Leaving no stone unturned: Allosteric targeting of SARS-CoV-2 spike protein at putative druggable sites disrupts human angiotensin-converting enzyme interactions at the receptor binding domain

Abstract

The systematic entry of SARS-CoV-2 into host cells, as mediated by its Spike (S) protein, is highly essential for pathogenicity in humans. Hence, targeting the viral entry mechanisms remains a major strategy for COVID-19 treatment. Although recent efforts have focused on the direct inhibition of S-protein receptor-binding domain (RBD) interactions with human angiotensin-converting enzyme 2 (hACE2), allosteric targeting remains an unexplored possibility. Therefore, in this study, for the first time, we employed an integrative meta-analytical approach to investigate the allosteric inhibitory mechanisms of SARS-CoV-2 S-protein and its association with hACE2. Findings revealed two druggable sites (Sites 1 and 2) located at the N-terminal domain (NTD) and S2 regions of the protein. Two high-affinity binders; ZINC3939013 (Fosaprepitant - Site 1) and ZINC27990463 (Lomitapide - Site 2) were discovered via site-directed high-throughput screening against a library of ~1500 FDA approved drugs. Interestingly, we observed that allosteric binding of both compounds perturbed the prefusion S-protein conformations, which in turn, resulted in unprecedented hACE2 displacement from the RBD. Estimated ΔG _binds for both compounds were highly favorable due to high-affinity interactions at the target sites. In addition, Site 1 residues; R190, H207, K206 and K187, I101, R102, I119, F192, L226, V126 and W104 were identified for their crucial involvement in the binding and stability of ZINC3939013. Likewise, energy contributions of Q957, N953, Q954, L303, Y313, Q314, L858, V952, N953, and A956 corroborated their importance to ZINC27990463 binding at the predicted Site 2. We believe these findings would pave way for the structure-based discovery of allosteric SARS-CoV-2 S-protein inhibitors for COVID-19 treatment.

Keywords: Allosteric targeting; High-affinity binding; Receptor binding domain; SARS-CoV-2; Spike protein; Virtual high-throughput screening.

Fig. 1

Structural architecture of SARS-CoV-2 S-protein and target human ACE2 (protease domain). A. Structural topology of the SARS-CoV-2 S-protein showing its different components. NTD, N-terminal domain; RBD, receptor binding domain; SD1, subdomain 1, SD2, subdomain 2; S1/S2, furin cleavage site 1; UH, upstream helix; L, linker region; S2’, furin cleavage site 2; FP, fusion peptide; HR1, heptad repeat 1; CH, central helix; CD, connector domain; HR2, heptad repeat 2; TM, transmembrane region, CT, cytoplasmic tail. B. 3D structure of the prefusion (S1/S2) S-protein and the interacting (protease) domain of the host hACE2 (grey).

Fig. 2

Structural depiction of the modeling approach employed herein for obtaining the prefusion S-protein-hACE2 complex. A. 3D structure of the prefusion SARS-CoV-2 S-protein as retrieved from PDB (ID 6VSB)B. Retrieved 3D structure of the truncated S-protein RBD and hACE2 (PDB ID 6M0J) C. Modelled complex of prefusion SARS-CoV-2 S-protein and hACE2 as obtained via structural superposition of A and B, followed by removal of the truncated domain.

Fig. 3

Predicted allosteric sites and their locations on SARS-CoV-2 S-protein. A. 3D structure of the SARS-CoV-2 prefusion S-protein showing surface representation of the predicted Sites 1 and 2. B. Closer look at the predicted Site 1, constituent residues and starting orientation of ZINC3939013. C. Inset showing the predicted Site 2, constituent residues and binding orientation of ZINC27990463.

Fig. 4

Allosteric disruption of S-protein hACE2 binding at the RBD. Inset A-D shows the trajectorial motions of the RBD (red) and its associated hACE2 (grey) along the simulation period from the starting to the ultimate structure at 300ns. Inset A’-D’ shows systematic perturbations of the RBD and time-based displacement of the hACE2 as induced by the binding of ZINC3939013 (yellow surface) at the predicted Site 1.

Fig. 5

Systematic perturbations of the S-protein RBD. Inset A’-D’ shows time-based alterations of the RBD and displacement of the hACE2 as induced by the allosteric binding of ZINC27990463 (green surface) at the predicted Site 2.

Fig. 6

Relative structural stabilities of the S-protein and corresponding hACE2 among the unbound and bound (red and green) systems. A and A’ shows the overall RMSD and FE-RMSD for unbound (black), ZINC3939013-bound (red), and ZINC27990463-bound (green) S-protein. B and B’ shows overall RMSD and FE-RMSD for hACE2s associated to unbound (black), ZINC3939013-bound (red), and ZINC27990463-bound (green) S-proteins.

Fig. 7

Estimations of Cα motions at the RBD domain of unbound and allosterically-bound S-proteins relative to hACE2 interactions. A. Comparative Cα RMSD plot of unbound (black), ZINC3939013-bound (red), and ZINC27990463-bound (green) S-protein RBD. B. Comparative Cα RoG plot of unbound (black), ZINC3939013-bound (red), and ZINC27990463-bound (green) S-protein RBD. C. Visual analyses of structural alterations that occurred differentially at the RBDs of unbound (black), ZINC3939013-bound (red), and ZINC27990463-bound (green) S-protein. These depictions were obtained via structural superposition of their resulting average structures.

Fig. 8

Cα FE-RMSF plots showing disparate per-residue motions and conformational flexibility among unbound and ligand-bound S-proteins together with their corresponding hACE2s A. Comparative per-residue fluctuations in S-proteins of unbound (black), ZINC3939013-bound (red), and ZINC27990463-bound (green) S-protein B. Comparative per-residue fluctuations in their corresponding hACE2s.

Fig. 9

Binding orientations of the allosteric binders at the predicted S-protein sites over the terminal post-equilibrated time-frames. A-C. show ZINC3939013 orientations at Site 1 along time-frames 270-300ns A’-C’. shows the trans-domain orientations of Site 2-bound ZINC27990463 from 270-300ns of the simulated trajectory.

Fig. 10

Ligand-residue interactions and energy contributions at the predicted allosteric sites of SARS-CoV-2 S-protein. A. Per-residue energy decomposition plot for ZINC3939013 at S-protein Site 1 A’. Complementary interaction pattern mediated by ZINC3939013 at S-protein Site 1 with constituent residues. Bond distances are also shown for crucial residues indicative of their strength B. Energy plot showing contributions of individual residues of the predicted Site 2 towards the stability of ZINC27990463. Corresponding interactions are showed in B’.