Evaluating the structure-based virtual screening performance of SARS-CoV-2 main protease: A benchmarking approach and a multistage screening example against the wild-type and Omicron variants

Noha Galal; Botros Y. Beshay; Omar Soliman; Muhammad I. Ismail; Mohamed Abdelfadil; Mohamed El-Hadidi; Reem K. Arafa; Tamer M. Ibrahim

doi:10.1371/journal.pone.0318712

Abstract

COVID-19 still poses a worldwide health threat due to continuous viral mutations and potential resistance to vaccination. SARS-CoV-2 viral multiplication hindrance by inhibiting the viral main protease (Mpro) deemed propitious. Structure-based virtual screening (SBVS) is a conventional strategy for discovering new inhibitors. Nonetheless, the SBVS efforts against Mpro variants needed to be benchmarked. Herein, in the first stage of the study, we evaluated four docking tools (FRED, PLANTS, AutoDock Vina and CDOCKER) via an in-depth benchmarking approach against SARS-CoV2 Mpro of both the wild type (WTMpro) and the deadly Omicron P132H variant (OMpro). We started by compiling an active dataset of non-covalent small molecule inhibitors of the WTMpro from literature and the COVID-Moonshot database along with generating a high-quality benchmark set via DEKOIS 2.0. pROC-Chemotype plots revealed superior performance for AutoDock Vina against WTMpro, while both FRED and AutoDock Vina demonstrated excellent performance for OMPro. In the second stage, VS was performed on a focused library of 636 compounds transformed from the early-enriched cluster related to perampanel via a scaffold hopping approach. Subsequently, molecular dynamics (MD) simulation and MM GBSA calculations validated the binding potential of the recommended hits against both explored targets. This study provides an example of how to conduct an in-depth benchmarking approach for both WTMPro and OMPro variants and offering an evaluated SBVS protocol for them both.

Citation: Galal N, Beshay BY, Soliman O, Ismail MI, Abdelfadil M, El-Hadidi M, et al. (2025) Evaluating the structure-based virtual screening performance of SARS-CoV-2 main protease: A benchmarking approach and a multistage screening example against the wild-type and Omicron variants. PLoS ONE 20(2): e0318712. https://doi.org/10.1371/journal.pone.0318712

Editor: Ahmed A. Al-Karmalawy,, University of Mashreq, IRAQ

Received: November 27, 2024; Accepted: January 20, 2025; Published: February 19, 2025

Copyright: © 2025 Galal et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: "The SMILES of the active and decoy molecules used in docking benchmarking are within the Supporting Information files. Codes for docking benchmarking are available as follows: AudoDock Vina docking and file preparation [https://vina.scripps.edu/manual/; https://ccsb.scripps.edu/mgltools/documentation/], PLANTS docking [https://github.com/purnawanpp/plants?tab=readme-ov-file], and FRED docking and preparation [https://docs.eyesopen.com/applications/oedocking/fred/fred.html]. Please note, the CDOCKER module (for CDOCKER docking) is property of “Dassault Systèmes BIOVIA,” therefore access to it might be restricted. Codes for molecular dynamics GROMACS and gmx MMPBSA codes are available at https://manual.gromacs.org/2023.3/manual-2023.3.pdf, https://github.com/Valdes-Tresanco-MS/gmx_MMPBSA. The transformed focused library was downloaded from https://drugspacex.simm.ac.cn/help/ [DOI: https://doi.org/10.1093/nar/gkaa920]. All other relevant data are within the paper and its Supporting information files."

Funding: The author(s) received no specific funding for this work.

Competing interests: The authors have declared that no competing interests exist.

Introduction

With the advent of the emergence of the COVID-19 pandemic mediated by the SARS-CoV-2 virus, the scientific arena found itself facing an unprecedented global health threat posing enigmatic challenges and pending swift resolution to attend to the daunting figures of disease spread and mortalities[1]. This was further aggravated by the inherent rapid mutations of the virus leading to new variants, such as. Alpha (B.1.1.7) [2,3], Beta (B.1.351) [4,5], Gamma (P.1) [6], and Delta (B.1.617.2) [7,8]; and particularly Omicron (B.1.1.529) [9,10]. Importantly, the Omicron variant is associated with higher risk of transmission, potential drug resistance mutations, and possible resistance to vaccination. As previously identified for other coronaviruses, SARS-CoV-2 contains a set of structural proteins (SPs) that are important for producing a complete viral particle, additional non-structural proteins (NSPs) that act as enzymes or transcription/replication factors in the viral life cycle, and other numerous accessory proteins. One essential NSP for viral replication is the SARS-CoV-2 main protease (Mpro) (also known as 3Clpro, or NSP5) whose proteolytic cleavage of polyproteins generates functional polypeptides [11,12]. Yet, on the basis of consideration to the emergence of drug resistance to drug-based interventions, the scientific community is vigilantly monitoring antivirals’ potential mechanisms of drug resistance against Mpro, particularly that observed by the Omicron variant, such as the P132H mutant variant [13,14].

For intervening with the function of Mpro, the non-covalent, nonpeptidic inhibitors option is preferred in many drug discovery campaigns to avoid potential serious side effects arising from off-target irreversible interactions and altered pharmacokinetic profile [15]. Accordingly, extensive research employing virtual screening and design approaches revealed potent and non-peptidic inhibitors with optimized PK profiles such as ensitrelvir (S-217622) [16,17].

In the context of structure-based virtual screening (SBVS), benchmarking is an evaluation approach of the screening performance of docking tools. Benchmark datasets consist of bioactive molecules and structurally similar inactive decoys for a specific protein target [18,19]. The effectiveness of a docking tool in SBVS is measured by its capacity to prioritize known bioactive molecules over decoys. Therefore, benchmarking is a valuable method for identifying suitable virtual screening pipelines with improved predictive capabilities [20–22].

Rational of the study

The aim of the current study is to provide an in-depth benchmarking investigation for popular docking tools on both WTMpro and the mutant OMPro (P132H) variant, and accordingly, offering an example of a prospective virtual screening approach on both Mpro variants, as illustrated in Fig 1. We started by compiling an active data set of noncovalent, non-peptidomimetic inhibitors, emphasizing the diversity and potency of small inhibitors of the Mpro reported from both literature and the COVID-Moonshot to produce high quality DEKOIS 2.0 benchmark set. Subsequently, we probed the benchmarking performance of four docking tools employing structures of Mpro of both the wild-type (WTMpro) and the Omicron P132H variant (OMpro) to stand on the most appropriate tool for VS against both targets. Grounded on the benchmarking results and the chemotype analysis, we subsequently built a virtual new focused chemical library featuring the scaffold of the top cluster (perampanel, pyridone derivatives in this study) to propose new plausible Mpro inhibitors. Thereafter, the recommended hits from the VS campaign were further validated via MD simulations and MM GBSA analysis.

Download:

Fig 1. Logical flow of the planned work.

https://doi.org/10.1371/journal.pone.0318712.g001

Methods

Active set selection and generation of DEKOIS 2.0 benchmark set

Filtration and clustering of our active set were performed using Data-Warrior software [23] resulting in 19 clusters adjusting fragment fingerprint as a descriptor, and the highest similarity fell below 0.6. DEKOIS 2.0 protocol [18,19] was employed to the compiled 40 structures of Mpro bioactives retrieved from both COVID-Moonshot and literature to create 1200 decoys (1:30 ratio) [24–26].

Preparation of small molecules

All the small molecules were built and prepared using the Molecular Operating Environment (MOE) version 2019.01 [27]. The preparation procedure was applied as reported earlier [28]. More details can be found in S1 File.

The SD files were transformed and split into separate PDBQT files for docking studies with AutoDock Vina using OpenBabel [29]. For PLANTS, the molecules were converted to mol2 format and the compatible atom types were specified using SPORES software [30,31].

Preparation of the protein structures

Prior to the docking studies, the protein structure of SARS-CoV-2 WTMpro (PDB ID: 6LU7 and 7L14) and OMPro (PDB ID: 7TLL) were prepared using MOE, as reported earlier [28]. Further preparation details can be found in the S1 File.

Screening and docking experiments

For screening using AutoDock Vina (version 1.1.2) [32], the Python script (prepare receptor4.py) offered by the MGLTools package (version 1.5.4) [33]was used to convert protein files to PDBQT format. The efficiency of the search algorithm was retained at its default settings. The grid box docking dimensions were 22.88 Å × 29.62 Å × 25.12 Å, with a spacing of 1 Å to deal with all the possible conformations of the docked molecules.

For screening using PLANTS, “ChemPLP” was used as the scoring function for docking utilizing PLANTS [34] with the “screen” mode selected. The receptor atoms around the co-crystal ligand coordinates by 5 Å were included in the binding site. Then the best-scored pose of each ligand was retrieved for further analysis.

For screening using FRED, Multi-conformer file generation for the small molecule database was carried out using the OMEGA module of OpenEye software [35] with a default number of 200 conformers generated for each ligand. The protein structures were prepared using MakeReceptor 4.1.1.0 GUI module of OpenEye software. Virtual screening was carried out using FRED module of OpenEye software [36,37].

For the screening using CDOCKER [38], a grid-based docking program (version 5.5). The binding site sphere of Mpro was defined as the regions that come within 14 Å radius from the geometric centroid of the co-crystalized ligand to make sure that all geometries of the docked compounds are covered. Other docking parameters were kept at the default level. The same docking protocol was carried out for WTMpro and OMpro sequences of SARS-COV-2.

Benchmarking utilizing pROC calculations and pROC-Chemotype plotting

The score-ordered rank was used in calculating the pROC-AUC using “R-Snippet” component of KNIME [39], based on the following equation [40]:

The bioactives number is given by n, while Di is the decoys fraction that is ordered higher than ith bioactive detected. The ith is the number of bioactive in the rank.

The pROC-Chemotype plots were plotted by the “pROC-Chemotype plot” tool [41].

To assess the ability of the docking tool to recognize true positives, from the active set, in the score-ordered list compared to the random collection, enrichment factor (EF) was computed based on the following equation [42] :

Drug likeness and ADME/TOX

Drug likeness and ADMET date were checked using SwissADME (http://www.swissadme.ch/) after generation of SMILES for the compounds using mol files using the pkCSM webtool (https://biosig.lab.uq.edu.au/pkcsm/).

