Identification of Potential RdRp and 3CLpro Inhibitors among FDA-approved Drugs against SARS-CoV-2
Category: Virology, Drug Discovery, COVID-19 Research
Abstract
The RNA-dependent RNA polymerase (RdRp) and 3C-like protease (3CLpro) from SARS-CoV-2 play crucial roles in the viral life cycle and are considered the most promising targets for drug discovery against SARS-CoV-2. In this study, FDA-approved drugs were screened to identify the probable anti-RdRp and 3CLpro inhibitors by molecular docking approach.
The number of ligands selected from the PubChem database of NCBI for screening was 1760. Ligands were energy minimized using Open Babel. The RdRp and 3CLpro protein sequences were retrieved from the NCBI database. For Homology Modeling predictions, we used the Swiss model server. Their structure was then energetically minimized using SPDB viewer software and visualized in the CHIMERA UCSF software.
Molecular dockings were performed using AutoDock Vina, and candidate drugs were selected based on binding affinity (∆G). Hydrogen bonding and hydrophobic interactions between ligands and proteins were visualized using Ligplot and the Discovery Studio Visualizer v3.0 software.
Our results showed 58 drugs against RdRp, which had binding energy of −8.5 or less, and 69 drugs to inhibit the 3CLpro enzyme with a binding energy of −8.1 or less. Six drugs based on binding energy and number of hydrogen bonds were chosen for the next step of molecular dynamics (MD) simulations to investigate drug-protein interactions (including Nilotinib, Imatinib and dihydroergotamine for 3clpro and Lapatinib, Dexasone and Relategravir for RdRp). Except for Lapatinib, other drugs-complexes were stable during MD simulation.
Raltegravir, an anti-HIV drug, was observed to be the best compound against RdRp based on docking binding energy (−9.5 kcal/mole) and MD results. According to the MD results and binding energy, dihydroergotamine is a suitable candidate for 3clpro inhibition (−9.6 kcal/mol). These drugs were classified into several categories, including antiviral, antibacterial, anti-inflammatory, anti-allergic, cardiovascular, anticoagulant, BPH and impotence, antipsychotic, antimigraine, anticancer, and so on.
The common prescription-indications for some of these medication categories appeared somewhat in line with manifestations of COVID-19. We hope that they can be beneficial for patients with certain specific symptoms of SARS-CoV-2 infection, but they can also probably inhibit viral enzymes. We recommend further experimental evaluations in vitro and in vivo on these FDA-approved drugs to assess their potential antiviral effect on SARS-CoV-2.
Graphical Abstract
Docking drugs against the viral RdRp and 3CLpro enzymes. The candidate drugs were classified into several categories.
1. Introduction
The world is currently experiencing an emerging pandemic called COVID-19 (caused by SARS-CoV-2), to which no effective antiviral drugs or vaccines have been approved to date [1]. A recent hypothesis has proposed that COVID-19 may have three phases. Some of the drugs are probably more effective in each phase separately. These three phases are called the viral early infection phase, the pulmonary phase, and the hyper-inflammation phase [2]. In the early infection phase, antiviral drugs are probably the best option. In the second phase, due to the involvement of the immune system, the lungs become involved. Some symptoms, such as cough, shortness of breath, and hypoxia, are observed in this phase. Blood clots are also reported mostly in the second phase. In the hyper-inflammation phase, the cytokine storm is triggered by the activation of the immune system. The cytokine storm leads to more severe damage to the lungs, kidneys, heart, and other organs. In this phase, the anti-inflammatory category of drug candidates is probably better to be more investigated. Given that these phases overlap, no single drug is expected to be sufficient for all three phases, and a combination of drugs would probably be more efficient [2].
The rapid global spread of this virus has underscored the need to develop anti-Coronavirus therapies. Several approaches and strategies are typically used to detect a potential antiviral treatment against various infections, such as the new Coronavirus. One possible common approach is applying the existing broad-spectrum antiviral drugs using standard assays. Screening the previously approved chemical compounds by bioinformatics tools is another fast method in antiviral drug discovery. In this method, medications are evaluated for their potency to inhibit some essential elements of the new viruses [1], [3].
