key: cord-0859435-bv3u06t2
authors: Yuce, Merve; Cicek, Erdem; Inan, Tugce; Dag, Aslihan Basak; Kurkcuoglu, Ozge; Sungur, Fethiye Aylin
title: Repurposing of FDA‐approved drugs against active site and potential allosteric drug‐binding sites of COVID‐19 main protease
date: 2021-07-05
journal: Proteins
DOI: 10.1002/prot.26164
sha: 6cb0a17fe5d3923a5a4a5c66ee8d81acc9c9c3c3
doc_id: 859435
cord_uid: bv3u06t2

The novel coronavirus disease 2019 (COVID‐19) caused by severe acute respiratory syndrome coronavirus 2 (SARS‐CoV‐2) still has serious negative effects on health, social life, and economics. Recently, vaccines from various companies have been urgently approved to control SARS‐CoV‐2 infections. However, any specific antiviral drug has not been confirmed so far for regular treatment. An important target is the main protease (M(pro)), which plays a major role in replication of the virus. In this study, Gaussian and residue network models are employed to reveal two distinct potential allosteric sites on M(pro) that can be evaluated as drug targets besides the active site. Then, Food and Drug Administration (FDA)‐approved drugs are docked to three distinct sites with flexible docking using AutoDock Vina to identify potential drug candidates. Fourteen best molecule hits for the active site of M(pro) are determined. Six of these also exhibit high docking scores for the potential allosteric regions. Full‐atom molecular dynamics simulations with MM‐GBSA method indicate that compounds docked to active and potential allosteric sites form stable interactions with high binding free energy (∆G (bind)) values. ∆G (bind) values reach −52.06 kcal/mol for the active site, −51.08 kcal/mol for the potential allosteric site 1, and − 42.93 kcal/mol for the potential allosteric site 2. Energy decomposition calculations per residue elucidate key binding residues stabilizing the ligands that can further serve to design pharmacophores. This systematic and efficient computational analysis successfully determines ivermectine, diosmin, and selinexor currently subjected to clinical trials, and further proposes bromocriptine, elbasvir as M(pro) inhibitor candidates to be evaluated against SARS‐CoV‐2 infections.