Molecular dynamics simulations

Molecular dynamics simulations were conducted using GROMACS 2022 [43]. The solvation of each protein-ligand complex was carried out in a dodecahedron box of TIP3P explicit water model [44]. The preparation, equilibration and production settings were adapted as reported earlier [28], with an exception that the production runs were for 100 ns. Detailed procedures can be found in the S1 File.

The trajectories obtained from the MD simulations were analyzed via RMSD, RMSF, RoG, H-bond count and minimum distance analysis, using tools available in the GROMACS 2022 package. Graphical representations were made using Grace Software (https://plasma-gate.weizmann.ac.il/Grace/).

MM GBSA calculations

The binding free energy of the top hits and reference compounds were assessed via MM GBSA approach by applying gmx_MMPBSA tool v1.5.2 [45]. The tool was conjugated with GROMACS to conduct AMBER’s based calculations [46].

Results and discussion

Structural insights on Mpro

Mpro is a homodimer composed of three domains, the N-terminal domain I (residues 10–99), domain II (residues 100–182), where both presenting the catalytic domains, and the C-terminal helical domain III (residues 198–303) [12], as illustrated in Fig 2. Mpro binding site (Fig 2B) constructs vital subsites near the cleavage region of the substrate, named S1, S2, S3, S4 and S5 [11]. They correspond to the positions P1, P2, P1′ that are essential for substrate specificity, and P3, P4 which are essential for substrate binding stability (Fig 2C) [11]. S1 holds the oxyanion hole loop (Gly138 – Gly146), while S2 holds a catalytic dyad between the interface of domain I with HIS41 and CYS145 from domain II [47]. The catalytic loop is formed via residues Ile43 – Tyr54 together with the residues Glu166–His172 from P1 position [47]. Finally, the S4 linker loop is composed of residues Asp187 – Ala193 [47].

Download:

Fig 2. The crystal structure of dimer SARS-CoV-2 Mpro Omicron mutant P132H (PDB ID 7TLL).

(A) The three different domains I, II, and III of the protein structure. (B) Structure of the catalytic pocket loops showing the site of P132H mutation; (C) the substrate-binding subpockets of Mpro.

https://doi.org/10.1371/journal.pone.0318712.g002

Although the OMpro P132H mutation is not directly at Mpro’s catalytic site and being 22 Å away (Fig 2A and B), it was found to cause resistance to the therapeutic effects of drugs [48]. A recent comparative molecular dynamics (MD) study shed light on the P132H mutation resistance mechanism proposing the possibility of this mutation’s ability to alter the enzyme’s conformational flexibility [49], despite an unchanged catalytic efficiency and a compromised thermal stability that was previously reported [48].

In this study, we used the Mpro structure (PDB ID: 6LU7) in our benchmarking analysis, VS and MD studies since it was frequently employed in different studies [26,50,51]. Furthermore, we selected another WT structure (PDB ID: 7L14) for the VS and MD investigation since it reflects a non-covalent binding scenario for small molecules to Mpro. Moreover, the mutant OMPro P132H (PDB ID: 7TLL) was utilized for the cross-benchmarking investigations, and subsequent VS and MD experiments.

Selection of the active set molecules

To build our actives set, we downloaded Mpro inhibitors that were available at the COVID-Moonshot website (https://dndi.org/research-development/portfolio/covid-moonshot/). Of those compounds, we excluded all compounds with covalent warheads and fragments. Additionally, we manually collected 35 non-peptide, non-covalent compounds from the literature [24–26] to attain a maximum diversity in the chemotype clusters. Subsequently, we merged the inhibitors, and filtration was carried out to exclude ligands with IC₅₀ values above 2 µ M or containing chiral centers. At the end, we selected 40 bioactive compounds comprising 7 co-crystal ligands with Mpro protein structures, with PDB codes: 7L13, 7M8P, 7M91, 7M8N, 7L11, 7L14, and 6M2N. Upon clustering, these compounds were found to collectively represent 19 scaffold chemotype classes as presented in Table A in S1 File. The complete list of the 40 bioactive compounds can be found in Table B in S1 File. For decoys generation, these 40 candidates were subjected to the DEKOIS 2.0 protocol [18,19] which creates 30 structurally related decoys per compound. Eventually, we compiled a challenging decoy set of 1200 compounds where the whole set of bioactives and decoys was used to assess the performance of the four docking tools investigated.

Benchmarking analysis

A docking study of the 1240 actives plus decoys was launched on the four docking tools AutoDock Vina, FRED, PLANTS, and CDOCKER to assess their performance. The results obtained from this docking investigation reflected that all docking tools exhibited better-than-random performance judged by having a pROC-AUC value of more than 0.43 for the two assessed Mpro sequences. Noteworthy is that pROC-AUC is a modified version plotted in semi-logarithmic scale instead of a linear receiver operating characteristics (ROC) curve [40]. Therefore, the calculation of the area under the curve (AUC) is biased to emphasize the detection of early enriched actives [40]. For the WTMpro sequence (PDB ID: 6LU7), AutoDock Vina showed the best screening performance with pROC-AUC value of 0.98, followed by FRED with pROC-AUC value of 0.68, as seen in (Fig 3). However, cross-benchmarking experiments for the OMPro mutant recommended both FRED and AutoDock Vina for the best screening performance by virtue of their pROC-AUC values of 1.2, and 0.92, respectively (Fig 3). Collectively, AutoDock Vina and FRED showed superior predictive power compared to the other two tools. This directed us to utilize AutoDock Vina in the subsequent experimental explorations adopted in this study.

Download:

Fig 3. pROC plots illustrating the screening performance of AutoDock Vina, FRED, PLANTS and CDOCKER.

The left panel is for the SARS-CoV-2 WTMpro (PDB ID: 6LU7). The right panel is for SARS-CoV-2 OMpro P132H mutant.

https://doi.org/10.1371/journal.pone.0318712.g003

Chemotype enrichment

To investigate the highest-ranking cluster/chemotype, an in-depth performance analysis via pROC-Chemotype plots was conducted [41]. The results illustrate a high enrichment of the best potent bioactives in the early docking ranks. This was done for the docking results obtained from AutoDock Vina against both the WT and Omicron Mpro structures and FRED against Omicron Mpro, in Figs 4–6, respectively. It is worth mentioning that there is no reported data available about the inhibitory effect of the active set compounds against the OMpro variant. Therefore, the cross-benchmarking experiment against the OMpro using the benchmark set of WTMpro can provide insights on how the docking tool behave against a closely similar protein structure with completely identical binding site residues.

Download:

Fig 4. pROC-Chemotype plot for the benchmarking using AutoDock Vina against SARS-CoV-2 WTMpro.

(A) The pROC-Chemotype plot where the docking information corresponds to the chemotype described by the cluster number and the bioactivity information. (B) Box plot of the fitness vs. chemotype clusters. (C) 3D representation of the best ranked bioactive molecule in the binding site of Mpro.

https://doi.org/10.1371/journal.pone.0318712.g004

Download:

Fig 5. pROC-Chemotype plot for the benchmarking using FRED against SARS-CoV-2 OMpro (P132H).

(A) The pROC-Chemotype plot where the docking information corresponds to the chemotype described by the cluster number and the bioactivity information. (B) Box plot of the fitness vs. chemotype clusters. (C) 3D representation of the best ranked bioactive molecule in the binding site of Mpro.

https://doi.org/10.1371/journal.pone.0318712.g005

Download:

Fig 6. pROC-Chemotype plot for the benchmarking using AutoDock Vina against SARS-CoV-2 OMpro (P132H).

(A) The pROC-Chemotype plot where the docking information corresponds to the chemotype described by the cluster number and the bioactivity information. (B) Box plot of the fitness vs. chemotype clusters. (C) 3D representation of the best ranked bioactive molecule in the binding site of Mpro.

https://doi.org/10.1371/journal.pone.0318712.g006

The diversity of the bioactive chemotype clusters (19 distinct scaffolds) underlines the challenge inherent in evaluating the performance of the utilized docking tools. Their bioactivity profiles are symbolized by the level of activity (LOA), extending from 10⁻⁶ to 10⁻⁸ M, and reported as IC₅₀ as a type of data (TOD), as displayed in (Fig 4A). The pROC-Chemotype plot visualized that AutoDock Vina can identify potent binders at early enrichment. For instance, at 1% of the score-ordered library, many bioactive molecules were recognized, leading to an Enrichment Factor (EF 1%) of 18.08 in case of the WTMpro (Fig 4A), especially at 1% of the score-ordered library. Furthermore, the best four potent bioactive compounds (with the bioactivity rank = 1-4) were recognized at the same cutoff. The enrichment factor (EF 1%) is 18.08, emphasizing a promising predictive capability of the docking tool. This signifies that AutoDock Vina with WTMpro is able to recognize active molecules more than the random performance by 18-fold at early enrichment. Furthermore, Fig 4B exhibits the docking fitness (fitness = docking score multiplied by −1) allocation of the bioactive molecules. The docking score is ranging from −9.8 (best score) to −6.7 (lowest score) and presented as fitness values of 9.8 to 6.7. Furthermore, molecules demonstrating cluster 1 were found to lie in a superior area of fitness (i.e., fitness > 9) compared to other scaffolds (Fig 4A). Docking pose of the best ranked compound (docking rank 1) from the active set in the binding site (PDB ID: 6LU7) is shown in Fig 4C, where key interactions with residues; His163, Cys145, Thr26 and His41 were mapped.

For the cross-benchmarking against OMpro, both FRED and AutoDock Vina exhibited the best performance with pROC-AUC values of 1.2 and 0.92, respectively. FRED chemotype analysis indicated an EF 1% = 20.4 (Fig 5A). The fitness distribution of the bioactive molecules exhibited that scaffolds of cluster 6 and 7 were best recognized at the early rank, as shown in Fig 5B. The best-scored docking pose (compound 12) presented the key interactions with Glu166, His163 and Phe140 (Fig 5C).

Unlike FRED, AutoDock Vina was able to enrich the most potent active (compound 1, bioactivity rank = 1) as the best scored docking pose (docking rank = 1), as well as the subsequent four potent molecules (bioactivity rank = 2-5), as shown in Fig 6A. Notably, scaffold of cluster 1 is entirely enriched at the early enrichment (at 1%), as revealed in the fitness distributions of the clusters in Fig 6B. Elucidating the interactions of the best ranked docking pose in the binding site of OMPro, compound 1 presented the key interactions with Cys145, His163 and Thr26 (Fig 6C).