The 3CLpro is the prime enzyme responsible for proteolysis. It cleaves the viral polyprotein into distinct functional components [4]. The essential value of 3CLpro in the virus life cycle makes it a suitable target for developing effective antiviral drugs against different Coronaviruses [5], [6]. 3Clpro offers unconventional Cys catalytic residues with a unique diversification. Differently from other chymotrypsin-like enzymes and many SER (or Cys) hydrolases, including catalytic Cys-His Dyad instead of a canonical Ser (Cys)-His-Asp (Glu) triad8. The Cys145 and His41 catalytic residues in 3Clpro are entombed on the protein surface in an active site cavity. This cavity can contain four substrates in P1′ to P4 positions and is flanked by both Domains I and II residues [7]. Another essential non-structural protein of the Coronavirus is the RNA-dependent RNA polymerase (RdRp, also known as nsp12) [8]. RdRp catalyzes the viral RNA synthesis and thus plays a pivotal role in the SARS-CoV-2 replication and transcription process, probably along with nsp7 and nsp8 as co-factors [9], [10]. Among coronaviruses, particularly in SARS-CoV-2, essential sites such as template entry and binding, polymerase activity reaction site followed by the exit through the tunnel (thumb) are highly conserved. Tyr618, Cys622, Asn691, Asn695, Met755, Ile756, Leu757, Leu758, Ser759, Asp760, Asp761, Ala762, Val763, Glu811, Phe812, Cys813 and Ser814 are the critical residues of interaction in the RDRP active site. The residue of active sites are adjoining aspartates, i.e. Asp761 and Asp762, participate in specific RdRp enzyme reactions [11].
Different anti-RNA polymerase drugs currently on the market have been previously approved for use against various viruses, including Ribavirin [12], Remdesivir [13], Galidesivir [14], and Tenofovir [15]. They are presently being examined against SARS-CoV-2 RNA-dependent RNA polymerase (RdRp). For the 3CLpro target, several studies and current clinical trials have proposed the Lopinavir [16], Ritonavir [17], Darunavir [18], Ganovo [19], ASC09F [20], and Cobicistat [21]. Ritonavir/Lopinavir (LPV) is one of the most commonly reported clinical trials for COVID-19. Even though some data indicate somewhat efficacy for LPV, its severe side effects are considerable [22], [23]. These confirm that RdRp and 3CLpro can be recommended as valuable targets for drug design against SARS-CoV-2, and inhibition of their activity seems a promising strategy to cure SARS-CoV-2 infection. In this study, we used a target-based virtual screening approach to identify novel inhibitors of SARS-CoV-2 RdRp and 3CLpro.
2. Methods
2.1. Retrieving drugs from databases and ligand minimizations
Drug repurposing using virtual screening (VS) techniques is one of the rapid and most promising strategies to candidate drugs against the Coronavirus [24].
In this study, 3D structures of 1760 FDA-approved drugs were retrieved from the NCBI PubChem database [25]. In fact, there were three-dimensional structures for approximately 2500 approved small molecule drugs (not proteins, etc); Therefore, We first removed some of the structures from our selection set including, the two-component structures, tiny compounds weighing less than 100 kDa, and the large-complex compounds with a high number of rotatable bonds. The remaining small molecules were filtered and selected for further docking analysis, including the 1760 small molecule drugs. The conjugate gradient geometry optimization was performed using Open Babel [26] and MMFF94 force fields for each drug geometry [27].
2.2. Molecular modeling and energy minimization of targets
RdRp and 3c-like proteinase (3CLpro) (from reference sequence of Accession number NC_045512) protein sequences were retrieved from the NCBI database. Then, homology modeling predictions were carried out using the Swiss model server (https://swissmodel.expasy.org/). The structures were energetically minimized using SPDB viewer software [28] and visualized by the CHIMERA UCSF v1.14 software [29]. The binding sites (active sites) in target proteins were identified by evaluating protein grooves in CHIMERA UCSF software 22 [30] and considering the previous studies [19], [31]. Since recently crystallography structures of the proteins were reported in PDB databank, we performed superimposing to check our homology modeling similarity with the crystallography results. Superimposing of the modeled structure with deposited crystallography structures available in PDB (Protein Databank) revealed the root mean square deviation (RMSD) value of < 2 angstroms (among 0.3–1.5 angstrom), which meant a perfect fit. Therefore, modeling has insignificant impacts on our overall results compared to using crystallography structures.
2.3. Preparation of protein structures for docking analysis
All nonpolar hydrogens were merged. Partial atomic charges were then assigned using the Gasteiger-Marsili approach for accurate ionization and tautomeric states of residues. Besides, charges were added to models, and Kollman United Atom charges and atomic salvation parameters were performed.
2.4. Molecular docking
Molecular docking was carried out to evaluate possible energy of interactions, hydrogen bonds, non-hydrogen bonds, and binding mode of FDA ligand datasets against RdRp and 3c-like proteinase binding sites. The docking studies were performed using AutoDock Vina v1.1.2 software in the PyRx v0.9.8 platform [32], [33]. In docking, targets were considered semi-rigid while ligands were flexible. To perform the suitable docking for each ligand, we set the search space box parameters on 32–37–39 Å (direction, x, y, and z), centered at (− 8, 15, and 67) Å, for 3c-like proteinase, and upon 35–39–42 Å, centered at (144, 133, 158) Å, for RdRp.