2019 4 and spread widely around the world, and so declared as a pandemic on March 2020 by the World Health Organization (WHO). 5 The viral life cycle involves entry, replication of genetic material, protein translation, assembly, and release from the host cell. Strategies for drug development target viral proteins and host receptors to interfere with different stages of the CoV life cycle. 6 Spike (S) protein plays a vital role in viral entry into the host cell, hence it is an attractive target for blocking SARS-CoV-2 infection. 7 Several theoretical studies have been performed to identify inhibitors against SARS-CoV-2 spike protein. [8] [9] [10] RNA-dependent RNA polymerase (RdRp) is another important target for SARS-CoV-2 due to its crucial role in replicating the positive sense viral RNA. 6 Promising drug candidates against SARS-CoV-2 RdRp have been reported in several in silico drug repurposing studies. [11] [12] [13] [14] One of the best-characterized drug targets among CoV is the M pro (also called 3CL pro ) that is functional as a homodimer, where each monomer contains the catalytic dyad defined by H41 and C145 residues. 15 This enzyme is essential for processing the polyproteins that are translated from the viral RNA 16 and mediating the maturation of the non-structural proteins, which is the main step in the replication of the virus. As a vital enzyme of SARS-CoV-2, inhibiting the activity of M pro would block viral replication and transcription. 17 In addition to kinetic studies indicating that the active form of the M pro corresponds to a homodimer, 18 significant conformational differences between the monomer and dimeric states have been reported in recent studies. 19, 20 To analyze structural and dynamic properties of M pro , all-atom MD simulations of various M pro mutants have been performed by Amamuddy et al 21 emphasizing that mutations located near the active site control the bending motions needed for catalysis, so they may influence enzymatic activity. All these studies suggest that drug discovery combining docking and MD simulations should be performed using the homodimeric conformation instead of the monomer. Previous molecular docking and MD simulation studies targeting SARS-CoV-2 M pro also reported plausible molecules to test experimentally. [22] [23] [24] [25] [26] To tackle this pandemic, overall efforts have been made to develop effective and safe therapies (including vaccines) for COVID- 19 . In the earlier studies, crystal structure of M pro in complex with peptide-like irreversible inhibitors (Michael acceptor N3 and its carbonyl derivative α-ketoamide) are used to design N3-analogues as potential inhibitors against M pro . 27, 28 Covalent inhibitors provide many pharmacological advantages such as enhanced biochemical and cellular potency, selectivity, and prolonged duration of action. 29, 30 However, their potential for nonspecific reactivity, toxicity, and undesired side-effects, rapid in vivo metabolism and reduced oral bioavailability make the irreversible inhibitors less efficient as therapeutic agents. [31] [32] [33] [34] On the other hand, drug repositioning or repurposing is a faster and less costly solution to propose potential effective drugs useful to control emerged infectious outbreaks immediately, as new drug development takes more than 10 years. 35, 36 Currently, there are no clinically approved therapeutics available. At this point, already known and FDA-approved potential candidate drugs can be screened and re-evaluated based on antiviral effects for alternative treatments. 37 In this regard, enzyme kinetics and cellular antiviral assays reveal that boceprevir, 38 GC-376, 39 calpain inhibitors II, XII, 40 and bepridil 41 inhibit SARS-CoV-2 M pro in the micromolar and nanomolar range. Several drugs employed for various diseases are also being tested in numerous clinical trials, including remdesivir, 42 favipiravir, 43 chloroquine, 44 dexamethasone, 45 nafamostat, 46 and ivermectin. 47 Furthermore, screening of approved and clinical drugs 48 and covalent and non-covalent fragments 49 have experimentally confirmed two allosteric binding sites on M pro , suggesting a novel multi-cyclindependent kinase inhibitor with moderate antiviral activity.

In the light of the reports stated above, it is evident that a search for effective drugs, having the potential to inhibit SARS-CoV-2, has become a global pursuit. In this study, we investigate the repurposing of the existing FDA-approved drugs to target COVID-19 virus M pro using a systematic computational approach combining different methods, namely elastic network theory, molecular docking, and MD simulations. We apply residue network and Gaussian network models (GNM) to identify potential allosteric sites on M pro , which may be evaluated to regulate the enzymatic activity of this protease. Proteins are all considered to use allostery to accomplish their function, whether or not they undergo a large conformational change. 50 Allosteric sites provide alternative drug binding regions on the same proteins, which improves the likelihood of effective drugs with greater specificity and regulates the activity of proteins by remotely affecting their active sites. 51, 52 Molecular docking is one of the most common computational approaches to determine potential drugs regarding their binding pose as well as scoring their binding affinity. 53 We then employ molecular docking of FDA-approved drugs against active site of M pro and its potential allosteric sites. We use AutoDock Vina for molecular docking studies of the constructed library comprising over 2400 molecules. We rank compounds based on their binding scores and consider promising hit compounds from docking experiments with good vina scores in 50-ns MD simulations using AMBER16 54 This study reveals unknown potential allosteric sites of M pro , suggests FDA-approved drugs to target M pro from its critical sites, and describes a systematic methodology that can be also used for drug repurposing for other diseases that urgently wait efficient therapeutics.

In the present study, 2447 FDA-approved drugs are screened with a molecular docking approach against the active site and potential allosteric sites of M pro . We employ GNM and residue network model to identify potential allosteric sites of M pro that can serve as drug-binding regions, besides the active sites. Based on the vina docking score rankings and the active site interaction patterns, MD simulations of the most promising hit compounds bound to active site and potential allosteric sites of M pro are performed. In order to understand the behavior of ligand molecules and their interactions with the active site and potential allosteric sites of the enzyme, 50 ns MD simulations are carried. The binding free energy analysis of M pro complexed with the most promising hit compounds is performed using the MM/GBSA method. 55 All findings are evaluated together to propose FDA-approved drugs to target the M pro . The schematic representation of the approach followed for repurposing FDA-approved drugs is provided in Figure 1 .

