key: cord-0876618-u54qh0zm
authors: Parvez, Md. Sorwer Alam; Karim, Md. Adnan; Hasan, Mahmudul; Jaman, Jomana; Karim, Ziaul; Tahsin, Tohura; Hasan, Md. Nazmul; Hosen, Mohammad Jakir
title: Prediction of potential inhibitors for RNA-dependent RNA polymerase of SARS-CoV-2 using comprehensive drug repurposing and molecular docking approach
date: 2020-09-17
journal: Int J Biol Macromol
DOI: 10.1016/j.ijbiomac.2020.09.098
sha: dfb04e93d4198d4b1133a9c42c29d059a192fcf3
doc_id: 876618
cord_uid: u54qh0zm

The pandemic prevalence of COVID-19 has become a very serious global health issue. Scientists all over the world have been seriously attempting in the discovery of a drug to combat SARS-CoV-2. It has been found that RNA-dependent RNA polymerase (RdRp) plays a crucial role in SARS-CoV-2 replication, and thus could be a potential drug target. Here, comprehensive computational approaches including drug repurposing and molecular docking were employed to predict an effective drug candidate targeting RdRp of SARS-CoV-2. This study revealed that Rifabutin, Rifapentine, Fidaxomicin, 7-methyl-guanosine-5′-triphosphate-5′-guanosine and Ivermectin have a potential inhibitory interaction with RdRp of SARS-CoV-2, and could be effective drugs for COVID-19. In addition, virtual screening of the compounds from ZINC database also allowed the prediction of two compounds (ZINC09128258 and ZINC09883305) with pharmacophore features that interact effectively with RdRp of SARS-CoV-2, indicating their potentiality as effective inhibitors of the enzyme. Furthermore, ADME analysis along with analysis of toxicity was also undertaken to check the pharmacokinetics and drug-likeness properties of the two compounds. Comparative structural analysis of protein-inhibitor complexes revealed that the amino acids Y32, K47, Y122, Y129, H133, N138, D140, T141, S709 and N781 are crucial for drug surface hotspot in the RdRp of SARS-CoV-2.

The pandemic Corona Virus Disease 19 has become a critical, rapidly emerging public health issue for the world. It is caused by the outbreak of Severe Acute Respiratory Syndrome Corona Virus 2 (SARS-CoV-2), the disease characterized by fever, cough, and severe shortness of breathing, nausea, vomiting and diarrhea [1] . As of 24 August 2020, SARS-CoV-2 infection has been reported in 188 countries with 23.4 million confirmed cases and 808,000 total deaths [2, 3] . Epidemiological data have determined person to person transmission as the route of the rapid outbreak of COVID-19, which has become a major obstruction in combating the virus [4, 5] .Clinical studies reported that older patients have a higher case of fatality rate (CFR) than the young, and males have a higher CFR than female [6] . Apart from acute respiratory distress, COVID-19 patients have been diagnosed with higher rate of renal impairment, indicating the development of kidney dysfunction [7] . Unfortunately, there are no proven drugs, vaccines or therapies available to fight against COVID-19.

SARS-CoV-2 is an enveloped, positive-sense, single-stranded RNA virus similar to SARS and MERS (Middle East Respiratory Syndrome) coronavirus [8] . Along with structural proteins (like spike glycoprotein and accessory proteins), the viral genome also encodes non-structural proteins, including 3-chymotrypsin-like protease, papain-like protease, helicase and RNAdependent RNA polymerase (RdRp) [9] . RdRp is an essential enzyme involved in the replication of RNA viruses including SARS-CoV-2. Several anti-viral drugs have been developed targeting this enzyme for treating infections like Hepatitis C, Zika and other coronaviruses [10] .

