Coevolution, Dynamics and Allostery Conspire in Shaping Cooperative Binding and Signal Transmission of the SARS-CoV-2 Spike Protein with Human Angiotensin-Converting Enzyme 2

Abstract

Binding to the host receptor is a critical initial step for the coronavirus SARS-CoV-2 spike protein to enter into target cells and trigger virus transmission. A detailed dynamic and energetic view of the binding mechanisms underlying virus entry is not fully understood and the consensus around the molecular origins behind binding preferences of SARS-CoV-2 for binding with the angiotensin-converting enzyme 2 (ACE2) host receptor is yet to be established. In this work, we performed a comprehensive computational investigation in which sequence analysis and modeling of coevolutionary networks are combined with atomistic molecular simulations and comparative binding free energy analysis of the SARS-CoV and SARS-CoV-2 spike protein receptor binding domains with the ACE2 host receptor. Different from other computational studies, we systematically examine the molecular and energetic determinants of the binding mechanisms between SARS-CoV-2 and ACE2 proteins through the lens of coevolution, conformational dynamics, and allosteric interactions that conspire to drive binding interactions and signal transmission. Conformational dynamics analysis revealed the important differences in mobility of the binding interfaces for the SARS-CoV-2 spike protein that are not confined to several binding hotspots, but instead are broadly distributed across many interface residues. Through coevolutionary network analysis and dynamics-based alanine scanning, we established linkages between the binding energy hotspots and potential regulators and carriers of signal communication in the virus-host receptor complexes. The results of this study detailed a binding mechanism in which the energetics of the SARS-CoV-2 association with ACE2 may be determined by cumulative changes of a number of residues distributed across the entire binding interface. The central findings of this study are consistent with structural and biochemical data and highlight drug discovery challenges of inhibiting large and adaptive protein-protein interfaces responsible for virus entry and infection transmission.

Keywords: ACE2; SARS-CoV spike protein; alanine scanning; allosteric interactions; binding free energy; coevolution; molecular dynamics; signal transmission.

Figure 1

Structural organization of the SARS-CoV-RBD (severe acute respiratory syndrome coronavirus receptor binding domain) and SARS-CoV-2-RBD (severe acute respiratory syndrome coronavirus 2 receptor binding domain) complexes with human ACE enzyme. (A) A general overview of the SARS-CoV-RBD complex with ACE2 (pdb id 2AJF). The SARS-CoV-RBD is in cyan ribbons and the bound ACE2 enzyme is in pink ribbons. The crystal structure of the unbound native human ACE2 enzyme (pdb id 1R42) is superimposed and shown in red ribbons. (B) A general overview of the SARS-CoV-2-RBD complex with ACE2 (pdb id 6M0J). The SARS-CoV-2-RBD is shown in blue ribbons, the bound ACE2 enzyme is in pink ribbons, and the unbound form of ACE2 is in green. (C) Structural superposition of the unbound SARS-CoV-RBD (pdb id 2GHV) (in orange), bound SARS-CoV-RBD (pdb id 2AJF) (in cyan), and SARS-CoV-2-RBD (pdb id 6M0J) (in blue). Structural similarity of the RBM motif is indicated by an arrow and annotated. A moderate local mobility of the flexible ridge loop at the tip of the RBM in the unbound and bound SARS-CoV-RBD forms is evident. (D) A general overview of the secondary structure elements and binding interface in the SARS-CoV-RBD complex with human ACE2. The SARS-CoV-RBD is shown in cyan and secondary structure elements are annotated. The RBM region in SARS-CoV-RBD (residues 424–494) that provides the contact interface with ACE2 is highlighted in blue ribbons and annotated. The subdomain I of human ACE2 is shown in red ribbons and the subdomain II is shown in green ribbons. (E) A general overview of the secondary structure elements and binding interface in the SARS-CoV-2 RBD complex with human ACE2. The SARS-CoV-RBD is shown in cyan and secondary structure elements are annotated. The RBM region is in blue ribbons and annotated. The subdomains I and II of human ACE2 are shown in red and green ribbons, respectively.