GNM is an elastic network model (ENM), which describes the protein structure as a network of connected nodes. Nodes are usually placed at Cα positions of the amino acids. Neighboring amino acid pairs within a cut-off distance of 7 Å are then linked by uniform elastic springs. 56 The total potential energy of the constructed network of N nodes is given as,

where γ is the spring constant, R ij and R ij 0 are the instantaneous and equilibrium distances between the i th and j th nodes, respectively. Γ ij is the ij th element of the Kirchhoff matrix Γ (N Â N) containing the connectivity information of the nodes. The singular decomposition of Γ = UΛU T gives the eigenvalues and eigenvectors corresponding to coupled motions of the nodes. Here, U is the orthogonal eigenvector matrix with elements u k indicating the k th eigenvector. Λ is the diagonal matrix of eigenvalues λ k . N À 1 eigenvalues give the vibrational frequencies of N À 1 modes, while one eigenvalue is zero indicating the rigid body motion.

Cross-correlation between i th and j th nodes is defined by,

and their relative fluctuations are calculated as,

Cross-correlations calculated for the low-frequency motions (slow modes) reveal dynamic domains of proteins where groups of amino acids fluctuate in the same direction. Low-frequency motions of the proteins correspond to their functional globular motions. Therefore, they are highly useful to understand the functional mechanisms of the protein. 57 On the other hand, ⟨ΔR ij 2 ⟩ describes the relative mobility of i th and jth nodes. The high-frequency motions in fast modes calculated with ⟨ΔR ij 2 ⟩ serve to find critical residues related to folding core 58, 59 or binding, such as protein, DNA, or drug binding sites. 59, 60 Here, we analyze the six slowest modes that give information about collective functional motions of the main protease with the highest contribution. Also, we consider the 20 fastest modes to detect the residues with high-frequency fluctuations. These residues have a high capacity to alter the energy landscape such as after binding a drug molecule, thus, highlight plausible drug target regions besides the active site of the main protease. The residue network model is similar to GNM; a network of connected nodes is constructed based on the protein structure. Nodes are placed at C α atoms, and two nodes (amino acids) are linked if they have atom-atom neighboring within a cut-off distance of 4.5 Å.

This cut-off distance includes van der Waals and electrostatic interactions. The local interaction strength of a (i, j) node-pair is calculated as,

Here, N ij is the total number of heavy atom pairs of ı th and ȷ th nodes.

The biasing effect of the amino acid size is eliminated by weighting N ij using their total number of atoms N i and N j . 61 The centrality measure of betweenness reveals the frequently visited nodes or "hubs" located on the shortest paths that are calculated for the network. The betweenness (C B ) value is determined as 63 ;

Here, σ ij is the shortest number of routes between nodes i and j, σ ij (l)

is the shortest number of routes between nodes i and j passing through node l. In this line, the nodes with high (C B ) values in the residue network have a high potential to reside on the allosteric communication paths 62,64 that can be evaluated as novel drug targets.

The crystal structure of M pro in complex with N3 at 2.1 Å resolution is retrieved from the Protein Data Bank (PDB ID:6lu7). 16 The dimeric M pro structure is hydrogenated at pH 7.0 using AMBER16 54 

The parameters, coordinate, and topology input files are generated for all compounds, the crystal structure of the protein (PDB ID:6lu7), and protein-ligand complexes using tLeaP module implemented in AMBER16. For the ligand parameterization, the general Amber force field (GAFF), 73 and the antechamber and parmchk2 modules are used with AM1-BCC charges. 74 The parameters for the protein are described using the AMBER ff14SB force field. 75 

The values of the free energy of binding (ΔG bind ) for a ligand can be calculated according to the equation 55 :

where ΔE MM is the molecular mechanics energy of the molecule expressed as the sum of the internal energy of the molecule and the electrostatic and van der Waals interactions. ΔG sol denotes the solvation energy which is composed of the polar (ΔG PB/GB ) and nonpolar contributions (ΔG SA ). The entropic contribution can be neglected because of similar types of ligands bind to the receptor, and enthalpic contribution is sufficient to compare different ligands. The protein-ligand interactions and the binding free energies are obtained using the MMPBSA.py 80 and Sander modules of AMBER16, respectively.

