key: cord-1042537-zyr33n6r
authors: Kumar, Sumit; Sharma, Prem Prakash; Upadhyay, Charu; Kempaiah, Prakasha; Rathi, Brijesh; Poonam,
title: Multi-targeting approach for nsp3, nsp9, nsp12 and nsp15 proteins of SARS-CoV-2 by diosmin as illustrated by molecular docking and molecular dynamics simulation methodologies
date: 2021-02-25
journal: Methods
DOI: 10.1016/j.ymeth.2021.02.017
sha: 9de529448c2adf332db75549de1c0e32a58f287e
doc_id: 1042537
cord_uid: zyr33n6r

Novel coronavirus SARS-CoV-2continues tospread rapidly worldwide and causing serious health and economic loss. In the absence of any effective treatment, various in-silico approaches are being explored towards the therapeutic discovery against COVID-19. Targeting multiple key enzymes of SARS-CoV-2 with a single potential drug could be an important in-silico strategy to tackle the therapeutic emergency. A number of Food and Drug Administration (FDA) approved drugs entered into clinical stages were originated from multi-target approaches with an increased rate, 16-21% between 2015-2017. In this study, we selected an FDA-approved library (Prestwick Chemical Library of 1520 compounds) and implemented in- silico virtual screening against multiple protein targets of SARS-CoV-2 on the Glide module of Schrödinger software (release 2020-1). Compounds were analyzed for their docking scores and the top-ranked against each targeted protein were further subjected to Molecular Dynamics (MD) simulations to assess the binding stability of ligand-protein complexes. A multitargeting approach was optimized that enabled the analysis of several compounds’ binding efficiency with more than one protein targets. It was demonstrated that Diosmin (6) showed the highest binding affinity towards multiple targets with binding free energy (kcal/mol) values of -63.39 (nsp3); -62.89 (nsp9); -31.23 (nsp12); and -65.58 (nsp15). Diosmin (6) possessing multi-targeting ability against non-structural proteins of SARS-CoV-2, and thus it could be considered for further validation experiments before using as therapeutic against COVID-19 disease.

Novel coronavirus (SARS-CoV-2) has emerged in late 2019 from a cluster of pneumonia cases in people associated with sea-food market in Wuhan, China [1, 2] . SARS-CoV-2 caused a viral infection in human and the disease was named as COVID-19 [3, 4] . As of February 2, 2021, the virus spread has infected over 102.3 million with a fatality of over 2.21 million people in the world [5] . Of particular note, United States of America, India, Brazil and Russia have officially reported the highest number of COVID-19 cases [6] . In contrast to the previously reported SARS (Sever Acute Respiratory Syndrome) in 2003, the SARS-CoV-2 has much lower estimated mortality ratio (~2.96%) [7] . However, the number of cases is high and have an immense growth rate in case of the SARS-CoV-2. The majority of the infection cases are asymptomatic, and thus raising serious health concerns as the virus continues infecting others [8] . Currently, researchers from academia and industries are in the dire need for finding treatments against COVID-19 including vaccines to limit the spread of the virus [9] .

Discovering therapeutics [10, 11] based on targeting multiple enzymes represents an important approach towards the anti-COVID drug discovery. As such, applications of in-silico methods are easily explorable and time saving to accomplish search of potential molecules against multiple protein targets [12] [13] [14] . Interestingly, multiple target strategy has been successfully demonstrated to develop therapeutics for various diseases including Cancer, Alzheimer's and Parkinson's with complex causative factors [15] , and few drugs have entered into the clinical trials [15, 16] .

FDA approved Prestwick Chemical Library (PCL) of 1520 compounds [17] was tested for antiviral activity against two previously reported coronaviruses, SARS-CoV and Middle East respiratory syndrome coronavirus (MERS-CoV) [18] . Frieman et al. [18] demonstrated 27 compounds active against both SARS-CoV and MERS-CoV. Lately, Frieman and his coworkers [19] analyzed the top 20 FDA approved drug out of 27 compounds against SARS-CoV-2, which showed in-vitro antiviral activity against SARS and MERS coronaviruses. Out of 20 compounds, 17 were successfully investigated against SARS-CoV-2 with IC50 values of 2.36 -43.02 µM. Likewise, another study presented the potency of the PCL compounds in inhibiting the replication of SARS-CoV-2 [20] . Encouraged with this, we selected PCL compounds and decided to perform multi-targeting virtual screening against non-structural proteins (nsp-3, nsp-9, nsp-12 and nsp-15) of SARS-CoV-2. The non-structural proteins are responsible for the survival of the virus and its protection in host environment. Molecular docking results were further substantiated with MD simulations.

The ADP ribose phosphatase (nsp3), replicase (nsp9), RNA dependent RNA polymerase (nsp12), and endonuclease (nsp15) protein structures were collected from the RCSB website (https://www.rcsb.org) having their PDB IDs 6W02, 6W4B, 6M71 and 6W01, respectively.

