G) Comparison of ACE2/RBD affinities for RBDs from different SARS-CoV-2 variants. See also Figures 5, 6, and S2 and Tables S2 and S3.

Figure 5

Structure of BA.2.86 ACE2 complex (A) Binding pose of ACE2 (green) to BA.2.86 RBD (gray) compared with binding pose of ACE2 (pale cyan) to Wuhan (left panel) and BA.2.75 (middle panel) RBD (pink). The right panel shows loss of direct contacts of ACE2 to residue 486 due to F486P mutation in BA.2.86 RBD. (B) Structural differences at the left shoulder between BA.2.86 (gray) and BA.2.75 (pink) RBDs due to V483 deletion in BA.2.86. (C) Electrostatic surfaces of the ACE2-RBD interface. See also Figures 4 and S2.

Figure 6

Structures of Delta-RBD/XBB-2, BA.2.12.1-RBD/XBB-4, delta-RBD/XBB-6, BA.2.86-RBD/XBB-7, and Delta-RBD/XBB-9 complexes (A–E) Binding pose of (A) XBB-2, (B) XBB-4, (C) XBB-6, (D) XBB-7, and (E) XBB-9 on the RBD, respectively. Only VH (red) and VL (blue) domains of the Fab are shown as ribbons for clarity. RBD is drawn as a gray surface representation with mutation sites common to XBB.1.5 and BA.2.86 highlighted in magenta, different or additional mutation sites in BA.2.86 in cyan. (F–I) (F) Positions of CDRs which have direct contacts (≤4.0 Å) with RBD, (G) details of Fab and RBD interactions for XBB-2, and (H and I) for XBB-4. The side chains of the RBD, Fab HC, and LC are shown as gray, red, and blue sticks, respectively. The yellow broken bonds represent hydrogen bonds or salt bridges. (J) Position of XBB-6 CDRs. (K) Structural changes of RBD left shoulder (gray) upon binding of XBB-6 (red) compared with the RBD (teal) bound with XBB-2 (brown). (L) Details of XBB-6 and RBD interactions. (M) Positions of XBB-7 CDRs that have direct contacts with the RBD. (N) Structural changes of BA.2.86 RBD (gray) due to deletion of V483 compared with XBB-2 bound Delta-RBD (teal). (O) Details of XBB-7 and RBD interactions. (P) Position of XBB-9 CDRs. (Q) Details of XBB-9 and RBD interactions. The drawing style and color scheme in (I), (L), (O), and (Q) are as in (G). See also Table S4.