EK1 with dual Q1004E/N1006I mutation: a promising fusion inhibitor for the HR1 domain of SARS-CoV-2

Dear Editor,

This journal has previously published commentaries on the consequences and severity of coronavirus disease 2019 (COVID-19) and the urgent need for novel therapy¹. The epidemic of COVID-19 has posed a great threat to many aspects including human health and the global economy. However, there is still no effective strategy to combat this disease caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), and the number of infections by this virus is still increasing, as recorded in an online website: https://www.worldometers.info/. It is reported that spike (S) protein is the main player of SARS-CoV-2 invasion into the host, and its structure can be divided into two domains (S1 and S2) that perform different functions². The S1 domain mainly promotes the binding of SARS-CoV-2 to the host through the interaction between its receptor binding domain (RBD) and angiotensin-converting enzyme 2 (ACE2), and the interaction mechanism between RBD and ACE2 has been discussed in detail^{3 , 4}. After the binding of S1 domain to ACE2, the S2 domain will initiate the fusion of the viral and host cell membranes. During this fusion process, the trimeric heptad repeat 1 (HR1) region of S2 domain will interact with the trimeric heptad repeat 2 (HR2) region of S2 domain to form a stable six-helix bundle (6-HB) structure to facilitate the close proximity of the membrane⁵. The 6-HB core plays a crucial role in the membrane fusion of the coronavirus and the host. In addition, HR1, HR2 and their interaction mode are conserved in various coronaviruses. Thus, HR1 and HR2 are considered to be promising targets for the development of fusion inhibitors against coronaviruses. Recently, Xia et al. have proved that the fusion inhibitor EK1 has inhibitory activity against SARS-CoV-2, and it can inhibit the formation of 6-HB fusion core, thereby hindering the fusion of viral and host cell membranes^{5 , 6}. The confirmation of this inhibitor provides novel ideas and insights for the future design of inhibitors against SARS-CoV-2.

Our focus here is to enhance the inhibition ability of EK1 through further optimization, and the binding affinity of the newly optimized inhibitors to HR1 domain of SARS-CoV-2 is stronger than that of EK1 to HR1. On the basis of the inhibitor EK1, we conducted a comprehensive screening of potential inhibitors for the HR1 domain of SARS-CoV-2 through a combination of residue mutation, molecular dynamics (MD) simulation and binding energy calculation.

The trimeric structure of HR1-EK1 complex was acquired from the protein data bank (PDB ID: 7C53), as shown in Fig. 1 A. The inhibitor EK1 is displayed in cartoon mode, and three EK1s in the trimer are represented by chains A, B and C, respectively. To explore the mutant residues that can enhance the binding affinity of EK1 to HR1, all the residues on EK1 in the trimer were mutated to 19 residues other than itself, and the changes in binding affinity resulting from 1824 possible mutations were preliminarily predicted by mCSM-PPI2⁷ (Fig. 1B). Among these mutations, the mutations Q1004E and Q1004D in chain A, Q1004E and Q1004D in chain B, and Q1004Y and N1006I in chain C could more significantly enhance the binding ability of EK1 to HR1, resulting in an increase in binding affinity by more than 1 kcal/mol. Although the performance of mCSM-PPI2 is better than that of previously developed tools including FoldX, mCSM-PPI2 still has certain shortcomings. For example, it can only predict single mutation, not multiple mutations. In the following work, the binding properties of trimeric HR1 to EK1s with single mutation (Q1004E, Q1004D, Q1004Y and N1006I) and dual mutation (Q1004E/N1006I, Q1004D/N1006I and Q1004Y/N1006I) were deeply analysed to overcome this problem.

Fig. 1.