Figure 2

Sequence alignment of the SARS-CoV and SARS-CoV-2 RBD proteins. (Top panel) The sequence alignment of the SARS-CoV and SARS-CoV-2 spike proteins. (Bottom panel). The sequence comparison of the RBM residues. SARS-CoV-RBD residues are in blue and SARS-CoV-2 RBD residues are in red. The close-up of the RBM alignment highlights presence of evolutionary conserved positions even in the most variable interface region. The emergence of a significant number of amino acid modifications between the SARS-CoV and SARS-CoV-2 RBD proteins can be also seen.

Figure 3

An overview of the binding interface residues in the SARS-CoV-RBD and SARS-CoV-2-RBD complexes with human ACE enzyme. The minimized average equilibrium structures obtained from MD simulations are used to depict the binding interface regions and highlight small conformational changes. (A) The N-terminus segment of the interface is shown. The SARS-CoV-RBD residues T487, Y491, N479, Y442, Y475, and L472 from the complex with ACE2 are shown in cyan sticks. The corresponding SARS-CoV-2-RBD residues N501, Y505, Q493, L455, Y489, and F486 are shown in blue sticks. (B) Another binding interface region is shown with SARS-CoV-RBD residues T486, G488, T433, G482, Y484, Y436, and D480 (cyan sticks). The corresponding SARS-CoV-2-RBD residues T500, G502, G446, G496, Q498, Y449, and S494 are shown in blue sticks. (C) The central segment of the binding interface with SARS-CoV-RBD residues N479, V404, Y442, L443, F460, and Y475 (cyan sticks). The SARS-CoV-2-RBD residues are Q493, K417, L455, F456, Y473, and Y489 (blue sticks). (D) The binding interface flexible ridge with SARS-CoV-RBD residues Y475, N473, and L472 (cyan sticks) and SARS-CoV-2-RBD residues Y489, N487, and F486 (in blue sticks). In all panels ACE2 interacting residues are annotated and shown in red sticks for the SARS-CoV-RBD complex and green sticks for the complex with SARS-CoV-2-RBD.

Figure 4

Sequence conservation profiles of the SARS CoV and ACE2 proteins. (A) The normalized ConSurf conservation scores for SARS-CoV-RBD spike protein are projected onto the SARS-CoV-RBD complex with ACE2 (residue numbering as in pdb id 2AJF). The ConSurf profiles are shown in maroon bars. The negative ConSurf scores correspond to highly conserved sites (with ConSurf score < 0), and high positive scores (Consurf score > 1.0) depict highly variable positions. Consurf score = 1.0 is defined as a threshold for differentiating moderately and highly variable positions. The binding interface residues with ACE2 are highlighted in orange filled diamonds. The position of the binding interface residues that differ between SARS-CoV-RBD and SARS-CoV-2-RBD are shown in smaller magenta filled circles. (B) The KL conservation score for SARS-CoV-RBD spike protein. High KL scores indicate highly conserved sites (above threshold of KL=1.0) and low scores correspond to more variable positions. The annotation of the binding interface residues is as in panel A. (C) The normalized ConSurf conservation scores for ACE2 residues (residue numbering corresponds to ACE2 structure in complexes with SARS-CoV (pdb id 2AJF) and SARS-CoV-2 (pdb id 6M0J). The ACE2 binding interface residues are highlighted in orange filled diamonds. (D) The KL conservation score for ACE2 residues. The ACE2 binding interface residues are highlighted in orange filled diamonds.