The computational work was performed using Schrödinger software (release 2020-1).

The SARS-CoV-2 virus protein structures were downloaded from Protein Data Bank at RCSB website (https://www.rcsb.org) and then prepared in the protein preparation wizard and prime module of Schrödinger suite to remove any associated defects prior to docking [21, 22] . The strained bonds and any other steric clashes were removed using restrained energy minimization, allowing movement in heavy atoms up to 0.3 Å. These structural errors were removed from downloaded protein structures as per the reported literature [8] .

After protein preparation, the PCL containing 1520 compounds was also prepared on the Schrödinger software (release 2020-1). The ligand structures of all compounds under study were prepared using Schrödinger's Ligprep tool, where all possible combination of chirality was generated for each compound prior to docking [23] . These structures were energetically minimized by using forcefield OPLS3e. The ionization state was determined for ligands at pH 7 ± 2 as well as all tautomers were generated using the Epike module of Schrödinger software [24] .

Molecular docking is one of the key methods used in computer-assisted drug design (CADD).

The main objective is it to predict the interaction of a ligand with the known crystal structure of the protein. Hence, docking was performed to find the most suitable interactions (hydrogen bonds, water bridge, hydrophobic interactions and salt bridge) and orientation of each ligand at the protein's active site [25] . These steps help in shortlisting the large compound library to smaller ones based on scoring function, which ranks the compound among library of compounds. Thus, the screening of 1520 compounds was carried out in extra precision (XP) mode to get their respective binding pose and docking score [26] . It was performed in three successive modes provided by the Glide module [27] . The first screening was performed in high throughput virtual screening (HTVS) mode where all compounds were screened and only top 10% of structures were further screened by standard precision (SP) mode. Then again only top 10% structures from SP mode were selected into the next stage of screening by XP mode.

The Van der Waals radii scaling factor and partial charge cutoff of 0.8 and 0.15, respectively, were used for docking. Finally, compounds were ranked based on their docking score, XP Gscore, and binding free energy. The binding free energy for these complexes were also calculated by prime MMGBSA [28] .

MD is a computer simulation method used for the analysis of physical movements of atoms and molecules. The atoms and molecules of the ligand under study are allowed to interact with the targeted protein for a fixed simulation period. It helps to calculate changes in the binding free energy of a compound before simulation and after it, which ultimately indicates the stability of compound within binding proximity of the targeted protein [29, 30] . Thus, MD simulation was carried out to validate docking observations for the top ranked ligand-protein complexes. The simulation was performed for a 100 nano seconds (ns) time period on the Desmond software (D. E. Shaw Research, New York, NY, 2015) to visualize the binding and conformational stability of compound within binding pocket of the respective protein [31] . The complex was solvated TIP3P water model prior to simulation, where 0.15 M NaCl added to mimic a physiological ionic concentration.

The main motive of this study is to find a multi-targeting inhibitor against several nonstructural proteins of SARS-CoV-2. The observation from the study was that four compounds out of 1520 screened compounds showed best results against four different proteins on the basis of molecular docking experiments. MD simulation was also carried out on these four compounds with respective proteins, which revealed the stability of ligand-protein complexes.

Then, other hit compounds were also analyzed against each targeted protein and only Diosmin (6) was found successful in targeting all the listed proteins. The overall study flow chart is shown in Figure 1 .

The objective of our study was centered at the discovery of multi-targeting inhibitors against key protein targets of SARS-CoV-2 i.e. nsp3 ADP ribose phosphatase, nsp9 replicase, nsp12 RNA dependent RNA polymerase and nsp15-endoribonuclease anticipating their applications as therapeutics against COVID-19. The docking results and MD simulations against individual protein targets are described separately. The crystal structures of selected target proteins were downloaded from the website. The number of amino acid (aa) present in these selected proteins for SARS-CoV-2 are found to be 170 aa for ADP ribose phosphatase (nsp3), 117 aa for nsp9, 942 aa for RNA-dependent RNA polymerase (RdRp; nsp12) and 370 aa for endoribonuclease (nsp15). The FDA-approved chemical library of 1520 compounds was downloaded from Prestwick database [17] .

The non-structural protein, nsp3 is a papain-like proteinase protein with a long sequence, it possesses several conserved domains: ssRNA binding, ADP ribose binding, G-quadruplex binding, protease (papain-like protease, and nsp4 binding), and transmembrane domain. It also exhibits deubiquitinating (DUB) and deISGylating (deISG) activities and interacts with nsp4 and nsp6 [32, 33] . Among 16 non-structural proteins known in SARS-CoV-2, only nsp3, nsp4, and nsp6 have transmembrane domains [34] . ADP ribose phosphatase domain ( Figure S1A ;

supporting information) of nsp3 is known to interfere with the immune response, possibly due to its ability to remove ADP-ribose from ADP-ribosylated proteins and RNA, which is part of the host's immune response's intricacies [35] . Considering this important protease activity to release essential proteins for viral activity, the inhibition of nsp3 protease activity is an important target for antiviral activity [32] .