The most promising compounds are selected following these steps: 

Since there is an urgent need to alleviate the COVID-19 pandemic, the repurposing of existing FDA-approved drugs is a highly effective strategy that reduces time cost, investment, and risks compared to traditional drug development strategies. 81, 82 Herein, we aim to iden- Ivermectin and paritaprevir exhibit relatively higher binding affinity. The hydrogen bond interaction with H41 residue ( Figure 3F ) for ivermectin, and several H-bond, π-π stacking, π-sulfur π-alkyl, and alkyl interactions ( Figure S2 ) with domain two residues especially with C145 residue of the catalytic dyad ( Figure 3G-H) paritaprevir. Although elbasvir displays a relatively lower binding affinity towards the active site with À8.8 kcal/mol compared to other FDA-approved drugs, it shows several H-bond, π-sulfur, π-alkyl, and alkyl interactions ( Figure S2 ) with domain 2 residues especially with C145 residue of catalytic dyad ( Figure 3G-H interactions with C145 residue of catalytic dyad (Figure 3 ). Digitoxin and antrafenine interact with the active site by forming several Hbond, π-π stacked, π-cation, π-anion, π-alkyl, alkyl, and halogen interactions ( Figure 3 ). Considering all analyses, ligands mostly interact with T26, H41, H163, H164, E166, R188, Q189, and T190 residues by forming H-bond and halogen interactions, with F305 residue of subunit B by π À π stacked interaction and with M49, C145, M165, and P168 residues by π-alkyl, π-sulfur, and alkyl interactions, in accordance with those stabilizing N3 inhibitor inside the substrate-binding pocket. 16 These residues are also reported in previous docking studies targeting M pro . [22] [23] [24] The number of interactions and interaction types of all compounds mentioned above are given in Figure S2 . We posit that these Figure 4A also shows amino acids with high relative mobility ⟨ΔR ij 2 ⟩ in the 20 fastest modes. They have a high potential to change the conformational energy landscape of the protein upon ligand binding. Therefore, the amino acids both having high relative mobility in the fastest modes and involving in the hinge regions of the slowest modes can be evaluated as potential drug binding sites, as was recently shown for the bacterial ribosome. 83 At the same time, they plausibly highlight allosteric regions that can affect the activity of the protein upon perturbation. Accordingly, M6-P9, V13, V18, N28, G29, C38, P39, A116, C117, G146, S147, Y161, M162, and H164 are the residues with high relative mobility located at the hinge regions.

Indeed, P39, G146, S147, Y161, M162, H164 are located next to the active sites, as shown in Figure 4B . Recent MD simulations also indicated that mutations at residues A7 and A116 increase the proximal interactions, which in turn affect the dynamics and the dimer stability. 21 We propose the region pinned by A7-P9 as a potential allosteric site, since it is located on a cavity at the subunit interface, and is solvent accessible.

Then, we employ the residue network model and centrality measurement of betweenness to determine hub residues with high potential to assist the flow of information in the structure. 62 We use the crystal structure of M pro (PDB ID:6lu7), and calculate betweenness values CB. The top 0.05 quantile (CB > 0.0485) is designated to hub residues as in 62 ( Figure 4C ). All findings are given in Table S4 Table S1 , E14, and G15 are solvent-accessible and are located at the potential allosteric region predicted by GNM ( Figure 4B highlighted due to their H-bond, π-cation, π-anion, π-alkyl, and alkyl interactions with the docked ligands ( Figure S4 ). In addition, we determine two other hits for the potential allosteric sites. Interestingly, elbasvir has a higher binding affinity for the potential allosteric sites 1 (À10.8 kcal/mol) and 2 (À11.1 kcal/mol) compared to both active sites of M pro (À8.8 and À 8.4 kcal/mol). Selinexor has a slightly lower binding affinity for the potential allosteric sites (À8.2 and À 9.5 kcal/ mol) as compared to the active sites (À10.1 kcal/mol). Evidently, the potential allosteric sites on M pro can be indeed novel target sites, which motivates us to further investigate the stability of these 8 hit compound-M pro complexes with 50-ns long MD simulations.