Virtual screening of a focused library based on benchmarking results

Based on the chemotype behavior in Figs 4 and 6, the most potent bioactive cluster (cluster 1 -N-aryl pyridone derivatives) were retrieved by AutoDock Vina at early enrichment. Interestingly, the literature reported that perampanel derivatives (cluster 1 scaffold) display a promising skeleton for novel inhibitors for WTMpro [26,52]. Therefore, this top enriched chemotype class encouraged us to integrate the parent drug (perampanel) into a scaffold hopping transformation strategy using the DrugspaceX database https://drugspacex.simm.ac.cn/ that enables expert-defined transformations of approved drugs for virtual screening with lingering structural novelty. We hence generated a focused library of 636 compounds based on class 1 scaffold for prospective virtual screening against the WTMpro. Furthermore, we selected a non-covalent co-crystal structure (PDB ID: 7L14) for the VS effort complexed with compound number 9 (cluster 1) in Table B in S1 File. It is noteworthy that upon docking of the 3 top ranked compounds against 6LU7on 7L14, we obtained similar docking ranks where compound 203 came first, then compound 386, then compound 329, as shown in (Table 1).

Download:

Table 1. The best enriched compounds against WTMpro and OMpro.

https://doi.org/10.1371/journal.pone.0318712.t001

The generated library was screened against OMpro (PDB ID: 7TLL) complexed with nirmatrelvir using both FRED and AutoDock Vina. We focused on the top 10% of the best scored compounds using both docking tools and explored the consensus between both of them. Moreover, focusing on detecting the compounds with comparable poses and similar important interactions, we were able to identify 3 compounds as best potential ligands for OMpro viz. compound 401 with scores of −9.4, −12.4; compound 541 with scores of −9.4, and −10.6; and compound 385 with scores of −10.15 and −9.9 for AutoDock Vina and FRED, respectively (Table 1 and Table C in S1 File).

Description of the postulated binding poses

The docking pose of compound 386, transformed from perampanel by replacing the phenyl ring with a bioisosteric tetrazole ring, showed multiple interactions including five anticipated hydrogen bonds between the ligand and the WTMpro (Fig 7A). The pyridone carbonyl oxygen showed two H-bonding interactions with Gly143 and Asn142, and a hydrophobic contact with Cys145. The cyanophenyl ring displayed a hydrophobic contact with Glu166 while forming H-bonding interaction with Cys145. The tetrazole ring formed a H-bonding interaction with the backbone of Thr26. Finally, the pyridine ring showed hydrophobic interaction with Met165 of S4 pocket.

Download:

Fig 7. The docking poses in binding site of the WTMpro (PDB ID: 7L14) represented by 3D and 2D.

(A) The pose of compound 386. (B) The pose of compound 329. (C) The pose of compound 203.

https://doi.org/10.1371/journal.pone.0318712.g007

The pose of compound 329, transformed from the perampanel by changing the phenyl ring to hydroxytriazoline, demonstrated good binding with the active site of WTMpro. The hydroxytriazoline group formed H-bonding interaction with Ser46, and the pyridone carbonyl oxygen exhibited also H-bonding interactions with Gly143 and Asn142. Additionally, Cys145 was involved in a hydrophobic contact with the same ring (Fig 7B).

The docking pose of compound 203 (Fig 7C) showed that the chromeno pyridine group established three H-bonding interactions with Glu166 (P1 catalytic loop), Gln189 (S4 sub pocket) and Thr25 (S1’ pocket).

The postulated binding pose of compound 401 exhibited three H-bond interactions with Cys145, His163 and Glu166 (S1 pocket). Furthermore, favorable interactions with Met49 (S2 pocket), and Met165 and Gly143 (S4 pocket) were observed (Fig 8A).

Download:

Fig 8. The docking poses in binding site of the OMpro (PDB ID: 7TLL) represented by 3D and 2D.

(A) The pose of compound 401. (B) The pose of compound 541. (C) The pose of compound 385.

https://doi.org/10.1371/journal.pone.0318712.g008

The pose of compound 541 showed four H-bonding interactions with Thr26, Met49 (S2 pocket), and His163, Cys145 (S1 pocket), as shown in Fig 8B [53]. Finally, the pose of compound 385 displayed H-bonding interaction with Cys145, Glu166 and Met165 (Fig 8C).

ADME/Tox and drug-likeness filtration applying expert systems

To assess the pharmacokinetic (PK) profile of the top hits, pkCSM (http://biosig.unimelb.edu.au/ pkcsm/) online server was used. Starting with the absorption parameter as represented by the Intestinal absorption parameter, the value of which must be greater than 30%, compounds 203, 401, 541 and 385 showed 100% values, while compounds 386 and 329 showed 93% and 72% values, respectively. For BBB permeability, if a compound has a log BB value greater than 0.3, it is anticipated to readily cross the BBB, while if the value is less than −1, it will poorly be distributed to the brain. For the evaluated compounds, only compounds 401, and 541 were found to demonstrate possible BBB permeation ability. This observation predicts the lack of potential CNS side effects of the majority of the compounds. The toxicity prediction showed that only compound 203 has positive AMES toxicity (Table 2). To evaluate the drug-likeness and PK profile of the compounds we used Swiss ADME (http://www.swissadme.ch/). Interestingly, no compound showed a violation of Lipinski’s rule. All the data related to the number of H bond donors and acceptors, consensus Log Po/w, TPSA, and bioavailability are shown in (Table 3). Collectively, all compounds carried acceptable criteria, except compounds 203, and 385. Therefore, we excluded those two compounds from further investigation due to their potential toxicity, and poor solubility profiles.

Download:

Table 2. ADME/Tox data for hits against WTMpro and OMpro.

https://doi.org/10.1371/journal.pone.0318712.t002

Download:

Table 3. Drug likeness, Lipinski’s rule, and PK properties for hits against WTMpro and OMpro.

https://doi.org/10.1371/journal.pone.0318712.t003

Molecular dynamics simulations

The two top-enriched ligands from DrugSpaceX for each of the WTMpro (compounds 386 and 329) and OMpro (compounds 401 and 541) were exposed to 100 ns molecular dynamics (MD) simulations to assess their stability in the binding sites in a time-dependent manner. Furthermore, for comparison purposes, additional MD runs were conducted for the non-liganded (apo) proteins and co-crystallized systems for both PDB ID 7L14, and 7TLL pertaining to the WTMpro and OMpro, respectively. This results in a total of eight MD runs, with a 100 ns each.

Root mean square deviation (RMSD)

For the WTMpro complex systems, the root mean square deviation (RMSD) for ligand’s heavy atoms showed high fluctuation between 2-3 nm for the co-crystallized ligand (blue). While both compounds 386 and 329 exhibited stable fluctuation with RMSD less than 0.5 nm (Fig 9A). The stability of the protein dynamics was evaluated via RMSD which was calculated on the alpha carbon atoms. At the first 20 ns; the apoprotein (black) exhibited RMSD values between 0.1 to 0.4 nm. Compound 329 (green) was able to stabilize the protein structure showing an RMSD value of less than 0.2 nm, then exhibited a higher fluctuation around an RMSD value of 0.3 nm. Likewise, the co-crystal ligand (blue) and compound 386 (red) showed acceptable RMSD values around 0.3 nm especially after 20 ns of time until the end of 100 ns simulation time (Fig 9B). Intermolecular hydrogen bonding can be utilized as a metric to evaluate the degree of protein-ligand binding as well as the stability of the complex. During a 100 ns simulation, the total number of hydrogen bonds formed by the two hit compounds ranged from 1 to 8, compared to a maximum of 4 H-bonding count for the co-crystal ligand, as depicted in (Fig 9C). This indicates the potential stable binding of these compounds.

Download:

Fig 9. The RMSD and H-bonding count analysis of the MD simulation for the WTMpro complex systems.

(A) Ligand RMSD. (B) Protein RMSD. (C) H-bonding count. Blue line is for co-crystal ligand, while red is for compound 386 and green is for compound 329, in all graphs.

https://doi.org/10.1371/journal.pone.0318712.g009

For OMpro complex systems, the ligand’s heavy atoms RMSD was evaluated. At the first 40 ns the co-crystallized ligand showed fluctuations of the RMSD values between 0.1-0.5 nm then stabilized between 0.1-0.3 nm afterwards. Similarly, compounds 401 and 541 showed fluctuations around RMSD values 0.1-0.3 nm then 0.2-0.35 nm till the end of the simulation (Fig 10A). Assessing the protein dynamics stability via RMSD (Fig 10B) indicated acceptable fluctuations for all systems. For instance, RMSD values showed a range from 0.2 to 0.4 nm of co-crystal complex (blue) and compound 401 (red), and from 0.1 to 0.3 nm for compound 541 (green). Furthermore, the unliganded (apo) showed minor fluctuations ranging from RMSD values of 0.2 to 0.3 nm. The H-bond count for the complex systems (Fig 10C) showed good stability for both compound 401 (red) and compound 541 (green), with at least 2 to 3 H-bond count along the simulation time. The co-crystal ligand exhibited 4 H-bond count on average, especially after 40 ns of simulation time.

Download:

Fig 10. The RMSD and H-bonding count analysis of the MD simulation for the OMpro complex systems.

(A) Ligand RMSD. (B) Protein RMSD. (C) H-bonding count. Blue line is for the co-crystal ligand, while red is for compound 401 and green is for compound 541, in all graphs.

https://doi.org/10.1371/journal.pone.0318712.g010

Radius of gyration

The Radius of gyration (RoG) is a general metric for the compactness of the protein during the simulation time. The WT-complex systems displayed acceptable RoG value ranges for all systems. For instance, the unliganded system (black) exhibited RoG values from 2.2 to 2.3 nm (Fig 11A). Likewise, co-crystal (blue), compound 386 (red) and compound 329 (green) complex systems showed a narrow range of RoG values, from 2.2 to 2.25 nm. This indicates the acceptable compactness of the simulated systems and the absence of major conformational discrepancies. The same outcome can be concluded from the OMpro complex systems (Fig 11C).

Download:

Fig 11. The RoG and RMSF analysis of the MD simulation for the WTMpro complex systems.