An extensive, in-silico study of the FDA-approved chemical library was carried out using Schrödinger software (release 2020-1). First, all the compounds were docked with targeted nsp3 protein and hit compounds were selected on the basis of docking score, XP Gscore and binding free energy (Table S1, entry 1-8; supporting information). One of the top hits compound was Hesperidin (1) (PubChem ID: 10621) that displayed a docking score of -12.967 and the binding free energy of -73.20 kcal/mol (Table S1 , entry 1; supporting information).

Hesperidin (1) is a flavanone glycoside isolated in the year 1828 from citrus peels [36] . It is known to treat hemorrhoids (blood vessel condition), alone or in combination with other citrus bioflavonoids [37] . Recently published articles suggested lopinavir as a known inhibitor of the nsp3 protein [38, 39] . So, our molecular docking result were also compared with known inhibitor, which displayed docking score of -7.478 and binding free energy of -65.24 kcal/mol (Table S1 , entry 9; supporting information). The result was lower than top-ranked Hesperidin (1) ( Table S1 , entry 1; supporting information) as well other shortlisted compounds (Table S1 , entry 2-8; supporting information). Hesperidin (1) showed six H-bond interactions with the amino acid residues Asp22, Lys44, Gly47, Leu126, Ser128 and Ala129 of targeted nsp3 protein (Figure 2) , was further analyzed for conformational stability using molecular dynamic (MD) simulation methodology.

To validate the docking result for Hesperidin (1), MD simulation was performed for a 100ns.

The root mean square deviation (RMSD) of Cα of nsp3 protein in complex with Hesperidin (1) remained stable within the range from 0.9 to 1.2 Å throughout the simulation period ( Figure   S2A ; supporting information). There was fluctuation noticed in ligand RMSD (up to 7 Å) for the first 60ns after which it became stable and fluctuated within an acceptable limit (i.e. 3 Å).

This change in RMSD is mainly due to both pyran rings of Hesperidin (1). The ligand root mean square fluctuation (RMSF) was also noticed due to the presence of pyran rings while less fluctuation for those ligand atoms, which are in contact with protein ( Figure S2B ; supporting information). Overall, the compound remained in the binding pocket of the targeted protein.

The RMSF plot of Cα of nsp3 protein showed fluctuations in Cα residues as well as with other residues interacting with compound, indicated by green color line ( Figure S2C ; supporting information). Hesperidin (1) displayed several interaction with nsp3 protein residues such as: H-bond interaction (Asp22, Ala38, Asn40, Lys44, Gly46, Gly47, Gly48, Val49, Leu126, Ser128, Ala129, Ile131, Phe156, Asp157), hydrophobic interaction (Ala38, Val49, Ala50, Ala52, Ala129, Ile131, Phe132, Phe156), water bridge (Asp22, Ile23, Val24, Asn37, Ala39, Asn40, Lys44, His45, Gly46, Gly47, Gly48, Val49, Lys55, Val95, Gly97, Leu126, Ser128, Ala129, Ile131, Ala154, Phe156, Asp157), and salt bridge (Asp22, Gly48, Asp157) ( Figure   S2D ; supporting information). There were several residues identified with more than one type of interactions.

Targeted protein, nsp9 is a non-structural protein, encoded by ORF1a and related to the viral RNA synthesis. It contains a single 5-folded β-barrel, which is unique, unlike the other single protein domain proteins. The crystal structure of nsp9 replicase is a dimeric protein ( Figure   S1B ; supporting information). The protein replicase specifically binds to the RNA, which further interacts with the nsp8 protein and activates the mechanism of action for its functioning [40, 41] . It plays a crucial role in viral replication, but the mechanism of the RNA binding within nsp9 protein is still not known [42] . The induction of disturbance to the dimerization of nsp9 can be a way to overcome SARS-CoV-2 [43, 44] . Thus, it acts as potential drug target against SARS-CoV-2, providing a hope for new drugs inhibiting the viral progression.

Diosmin (6) (PubChem ID: 5281613) [45] was found to have significant better docking score (-9.409) and binding free energy (-62.89 kcal/mol) among screened FDA-approved chemical library (Table S1 , entry 10; supporting information). Diosmin (6) is approved for the treatment of hemorrhoids or chronic venous diseases. It is a flavone extracted from natural occurring plant Teucrium gnaphalodes [46] . There are some scaffolds reported to bind with nsp9 protein of SARS-CoV-2 such as Withanone (PubChem ID: 21679027) by Maurya et al. [47] in which authors have studied the molecule using molecular docking methodology on AutoDock Vina software. Therefore, our docking results were compared with Withanone that displayed a docking score of -4.673 with the binding free energy of -44.43 kcal/mol (Table S1, entry 19; supporting information). Interestingly, Diosmin (6) showed improved docking results in comparison to Withanone (Table S1 , entry 10; supporting information). Diosmin (6) displayed three important H-bond interactions with residues Arg40, Val42 and Ser60 of nsp9 protein as represented in Figure 3 .