The main purpose of this study is to determine FDA-approved drugs that exhibit high binding affinity to active and/or potential allosteric (Table S3) and remain stable in these neighboring sites. It should be noted that these two residues are on the second shell of the predicted allosteric site.

Other than indicated, ligands are stable in their initial docking sites, that is, active sites, potential allosteric sites 1 and 2. Tables S5 and S6 for subunits A and B, respectively.

Vina docking scores and ΔG bind values of hit compounds at both subunits are displayed in Figure 5A . Vina score ranking is in agreement with MM/GBSA binding energy rankings for most of the ligands.

Dihydroergotamine has the highest vina docking score and also has a high binding free energy in the active site. On the other hand, elbasvir exhibits a lower binding affinity towards the active sites, which is con- Figure 5B ).

H-bond interactions between the potential allosteric site 1 and 2 residues are detailed in Table S3 and 3.3.1 | Per-residue free-energy decomposition Figure 5A ). Similarly, for bromocriptine, the interactions on both active sites involve the same hydrophobic and polar residues that stabilize the ligand in the binding sites ( Figure 6B ). For the diosmin-M pro complex, interacting residues with the ligand differ ( Figure 6C ).

Hydrophobic residues L50, L167, P168, and A191 contribute to stabilizing diosmin at subunit A, whereas polar residues T25, N142, S144, (Figure 7) . Notably, catalytic residues H41 and C145, and M49, M165, Q189 consistently involve in binding of the ligands, as also reported by previous studies on M pro . 16, 17, [84] [85] [86] These residues are also reported for N3 16 complexed with M pro crystal structure, and for the proteolysis reaction catalyzed by M pro investigated using QM/MM molecular dynamics simulations. [84] [85] [86] [87] The same analysis is carried for elbasvir docked to potential allosteric sites 1 and 2. Elbasvir has a high-binding affinity for both potential allosteric sites; À51.08 and À 42.37 kcal/mol for allosteric sites 1 and 2, respectively. Potential allosteric sites are located at the subunit interfaces, therefore, residues of both subunits simultaneously contribute to ligand stabilization. Elbasvir binds to potential allosteric and positively charged K12, K97 residues of both subunits, whereas hydrophobic residues W207, L282, G283, L286 of both subunits stabilize its binding at potential allosteric site 2 ( Figure 8A ). We also investigate the per-residue free-energy decomposition of elbasvir on the active site. Its binding free energy at the active sites is less than that at the potential allosteric sites (approximately À32 kcal/mol), where hydrophobic and polar residues of subunit A, and polar residues (C44, T45, S46, N142, S144, and Q189) of subunit B stabilize the ligand ( Figure 8B) . Moreover, S301, G302, and V303 residues of subunit A are noted to take a role in the stabilization of elbasvir at subunit B.

In general, hydrophobic residues of both subunits contribute to 97 Considering the importance of natural products as antiviral agents recently, 98,99 diosmin could be a promising drug to treat COVID-19. Worth to note that selinexor and diosmin have not been tested against M pro ; yet, they can be evaluated as potent M pro inhibitors. The inhibitory effects of bromocriptine, an ergot-derived dopamine receptor agonist, on zika virus and dengue virus replication have been reported, [100] [101] [102] so it has also a high potential for the treatment of SARS-CoV-2 infected patients. Ivermectin is an anti-parasitic drug the usage of which extended from veterinary medicine to humans. 103 In addition to in silico studies 104 involving the interactions between ivermectin and SARS-CoV-2 3CL pro , an in vitro study is performed by Caly et al. 92 indicating ivermectin's capability to reduce viral RNA.