(A) Protein radius of gyration (RoG). (B) Per-residue root mean square fluctuation (RMSF). For OMpro (C) Protein radius of gyration (RoG). (D) Per-residue root mean square fluctuation (RMSF). The color scheme is represented as: blue line is for the co-crystal ligand in both WT and Omicron. The red line is for compound 386-WTMpro and compound 401-OMpro complex systems. The green line is for compounds 329-WTMpro and 541-OMpro complex systems.

https://doi.org/10.1371/journal.pone.0318712.g011

RMSF analysis

Root mean square fluctuation (RMSF) quantifies the conformational perturbations per-residue during the simulation time. We noticed shared flexibility behavior in all systems in certain regions, for the complex systems with WTMpro (Fig 11B), and the complex systems with the OMpro (Fig 11D). For Instance, the P132H mutation region of OMpro, the co-crystal and compound 401 showed more stable fluctuation than the apo structure. Comparatively, both showed fluctuation lower than 0.2 nm except in residues around 225 and 275. However, compound 451- OMpro (green) displayed high flexibility in most regions. Some high fluctuations in the RMSF values are reflected in the flexibility of structural loop regions for the Mpro protein. The flexibility of these regions was greatly minimized in liganded systems. Interestingly, complex systems with compound 386 (red), co-crystal ligand (blue), and compound 329 (green) exhibited lower RMSF at the binding site amino acids compared to the unliganded system. This suggests a pronounced stabilization effect of the complexed ligands and their influence in reducing the conformational perturbations of the binding site residues.

Binding free energy (ΔG) via MM GBSA calculation

Binding free energy (ΔG) for all the 4 complexes compound 386, compound 329 with the WTMpro, and compound 401 and compound 541 with the OMpro were calculated for the whole 100 ns of the simulation and averaged. In addition, the two co-crystal ligands were calculated for comparison purposes. This would provide insights to estimate the overall binding strength. Table 4 shows the binding free energy (ΔG) and its components for all complex systems. The free energies are calculated based on two main terms, the Gibbs free energy of the gas (ΔG_GAS) and solvation free energy (Gsolv). The first term ΔG_GAS is composed of the sum of electrostatic energy (Δ_ELE) and van der Waals energy (_VDWAALS)). The second term (Gsolv) is represented by the sum of generalized Born Polar solvation energy (E_GB) and the non-polar solvation energies ΔE_SURF [54]. For the WT complex systems, compound 329 complex showed superior binding compared to both the co-crystal ligand and compound 386, with a ΔG value of −31.32 kcal/mol, compared to −21.15 kcal/mol and −19.16 kcal/mol for the co-crystal ligand and compound 386, respectively. The major contributing component to this superior binding is the van der Waals energy (_VDWAALS) (Table 4). This highlights the importance of the hydrophobic groups and aromatic groups of the proposed binders. For the Omicron complex systems, both nirmatrelvir (co-crystal ligand) and compound 401 showed the best and comparable ΔG values of −37.88 kcal/mol and −37.93 kcal/mol, respectively. These results indicated that hit compounds had an acceptable range of ΔG and formed a stable complex with the WTMpro and its OMpro. Again, the van der Waals energy is the main contributing term for the superior binding to OMpro of compound 401. To provide more insights, the total energy components for the best two binders were plotted as shown in Fig 12A and B for compounds 329, and 401, respectively. The energy composition analysis verified the strong interaction of compound 329 with Met49, Met165, Ser46, Asn142, Gly143, Thr45, and Cys145 of the WTMpro, as displayed in the individual energy contribution graph (Fig 12C). Also, the heatmap showed some residues to maintain the individual energy contribution throughout the 100 ns simulation’s time, specifically, Met49 and Ser46 (Fig 12E). For OMpro, the individual energy contribution graph of compound 401 showed favorable interactions with Met165, Glu166, Met49, Asn142, His41, Thr25, and Ser144. Interestingly, the heatmap (Fig 12F) confirmed the maintenance of the energy contribution through the simulation’s time.

Download:

Table 4. Binding free energy of ligand-Mpro complexes.

https://doi.org/10.1371/journal.pone.0318712.t004

Download:

Fig 12. Binding free energy (

ΔG) analysis using MM GBSA. (A) Free energy components of compound 329 bound to WTMpro. (B) Energy components for compound 401 bound to OMpro. (C) Individual residue’s energy contribution for compound 329. (D) Individual residue’s energy contributions for compound 401. (E), (F). Heatmap showing the energy contribution of individual residues plotted against time for compounds 329, and 401, respectively.

https://doi.org/10.1371/journal.pone.0318712.g012

Minimum distance analysis

Elucidating more comprehension into the poses of compound 329-WTMpro complex during the MD simulation, we calculated the minimum distance between the interacting ligand and protein atoms (Fig 13). Initially, the O atom of compound 329 showed H-bonding interactions with the HN atom of Gly143. This favorable interaction is mirrored in distances of ~ 0.3 to 0.4 nm (i.e., 3–4 Å), till near 60 ns and then the distance graph proposes that some dynamics affected the distance to reach ~ 1 nm. Interestingly, the position of the Met49 residue allowed the new interactions to appear from 40 ns between SD atom and two carbon atoms (C9 and C17) of compound 329, reflecting the minimum distance starting from ~ 0.6 to 0.8 nm (i.e., 6–8 Å) (for C17) and then stabilized at 70 ns till the end. Likewise, ligand’s O2 atom and NH atom of Ser46 displayed H-bonding interactions with comparable range of distances ~ 0.2 to 0.6 nm (i.e., 2–6 Å). Overall, this highlights the ability of compound 329 to exhibit mainly one pose during the simulation time until 60/70 ns, and another pose from 70 ns until the end. For the complex of compound 401 and OMpro (Fig 14), the ligand adapted one major pose along the simulation time. This is evident from the consistent distances (around ~ 0.2 nm) between the interacting atoms of the pyridone core (carbonyl oxygen atom) and the protein (H atom of Glu166). Likewise, stable distances indicated by the interacting atoms around ~ 0.2-0.3 nm were observed between SD of Met49 and H of Ser144 with C11 and N2 of compound 401, respectively. Another stable distance between OD1 atom of Asn142 and C9 of compound 401 (around ~ 0.6 nm) indicates one major pose of the ligand.

Download:

Fig 13. The minimum distance graph of compound 329-WTMpro complex system.

Upper panel: The minimum distance graph throughout the MD simulation. Lower panel: snapshots at various simulation times.

https://doi.org/10.1371/journal.pone.0318712.g013

Download:

Fig 14. The minimum distance graph of compound 401-OMpro complex system.

Upper panel: The minimum distance graph throughout the MD simulation. Lower panel: snapshots at various simulation times.

https://doi.org/10.1371/journal.pone.0318712.g014

Conclusion

The aim of this study was a biphasic one, first to carry out an in-depth benchmarking investigation to recommend the best performing docking tool against SARS-CoV-2 Mpro on both wild type and the Omicron P132H variant. In order to do that, we assembled an active data set of small molecule inhibitors of the wild-type Mpro from the COVID-Moonshot database and literature, that was used to assess the performance of four docking tools, AutoDock Vina, FRED, PLANTS, and CDOCKER, we consequently created a top-notch DEKOIS 2.0 benchmark set for challenging these tools to discriminate between the actives and decoys. Benchmarking analysis assisted by pROC-Chemotype plots demonstrated that the most potent bioactives were highly enriched in the early docking ranks using AutoDock Vina for WTMPro and by both FRED and AutoDock Vina in the case of OMPro. The best enriched chemotype was detected to be cluster 1 (N-aryl pyridones), a well-known inhibitor belonging to which is Perampanel. In light of the benchmarking results, the SBVS was conducted for a targeted library transformed from Perampanel by scaffold hopping. The library was subjected to VS against both WTMpro and OMpro. Identifying two very well performing hits for each target, their stable binding was verified by MM GBSA calculations and molecular dynamics simulations. We thus identified compound 329 for WTMpro, and compound 401 for OMpro as very promising inhibitors and potentially good candidates for eliciting antiviral effect against SARS-CoV-2. Our work in progress is synthesizing those compounds for further extensive in vitro testing in the near future. In summary, this work provides an example of how to employ the power of benchmarking and SBVS to identify chemotypes for computer-assisted discovery of MPro inhibitors for the management of SARS-CoV-2.

Supporting information

S1 File. SI_file01 contains detailed description of the methodology of: Preparation of small molecules, preparation of the protein structures, and molecular dynamics simulations.

https://doi.org/10.1371/journal.pone.0318712.s001

It also contains Table A for the active set chemotype clusters used for benchmarking, and Table B for the active set compounds used for benchmarking. It includes Table C showing the structures of Table 1. SI_file02 contains the SMILES of the active and decoy molecules.(ZIP)

Acknowledgments

TMI acknowledges the cluster of Bibliotheca Alexandrina High-Performance Computing (BA-HPC) for awarding access to perform some calculations, and the lab of Prof. Frank Boeckler (Tuebingen University) for granting access to some computational tools. The authors acknowledge the support of OpenEye Scientific Software Inc. for offering a free academic license.