Binding characteristics of WT/mutant EK1s to HR1. (A) The trimeric structure of HR1-EK1 complex, in which HR1 and inhibitor EK1 are shown in surface and cartoon modes, respectively. (B) Workflow for predicting mutations that may enhance the binding affinity of EK1 to HR1 by mCSM-PPI2. (C) The RMSD values of all complexes during the entire MD simulations. (D-G) The binding free energies between HR1 and the WT/mutant EK1s calculated by MM-PBSA (D and E) or SIE (F and G) method, in which the conformational interval used for calculation is 1 frame (D/F) or different frames (E/G). Dashed lines represent the mean of the calculated results at different intervals. (H) Hierarchical clustering tree for the energy contributions of individual residues. Orange and blue represent favorable and unfavorable energy contributions of residues, and white indicates the corresponding residues have no contribution to the binding. The energy contributions of residues on EK1s of WT, Q1004E, Q1004D, Q1004Y, N1006I, Q1004E/N1006I, Q1004D/N1006I and Q1004Y/N1006I mutants are arranged in order from the inside to the outside, as indicated by the blue arrow.

MD simulation has been proven to be an effective tool for exploring the binding properties of receptors and ligands^{8 , 9}. MD simulations of the above complexes were performed mainly by the following steps: (1) preparing the system, including adding hydrogen atoms and counterions, and setting the force field; (2) performing optimization, heating and dynamic equilibrium; (3) running long-time (200 ns) MD simulations without any restrictions. Through root mean square deviation (RMSD) analysis, it is found that all systems have reached equilibrium after 120 ns MD simulations, and their RMSD values are basically less than 2.5 Å, indicating that all systems are already in a stable state and can be used for subsequent calculations (Fig. 1C).

To investigate the binding ability of the wild-type (WT)/mutant EK1s to HR1, molecular mechanics Poisson-Boltzmann surface area (MM-PBSA) and solvated interaction energy (SIE) methods were selected to predict the binding free energy (ΔG) between them. First, 10,000 conformations were extracted from the equilibrium MD trajectory with a conformation interval of 1 frame for calculation. From the calculated results of the MM-PBSA method (Fig. 1D), it can be seen that all the selected single mutations (Q1004E, Q1004D, Q1004Y and N1006I) on EK1s can enhance the binding affinity of EK1 to HR1, which is consistent with the predicted results of mCSM-PPI2. Notably, the binding affinity of the mutant EK1s with dual mutation (Q1004E/N1006I, Q1004D/N1006I and Q1004Y/N1006I) to HR1 was also significantly enhanced compared to that of WT EK1 to HR1, and the mutant EK1 with Q1004E/N1006I showed the strongest binding ability to HR1. The similar results was also obtained by the SIE method (Fig. 1F). Next, different conformational intervals were applied to calculate the binding free energy to confirm these results (Fig. 1E and 1G). It is found that the mean of the binding free energies calculated at different intervals was largely consistent with the result shown in the above figures (Fig. 1D and 1F), which further suggests that the selection of the conformation is reasonable. Finally, the hotspot residues that promote the binding of WT/mutant EK1s to HR1 were also found through free energy decomposition and hierarchical clustering analyses (Fig. 1H). As shown in Fig. 1H, all EK1s in the trimer have similar clustering results. According to the energy contribution of each residue, the residues on EK1 are mainly divided into three clusters (a, b and c), and the residues with higher energy contribution to the binding are mainly distributed in cluster b.

Through the above analyses, it can be concluded that the mutant EK1 with dual Q1004E/N1006I mutation is the most competitive inhibitor among WT/mutant EK1s. Therefore, it may be more effective than EK1 in combating SARS-CoV-2. Meanwhile, it is found that twelve residues (L1002, I1005, V1007, F1009, L1010, L1012, L1019, I1023, L1026, Y1030, I1031 and L1033) in cluster b are the main contributors of WT/mutant EK1s binding to HR1, which would be helpful to understand their inhibitory mechanisms. We hope that the current study will promote the development of effective inhibitors targeting the conserved HR1 domain of SARS-CoV-2.

Declaration of Competing Interest

All authors claim no conflict of interest.