Ivermectin (NCT04668469) has recently been evaluated for the clinical trials. 47 Elbasvir is another antiviral drug in our list that displays a higher binding affinity towards potential allosteric sites than the active site of M pro , agreeing with previous in silico studies for M pro . 105, 106 Hit compounds identified as potent M pro inhibitors in computational studies should be experimentally tested whether they are promiscuous inhibitors or not. 107, 108 Regarding the immediate need for therapeutics against SARS-CoV-2, bromocriptine and elbasvir should be evaluated for advanced experimental research to cure COVID-19.

In this study, we conduct a systematic approach to suggest potent inhibitory compounds to inhibit the activity of M pro by combining different computational methods. First, we predict two potential alloste- interactions, such as H-bond, π-alkyl, π-sulfur, and alkyl interactions with the binding pockets, which traces a clear template for pharmacophore design (Figure 7) .

Bromocriptine, diosmin, selinexor, ivermectin, elbasvir, nilotinib, entrectinib, rutin, dihydroergotamine, and digitoxin are determined to have a high affinity for M pro . These FDA-approved drugs are suggested as anti-COVID-19 therapeutics to be further evaluated in vitro and in vivo testing for viral activity. Indeed, our calculations successfully determine diosmin, ivermectin, and selinexor that have already been subjected to clinical trials. Therefore, our systematic approach followed in this study serves as a guideline to propose effective compounds that can be rapidly tested in clinical trials for the treatment of various diseases, including COVID-19.

Coronavirusesdrug discovery and therapeutic options

Severe acute respiratory syndrome coronavirus as an agent of emerging and reemerging infection

Middle East respiratory syndrome coronavirus: another zoonotic betacoronavirus causing SARS-like disease

A novel coronavirus from patients with pneumonia in China

WHO declares COVID-19 a pandemic

Tackling SARS-CoV-2: proposed targets and repurposed drugs

A comprehensive review on promising anti-viral therapeutic candidates identified against main protease from SARS-CoV-2 through various computational methods

Molecular docking study of potential phytochemicals and their effects on the complex of SARS-CoV2 spike protein and human ACE2

Dynamics of the ACE2-SARS-CoV-2/SARS-CoV spike protein interface reveal unique mechanisms

Supervised molecular dynamics for exploring the druggability of the SARS-CoV-2 spike protein

In silico identification of widely used and well-tolerated drugs as potential SARS-CoV-2 3C-like protease and viral RNA-dependent RNA polymerase inhibitors for direct use in clinical trials

Potential RNAdependent RNA polymerase inhibitors as prospective therapeutics against SARS-CoV-2

Predicting inhibitors for SARS-CoV-2 RNA-dependent RNA polymerase using machine learning and virtual screening. arXiv (preprint)

Prediction of potential inhibitors for RNA-dependent RNA polymerase of SARS-CoV-2 using comprehensive drug repurposing and molecular docking approach

Activation and maturation of SARS-CoV main protease

Structure of Mpro from SARS-CoV-2 and discovery of its inhibitors

Crystal structure of SARS-CoV-2 main protease provides a basis for design of improved a-ketoamide inhibitors

SARS CoV main proteinase: the monomer-dimer equilibrium dissociation constant

Prediction of potential inhibitors of the dimeric SARS-CoV2 main proteinase through the MM/GBSA approach

Computational studies of SARS-CoV-2 3clpro: insights from md simulations

Impact of early pandemic stage mutations on molecular dynamics of SARS-CoV-2 M Pro

Computational determination of potential inhibitors of SARS-CoV-2 Main protease

Structure-based virtual screening to discover potential Lead molecules for the SARS-CoV-2 Main protease

Discovery of new Hydroxyethylamine analogs against 3CLproProtein target of SARS-CoV-2: molecular docking, molecular dynamics simulation, and structureactivity relationship studies