Figure 5

Coevolutionary profiles of the SARS CoV and ACE2 proteins. (A) The structure-based pMI profile for SARS-CoV-RBD spike protein is projected onto the SARS-CoV-RBD complex with ACE2 (residue numbering as in pdb id 2AJF). The binding interface residues that make contacts with ACE2 are highlighted in orange filled diamonds. The position of the binding interface residues that differ between SARS-CoV-RBD and SARS-CoV-2-RBD are shown in smaller magenta filled circles. (B) The pMI profile for ACE2 residues (pdb id 2AJF). The annotation of the binding interface residues is as in panel A. (C) The sequence-based cMI profile for SARS-CoV-RBD is projected onto the SARS-CoV-RBD complex with ACE2 (pdb id 2AJF). (D) The cMI profile for ACE2 residues (residue numbering corresponds to ACE2 structure in the complex with SARS-CoV (pdb id 2AJF). The annotation of the binding interface residues is as in panel A. The horizontal lines on the graphs are materialized in cyan to indicate thresholds used to differentiate moderate and high values of pMI and cMI scores.

Figure 6

Structural mapping of the SARS-CoV-RBD and ACE2 residues with high pMI and cMI values. (A) A general overview of the SARS-CoV-RBD complex with ACE2. The high pMI residues serving as regulators of coevolutionary network are shown in pink spheres. The highly coevolving residues with high cMI values are shown in red sticks. Note a consolidation of regulatory high pMI sites in the central segment of the binding interface and clusters of highly coevolving residues near the flexible loop regions of the binding interface. (B) A close-up of the binding interface between SARS-CoV-RBD and ACE2. The high pMI residues (pink spheres) and high cMI residues (red sticks) are shown and annotated for the key positions that control coevolutionary couplings in the complex.

Figure 7

All-atom conformational dynamics of the unbound and bound SARS-CoV structures. The root mean square fluctuations (RMSF) profiles are obtained from all-atom MD simulations of the crystal structures. (A) Conformational dynamics profiles of the unbound SARS-CoV (pdb id 2GHV) (in orange lines) and SARS-CoV-RBD in the complex with ACE2 (pdb id 2AJF) (in maroon lines). The positions of the SARS binding interface residues are highlighted in magenta filled circles. (B) The RMSF dynamics profiles of the unbound SARS-CoV-2 (in orange lines) and SARS-CoV-2 RBD in the complex with ACE2 (pdb id 6M0J) (in maroon lines). The positions of the SARS binding interface residues are highlighted in magenta filled circles. (C) The RMSF dynamics profiles of the unbound ACE2 (pdb id 1R42) (in orange lines) and in the complexes with SARS-CoV-RBD (in maroon lines). The ACE2 binding interface residues are shown in magenta filled circles. (D) The RMSF dynamics profiles of the unbound ACE2 (pdb id 1R42) (in orange lines) and in the complexes with SARS-CoV-2 RBD (in maroon lines). The ACE2 binding interface residues are shown in magenta filled circles. The horizontal lines on all graph panels are materialized in cyan to indicate thresholds used to differentiate small and higher thermal motions.

Figure 8

CG conformational dynamics of the unbound and bound SARS-CoV structures. The RMSF profiles are obtained from CG simulations of the crystal structures. (A) Conformational dynamics profiles of the unbound SARS-CoV (pdb id 2GHV) (in orange lines) and SARS-CoV-RBD in the complex with ACE2 (pdb id 2AJF) (in maroon lines). The positions of the SARS binding interface residues are highlighted in magenta filled circles. (B) The RMSF dynamics profiles of the unbound SARS-CoV-2 (in orange lines) and SARS-CoV-2 RBD in the complex with ACE2 (pdb id 6M0J) (in maroon lines). The positions of the SARS binding interface residues are highlighted in magenta filled circles. (C) The RMSF dynamics profiles of the unbound ACE2 (pdb id 1R42) (in orange lines) and in the complexes with SARS-CoV-RBD (in maroon lines). The ACE2 binding interface residues are shown in magenta filled circles. (D) The RMSF dynamics profiles of the unbound ACE2 (pdb id 1R42) (in orange lines) and in the complexes with SARS-CoV-2 RBD (in maroon lines). The ACE2 binding interface residues are shown in magenta filled circles. The horizontal lines on all graph panels are materialized in cyan to indicate thresholds used to differentiate small and higher thermal motions.