Although not yet extensively explored, some of these drugs also target the SARS-CoV-2 RdRp or its catalyzed polymerization process [11] [12] [13] [14] . A recent study suggested that two known anti-viral drugs Remdesivir and Favipiravir which are used for the treatment of variety of RNA virus diseases, by targeting RNA polymerase and RdRp respectively, could successfully inhibit SARS-well as in figuring out drug surface hotspot, which are important for the discovery of an effective drug [18, 19] . In addition, drug repurposing approach is used for the identification of existing drugs for one disease, for the treatment of another disease. As the long-term solution, vaccine, will take years to be marketable, effective repurposing of existing drug remains the only alternative way to fight against an emerging disease like COVID-19 [20] [21] [22] [23] . Thus, in the presented study, drugs with proven anti-viral activity were analyzed using molecular docking and pharmacophore modeling technique to target RdRp of SARS-CoV-2. Our findings can open a new avenue to fight against COVID-19.

The RdRp structure of SARS-CoV-2 was retrieved from the Protein Data Bank (PDB) with PDB ID: 6M71 [24] . Furthermore, 44 drug candidates having inhibitory activity against RNA polymerase were selected by comprehensive literature study, and their PDB structures were retrieved from the Drug Bank Database (Supplementary Table 1 ) [25] . RdRp inhibitor drugs were prioritized; but drugs that are DNA-dependent RNA polymerase inhibitors having the antiviral activity were also included.

The AutoDock Vina software of molecular docking approaches was employed for the screening of the drugs against RdRp of SARS-CoV-2 [26] . At first, the crystal structure of RdRp was retrieved from the PDB, processed by removing water and complex molecules using PyMOL [27] . After preparing the PDB structures of the inhibitors, they were exposed to the RdRp polymerase of SARS-CoV-2 for analyzing the lowest binding energy and interactive amino acids. The grid box parameters were set to size 80Å x 95Å x 95Å (x, y and z) and centre 121.253Å x 121.376Å x 120.149Å (x, y and z). The 2D ligand-protein interaction diagrams were generated by LigPlot+ to find out the involved amino acids with their interactive position in the docked molecule [28] . Discovery Studio and PyMOL were used to visualize and analyze the ligand molecules' interactions with the viral proteins [29] . Additionally, the Protein-Ligand Interaction Profiler (PLIP) was also used to analyze the total number of non-covalent interactions J o u r n a l P r e -p r o o f Journal Pre-proof (hydrogen bonds, water bridges, salt bridges, halogen bonds, hydrophobic interactions, πstacking, π-cation interactions and metal complexes) in protein-ligand complexes [30] .

LigPlot+, Discovery Studio and PyMOL were used to figure out the drug surface hotspot from the docked structures of RdRp with the top-most polymerase inhibitors. Remdesivir and Favipiravir were used as positive control, as they were reported to be effective for COVID-19 by several recent studies [13, 14] .

PharmaGIST was used for the modeling of the pharmacophore features that are essential for the interaction of RNA polymerase inhibitors with RdRp of SARS-CoV-2. In this study, the topmost inhibitors along with Remdisivir were used for the pharmacophore modeling [31, 32] .

ZINCPharmer was used to import the generated pharmacophore from PharmaGISTfor the virtual screening of the novel compounds from the ZINC database [33, 34] . These novel compounds were further used for the screening of new inhibitors of RdRp of SARS-CoV-2. The validity of the screened compounds was checked by Molecular docking approaches.

SwissADME server was used to assess the Absorption, Distribution, Metabolism and Excretion (ADME) properties of the compounds screened from ZINC database [35] . This server is well known for successfully evaluating the pharmacokinetics, drug-likeness and medicinal chemistry friendliness of potential drug candidates. In this study, the physicochemical parameters (Formula inhibitor, CYP2D6 inhibitor, CYP3A4 inhibitor, Log Kp; skin permeation) and water solubility (Log S: SILICOS-IT, Solubility) were checked for the screened compounds considering all default parameters [36, 37] . Additionally, OSIRIS Property Explorer and admetSAR were used to investigate the undesired effects of these compounds like mutagenicity, tumerogenecity and J o u r n a l P r e -p r o o f Journal Pre-proof toxicity [38] [39] [40] [41] . No ADME or toxicity analysis was required for the top selected drugs since the drugs were previously tested for FDA approval.