Interactive molecular dynamics in virtual reality is an effective tool for flexible substrate and inhibitor docking to the SARS-CoV-2 Main protease

Structure and inhibition of the SARS-CoV-2 main protease reveal strategy for developing dual inhibitors against M pro and cathepsin L

Antiviral agents against COVID-19: structurebased design of specific peptidomimetic inhibitors of SARS-CoV-2 main protease

Mechanism of inhibition of SARS-CoV-2 M pro by N3 peptidyl Michael acceptor explained by QM/MM simulations and design of new derivatives with tunable chemical reactivity

The resurgence of covalent drugs

Drug discovery for a new generation of covalent drugs

Designing irreversible inhibitors -worth the effort?

An overview of severe acute respiratory syndrome-coronavirus (SARS-CoV) 3CL protease inhibitors: Peptidomimetics and small molecule chemotherapy

Covalent versus noncovalent enzyme inhibition: which route should we take? A justification of the good and bad from molecular Modelling perspective

Drug targets, mechanisms of drug action, and therapeutics against SARS-CoV-2

Repurposing of clinically developed drugs for treatment of Middle East respiratory syndrome coronavirus infection

Utilizing drug repurposing against COVID-19-efficacy, limitations, and challenges

COVID-19 treatment: close to a cure? A rapid review of pharmacotherapies for the novel coronavirus (SARS-CoV-2)

Both Boceprevir and GC376 efficaciously inhibit SARS-CoV-2 by targeting its main protease

Feline coronavirus drug inhibits the main protease of SARS-CoV-2 and blocks virus replication

GC-376, and calpain inhibitors II, XII inhibit SARS-CoV-2 viral replication by targeting the viral main protease

Bepridil is potent against SARS-CoV-2 in vitro

Remdesivir for the treatment of Covid-19 -final report

Favipiravir versus Arbidol for COVID-19: a randomized clinical trial. medRxiv

National Institute for the infectious diseases "L. Spallanzani", IRCCS. Recommendations for COVID-19 clinical management

Dexamethasone in hospitalized patients with Covid-19

Nafamostat mesylate blocks activation of SARS-CoV-2: new treatment option for COVID-19

Effect of Ivermectin on time to resolution of symptoms among adults with mild COVID-19: a randomized clinical trial

X-ray screening identifies active site and allosteric inhibitors of SARS-CoV-2 main protease

Crystallographic and electrophilic fragment screening of the SARS-CoV-2 main protease

The origin of allosteric functional modulation: multiple pre-existing pathways

Docking to large allosteric binding sites on protein surfaces

Allosteric sites: remote control in regulation of protein activity

Molecular docking: a powerful approach for structure-based drug discovery

Assessing the performance of the MM/PBSA and MM/GBSA methods. 1. The accuracy of binding free energy calculations based on molecular dynamics simulations

Gaussian dynamics of folded proteins

Global dynamics of proteins: bridging between structure and function

Unfolding events of chymotrypsin inhibitor 2 (CI2) revealed by Monte Carlo (MC) simulations and their consistency from structure-based analysis of conformations

Hot spots in a network of functional sites

DNABINDPROT: fluctuation-based predictor of DNA-binding residues within a network of interacting residues

A network representation of protein structures: implications for protein stability

Exploring allosteric communication in multiple states of the bacterial ribosome using residue network analysis

Residue geometry networks: a rigidity-based approach to the amino acid network and evolutionary rate analysis

Identification of potential allosteric communication pathways between functional sites of the bacterial ribosome by graph and elastic network models

PROPKA3: consistent treatment of internal and surface residues in empirical pKa predictions

ZINC -a free database of commercially available compounds for virtual screening

DrugBank: a comprehensive resource for in silico drug discovery and exploration

Gaussian 09, Revision B.01

Optimization of parameters for semiempirical methods I Method

Open babel: an open chemical toolbox

AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading

Development and testing of a general Amber force field

Fast, efficient generation of highquality atomic charges. AM1-BCC model: II. Parameterization and validation

ff14SB: improving the accuracy of protein side chain and backbone parameters from ff99SB