The RMSD value of Cα within the complex of Diosmin (6)-nsp9 increased for the initial 20ns, and remained stable during the simulation period (100ns) (Figure 4A) . Likewise, ligand (6) RMSD was also found to be stable throughout the simulation period, indicating that ligand remained within the proximity of the binding site as shown in Figure 4B . The RMSF plot of Cα residues lie in the loop region of nsp9 protein and showed more fluctuations than those residues, which lie in secondary structure ( Figure 4C ). Residues interacting within the binding site with ligand are shown in green color. Diosmin (6) showed three types of interactions with the amino acid residues. The residues involved in interactions are present either loop formation or sheet formation. Three types of interaction are: 1) H-bond (Ser14, Thr36, Lys37, Arg40, Val42, Ser60, Asp61, Thr68); 2) hydrophobic interaction (Met13, Arg40, Phe41, Phe57, Pro58, Ile66); and 3) water bridge (Ser14, Ala16, Asp27, Thr36, Lys37, Gly38, Arg40, Val42, Arg56, Lys59, Ser60, Asp61, Tyr67, Thr68) shown in Figure 4D . There are several residues which showed more than one type of interaction. It can be observed that water is playing important role in ligand stability within the binding pocket.

The other target protein nsp12, also known as RNA dependent RNA polymerase (RdRp), catalyzes the synthesis of viral RNA and hence plays an important role in the replication process of SARS-CoV-2 virus proliferation ( Figure S1C ; supporting information). The replication process is achieved possibly with the assistance of nsp7 and nsp8 as co-factors [48] .

Because of the crucial role of RdRp in ssRNA replication, it has been targeted for drug design and development for various viral infections such as hepatitis C virus (HCV) [49] , zika virus (ZIKV) [50] and coronaviruses (CoVs) [51] . The nucleoside triphosphate (NTP) entry channel for RdRp polymerase of SARS-CoV-2 was known to be formed by a 759 to 761 amino acid stretch and other amino acid residues such as Lys545, Arg553 and Arg555 [52] . Hence, this active site stretches of amino acids of RdRp pose a potential inhibitory drug target for SARS-CoV-2.

The library under study was screened virtually against nsp12 protein and top 4 hit compounds based upon docking score, XP Gscore and binding free energy are given in Table S1 , entry 20-23; supporting information. Catenulin (13) (PubChem ID: 4689) [53] was identified as best among 4 hit compounds with high docking score, XP Gscore, and binding free energy of -13.819, -14.449, and -50.14 kcal/mol respectively (Table S1 , entry 20; supporting information). Catenulin (13) is generic name for Paromomycin, which is an antiprotozoal and antibacterial agent [54] . Next, the observed docking results were compared with known inhibitor of RdRp, Remdesivir triphosphate (remdesivirTP) [55, 56] using similar methodology. RemdesivirTP displayed docking score, XP Gscore, and binding free energy values, -8.199, -9.494, and -25.26 kcal/mol, respectively (Table S1 , entry 24; supporting information); which is not up to scratch as compared to top-ranked Catenulin (13) (Table S1 , entry 20; supporting information) and other hit compounds of nsp12 (Table S1, entry 21-23; supporting information). Catenulin (13) was later analyzed for its interaction with amino acid residues of nsp12 protein and found to form six crucial H-bond interactions (Tyr619, Asp623, Asn691, Asp760, Asp761, and Glu811) and three salt bridge interaction (Asp760, Asp761, and Glu811) needed for the stability of the ligand with the protein (Figure 5 ).

The RMSD of Cα of nsp12 protein on complex with Catenulin (13) increased initially for initial 40ns after that it became stable and remained stable at an average RMSD value of 3.5 Å during simulation period (100ns) (Figure S3A; supporting information) . A similar trend was observed for ligand RMSD which increased initially from 0.9 Å and became stable at an average RMSD value of 4.8 Å. The ligand (13) remained within the proximity of the binding site (Figure S3B; supporting information). The RMSF plot of Cα of nsp12 protein shown high fluctuations for the residues present in the loop region as compared to other residues in secondary structure.

