An ACE2 Triple Decoy that neutralizes SARS-CoV-2 shows enhanced affinity for virus variants

Abstract

The SARS-CoV-2 variants replacing the first wave strain pose an increased threat by their potential ability to escape pre-existing humoral protection. An angiotensin converting enzyme 2 (ACE2) decoy that competes with endogenous ACE2 for binding of the SARS-CoV-2 spike receptor binding domain (S RBD) and inhibits infection may offer a therapeutic option with sustained efficacy against variants. Here, we used Molecular Dynamics (MD) simulation to predict ACE2 sequence substitutions that might increase its affinity for S RBD and screened candidate ACE2 decoys in vitro. The lead ACE2(T27Y/H34A)-IgG₁F_C fusion protein with enhanced S RBD affinity shows greater live SARS-CoV-2 virus neutralization capability than wild type ACE2. MD simulation was used to predict the effects of S RBD variant mutations on decoy affinity that was then confirmed by testing of an ACE2 Triple Decoy that included an additional enzyme activity-deactivating H374N substitution against mutated S RBD. The ACE2 Triple Decoy maintains high affinity for mutated S RBD, displays enhanced affinity for S RBD N501Y or L452R, and has the highest affinity for S RBD with both E484K and N501Y mutations, making it a viable therapeutic option for the prevention or treatment of SARS-CoV-2 infection with a high likelihood of efficacy against variants.

Figure 1

Biolayer Interferometry (BLI) of ACE2-IgG₁F_C, ACE2-IgAF_C, and mutant decoy binding to the spike receptor binding domain; and live virus neutralization. The (a) ACE2-IgG₁F_C decoy and (b) dimeric ACE2-IgAF_C decoy fused via a J-chain are shown. BLI kinetics analysis of (c) 1:1 binding and (d) binding with avidity for the ACE2-IgG₁F_C decoy; (e) BLI binding with avidity for the ACE2-IgAF_C decoy; (f) ACE2(T27Y)-IgG₁F_C, (g) ACE2(H34A)-IgG₁F_C, and (h) ACE2(T27Y/H34A)-IgG₁F_C decoys are shown with K_D values. (i) Virus neutralization as percent for concentrations of decoy is shown with IC50s for ACE2(WT)-IgG₁F_C and ACE2(T27Y/H34A)-IgG₁F_C. Negative controls: media and higG1 (human IgG1).

Figure 2

Molecular effects of T27Y and H34A ACE2 mutations predicted by MD simulation. (a) Spike (S) occurs as a trimer on the viral surface (orange projections), with the receptor binding domain (RBD) being on the outermost surface. The interface between S RBD and ACE2 is within the dashed box. Simulation models are shown for (b) ACE2(T27Y)-, (c) ACE2(H34A)-, and (d) ACE2(T27Y/H34A)-S RBD interactions. S RBD residues are labeled in the yellow boxes and ACE2 residues in blue boxes.

Figure 3

BLI of ACE2(WT) or Triple Decoy for mutated RBD and inhibition in the neutralization assay. (a–r) BLI of WT or Triple Decoy ACE2 to WT or mutated S RBD are shown side-by-side. (s) The percent inhibition of RBD binding to ACE2 in the surrogate neutralization assay is shown for the ACE2 Triple Decoy with S RBD WT and listed variants. RBD concentrations were 25 μg/mL. The negative control is no decoy. Inhibition of ≥ 30% (dashed line) correlates with neutralization of the virus. Data graphed as mean with SEM. Statistics: one-way ANOVA and Tukey’s post-hoc analysis comparing Triple Decoy binding to RBD WT/variants. For RBD K417N vs WT, p = 0.0495; vs L452R, p = 0.0451; and vs E484K/K417N/N501Y, p = 0.0128.

Figure 4

MD simulation predicts highest affinity for the T27Y/H34A decoy to S RBD WT and B.1.351. The free energy surfaces (FES) of wild type (WT) ACE2 upon interaction with (a) WT RBD or (b) B.1.351 RBD; and FES for the ACE2 T27Y/H34A decoy and (c) B.1.351 RBD or (d) WT RBD are shown. Darker purple represents lower free energy (ΔA_FES, scale at right of each panel). The free energy is a function of the number of intramolecular contacts (x-axis) and the distance between the centers of mass (COM, y-axis) of the interface regions. Binding free energy (ΔA_bind) is estimated by integrating the FES using Eq. (1) in “Methods”.