References

1. Andersen KG, Rambaut A, Lipkin WI, Holmes EC, Garry RF. The proximal origin of SARS-CoV-2. Nat Med. 2020;26(4):450–2. Epub 2020/04/15. pmid:32284615
- View Article
- PubMed/NCBI
- Google Scholar
2. Flannagan J, Twohig KA, Carter E, Chudasama DY, Lamagni T, Dabrera G. Increased residential clustering of COVID-19 cases associated with SARS-CoV-2 variant of concern B.1.1.7. Epidemiology. 2022;33(4):465–9. Epub 2022/05/06. pmid:35512232; PMCID: PMCPMC9148639
- View Article
- PubMed/NCBI
- Google Scholar
3. Davies NG, Abbott S, Barnard RC, Jarvis CI, Kucharski AJ, Munday JD, et al. Estimated transmissibility and impact of SARS-CoV-2 lineage B.1.1.7 in England. Science. 2021;372(6538):eabg3055. Epub 2021/03/05. pmid:33658326; PMCID: PMCPMC8128288
- View Article
- PubMed/NCBI
- Google Scholar
4. Wang P, Nair MS, Liu L, Iketani S, Luo Y, Guo Y, et al. Antibody resistance of SARS-CoV-2 variants B.1.351 and B.1.1.7. Nature. 2021;593(7857):130–5. Epub 2021/03/09. pmid:33684923
- View Article
- PubMed/NCBI
- Google Scholar
5. Madhi SA, Baillie V, Cutland CL, Voysey M, Koen AL, Fairlie L, et al.; NGS-SA Group. Efficacy of the ChAdOx1 nCoV-19 Covid-19 vaccine against the B.1.351 variant. N Engl J Med. 2021;384(20):1885–98. Epub 2021/03/17. pmid:33725432; PMCID: PMCPMC7993410
- View Article
- PubMed/NCBI
- Google Scholar
6. Wang P, Casner RG, Nair MS, Wang M, Yu J, Cerutti G, et al. Increased resistance of SARS-CoV-2 variant P.1 to antibody neutralization. Cell Host Microbe. 2021;29(5):747–51.e4. Epub 2021/04/23. pmid:33887205
- View Article
- PubMed/NCBI
- Google Scholar
7. Ng OT, Koh V, Chiew CJ, Marimuthu K, Thevasagayam NM, Mak TM, et al. Impact of delta variant and vaccination on SARS-CoV-2 secondary attack rate among household close contacts. Lancet Reg Health West Pac. 2021;17:100299. Epub 2021/11/09. pmid:34746899; PMCID: PMCPMC8560026
- View Article
- PubMed/NCBI
- Google Scholar
8. Allen H, Vusirikala A, Flannagan J, Twohig KA, Zaidi A, Chudasama D, et al. COVID-19 Genomics UK (COG-UK Consortium). Household transmission of COVID-19 cases associated with SARS-CoV-2 delta variant (B.1.617.2): national case-control study. Lancet Reg Health Eur. 2022;12:100252. Epub 2021/11/04. pmid:34729548; PMCID: PMCPMC8552812
- View Article
- PubMed/NCBI
- Google Scholar
9. Hastert FD, Hein S, von Rhein C, Benz NI, Husria Y, Oberle D, et al. The SARS-CoV-2 variant omicron is able to escape vaccine-induced humoral immune responses, but is counteracted by booster vaccination. Vaccines (Basel). 2022;10(5). Epub 2022/05/29. pmid:35632550; PMCID: PMCPMC9144538
- View Article
- PubMed/NCBI
- Google Scholar
10. Dejnirattisai W, Shaw RH, Supasa P, Liu C, Stuart AS, Pollard AJ, et al.; Com-COV2 study group. Reduced neutralisation of SARS-CoV-2 omicron B.1.1.529 variant by post-immunisation serum. Lancet. 2022;399(10321):234–6. Epub 2021/12/24. pmid:34942101; PMCID: PMCPMC8687667
- View Article
- PubMed/NCBI
- Google Scholar
11. Jin Z, Du X, Xu Y, Deng Y, Liu M, Zhao Y, et al. Structure of M(pro) from SARS-CoV-2 and discovery of its inhibitors. Nature. 2020;582(7811):289–93. Epub 2020/04/10. pmid:32272481
- View Article
- PubMed/NCBI
- Google Scholar
12. Zhang L, Lin D, Sun X, Curth U, Drosten C, Sauerhering L, et al. Crystal structure of SARS-CoV-2 main protease provides a basis for design of improved α-ketoamide inhibitors. Science. 2020;368(6489):409–12. Epub 2020/03/22. pmid:32198291; PMCID: PMCPMC7164518
- View Article
- PubMed/NCBI
- Google Scholar
13. Iketani S, Mohri H, Culbertson B, Hong SJ, Duan Y, Luck MI, et al. Multiple pathways for SARS-CoV-2 resistance to nirmatrelvir. Nature. 2023;613(7944):558–64. Epub 2022/11/10. pmid:36351451; PMCID: PMCPMC9849135
- View Article
- PubMed/NCBI
- Google Scholar
14. Vangeel L, Chiu W, De Jonghe S, Maes P, Slechten B, Raymenants J, et al. Remdesivir, Molnupiravir and Nirmatrelvir remain active against SARS-CoV-2 Omicron and other variants of concern. Antiviral Res. 2022;198:105252. Epub 2022/01/28. pmid:35085683
- View Article
- PubMed/NCBI
- Google Scholar
15. Mengist HM, Mekonnen D, Mohammed A, Shi R, Jin T. Potency, safety, and pharmacokinetic profiles of potential inhibitors targeting SARS-CoV-2 main protease. Front Pharmacol. 2020;11:630500. Epub 2021/02/19. pmid:33597888; PMCID: PMCPMC7883113
- View Article
- PubMed/NCBI
- Google Scholar
16. Unoh Y, Uehara S, Nakahara K, Nobori H, Yamatsu Y, Yamamoto S, et al. Discovery of S-217622, a noncovalent Oral SARS-CoV-2 3CL protease inhibitor clinical candidate for treating COVID-19. J Med Chem. 2022;65(9):6499–512. Epub 2022/03/31. pmid:35352927
- View Article
- PubMed/NCBI
- Google Scholar
17. Zeyaullah M, Khan N, Muzammil K, AlShahrani AM, Khan MS, Alam MS, et al. In-silico approaches for identification of compounds inhibiting SARS-CoV-2 3CL protease. PLOS ONE. 2023;18(4):e0284301. pmid:37058496
- View Article
- PubMed/NCBI
- Google Scholar
18. Bauer MR, Ibrahim TM, Vogel SM, Boeckler FM. Evaluation and optimization of virtual screening workflows with DEKOIS 2.0--a public library of challenging docking benchmark sets. J Chem Inf Model. 2013;53(6):1447–62. Epub 2013/05/28. pmid:23705874
- View Article
- PubMed/NCBI
- Google Scholar
19. Ibrahim TM, Bauer MR, Boeckler FM. Applying DEKOIS 2.0 in structure-based virtual screening to probe the impact of preparation procedures and score normalization. J Cheminform. 2015;7:21. Epub 2015/06/03. pmid:26034510; PMCID: PMCPMC4450982
- View Article
- PubMed/NCBI
- Google Scholar
20. Vogel SM, Bauer MR, Boeckler FM. DEKOIS: demanding evaluation kits for objective in silico screening--a versatile tool for benchmarking docking programs and scoring functions. J Chem Inf Model. 2011;51(10):2650–65. Epub 2011/07/22. pmid:21774552
- View Article
- PubMed/NCBI
- Google Scholar
21. Boeckler FM, Bauer MR, Ibrahim TM, Vogel SM. Use of DEKOIS 2.0 to gain insights for virtual screening. J Cheminform. 2014;6(Suppl 1):O24. Epub 2014/04/26. pmid:24765123; PMCID: PMCPMC3980182
- View Article
- PubMed/NCBI
- Google Scholar
22. Hassan HHA, Ismail MI, Abourehab MAS, Boeckler FM, Ibrahim TM, Arafa RK. In silico targeting of fascin protein for cancer therapy: benchmarking, virtual screening and molecular dynamics approaches. Molecules. 2023;28(3):1296. Epub 2023/02/12. pmid:36770963; PMCID: PMCPMC9921211
- View Article
- PubMed/NCBI
- Google Scholar
23. Sander T, Freyss J, von Korff M, Rufener C. DataWarrior: an open-source program for chemistry aware data visualization and analysis. J Chem Inf Model. 2015;55(2):460–73. Epub 2015/01/07. pmid:25558886
- View Article
- PubMed/NCBI
- Google Scholar
24. Yang J, Lin X, Xing N, Zhang Z, Zhang H, Wu H, et al. Structure-based discovery of novel nonpeptide inhibitors targeting SARS-CoV-2 M(pro). J Chem Inf Model. 2021;61(8):3917–26. Epub 2021/07/20. pmid:34279924
- View Article
- PubMed/NCBI
- Google Scholar
25. Su HX, Yao S, Zhao WF, Li MJ, Liu J, Shang WJ, et al. Anti-SARS-CoV-2 activities in vitro of Shuanghuanglian preparations and bioactive ingredients. Acta Pharmacol Sin. 2020;41(9):1167–77. Epub 2020/08/02. pmid:32737471; PMCID: PMCPMC7393338
- View Article
- PubMed/NCBI
- Google Scholar
26. Zhang CH, Stone EA, Deshmukh M, Ippolito JA, Ghahremanpour MM, Tirado-Rives J, et al. Potent noncovalent inhibitors of the main protease of sars-cov-2 from molecular sculpting of the drug perampanel guided by free energy perturbation calculations. ACS Cent Sci. 2021;7(3):467–75. Epub 2021/04/01. pmid:33786375; PMCID: PMCPMC7931627
- View Article
- PubMed/NCBI
- Google Scholar
27. Molecular operating environment (MOE). Montreal, QC, Canada: Chemical Computing Group Inc.; 2019.
28. Elghoneimy LK, Ismail MI, Boeckler FM, Azzazy HME, Ibrahim TM. Facilitating SARS CoV-2 RNA-dependent RNA polymerase (RdRp) drug discovery by the aid of HCV NS5B palm subdomain binders: in silico approaches and benchmarking. Comput Biol Med. 2021;134:104468. Epub 2021/05/21. pmid:34015671; PMCID: PMCPMC8111889
- View Article
- PubMed/NCBI
- Google Scholar
29. O’Boyle NM, Banck M, James CA, Morley C, Vandermeersch T, Hutchison GR. Open babel: an open chemical toolbox. J Cheminform. 2011;3:33. Epub 2011/10/11. pmid:21982300; PMCID: PMCPMC3198950
- View Article
- PubMed/NCBI
- Google Scholar
30. ten Brink T, Exner TE. Influence of protonation, tautomeric, and stereoisomeric states on protein-ligand docking results. J Chem Inf Model. 2009;49(6):1535–46. Epub 2009/05/21. pmid:19453150
- View Article
- PubMed/NCBI
- Google Scholar
31. ten Brink T, Exner TE. pK(a) based protonation states and microspecies for protein-ligand docking. J Comput Aided Mol Des. 2010;24(11):935–42. Epub 2010/10/01. pmid:20882397
- View Article
- PubMed/NCBI
- Google Scholar
32. Trott O, Olson AJ. AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. J Comput Chem. 2010;31(2):455–61. Epub 2009/06/06. pmid:19499576; PMCID: PMCPMC3041641
- View Article
- PubMed/NCBI
- Google Scholar
33. Sanner MF. Python: a programming language for software integration and development. J Mol Graph Model. 1999;17(1):57-61. Epub 2000/02/08. pmid:10660911
- View Article
- PubMed/NCBI
- Google Scholar
34. Korb O, Stützle T, Exner TE. Empirical scoring functions for advanced protein-ligand docking with plants. J Chem Inf Model. 2009;49(1):84–96. Epub 2009/01/08. pmid:19125657
- View Article
- PubMed/NCBI
- Google Scholar
35. Hawkins PC, Skillman AG, Warren GL, Ellingson BA, Stahl MT. Conformer generation with OMEGA: algorithm and validation using high quality structures from the Protein Databank and Cambridge Structural Database. J Chem Inf Model. 2010;50(4):572–84. Epub 2010/03/20. pmid:20235588; PMCID: PMCPMC2859685
- View Article
- PubMed/NCBI
- Google Scholar
36. McGann M. FRED pose prediction and virtual screening accuracy. J Chem Inf Model. 2011;51(3):578–96. Epub 2011/02/18. pmid:21323318
- View Article
- PubMed/NCBI
- Google Scholar
37. McGann M. FRED and HYBRID docking performance on standardized datasets. J Comput Aided Mol Des. 2012;26(8):897–906. Epub 2012/06/07. pmid:22669221
- View Article
- PubMed/NCBI
- Google Scholar
38. Dassault Systèmes BIOVIA. Discovery studio modeling environment. San Diego: Dassault Systèmes; 2019.
39. Berthold MR, Cebron N, Dill F, Gabriel TR, Kötter T, Meinl T, et al. KNIME: The Konstanz Information Miner. Berlin, Heidelberg: Springer Berlin Heidelberg; 2008.
40. Clark RD, Webster-Clark DJ. Managing bias in ROC curves. J Comput Aided Mol Des. 2008;22(3-4):141-6. Epub 2008/02/08. pmid:18256892
- View Article
- PubMed/NCBI
- Google Scholar
41. Ibrahim TM, Bauer MR, Dörr A, Veyisoglu E, Boeckler FM. pROC-Chemotype plots enhance the interpretability of benchmarking results in structure-based virtual screening. J Chem Inf Model. 2015;55(11):2297–307. Epub 2015/10/06. pmid:26434782
- View Article
- PubMed/NCBI
- Google Scholar
42. Wei BQ, Baase WA, Weaver LH, Matthews BW, Shoichet BK. A model binding site for testing scoring functions in molecular docking. J Mol Biol. 2002;322(2):339–55. Epub 2002/09/10. pmid:12217695
- View Article
- PubMed/NCBI
- Google Scholar
43. Mark James A, Teemu M, Roland S, Szilárd P, Jeremy CS, Berk H, et al. GROMACS: high performance molecular simulations through multi-level parallelism from laptops to supercomputers. SoftwareX. 2015;1–2:19–25. pmid:201519
- View Article
- PubMed/NCBI
- Google Scholar
44. Mark P, Nilsson L. Structure and dynamics of liquid water with different long-range interaction truncation and temperature control methods in molecular dynamics simulations. J Comput Chem. 2002;23(13):1211–9. Epub 2002/09/05. 12210146
- View Article
- PubMed/NCBI
- Google Scholar
45. Valdés-Tresanco MS, Valdés-Tresanco ME, Valiente PA, Moreno E. gmx_MMPBSA: a new tool to perform end-state free energy calculations with GROMACS. J Chem Theory Comput. 2021;17(10):6281–91. Epub 2021/09/30. pmid:34586825
- View Article
- PubMed/NCBI
- Google Scholar
46. Miller BR 3rd, McGee TD Jr., Swails JM, Homeyer N, Gohlke H, Roitberg AE. MMPBSA.py: an efficient program for end-state free energy calculations. J Chem Theory Comput. 2012;8(9):3314–21. Epub 2012/09/11. pmid:26605738
- View Article
- PubMed/NCBI
- Google Scholar
47. Lee J, Worrall LJ, Vuckovic M, Rosell FI, Gentile F, Ton AT, et al. Crystallographic structure of wild-type SARS-CoV-2 main protease acyl-enzyme intermediate with physiological C-terminal autoprocessing site. Nat Commun. 2020;11(1):5877. Epub 2020/11/20. pmid:33208735; PMCID: PMCPMC7674412
- View Article
- PubMed/NCBI
- Google Scholar
48. Sacco MD, Hu Y, Gongora MV, Meilleur F, Kemp MT, Zhang X, et al. The P132H mutation in the main protease of Omicron SARS-CoV-2 decreases thermal stability without compromising catalysis or small-molecule drug inhibition. Cell Res. 2022;32(5):498–500. Epub 2022/03/17. pmid:35292745; PMCID: PMCPMC8923085
- View Article
- PubMed/NCBI
- Google Scholar
49. Bhat ZA, Khan MM, Rehman A, Iqbal J, Sanjeev BS, Madhumalar A. MD simulations indicate Omicron P132H of SARS-CoV-2 M(pro) is a potential allosteric mutant involved in modulating the dynamics of catalytic site entry loop. Int J Biol Macromol. 2024;262(Pt 2):130077. Epub 2024/02/13. pmid:38346625
- View Article
- PubMed/NCBI
- Google Scholar
50. Llanos MA, Gantner ME, Rodriguez S, Alberca LN, Bellera CL, Talevi A, et al. strengths and weaknesses of docking simulations in the SARS-CoV-2 era: the main protease (Mpro) case study. J Chem Inf Model. 2021;61(8):3758–70. Epub 2021/07/28. pmid:34313128
- View Article
- PubMed/NCBI
- Google Scholar
51. Hung YP, Lee JC, Chiu CW, Lee CC, Tsai PJ, Hsu IL, et al. Oral nirmatrelvir/ritonavir therapy for COVID-19: the dawn in the dark? Antibiotics (Basel). 2022;11(2):220. Epub 2022/02/26. pmid:35203821; PMCID: PMCPMC8868411
- View Article
- PubMed/NCBI
- Google Scholar
52. Deshmukh MG, Ippolito JA, Zhang CH, Stone EA, Reilly RA, Miller SJ, et al. Structure-guided design of a perampanel-derived pharmacophore targeting the SARS-CoV-2 main protease. Structure. 2021;29(8):823–33.e5. Epub 2021/06/24. pmid:34161756
- View Article
- PubMed/NCBI
- Google Scholar
53. Shaqra AM, Zvornicanin SN, Huang QYJ, Lockbaum GJ, Knapp M, Tandeske L, et al. Defining the substrate envelope of SARS-CoV-2 main protease to predict and avoid drug resistance. Nat Commun. 2022;13(1):3556. Epub 2022/06/22. pmid:35729165; PMCID: PMCPMC9211792
- View Article
- PubMed/NCBI
- Google Scholar
54. Genheden S, Ryde U. The MM/PBSA and MM/GBSA methods to estimate ligand-binding affinities. Expert Opin Drug Discov. 2015;10(5):449–61. Epub 2015/04/04. pmid:25835573; PMCID: PMCPMC4487606
- View Article
- PubMed/NCBI
- Google Scholar