The amino acid residues showing interaction with the ligand are represented in green color ( Figure S3C ; supporting information). The residues with which Catenulin (13) showed interaction, were involved either in loop formation or sheet formation. Catenulin (13) displayed several type of interaction with the amino acid residues of targeted protein such as: H-bond (Asp618, Lys621, Asp623, Asp760, Asp761, Lys798, Trp800, His810, Glu811, Ser814), water bridge (Lys551, Arg553, Arg555, Trp617, Asp618, Tyr619, Lys621, Cys622, Asp623, Thr687, Ser759, Asp760, Asp761, Ala762, Lys798, Trp800, His810, Glu811, Cys813, Ser814), and salt bridge interaction (Lys545, Lys551, Arg553, Arg555, Asp618, Lys621, Asp623, Ser759, Asp760, Asp761, Lys798, His810, Glu811, Cys813, Ser814) as depicted in Figure S3D ; supporting information. There were several residues which were showing more than one type of interaction, but no hydrophobic interaction observed between the target protein and the ligand. The common residues in molecular docking and MD simulation were Asp623, Asp760, Asp761, and Glu811 displaying H-bond; and Asp760, Asp761and Glu811 displaying salt bridge interaction. Overall, MD simulation studies strongly validated the molecular docking result for Catenulin (13)-nsp12 interactions.

The nsp15 is non-structured protein of SARS-CoV-2, uridylate-specific endoribonuclease structure. It was reported and suggested as a drug target based on its high sequence similarity with nsp15 protein of SARS and MERS [57] . The catalytic function of nsp15 resides in the Cterminal NendoU domain (Figure S1D, supporting information) . The active site is located in a shallow groove between two β-sheets which carries six key amino acid residues conserved among SARS-CoV-2, SARS-CoV and MERS-CoV proteins: His235, His250, Lys290, Thr341, Tyr343, and Ser294 [58] . The protein is reported to be involved in the RNA replication and processing of sub-genomic RNAs but the exact function is still not clearly understood [57] .

Saikosaponins were reported to show potency against nsp15 of SARS-CoV-2, of which Saikosaponin V (PubChem ID: 100958093) was the top hit with a docking score of -8.538

using molecular docking methodology [59] . Thus, Saikosaponin V was selected as positive control for chosen library of compound. All the compounds displaying better result as compared to Saikosaponin V were shortlisted as hit analogs. A total of 6 compounds (Table   S1 , entry 25-31; supporting information) showed better result, of which compound 16 (Acarbose; PubChem ID: 41774) [60] was the top most on the basis of docking score and binding free energy respectively; -13.527 and -63.01 kcal/mol (Table S1 , entry 25; supporting information). Acarbose (16) is a pseudotetrasaccharide displaying antihyperglycemic activity, which also inhibit alpha-glucosidase, preventing the breakdown of essential larger carbohydrates [60] . Acarbose showed seven H-bond interactions with amino acid residues (His235, Asp240, Gln245, Gly248, Val292, Glu340, and Tyr343) of nsp15 protein as shown in Figure 6 . The compound was further studied for MD simulation to evaluate the stability of Acarbose (16)-nsp15 complex.

The RMSD of Cα of nsp15 protein within complex with Acarbose (16) increased initially from 0.8 Å and stabilized at 2.6 Å, which remain stable throughout simulation period of a 100ns as displayed in Figure S4A ; supporting information. Similarly, ligand RMSD was also increased initially from 0.8 Å and became stable at 50ns. The ligand (16) remained within the proximity of the binding site of nsp15 protein ( Figure S4B ; supporting information) till the simulation period. The RMSF plot of Cα of nsp15 protein showed high fluctuations for the amino acid residues, which lie in loop region as compared to residues lie in secondary structure. The amino acid residues showing interaction with Acarbose (16) represented as green line (Figure S4C; supporting information). Acarbose (16) showed four types of interaction with different amino acid residues of nsp15 protein namely; 1) H-bond interaction (His235, Asp240, Gln245, Gly248, His250, Lys290, Val292, Ser294, Glu340, Thr341, Tyr343); 2) hydrophobic interaction (His243, Tyr343); 3) water bridge (His235, Asp240, Ser242, His243, Gln245, Leu246, Gly248, His250, Asn278, Lys290, Val292, Ser294, Lys335, Glu340, Thr341, Tyr343, Pro344, Lys345, Leu346); and 4) salt bridge interaction (Asp240, Tyr343, Lys345, Leu346) shown in Figure S4D ; supporting information. Many residues showed more than one type of interaction. Overall, the Acarbose (16)-nsp15 complex was stable and strongly supported the molecular docking results.

The main objective of our study is to find out multi-targeting inhibitors against large number of proteins of SARS-CoV-2. The four top-ranked compounds against four selected proteins were: 1) Hesperidin (1) (nsp3; Table S1 , entry 1; supporting information); 2) Diosmin (6) (nsp9; Table S1, entry 10); 3) Catenulin (13) (nsp12; Table S1, entry 20); and 4) Acarbose (16) (nsp15; Table S1, entry 25). Thus, we have implemented the methodology of analyzing interactions of all the top-ranked compounds shortlisted using molecular docking against other proteins considered for this study. It was found that Hesperidin (1) also targeted two more protein other than nsp3 which were: (1) nsp9 (Table S1 , entry 14; supporting information); and

(2) nsp15 (Table S1, entry 28). Diosmin (6) targeted three more protein other than nsp9 which were: (1) nsp3 (Table S1, entry 6); (2) nsp12 (Table S1, entry 23); and nsp15 (Table S1, entry 27). Catenulin (13) targeted one more protein other than nsp12 which was nsp15 (Table S1 , entry 31; supporting information) and Acarbose (16) not targeted any other protein except nsp15. Overall, from 4 top-ranked compounds against 4 proteins considered for the study, 3 have shown multi-targeting capability. The best multi-targeting capacity was displayed by Diosmin (6) , which showed interaction with all the targeted proteins displayed in Figure 7 .