Comparison of simple potential functions for simulating liquid water

Particle mesh Ewald: an NÁlog (N) method for Ewald sums in large systems

Numerical integration of the cartesian equations of motion of a system with constraints: molecular dynamics of n-alkanes

Synchronization of trajectories in canonical molecular-dynamics simulations: observation, explanation, and exploitation

Py: an efficient program for end-state free energy calculations

Improving therapy of severe infections through drug repurposing of synergistic combinations

Potential therapeutic targets for combating SARS-CoV-2: drug repurposing, clinical trials and recent advancements

Local and global motions underlying antibiotic binding in bacterial ribosome

Identification of saquinavir as a potent inhibitor of dimeric SARS-CoV2 main protease through MM/GBSA

Drug repurposing against SARS-CoV-2 using E-pharmacophore based virtual screening, molecular docking and molecular dynamics with main protease as the target drug repurposing against SARS-CoV-2 using E-pharmacophore based virtual screening, molecular doc

Why are Lopinavir and ritonavir effective against the newly emerged coronavirus 2019? Atomistic insights into the inhibitory mechanisms

Revealing the molecular mechanisms of proteolysis of SARS-CoV-2 Mproby QM/MM computational methods

Identification of novel compounds against three targets of SARS CoV2 coronavirus by combined virtual screening and supervised machine learning

Phytotherapeutics and SARS-CoV-2 infection: potential role of bioflavonoids

Possible SARS-coronavirus 2 inhibitor revealed by simulated molecular docking to viral main protease and host toll-like receptor

Pharmacoinformatics and molecular dynamics simulation studies reveal potential covalent and FDA-approved inhibitors of SARS-CoV-2 main protease 3CL pro

The FDAapproved drug ivermectin inhibits the replication of SARS-CoV-2 in vitro

Drug repurposing of approved drugs Elbasvir, Ledipasvir, Paritaprevir, Velpatasvir, Antrafenine and ergotamine for combating COVID19

Cambridge Open Engage (preprint)

Identification of antiviral drug candidates against SARS-CoV-2 from FDA-approved drugs

The tyrosine kinase inhibitor nilotinib inhibits SARS-CoV-2 in vitro

Is hesperidin essential for prophylaxis and treatment of COVID-19 infection? Med Hypotheses

A spectrophotometric method for Diosmin determination

Antiviral and Virucidal activities of natural products

Recent FDA approval of sofosbuvir and simeprevir. Implications for current HCV treatment

Novel antiviral activity and mechanism of bromocriptine as a Zika virus NS2B-NS3 protease inhibitor

Novel antiviral activity of bromocriptine against dengue virus replication

Repurposing of bromocriptine for cancer therapy

The pharmacokinetics and interactions of ivermectin in humans -a mini-review

Molecular docking and molecular dynamic study of two viral proteins associated with SARS-CoV-2 with Ivermectin

Repurposing of FDA approved drugs targeting Main protease M Pro for SARS-CoV-2

Discovery of SARS-CoV-2 antiviral synergy between remdesivir and approved drugs in human lung cells

Disulfiram, Carmofur, PX-12, Tideglusib, and Shikonin are nonspecific promiscuous SARS-CoV-2 Main protease inhibitors

Dipyridamole, chloroquine, montelukast sodium, candesartan, oxytetracycline, and atazanavir are not SARS-CoV-2 main protease inhibitors

Repurposing of FDA-approved drugs against active site and potential allosteric drug-binding sites of COVID-19 main protease

The authors thank the National Center of High-Performance Computing (UHeM) at ITU (1007692020) for computational time. National Center of High-Performance Computing (UHeM) at ITU, Grant/Award Number: 1007692020.

The authors declare that they have no conflict of interest.

The authors confirm that the data supporting the findings of this study are available within the article or its supplementary material.

The peer review history for this article is available at https://publons.