[ref1] 1. Andersen KG, Rambaut A, Lipkin WI, Holmes EC, Garry RF. The proximal origin of SARS-CoV-2. Nat Med. 2020;26(4):450–2. Epub 2020/04/15. pmid:32284615
View Article
PubMed/NCBI
Google Scholar

[2] View Article

[3] PubMed/NCBI

[4] Google Scholar

[ref2] 2. Flannagan J, Twohig KA, Carter E, Chudasama DY, Lamagni T, Dabrera G. Increased residential clustering of COVID-19 cases associated with SARS-CoV-2 variant of concern B.1.1.7. Epidemiology. 2022;33(4):465–9. Epub 2022/05/06. pmid:35512232; PMCID: PMCPMC9148639
View Article
PubMed/NCBI
Google Scholar

[6] View Article

[7] PubMed/NCBI

[8] Google Scholar

[ref3] 3. Davies NG, Abbott S, Barnard RC, Jarvis CI, Kucharski AJ, Munday JD, et al. Estimated transmissibility and impact of SARS-CoV-2 lineage B.1.1.7 in England. Science. 2021;372(6538):eabg3055. Epub 2021/03/05. pmid:33658326; PMCID: PMCPMC8128288
View Article
PubMed/NCBI
Google Scholar

[10] View Article

[11] PubMed/NCBI

[12] Google Scholar

[ref4] 4. Wang P, Nair MS, Liu L, Iketani S, Luo Y, Guo Y, et al. Antibody resistance of SARS-CoV-2 variants B.1.351 and B.1.1.7. Nature. 2021;593(7857):130–5. Epub 2021/03/09. pmid:33684923
View Article
PubMed/NCBI
Google Scholar

[14] View Article

[15] PubMed/NCBI

[16] Google Scholar

[ref5] 5. Madhi SA, Baillie V, Cutland CL, Voysey M, Koen AL, Fairlie L, et al.; NGS-SA Group. Efficacy of the ChAdOx1 nCoV-19 Covid-19 vaccine against the B.1.351 variant. N Engl J Med. 2021;384(20):1885–98. Epub 2021/03/17. pmid:33725432; PMCID: PMCPMC7993410
View Article
PubMed/NCBI
Google Scholar

[18] View Article

[19] PubMed/NCBI

[20] Google Scholar

[ref6] 6. Wang P, Casner RG, Nair MS, Wang M, Yu J, Cerutti G, et al. Increased resistance of SARS-CoV-2 variant P.1 to antibody neutralization. Cell Host Microbe. 2021;29(5):747–51.e4. Epub 2021/04/23. pmid:33887205
View Article
PubMed/NCBI
Google Scholar

[22] View Article

[23] PubMed/NCBI

[24] Google Scholar

[ref7] 7. Ng OT, Koh V, Chiew CJ, Marimuthu K, Thevasagayam NM, Mak TM, et al. Impact of delta variant and vaccination on SARS-CoV-2 secondary attack rate among household close contacts. Lancet Reg Health West Pac. 2021;17:100299. Epub 2021/11/09. pmid:34746899; PMCID: PMCPMC8560026
View Article
PubMed/NCBI
Google Scholar