Figure 9

Structural mapping of the conformational mobility profiles in the SARS-CoV unbound structure and SARS-CoV-RBD/SARS-CoV-2 RBD complexes with ACE2 obtained from all-atom MD simulations. (A) Conformational dynamics profiles mapped on the unbound form of SARS-CoV-RBD (pdb id 2GHV). A ribbon-based protein representation is used with coloring (blue-to-red) according to the protein residue motilities (from more rigid-blue regions to more flexible-red regions). (B) Structural map of the conformational mobility profile of SARS-CoV-RBD in the complex with ACE2. The more stable regions are shown in blue and more flexible are in green-to-red coloring scale. (C) Structural map of dynamics profile of SARS-CoV-2-RBD in the complex with ACE2. The more stable regions are shown in blue and more flexible are in green-to-red coloring scale. For each panel, the position of the RBM region and the interfacial loop, which is an element of the RBM, are indicated with respective arrows. Note the redistribution of conformational mobility between the unbound and bound forms for SARS-CoV and SARS-CoV-2.

Figure 10

Alanine scanning of the key binding interface residues in the SARS-CoV and SARS-CoV-2 complexes with ACE2. (A) The binding free energy changes upon alanine mutations for the interface residues in the SARS-CoV-RBD complex with ACE2 (pdb id 2AJF). The binding energy changes for the ACE residues are shown in orange bars and for the SARS-CoV-RBD interacting residues in maroon bars. The computed values were obtained using BeAtMuSiC approach and were averaged over equilibrium samples from MD simulation. (B) The binding free energy changes upon alanine mutations for the interface residues in the SARS-CoV-2 RBD complex with ACE2 (pdb id 6M0J). The binding energy changes for the ACE residues are shown in orange bars and for the SARS-CoV-2 RBD interacting residues in maroon bars. (C) The binding free energy changes upon alanine mutations for the interface residues in the engineered chimera SARS-CoV-2 RBD bound with ACE2 (pdb id 6VW1). The binding energy changes for the ACE residues are shown in orange bars and for the SARS-CoV-2 RBD interacting residues in maroon bars. (D) The binding free energy changes upon alanine mutations for the interface residues in the SARS-CoV-2 RBD complex with ACE2 obtained from the cryo–electron microscopy structure of the full-length human ACE2 with SARS-CoV-2 RBD in the presence of the neutral amino acid transporter B⁰AT1 (pdb id 6M17). The binding energy changes for the ACE residues are shown in orange bars and for the SARS-CoV-2 RBD interacting residues in maroon bars. The binding free energy changes for each complex are based on the average binding energy estimates over MD trajectories. The error bars for the binding free energy changes were in a small range of 0.05–0.15 kcal/mol. The horizontal lines on the graphs are materialized in cyan to indicate a threshold of 1.5 kcal/mol used to differentiate moderate and large free energy changes and identify key binding energy hotspots.

Figure 11

Structural mapping of the binding energy hotspots in the SARS-CoV-RBD and SARS-Cov-2-RBD complexes with ACE2. (A) A general overview of the SARS-CoV-RBD complex with ACE2. SARS-CoV-RBD is shown in cyan ribbons and ACE2 is in green ribbons. The binding energy hotspots that experience significant loss of binding energy upon alanine mutations are shown in cyan spheres for SARS-CoV-RBD and in green spheres for ACE2. (B) A close-up of these binding energy hotspots in the SARS-CoV-RBD complex with ACE2. The SARS-CoV-RBD residues are annotated and shown in cyan sticks, ACE2 hotspot residues are in green sticks. (C) A general overview of the SARS-CoV-2-RBD complex with ACE2. SARS-Cov-2-RBD is shown in blue ribbons and ACE2 is in pink ribbons. The binding energy hotspots that experience significant loss of binding energy upon alanine mutations are shown in blue spheres for SARS-CoV-2-RBD and in pink spheres for ACE2. (D) A close-up of the binding energy hotspots in the SARS-CoV-2-RBD complex with ACE2. The SARS-CoV-RBD residues are annotated and shown in blue sticks, ACE2 hotspot residues are in pink sticks.