Computational pipeline provides mechanistic understanding of Omicron variant of concern neutralizing engineered ACE2 receptor traps

Abstract

The SARS-CoV-2 Omicron variant, with 15 mutations in Spike receptor-binding domain (Spike-RBD), renders virtually all clinical monoclonal antibodies against WT SARS-CoV-2 ineffective. We recently engineered the SARS-CoV-2 host entry receptor, ACE2, to tightly bind WT-RBD and prevent viral entry into host cells ("receptor traps"). Here we determine cryo-EM structures of our receptor traps in complex with stabilized Spike ectodomain. We develop a multi-model pipeline combining Rosetta protein modeling software and cryo-EM to allow interface energy calculations even at limited resolution and identify interface side chains that allow for high-affinity interactions between our ACE2 receptor traps and Spike-RBD. Our structural analysis provides a mechanistic rationale for the high-affinity (0.53-4.2 nM) binding of our ACE2 receptor traps to Omicron-RBD confirmed with biolayer interferometry measurements. Finally, we show that ACE2 receptor traps potently neutralize Omicron and Delta pseudotyped viruses, providing alternative therapeutic routes to combat this evolving virus.

Keywords: ACE2 receptor traps; Rosetta; SARS-CoV-2 Omicron variant; Spike; cryo-EM; protein therapeutics; pseudovirus neutralization.

Graphical abstract

Figure 1

Cryo-EM reconstruction of the S protein with computationally designed, CVD293 or linker variant of the affinity-matured variant, CVD432 (A and B) Cryo-EM reconstructions of the S protein with CVD293 or CVD432 showing the heterogeneity in distribution of all RBD-down, 1-RBD- or 2-RBD-up states and variable ACE2 occupancy. Also shown is schematic of the primary structure of CVD293 or CVD432 and the engineered mutations, colored by domain. S protein is shown in blue and ACE2 is shown in green fitted into 1-RBD- or 2-RBD-up states. Partial glycosylation could be resolved in our structures and glycans are shown as yellow spheres. See also Figures S1–S5 and Table S3.

Figure 2

Cryo-EM reconstruction of WT-Spike-RBD with engineered ACE2 Fc-fusions reveal contributions from hydrophobic interactions at RBD-ACE2 interface (A) WT-Spike-RBD/CVD293 and WT-Spike-RBD/CVD432 models colored by estimated per-residue Q-score ranging from 0 (red) to 0.7 (purple). The color bar shows corresponding estimated resolution in Å for each Q-score. Expected Q-score for 3.5 Å map is 0.49 and expected Q-score for 3.36 Å map is 0.52. (B and C) Cryo-EM reconstructions of WT-Spike-RBD with either CVD293 or CVD432 show favorable π–π stacking interactions between WT-Spike-RBD residue Y489 and engineered ACE2 residue F31. In addition, there are also hydrophobic interactions between WT-Spike-RBD residue L455 and CVD293 residue I34. Hydrogen bond interactions between WT-Spike-RBD residue Q493 and CVD293 or CVD432 residue Q35 are not apparent in the cryo-EM consensus model. (D) The Rosetta lowest energy model for CVD293 is overlaid with the cryo-EM model. Both models show hydrophobic and hydrogen bond interactions between CVD293 and WT-Spike-RBD residues that contribute to improved interface energy (REU) compared with the ACE2-WT-Spike-RBD interaction. See also Figures S1–S5 and Table S3.

Figure 3

Multi-model pipeline improves confidence of molecular interactions at the interface residues in cryo-EM derived models of WT-Spike-RBD with engineered ACE2 Fc-fusions (A) Multi-model pipeline with average Rosetta interface energy and average per-residue side-chain RMSD metrics for interface residue rotamer positions. (B) Average per-residue side-chain RMSD for interface helix residues of CVD293 and CVD432. (C and D) Superposition of critical interface residues of the top 80 selected cryo-EM based models for CVD293 and CVD432. (E) Average Rosetta interface energy for CVD293 design model, CVD293 cryo-EM based models, CVD432 design model, and CVD432 cryo-EM based models. Error bars show SD calculated from 80 cryo-EM models or 80 design models. See also Figures S6 and S7.

Figure 4

Binding to Omicron- and Delta-RBD and neutralization of Omicron- and Delta-SARS-CoV-2 VOCs by CVD293 and CVD432 (A and B) Left panel: Predictions based on Rosetta interface energy calculations suggest that Omicron-RBD binds CVD293 and CVD432 with high affinity. Residue-pair interactions of RBD residues with CVD293/CVD432 residue F31 (yellow), with residue I34/S34 (blue) and with residue Q35 (red) are shown. Right panel: Zoomed-in view of the interface of the models Omicron-RBD/CVD293 and Omicron-RBD/CVD432. Wheat-colored residues indicate RBD interactions ACE2 K/F31. Blue residues indicate RBD interactions with ACE2 H/I/S34. Magenta residues indicate RBD interactions with more than one engineered ACE2 residue. (C) Biolayer interferometry measurements for CVD293 or CVD432 interactions with Omicron- or Delta-RBD. (D) CVD293 and CVD432 potently neutralize vesicular stomatitis virus (VSV) pseudotyped with SARS-CoV2 Omicron- and Delta-Spike. Error bars represent SD over all technical replicates from two biological replicates. See also Figures S8 and S9, Tables S1, S2, and S4.