Final docked conformations were ranked based on binding energy (∆G) results, which meant the most favorable binding conformations had the lowest free energies. They were selected as suitable poses of binding and were then visually analyzed. Hydrogen bonds and the hydrophobic interactions between ligands and RdRp and 3c-like proteinase were analyzed (two-dimensionally) using LIGPLOT v.4.5.3 27 software [34]. Besides, the two-dimensional and three-dimensional structures of the selected ligands were analyzed using Discovery Studio Visualizer v3.0 software [35], [30].
2.5. Molecular dynamics simulations
An100 ns MD simulation for RdRp and 3clpro was used to confirm the docking results for identified candidate antiviral drugs. Molecular dynamics (MD) is a mathematical tool for analyzing the system dynamic structural behavior; in this process, atoms and molecules interact as a time-based function. The simulations of MD take the versatility of goals into account. The structural parameter RMSD and the number of intermolecular H-bonds have been used for determining the stability, dynamics and compactness of protein-drug complexes [36].
Six drugs were chosen for MD analysis based on binding energy and the number of hydrogen bonds in docking analysis. Six simulations were performed using the GROMACS 5.1.4 simulation suite for FDA-approved drugs containing Nilotinib, Imatinib, and dihydroergotamine for 3clpro and Lapatinib, Dexasone, and Relategravir for RdRp. The gromos54a7 force field was utilized for the complexes [37]. The ATB server was used for the preparation of the coordinates and topology of ligands [38]. The complexes were then solvated with TIP3P water molecules in a truncated octahedron periodic box with an 8 Å radius buffer zone of water molecules around the complexes using Gmx Editconf&Solvate softwares. Then counter ions have been added with the tool of Gromacs to neutralize the overall system charge. The surface charge of the structure was neutralized by adding several sodium ions. Reduction of energy on the structures was performed with 50,000 steps using the steepest descent method for eliminating van der Waals interactions and formation of hydrogen bonds between water molecules and the complex. In the next step, the system temperature was gradually increased from 0° to 310° K for 500 ps at constant volume, and then at constant pressure for 500 ps the system was equilibrated. Molecular dynamics simulations were performed at a temperature of 310 K and a duration of 100 nanoseconds. Non-bonded interactions with 10 Å intervals were calculated by the PME method. The SHAKE algorithm was used to limit the hydrogen atom bonds to increase computational speed. Finally, the simulation information was saved at 0.2 ps intervals for analysis.
3. Result
Molecular docking was performed on FDA-approved drugs to determine the potential drug candidates for inhibiting the SARS-CoV-2. The docking was based on the recognition of the binding pocket of Homology Modeled RdRp and 3CLpro enzymes. The SWISS online server modeled the viral proteins. The number of ligands selected from the PubChem database of NCBI for screening was 1760. All these drugs were docked against the two target enzymes of SARS-CoV-2 and ranked based on their binding affinity. The compounds with a binding affinity of − 8.5 or less were considered better compounds, possibly inhibiting the RdRp enzyme. The binding affinity of − 8.1 or less was considered the selection criterion against the 3CLpro protein. We used AutoDock Vina to dock the drugs to achieve more accurate medicines related to the two viral essential components. We first selected the top 100 medications for each viral target based on the order of their affinity energies. Depending on the rate of changes in the affinity energies among the drugs ordered, we selected 58 candidate drugs against the active site of the RdRp enzyme with an affinity of − 8.5 or less and 69 candidate drugs against the active site of the 3CLpro enzyme with affinity binding. − 8.1 or less. We observed that 20 drugs had binding affinity energy less than − 9 against the RdRp target. However, only seven drugs had binding affinity energy less than − 9 against the 3CLpro. They are likely to provide promising drugs against SARS-CoV-2. All the candidate drugs were then classified into several categories (Table 1, Table 2). We sought further studies on COVID-19 drugs to validate our identified drugs. We found that some of these candidate drugs have already been introduced or validated by various other studies, including in-silico, preclinical, and clinical trials. These verifying studies are available in Table 4. We also compared the two identified drug lists using the online Venn diagram tool. Supplementary Fig. S1 depicts the Venn diagram comparing the two drug lists against RdRp and 3CLpro. We found that 32 drugs were shared between the two drug lists. They seem to be promising since they would probably inhibit both of the essential viral components. Supplementary Table S1 lists these 32 shared drugs (vs. RdRp and 3CLpro).
📄 Download Full Text (PDF)
Identification-of-FDA-approved-drugs-against-SARS-CoV-2-RNA-_2021_Biomedicin
Author: Amirjafar Adibi
دیدگاه خود را درج کنید