The docking results were further evaluated using two in silico techniques -i) re-docking of the selected inhibitors with other two docking tools such as PatchDock and iGEMDOCK, and ii) site-specific docking of these inhibitors against another PDB structure of RdRp of SARS-CoV-2 with the predicted binding site. PatchDock is an advanced molecular docking algorithm in which unacceptable penetration of receptor's atoms to ligand's atoms was discarded [42] . This algorithm has three major stages such as Molecular Shape Representation, Surface Patch Matching and Scoring with Filtering. iGEMDOCK is a molecular docking tool which employed generic evolutionary method for docking and specialized for pharmacological interactions [43] .

In iGEMDOCK, the number of generation was set to 70 and the number of solution was 2.

Further, the best-pose was selected for interaction analysis.

Additionally, the selected drugs and zinc compounds along with Remdesivir were allowed to dock against another structure of RdRp of SARS-COV-2 in a site-specific manner. The previously predicted binding site was set for this site-specific docking, and was carried out by Autodock Vina. Another crystal structure of RdRp of SAR-CoV-2 was retrieved from the Protein Data Bank (PDB) with PDB ID: 7BV2. After cleaning the structure with PyMOL, the structure was prepared for docking by Autodock tools. The grid box parameters were set to size 20Å x 26Å x 18Å (x, y and z) and centre 117.203Å x 90.242Å x 76.951Å (x, y and z).

Molecular dynamics (MD) simulation was performed in order to analyze the docked poses. In MD simulation, the integration of Newton's low generates successive iterations where the results specify how the positions and velocities of each molecule vary over time in a system [44] . In this study, MD simulation was performed for the top predicted drug with lowest binding energy and the two predicted compounds from Zinc database. LARMD tools was adopted to run MD simulation considering the influence of water molecules and the time interval was set to 4ns [45] .

AMBER16 was used as the force field in this MD simulation where the Sander module was used J o u r n a l P r e -p r o o f Journal Pre-proof to perform the minimization in 4 steps before the simulation [46] . The 2000 steps steepest descent method along with the 3000 steps conjugated gradient method were used in all minimization processes and the system was heated from 10 to 300K in 30ps. Finally, periodic boundary condition was applied to relax all the atoms in 300K [45] .

Molecular docking of all the RNA polymerase inhibitors against RdRp of SARS-CoV-2 (PDB ID: 6M71) revealed Rifabutin, Rifapentine, Fidaxomicin, 7-methyl-guanosine-5'-triphosphate-5'guanosineand Ivermectin chronologically as the top-most effective RdRp inhibitors of SARS-CoV-2 with highest binding affinity and lowest free energy ( 

The molecular docking pattern and amino acid residues involved were further analyzed to reveal the common interactive sites of RdRp in the SARS-CoV-2. Thus, the binding pattern of five most effective compounds along with Remdesivir and Favipiravir, were analyzed to observe the common drug surface hotspot. It was found that the amino acids Y32, K47, Y129, H133, N138, C139, T141 and S709 in RdRp were involved in the interaction with Rifabutin. The amino acids Y32, K47, Y129, H133 and S709 along with D140 and N781 were also found crucial for the RdRp of SARS-CoV-2 to interact with Rifapentine, Fidaxomicin, 7-methyl-guanosine-5'triphosphate-5'-guanosine and Ivermectin (Table 1 and Fig. 2 & 3) . Most surprisingly, these amino acid residues were also found to be involved in the interaction of Remdesivir (K47, Y129, A130, H133, F134, D135, N138, C139, T141, S709, T710, D711, Q773 and N781) and

Favipiravir (Y129, H133, S709, K780 and N781).

The top listed RdRp inhibitors along with Remdesivir were further used for the pharmacophore modeling and screening of ZINC database. The pharmacophore modeling (predicted by the J o u r n a l P r e -p r o o f Journal Pre-proof PharmaGIST) revealed 6 spatial features (Aromatic-1, Hydrophobic-1 and Acceptors-4) ( Figure   3 ). This pharmacophore model was imported in the ZINCPharmer for the screening of the ZINC Database, which revealed four different hits (ZINC09128258, ZINC09883305, ZINC09883308 and ZINC11286235). In addition, molecular docking analysis revealed that two compounds (ZINC09128258 and ZINC09883305) which are structural isomer of each other, could also act as the inhibitors of RdRp of SARS-CoV-2 as they showed similar binding pattern as the top drugs and controls ( Table 2 and Fig. 5 & 6) . However, these two compounds showed lower binding affinity than the top selected RdRp inhibitors. Additionally, higher number of non-covalent interactions was found for ZINC09883305 (9 non-covalent interactions). The chemical diagrams of all compounds analyzed in this study were shown in Fig 7. 