Recently reported literature by Meiyanto et al. [61] also found Hesperidin (1) and Diosmin (6) as best binding molecules out of total analyzed against different proteins of SARSCoV-2 viz.

TMPRSS, 3CL-pro, S2-RBD, and PD-ACE2. These two also possess chemopreventive properties against cancer, viral infection and inflammatory symptoms that could be beneficial to combat the comorbidities of COVID-19 [61] . In our study, Hesperidin (1) and Diosmin (6) also showed strong binding affinity towards maximum number of proteins. These two drug compounds are under clinical trial phase 1 and explored as treatment option against COVID-19 [62] . Interestingly, nsp-15 protein was found to be a common target for all the four hit compounds, Hesperidin (1), Diosmin (6), Catenulin (13) and Acarbose (16) displayed remarkable interactions with key residues, Gln245 and Glu340 present on the active site of the protein (Table S2 ; supporting information). Therefore, nsp-15 represents an attractive target for the development of therapeutics against SARS-CoV-2.

Diosmin (6) showed a docking score, XP Gscore and binding free energy of -10.972, -10.972 and -63.39 kcal/mol respectively; (Table S1 , entry 6; supporting information) against nsp3 protein. It was analyzed for its interaction with nsp3 protein in which the compound displayed five H-bond with residues Asp22, Leu126, Ser128, Val155 and Asp157 as shown in Figure 8 . Figure 9B) . The RMSF plot of Cα of targeted nsp3 protein showed fluctuations for the residues which are present in loop region as well as in secondary structure. The plot also displayed fluctuation for the residues interacting with Diosmin (6) indicated by green color line ( Figure 9C) . The compound showed essential interaction with the residues of nsp3 protein during simulation period essential for the stability of ligand-protein complex were: 1) H-bond interaction (Asp22, Ile23, Gly48, Val49, Leu126, Ser128); 2) hydrophobic interaction (Ala38, Val49, Ala129, Ile131, Phe132, Leu160); and 3) water bridge (Asp22, Ala39, Gly48, Val49, Val95, Gly97, Ala129, Ile131, Ala154, Phe156, Asp157) and displayed in Figure 9D . Further, Diosmin (6) was stable throughout MD simulation process.

Diosmin (6) within nsp12 complex exhibited docking score, XP Gscore, and binding free energy of -9.531, -9.531, and -31.23, respectively (Table S1 , entry 23; supporting information).

Diosmin (6) showed seven important H-bond interactions with residues Asp623, Asp760, Glu811, Ser682, Tyr619, and Trp800 of nsp12 protein as depicted in Figure 10 . Next, Diosmin Figure 11A) . Similarly, ligand RMSD increased initially for first 50ns after which the compound achieved conformational stability at 17.5 Å as well as fluctuated within the range of 3Å which is in the acceptable range. The trajectory analysis revealed that the compound was shifted from its initial docking site to achieve energetically favorable conformational state where it interacted with the residue Lys438 (binding domain for nsp7) as well as with Asp760, which is a catalytic site for the polymerase activity. This interaction caused shifting in the binding domain residue (Asn416-Gly427) which may further affect binding stability of nsp2 to its cofactor nsp7. After 50ns, the ligand remained within the proximity of new binding site (Figure 11B) . The RMSF plot of the protein showed high fluctuations for the residues which lie in loop region as compared to those lie in secondary structure. The residues interacting with the compound are shown in green line ( Figure 11C ).

Diosmin (6) showed three types of important interaction with the residues of targeted nsp12 protein which were: (1) H-bond interactions (Lys438, Trp617, Asp618, Asp761, Trp800, Glu811, Asp833, Arg836); (2) hydrophobic interactions (Phe440, Ala550, Lys551, Arg555, Arg836); and (3) water bridge (Lys438, Ala550, Lys551, Asn552, Arg555, Asp618, Tyr619, Asp760, Asp761, Ala797, Trp800, His810, Glu811, Cys813, Arg814, Asp833, Arg836) ( Figure 11D ). In total, the MD study indicated the stability of the Diosmin (6) in complex with nsp12 protein.