[26] View Article

[27] PubMed/NCBI

[28] Google Scholar

[ref8] 8. Allen H, Vusirikala A, Flannagan J, Twohig KA, Zaidi A, Chudasama D, et al. COVID-19 Genomics UK (COG-UK Consortium). Household transmission of COVID-19 cases associated with SARS-CoV-2 delta variant (B.1.617.2): national case-control study. Lancet Reg Health Eur. 2022;12:100252. Epub 2021/11/04. pmid:34729548; PMCID: PMCPMC8552812
View Article
PubMed/NCBI
Google Scholar

[30] View Article

[31] PubMed/NCBI

[32] Google Scholar

[ref9] 9. Hastert FD, Hein S, von Rhein C, Benz NI, Husria Y, Oberle D, et al. The SARS-CoV-2 variant omicron is able to escape vaccine-induced humoral immune responses, but is counteracted by booster vaccination. Vaccines (Basel). 2022;10(5). Epub 2022/05/29. pmid:35632550; PMCID: PMCPMC9144538
View Article
PubMed/NCBI
Google Scholar

[34] View Article

[35] PubMed/NCBI

[36] Google Scholar

[ref10] 10. Dejnirattisai W, Shaw RH, Supasa P, Liu C, Stuart AS, Pollard AJ, et al.; Com-COV2 study group. Reduced neutralisation of SARS-CoV-2 omicron B.1.1.529 variant by post-immunisation serum. Lancet. 2022;399(10321):234–6. Epub 2021/12/24. pmid:34942101; PMCID: PMCPMC8687667
View Article
PubMed/NCBI
Google Scholar

[38] View Article

[39] PubMed/NCBI

[40] Google Scholar

[ref11] 11. Jin Z, Du X, Xu Y, Deng Y, Liu M, Zhao Y, et al. Structure of M(pro) from SARS-CoV-2 and discovery of its inhibitors. Nature. 2020;582(7811):289–93. Epub 2020/04/10. pmid:32272481
View Article
PubMed/NCBI
Google Scholar

[42] View Article

[43] PubMed/NCBI

[44] Google Scholar

[ref12] 12. Zhang L, Lin D, Sun X, Curth U, Drosten C, Sauerhering L, et al. Crystal structure of SARS-CoV-2 main protease provides a basis for design of improved α-ketoamide inhibitors. Science. 2020;368(6489):409–12. Epub 2020/03/22. pmid:32198291; PMCID: PMCPMC7164518
View Article
PubMed/NCBI
Google Scholar

[46] View Article

[47] PubMed/NCBI

[48] Google Scholar

[ref13] 13. Iketani S, Mohri H, Culbertson B, Hong SJ, Duan Y, Luck MI, et al. Multiple pathways for SARS-CoV-2 resistance to nirmatrelvir. Nature. 2023;613(7944):558–64. Epub 2022/11/10. pmid:36351451; PMCID: PMCPMC9849135
View Article
PubMed/NCBI
Google Scholar

[50] View Article

[51] PubMed/NCBI

[52] Google Scholar

[ref14] 14. Vangeel L, Chiu W, De Jonghe S, Maes P, Slechten B, Raymenants J, et al. Remdesivir, Molnupiravir and Nirmatrelvir remain active against SARS-CoV-2 Omicron and other variants of concern. Antiviral Res. 2022;198:105252. Epub 2022/01/28. pmid:35085683
View Article
PubMed/NCBI
Google Scholar

[54] View Article

[55] PubMed/NCBI

[56] Google Scholar

[ref15] 15. Mengist HM, Mekonnen D, Mohammed A, Shi R, Jin T. Potency, safety, and pharmacokinetic profiles of potential inhibitors targeting SARS-CoV-2 main protease. Front Pharmacol. 2020;11:630500. Epub 2021/02/19. pmid:33597888; PMCID: PMCPMC7883113
View Article
PubMed/NCBI
Google Scholar

[58] View Article

[59] PubMed/NCBI

[60] Google Scholar

[ref16] 16. Unoh Y, Uehara S, Nakahara K, Nobori H, Yamatsu Y, Yamamoto S, et al. Discovery of S-217622, a noncovalent Oral SARS-CoV-2 3CL protease inhibitor clinical candidate for treating COVID-19. J Med Chem. 2022;65(9):6499–512. Epub 2022/03/31. pmid:35352927
View Article
PubMed/NCBI
Google Scholar

[62] View Article

[63] PubMed/NCBI

[64] Google Scholar

[ref17] 17. Zeyaullah M, Khan N, Muzammil K, AlShahrani AM, Khan MS, Alam MS, et al. In-silico approaches for identification of compounds inhibiting SARS-CoV-2 3CL protease. PLOS ONE. 2023;18(4):e0284301. pmid:37058496
View Article
PubMed/NCBI
Google Scholar

[66] View Article

[67] PubMed/NCBI

[68] Google Scholar

[ref18] 18. Bauer MR, Ibrahim TM, Vogel SM, Boeckler FM. Evaluation and optimization of virtual screening workflows with DEKOIS 2.0--a public library of challenging docking benchmark sets. J Chem Inf Model. 2013;53(6):1447–62. Epub 2013/05/28. pmid:23705874
View Article
PubMed/NCBI
Google Scholar

[70] View Article

[71] PubMed/NCBI

[72] Google Scholar

[ref19] 19. Ibrahim TM, Bauer MR, Boeckler FM. Applying DEKOIS 2.0 in structure-based virtual screening to probe the impact of preparation procedures and score normalization. J Cheminform. 2015;7:21. Epub 2015/06/03. pmid:26034510; PMCID: PMCPMC4450982
View Article
PubMed/NCBI
Google Scholar

[74] View Article

[75] PubMed/NCBI

[76] Google Scholar

[ref20] 20. Vogel SM, Bauer MR, Boeckler FM. DEKOIS: demanding evaluation kits for objective in silico screening--a versatile tool for benchmarking docking programs and scoring functions. J Chem Inf Model. 2011;51(10):2650–65. Epub 2011/07/22. pmid:21774552
View Article
PubMed/NCBI
Google Scholar

[78] View Article

[79] PubMed/NCBI

[80] Google Scholar

[ref21] 21. Boeckler FM, Bauer MR, Ibrahim TM, Vogel SM. Use of DEKOIS 2.0 to gain insights for virtual screening. J Cheminform. 2014;6(Suppl 1):O24. Epub 2014/04/26. pmid:24765123; PMCID: PMCPMC3980182
View Article
PubMed/NCBI
Google Scholar

[82] View Article

[83] PubMed/NCBI

[84] Google Scholar

[ref22] 22. Hassan HHA, Ismail MI, Abourehab MAS, Boeckler FM, Ibrahim TM, Arafa RK. In silico targeting of fascin protein for cancer therapy: benchmarking, virtual screening and molecular dynamics approaches. Molecules. 2023;28(3):1296. Epub 2023/02/12. pmid:36770963; PMCID: PMCPMC9921211
View Article
PubMed/NCBI
Google Scholar

[86] View Article

[87] PubMed/NCBI

[88] Google Scholar

[ref23] 23. Sander T, Freyss J, von Korff M, Rufener C. DataWarrior: an open-source program for chemistry aware data visualization and analysis. J Chem Inf Model. 2015;55(2):460–73. Epub 2015/01/07. pmid:25558886
View Article
PubMed/NCBI
Google Scholar

[90] View Article

[91] PubMed/NCBI

[92] Google Scholar

[ref24] 24. Yang J, Lin X, Xing N, Zhang Z, Zhang H, Wu H, et al. Structure-based discovery of novel nonpeptide inhibitors targeting SARS-CoV-2 M(pro). J Chem Inf Model. 2021;61(8):3917–26. Epub 2021/07/20. pmid:34279924
View Article
PubMed/NCBI
Google Scholar

[94] View Article

[95] PubMed/NCBI

[96] Google Scholar

[ref25] 25. Su HX, Yao S, Zhao WF, Li MJ, Liu J, Shang WJ, et al. Anti-SARS-CoV-2 activities in vitro of Shuanghuanglian preparations and bioactive ingredients. Acta Pharmacol Sin. 2020;41(9):1167–77. Epub 2020/08/02. pmid:32737471; PMCID: PMCPMC7393338
View Article
PubMed/NCBI
Google Scholar

[98] View Article

[99] PubMed/NCBI

[100] Google Scholar

[ref26] 26. Zhang CH, Stone EA, Deshmukh M, Ippolito JA, Ghahremanpour MM, Tirado-Rives J, et al. Potent noncovalent inhibitors of the main protease of sars-cov-2 from molecular sculpting of the drug perampanel guided by free energy perturbation calculations. ACS Cent Sci. 2021;7(3):467–75. Epub 2021/04/01. pmid:33786375; PMCID: PMCPMC7931627
View Article
PubMed/NCBI
Google Scholar

[102] View Article

[103] PubMed/NCBI

[104] Google Scholar

[ref27] 27. Molecular operating environment (MOE). Montreal, QC, Canada: Chemical Computing Group Inc.; 2019.