The physico-chemistry, pharmacokinetics, medicinal chemistry friendliness and toxicity of these two screened compounds (ZINC09128258 and ZINC09883305) from ZINC database were analyzed by SwissADME, ADMETsar and OSIRIS Property Explorer. The physiochemical parameters, lipophilicity and water solubility of these compounds are described in Table 3 . Both compounds had some similar properties as they are structural isomer of each other. Water solubility was also studied in this study and it was found that all the compounds were moderately soluble. Other important properties such as Molecular weight (MW), molecular refractivity (MR) and topological polar surface area (TPSA), which are very useful for the estimation of ADME properties were also included in this study. Remarkably, none of the screened compounds showed any undesired effects such as mutagenicity, tumorigenicity, irritating and reproductive effects. However, these compounds showed CYP450 enzymes inhibition effects except CYP1A2. Lastly, BOILED-Egg model was employed to calculate the Blood-brain barrier (BBB) permeation that revealed no BBB permeate in the studied compounds [47] . (Table 4) . Additionally, site-specific docking against another structure with PDB ID 7BV2 revealed the binding affinity of all drugs were higher than Remdesivir and mostly similar as the previous blind docking against the structure with PDB ID 6M71 (Table 4 ). In this case, Fidaxomicin was found as the top most compound followed by Rifabutin and others.

Rifabutin and other two-screened compounds ZINC09128258 and ZINC09883305. Analysis of root mean square deviation (RMSD) showed that the receptor was in an equilibrium state before 1ns after the binding of the inhibitors and the RMSD was less than 2 Å in all three complexes (Fig.   8 ). However, there was a much fluctuation found in RdRp for the complex with ZINC09883305.

In case of Rifabutin-RdRp complex, both ligand and receptor sites amino acids are found in equilibrium state in 1ns. Additionally, the number of H-bonds involved in the interactions between the inhibitors and the RdRp was also investigated during this simulation. An average of 2 hydrogen bonds was observed when the RdRp of SARS-CoV-2 in complex with rifabutin and ZINC09883305, while an average of 3 hydrogen bonds were found for ZINC09128258 in this 4ns time interval (Fig. 9) . However, all three complexes were found to have a maximum number of 5 hydrogen bonds.

In this study, the estimation of free energies consists of electrostatic energy (ELE), Van der Waals contribution (VDW), total gas phase energy (GAS) and final estimated binding energy (deltaPB/deltaGB). MM/PB (GB) SA calculations showed that the value of free energy in the RdRp in complex with Rifabutin, deltaPB ≈ -10.44 kcal/mol. Meanwhile, ZINC09128258 had the value of deltaPB ≈ -8.15 kcal/mol and ZINC09883305 had deltaPB ≈ -6.81kcal/mol (Table   5 ).

Finally, Principal Component Analysis (PCA) was employed for the investigation of the effects of the conformational factors. This study found two distinct clusters for all three complexes in PC1 vs PC2 J o u r n a l P r e -p r o o f graph (Fig. 10) . This graph has blue to red colored frames over time and represents the active and inactive state of a protein.

At present, COVID-19 is a global challenge for the scientific communities as its pandemic attitude is dangerously affecting millions of people and taking thousands of lives everyday. But to date, no satisfactory breakthrough has been made in the treatment of COVID-19 [48] [49] [50] [51] .