Diosmin (6) in complex with nsp15 protein showed docking score, XP Gscore, and binding free energy of -10.356, -10.356, and -65.58, respectively (Table S1 , entry 27; supporting information). It showed 8 H-bond interactions with the amino acid residues (His235, Gln245, Gly248, His250, Lys290, and Glu340) and also displayed pi-pi interaction with Tyr343 residue (Figure 12) . The RMSD plot of Cα of nsp15 protein in complex with the compound increased initially from 1.0 Å and gets stabilized at 2.5 Å, which then remained stable till the end of simulation period (Figure 13A) . Similarly, the ligand RMSD initially increased upto 4.0 Å and became stable at 36ns and maintained its stable conformation. The major fluctuation observed in atoms 21, 25, 26, 30-32, 40-41, and 43 , which belongs to 2-methoxyphenyl moiety of the compound. Ligand remains within the proximity of the binding site ( Figure 13B) . The RMSF plot of protein showed high fluctuations for those residues, which lie in loop region as compared to those lie in secondary structure. The residues showing interaction with compound are represented as green line ( Figure 13C ). Diosmin (6) interacted with the residues by H-bond (His235, Asp240, Gln245, Gly248, His250, Cys334, Glu340), hydrophobic interaction (His235, Val315, Trp333, Lys335, Tyr343), water bridge (Glu234, His235, Asp240, His243, Gln245, Leu246, Gly248, His250, Lys290, Trp333, Cys334, Lys335, Val339, Glu340, Thr341, Tyr343), and salt bridge interaction (Glu340) (Figure 13D) . Inclusively, MD simulation study indicated that Diosmin (6) was stable in the binding proximity of targeted nsp15 protein.

The stereochemical geometry for all the amino acid residues of all 4 targeted proteins in complexes with their respective ligands were analyzed. All protein complexes have outlier residues either one or zero (Table S3 ; supporting information). The large percentage of the amino acid residues were found to be in favored, additional allowed and generously allowed regions for all the complexes, indicating stability of the complexes (Figure S5 ; supporting information).

The NCBI BLAST results showed that the sequence alignment for the nsp3 binding site has sequence similarity with the residues SAGIF in protein 3Q6Z, 3VFQ, and 2X47 found in humans. It is possible that the compound (6) may interfere with the functioning of these proteins ( Figure S6 ; supporting information). However, nsp9 protein showed very less query cover (21%), indicating the percentage of sequence aligned to the sequence in gene bank, particularly related to humans. This suggests that the absence of the proteins in human with binding site similar to nsp9 protein ( Figure S7 ; supporting information). We could not find similar sequence for the proteins, nsp12 and nsp15 binding site residues in the BLAST search ( Figure S8-S9 ; supporting information).

In the current work, we virtually validated the FDA-approved candidates against multiple targets, non-structural proteins of SARS-CoV-2. In-silico molecular docking aided with MD simulations suggested that four top-ranked compounds against each targeted protein include Flow chart of all the experiments carried out during the study. nsp-3, nsp-9, nsp-12 and nsp-15 protein of SARS-CoV-2 were screened with an approved library of 1520 compounds. The observation from the study revealed top compound against each targeted protein were: Hesperidin (1) (nsp3), Diosmin (6) (nsp9); Catenulin (13) (nsp12), and Acarbose (16) (nsp15) on the basis of molecular docking experiments. MD simulation was also carried out for these four compounds with respective protein, which revealed the stability of ligandprotein complex. Next, other hits compounds were also analyzed against each targeted protein and it was found that Diosmin was the one targeting all the proteins and hence considered for further studies. 

A Novel Coronavirus from Patients with Pneumonia in China

A new coronavirus associated with human respiratory disease in China

Coronavirus Disease COVID-19: A New Threat to Public Health

Corona virus SARS-CoV-2 disease COVID-19: Infection, prevention and clinical advances of the prospective chemical drug therapeutics: A review on Corona Virus Disease COVID-19, epidemiology, prevention, and anticipated therapeutic advances

Weekly operational update on COVID-19 -1

World Health Organisation: Coronavirus disease (COVID-2019) Weekly Epidemiological and Operational updates October

Discovery of New Hydroxyethylamine Analogs against 3CLpro Protein Target of SARS-CoV-2: Molecular Docking, Molecular Dynamics Simulation, and Structure-Activity Relationship Studies

Centers for Disease Control and Prevention: Vaccines & Immunizations; Vaccine Testing and the Approval Process

Therapeutics for COVID-19: from computation to practices-where we are, where we are heading to

Nitazoxanide/Camostat combination for COVID-19: An unexplored potential therapy

Discovery of Multitarget Antivirals Acting on Both the Dengue Virus NS5-NS3 Interaction and the Host Src/Fyn Kinases

Dual or multi-targeting inhibitors: The next generation anticancer agents

In silico identification of potent FDA approved drugs against Coronavirus COVID-19 main protease: A drug repurposing approach

A perspective on multitarget drug discovery and design for complex diseases

Network Analysis of Drug-target Interactions: A Study on FDA-approved New Molecular Entities Between