[ref28] 28. Elghoneimy LK, Ismail MI, Boeckler FM, Azzazy HME, Ibrahim TM. Facilitating SARS CoV-2 RNA-dependent RNA polymerase (RdRp) drug discovery by the aid of HCV NS5B palm subdomain binders: in silico approaches and benchmarking. Comput Biol Med. 2021;134:104468. Epub 2021/05/21. pmid:34015671; PMCID: PMCPMC8111889
View Article
PubMed/NCBI
Google Scholar

[107] View Article

[108] PubMed/NCBI

[109] Google Scholar

[ref29] 29. O’Boyle NM, Banck M, James CA, Morley C, Vandermeersch T, Hutchison GR. Open babel: an open chemical toolbox. J Cheminform. 2011;3:33. Epub 2011/10/11. pmid:21982300; PMCID: PMCPMC3198950
View Article
PubMed/NCBI
Google Scholar

[111] View Article

[112] PubMed/NCBI

[113] Google Scholar

[ref30] 30. ten Brink T, Exner TE. Influence of protonation, tautomeric, and stereoisomeric states on protein-ligand docking results. J Chem Inf Model. 2009;49(6):1535–46. Epub 2009/05/21. pmid:19453150
View Article
PubMed/NCBI
Google Scholar

[115] View Article

[116] PubMed/NCBI

[117] Google Scholar

[ref31] 31. ten Brink T, Exner TE. pK(a) based protonation states and microspecies for protein-ligand docking. J Comput Aided Mol Des. 2010;24(11):935–42. Epub 2010/10/01. pmid:20882397
View Article
PubMed/NCBI
Google Scholar

[119] View Article

[120] PubMed/NCBI

[121] Google Scholar

[ref32] 32. Trott O, Olson AJ. AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading. J Comput Chem. 2010;31(2):455–61. Epub 2009/06/06. pmid:19499576; PMCID: PMCPMC3041641
View Article
PubMed/NCBI
Google Scholar

[123] View Article

[124] PubMed/NCBI

[125] Google Scholar

[ref33] 33. Sanner MF. Python: a programming language for software integration and development. J Mol Graph Model. 1999;17(1):57-61. Epub 2000/02/08. pmid:10660911
View Article
PubMed/NCBI
Google Scholar

[127] View Article

[128] PubMed/NCBI

[129] Google Scholar

[ref34] 34. Korb O, Stützle T, Exner TE. Empirical scoring functions for advanced protein-ligand docking with plants. J Chem Inf Model. 2009;49(1):84–96. Epub 2009/01/08. pmid:19125657
View Article
PubMed/NCBI
Google Scholar

[131] View Article

[132] PubMed/NCBI

[133] Google Scholar

[ref35] 35. Hawkins PC, Skillman AG, Warren GL, Ellingson BA, Stahl MT. Conformer generation with OMEGA: algorithm and validation using high quality structures from the Protein Databank and Cambridge Structural Database. J Chem Inf Model. 2010;50(4):572–84. Epub 2010/03/20. pmid:20235588; PMCID: PMCPMC2859685
View Article
PubMed/NCBI
Google Scholar

[135] View Article

[136] PubMed/NCBI

[137] Google Scholar

[ref36] 36. McGann M. FRED pose prediction and virtual screening accuracy. J Chem Inf Model. 2011;51(3):578–96. Epub 2011/02/18. pmid:21323318
View Article
PubMed/NCBI
Google Scholar

[139] View Article

[140] PubMed/NCBI

[141] Google Scholar

[ref37] 37. McGann M. FRED and HYBRID docking performance on standardized datasets. J Comput Aided Mol Des. 2012;26(8):897–906. Epub 2012/06/07. pmid:22669221
View Article
PubMed/NCBI
Google Scholar

[143] View Article

[144] PubMed/NCBI

[145] Google Scholar

[ref38] 38. Dassault Systèmes BIOVIA. Discovery studio modeling environment. San Diego: Dassault Systèmes; 2019.

[ref39] 39. Berthold MR, Cebron N, Dill F, Gabriel TR, Kötter T, Meinl T, et al. KNIME: The Konstanz Information Miner. Berlin, Heidelberg: Springer Berlin Heidelberg; 2008.

[ref40] 40. Clark RD, Webster-Clark DJ. Managing bias in ROC curves. J Comput Aided Mol Des. 2008;22(3-4):141-6. Epub 2008/02/08. pmid:18256892
View Article
PubMed/NCBI
Google Scholar

[149] View Article

[150] PubMed/NCBI

[151] Google Scholar

[ref41] 41. Ibrahim TM, Bauer MR, Dörr A, Veyisoglu E, Boeckler FM. pROC-Chemotype plots enhance the interpretability of benchmarking results in structure-based virtual screening. J Chem Inf Model. 2015;55(11):2297–307. Epub 2015/10/06. pmid:26434782
View Article
PubMed/NCBI
Google Scholar

[153] View Article

[154] PubMed/NCBI

[155] Google Scholar

[ref42] 42. Wei BQ, Baase WA, Weaver LH, Matthews BW, Shoichet BK. A model binding site for testing scoring functions in molecular docking. J Mol Biol. 2002;322(2):339–55. Epub 2002/09/10. pmid:12217695
View Article
PubMed/NCBI
Google Scholar

[157] View Article

[158] PubMed/NCBI

[159] Google Scholar

[ref43] 43. Mark James A, Teemu M, Roland S, Szilárd P, Jeremy CS, Berk H, et al. GROMACS: high performance molecular simulations through multi-level parallelism from laptops to supercomputers. SoftwareX. 2015;1–2:19–25. pmid:201519
View Article
PubMed/NCBI
Google Scholar

[161] View Article

[162] PubMed/NCBI

[163] Google Scholar

[ref44] 44. Mark P, Nilsson L. Structure and dynamics of liquid water with different long-range interaction truncation and temperature control methods in molecular dynamics simulations. J Comput Chem. 2002;23(13):1211–9. Epub 2002/09/05. 12210146
View Article
PubMed/NCBI
Google Scholar

[165] View Article

[166] PubMed/NCBI

[167] Google Scholar

[ref45] 45. Valdés-Tresanco MS, Valdés-Tresanco ME, Valiente PA, Moreno E. gmx_MMPBSA: a new tool to perform end-state free energy calculations with GROMACS. J Chem Theory Comput. 2021;17(10):6281–91. Epub 2021/09/30. pmid:34586825
View Article
PubMed/NCBI
Google Scholar

[169] View Article

[170] PubMed/NCBI

[171] Google Scholar

[ref46] 46. Miller BR 3rd, McGee TD Jr., Swails JM, Homeyer N, Gohlke H, Roitberg AE. MMPBSA.py: an efficient program for end-state free energy calculations. J Chem Theory Comput. 2012;8(9):3314–21. Epub 2012/09/11. pmid:26605738
View Article
PubMed/NCBI
Google Scholar

[173] View Article

[174] PubMed/NCBI

[175] Google Scholar

[ref47] 47. Lee J, Worrall LJ, Vuckovic M, Rosell FI, Gentile F, Ton AT, et al. Crystallographic structure of wild-type SARS-CoV-2 main protease acyl-enzyme intermediate with physiological C-terminal autoprocessing site. Nat Commun. 2020;11(1):5877. Epub 2020/11/20. pmid:33208735; PMCID: PMCPMC7674412
View Article
PubMed/NCBI
Google Scholar

[177] View Article

[178] PubMed/NCBI

[179] Google Scholar

[ref48] 48. Sacco MD, Hu Y, Gongora MV, Meilleur F, Kemp MT, Zhang X, et al. The P132H mutation in the main protease of Omicron SARS-CoV-2 decreases thermal stability without compromising catalysis or small-molecule drug inhibition. Cell Res. 2022;32(5):498–500. Epub 2022/03/17. pmid:35292745; PMCID: PMCPMC8923085
View Article
PubMed/NCBI
Google Scholar

[181] View Article

[182] PubMed/NCBI

[183] Google Scholar

[ref49] 49. Bhat ZA, Khan MM, Rehman A, Iqbal J, Sanjeev BS, Madhumalar A. MD simulations indicate Omicron P132H of SARS-CoV-2 M(pro) is a potential allosteric mutant involved in modulating the dynamics of catalytic site entry loop. Int J Biol Macromol. 2024;262(Pt 2):130077. Epub 2024/02/13. pmid:38346625
View Article
PubMed/NCBI
Google Scholar

[185] View Article

[186] PubMed/NCBI

[187] Google Scholar

[ref50] 50. Llanos MA, Gantner ME, Rodriguez S, Alberca LN, Bellera CL, Talevi A, et al. strengths and weaknesses of docking simulations in the SARS-CoV-2 era: the main protease (Mpro) case study. J Chem Inf Model. 2021;61(8):3758–70. Epub 2021/07/28. pmid:34313128
View Article
PubMed/NCBI
Google Scholar

[189] View Article

[190] PubMed/NCBI

[191] Google Scholar

[ref51] 51. Hung YP, Lee JC, Chiu CW, Lee CC, Tsai PJ, Hsu IL, et al. Oral nirmatrelvir/ritonavir therapy for COVID-19: the dawn in the dark? Antibiotics (Basel). 2022;11(2):220. Epub 2022/02/26. pmid:35203821; PMCID: PMCPMC8868411
View Article
PubMed/NCBI
Google Scholar

[193] View Article

[194] PubMed/NCBI

[195] Google Scholar

[ref52] 52. Deshmukh MG, Ippolito JA, Zhang CH, Stone EA, Reilly RA, Miller SJ, et al. Structure-guided design of a perampanel-derived pharmacophore targeting the SARS-CoV-2 main protease. Structure. 2021;29(8):823–33.e5. Epub 2021/06/24. pmid:34161756
View Article
PubMed/NCBI
Google Scholar

[197] View Article

[198] PubMed/NCBI

[199] Google Scholar

[ref53] 53. Shaqra AM, Zvornicanin SN, Huang QYJ, Lockbaum GJ, Knapp M, Tandeske L, et al. Defining the substrate envelope of SARS-CoV-2 main protease to predict and avoid drug resistance. Nat Commun. 2022;13(1):3556. Epub 2022/06/22. pmid:35729165; PMCID: PMCPMC9211792
View Article
PubMed/NCBI
Google Scholar

[201] View Article

[202] PubMed/NCBI

[203] Google Scholar

[ref54] 54. Genheden S, Ryde U. The MM/PBSA and MM/GBSA methods to estimate ligand-binding affinities. Expert Opin Drug Discov. 2015;10(5):449–61. Epub 2015/04/04. pmid:25835573; PMCID: PMCPMC4487606
View Article
PubMed/NCBI
Google Scholar

[205] View Article

[206] PubMed/NCBI

[207] Google Scholar

Figures

Abstract

Introduction

Rational of the study

Methods

Active set selection and generation of DEKOIS 2.0 benchmark set

Preparation of small molecules

Preparation of the protein structures

Screening and docking experiments

Benchmarking utilizing pROC calculations and pROC-Chemotype plotting

Drug likeness and ADME/TOX

Molecular dynamics simulations

MM GBSA calculations

Results and discussion

Structural insights on Mpro

Selection of the active set molecules

Benchmarking analysis

Chemotype enrichment

Virtual screening of a focused library based on benchmarking results

Description of the postulated binding poses

ADME/Tox and drug-likeness filtration applying expert systems

Molecular dynamics simulations

Root mean square deviation (RMSD)

Radius of gyration

RMSF analysis

Binding free energy (ΔG) via MM GBSA calculation

Minimum distance analysis

Conclusion

Supporting information

S1 File. SI_file01 contains detailed description of the methodology of: Preparation of small molecules, preparation of the protein structures, and molecular dynamics simulations.

Acknowledgments

References