Several attempts have been made to treat this disease but these drug candidates remain questionable owing to low efficacy [52] . Computational approaches along with drug repurposing methods could be an effective approach to this COVID-19 challenge. In this study, several polymerase inhibitors targeting RdRp of SARS-CoV-2 were studied, as RdRp had already been shown to be an effective anti-viral drug target for various viral pathogens such as Hepatitis C Virus, HIV, Zika virus etc [53, 54] . Here, drug repurposing along with molecular docking was employed for the screening and analysis of the drug candidates against RdRp of SARS-CoV-2.

Moreover, the common drug surface hotspot was studied along with modeling of pharmacophore which is very important for drug discovery. Additionally, novel compounds from ZINC database were screened out which could be developed as a new drug to treat COVID-19.

RdRp plays indispensable roles in the life cycle of RNA viruses. RNA viruses initiate RNA synthesis by virus polymerase utilizing primer-independent and primer-dependent mechanism.

Moreover, RdRp based RNA synthesis doesn't occur in the mammalian cells offering an opportunity to design drugs specifically acting against RNA viruses. Additionally, the protein structure of RdRp in RNA viruses is found to be remarkably conserved. Various antiviral drugs have been developed targeting this enzyme for the treatment of infections caused by RNA viruses and they are working effectively [10, 55] .Therefore, this present study aimed to identify potential drugs targeting this enzyme for the treatment of COVID-19. [57, 58] . Moreover, the vast number of non-covalent interactions between these screened compounds with RdRp suggests that the protein-inhibitor complexes are very stable. In addition to these drugs, Remdesivir along with Favipiravir which has already been suggested for the treatment of COVID-19, was also found to be effective as the inhibitor of RdRp of SARS-CoV-2 [14] . Although their binding affinity was lower than the aforementioned top five drugs, the binding pattern was almost similar to these drugs. The drug surface hotspot study revealed that the molecular binding sites of all the five compounds were of a similar pattern that could be the hotspot of drug binding to the RdRp of SARS-CoV-2. This study suggested that amino acids Y32, K47, Y122, Y129, H133, N138, D140, T141, S709 and N781 in RdRp of SARS-CoV-2 could make effective interactions with drugs, though this needs to be validated in wet lab. Additionally, pharmacophore was designed using the top RdRp inhibitor drugs along with Remdesivir, which was used further for the screening of ZINC database. Molecular docking analysis revealed that two compounds among the four hits had interacted effectively with the RdRp of SARS-CoV-2 which is indicative of their potential as inhibitors of the enzyme. Although their binding affinity was lower than one control (Remdesivir), the vast number of interactions would give the complex its stability with this binding energy. Moreover, ADME and toxicity analysis of these compounds suggested that they could be used for the development of new drugs to treat COVID-19. However, the study of the inhibition of cytochromes P450 (CYP) isoforms revealed that there was a possibility that the suggested compounds could interact with CYP isoforms.

Evaluation of the docking performance is very crucial to make the prediction more strong. Redocking with other docking tools and site-specific docking with another structure are very effective ways of the evaluation of the docking results [59] . In this study, re-docking provided the mostly similar docking results and Rifabutin was found to have the highest binding affinity in all tools. All the drugs were found to have much higher binding affinity than the control. However, the scores were different as different methods use different scoring functions.

Additionally, site-specific docking against another structure of RdRp revealed the mostly same binding poses and the binding affinity.

The accurate prediction of protein-inhibitors complexes is very important for the drug development by the computational approaches. MD simulation could be used in the assessment of the docking pose predicted through the molecular docking analysis. It could be predicted whether a docked pose is stable or not in an aqueous environment [60] . In this study, MD simulation was also adopted to assess the stability of Rifabutin-RdRp, ZINC09128258-RdRp and https://doi.org/10.1186/s13578-020-00404-4. J o u r n a l P r e -p r o o f J o u r n a l P r e -p r o o f J o u r n a l P r e -p r o o f J o u r n a l P r e -p r o o f 

Similarity in Case Fatality Rates (CFR) of COVID-19/SARS-COV-2 in Italy and China

Coronavirus disease (COVID-19) Situation Dashboard

A familial cluster of pneumonia associated with the 2019 novel coronavirus indicating person-to-person transmission: a study of a family cluster

Epidemiological and clinical characteristics of 99 cases of 2019 novel coronavirus pneumonia in Wuhan, China: a descriptive study

Epidemiological and clinical features of the 2019 novel coronavirus outbreak in China

2019-nCoV Volunteers

Caution on Kidney Dysfunctions of 2019-nCoV Patients

Severe Acute Respiratory Syndrome Coronavirus-2 (SARS-CoV-2): An Update

Coronaviruses-drug discovery and therapeutic options

Tenofovir against SARS-CoV-2 RNA dependent RNA polymerase (RdRp): A molecular docking study

The potential chemical structure of anti-SARS-CoV-2 RNA-dependent RNA polymerase

2020) jmv.25761

Therapeutic options for the 2019 novel coronavirus (2019-nCoV)

Favipiravir (T-705), a novel viral RNA polymerase inhibitor

Remdesivir and chloroquine effectively inhibit the recently emerged novel coronavirus (2019-nCoV) in vitro

Coronavirus susceptibility to the antiviral remdesivir (GS-5734) is mediated by the viral polymerase and the proofreading exoribonuclease

Binding mechanism of remdesivir to SARS-CoV-2 RNA dependent RNA polymerase

Coronavirus puts drug repurposing on the fast track

In silico pharmacology for drug discovery: Methods for virtual ligand screening and profiling

Informatics and modeling challenges in fragment-based drug discovery

Drug repurposing approaches for the treatment of influenza viral infection: Reviving old drugs to fight against a long-lived enemy

Drug Repurposing for Viral Infectious Diseases: How Far Are We?

Network bioinformatics analysis provides insight into drug repurposing for COVID-2019

The RCSB protein data bank: integrative view of protein, gene and 3D structural information

DrugBank 5.0: a major update to the DrugBank database for 2018

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

Pymol: An open-source molecular graphics tool

LigPlot+: Multiple ligand-protein interaction diagrams for drug discovery

Discovery Studio Modeling Environment

PLIP -Fully automated Protein-Ligand Interaction Profiler</p>

Deterministic pharmacophore detection via multiple flexible alignment of drug-like molecules

PharmaGist: a webserver for ligand-based pharmacophore detection

ZINCPharmer: pharmacophore search of the ZINC database

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

SwissADME: A free web tool to evaluate pharmacokinetics, drug-likeness and medicinal chemistry friendliness of small molecules

Bioavailability prediction of phytochemicals present in Calotropis procera (Aiton) R. Br. by using Swiss-ADME tool

VITRO ADME STUDIES OF TUG-891

Molecular Properties Prediction -Osiris Property Explorer

Estimation of ADME properties with substructure pattern recognition

In silico prediction of chemical ames mutagenicity

In Silico Estimation of Chemical Carcinogenicity with Binary and Ternary Classification Methods

PatchDock and SymmDock : servers for rigid and symmetric docking

iGEMDOCK : a graphical environment of enhancing GEMDOCK using pharmacological interactions and post-screening analysis

Molecular modelling: principles and applications

LARMD : integration of bioinformatic resources to profile ligand-driven protein dynamics with a case on the activation of estrogen receptor

The Amber Biomolecular Simulation Programs

A BOILED-Egg To Predict Gastrointestinal Absorption and Brain Penetration of Small Molecules

What we know so far: COVID-19 current clinical knowledge and research

Sensitivity of Chest CT for COVID-19: Comparison to RT-PCR

Discovering drugs to treat coronavirus disease 2019 (COVID-19)

COVID-19: a recommendation to examine the effect of hydroxychloroquine in preventing infection and progression

The design of drugs for HIV and HCV

Purification of Zika virus RNA-dependent RNA polymerase and its use to identify small-molecule Zika inhibitors

RNA-Dependent RNA Polymerases and Their Emerging Roles in Antiviral Therapy

The FDA-approved Drug Ivermectin inhibits the replication of SARS-CoV-2 in vitro

Antiviral Chemotherapy for Infection with Human Immunodeficiency Virus

Orthopoxvirus targets for the development of antiviral therapies

Benchmarking of different molecular docking methods for protein-peptide docking

Molecular dynamics analysis to evaluate docking pose prediction