NEW LIBRARY OF 1520 APPROVED DRUGS

Repurposing of Clinically Developed Drugs for Treatment of Middle East Respiratory Syndrome Coronavirus Infection

Broad Anti-coronavirus Activity of Food and Drug Administration-Approved Drugs against SARS-CoV-2 In Vitro and SARS-CoV In Vivo

In vitro screening of a FDA approved chemical library reveals potential inhibitors of SARS-CoV-2 replication

Schrödinger Release 2020-1: Protein Preparation Wizard

Schrödinger Release 2020-1: Protein Preparation Wizard

Schrödinger Release 2020-1: LigPrep, Schrödinger, LLC

Schrödinger Release 2020-1: Epik, Schrödinger, LLC

Virtual screening, ADMET profiling, molecular docking and dynamics approaches to search for potent selective natural molecules based inhibitors against metallothionein-III to study Alzheimer's disease

In silico identification of inhibitors against Plasmodium falciparum histone deacetylase 1 (PfHDAC-1)

Schrödinger Release 2020-1: Glide, Schrödinger, LLC

Schrödinger Release 2020-1: Prime, Schrödinger, LLC

05 -Ligand-Based Approaches: Core Molecular Modeling

Methods for molecular dynamics simulations of protein folding/unfolding in solution

Schrödinger Release 2020-1: Desmond Molecular Dynamics System

The SARS-coronavirus papain-like protease: Structure, function and inhibition by designed antiviral compounds

A Review of the SARS-CoV-2 (COVID-19) Genome and Proteome

Two-amino acids change in the nsp4 of SARS coronavirus abolishes viral replication

The SARS-CoV-2 conserved macrodomain is a highly efficient ADP-ribosylhydrolase

Hesperetin 7-rutinoside (hesperidin) and taxifolin 3-arabinoside as germination and growth inhibitors in soils associated with the weed,Pluchea lanceolata (DC) C.B. Clarke (Asteraceae)

Hemorrhoids: diagnosis and treatment options

Computational studies of drug repurposing and synergism of lopinavir, oseltamivir and ritonavir binding with SARS-CoV-2 protease against COVID-19

In silico molecular docking analysis for repurposing therapeutics against multiple proteins from SARS-CoV-2

Structure-based drug repurposing for targeting Nsp9 replicase and spike proteins of severe acute respiratory syndrome coronavirus 2

The nsp9 replicase protein of SARS-coronavirus, structure and functional insights

Crystal Structure of the SARS-CoV-2 Non-structural Protein 9

Structural basis for dimerization and RNA binding of avian infectious bronchitis virus nsp9

The severe acute respiratory syndrome-coronavirus replicative protein nsp9 is a single-stranded RNA-binding subunit unique in the RNA virus world

Evaluation of protective effects of diosmin (a citrus flavonoid) in chemical-induced urolithiasis in experimental rats

Evaluation of Traditional Ayurvedic Preparation for Prevention and Management of the Novel Coronavirus (sars-cov-2) Using Molecular Docking Approach

One severe acute respiratory syndrome coronavirus protein complex integrates processive RNA polymerase and exonuclease activities

Hepatitis C virus (HCV) NS5A binds RNA-dependent RNA polymerase (RdRP) NS5B and modulates RNA-dependent RNA polymerase activity

Crystal structure of Zika virus NS5 RNA-dependent RNA polymerase

Ribavirin, Remdesivir, Sofosbuvir, Galidesivir, and Tenofovir against SARS-CoV-2 RNA dependent RNA polymerase (RdRp): A molecular docking study

Structure of the RNA-dependent RNA polymerase from COVID-19 virus

PubChem Compound Summary for CID 4689, Catenulin

Adenine C-Nucleoside (GS-5734) for the Treatment of Ebola and Emerging Viruses

Remdesivir is a direct-acting antiviral that inhibits RNA-dependent RNA polymerase from severe acute respiratory syndrome coronavirus 2 with high potency

Structure-based drug designing towards the identification of potential anti-viral for COVID-19 by targeting endoribonuclease NSP15

Crystal structure of Nsp15 endoribonuclease NendoU from SARS-CoV-2

An in-silico evaluation of different Saikosaponins for their potency against SARS-CoV-2 using NSP15 and fusion spike glycoprotein as targets

PubChem Compound Summary for CID 41774

The Chemopreventive Potential of Diosmin and Hesperidin for COVID-19 and Its Comorbid Diseases

Hesperidin and Diosmin for Treatment of COVID-19

Novel coronavirus SARS-CoV-2 continues to spread rapidly worldwide, causing serious health and economic loss. A methodology of targeting multiple enzymes of SARS-CoV-2 with a single potential drug could be an important in-silico strategy

A chemical library of 1520 compounds was selected for virtual screening against multiple proteins of SARS-CoV-2 using docking methodology

A multitargeting approach was optimized that enabled the analysis of several compounds' binding efficiency with more than one protein targets. Diosmin showed highest binding affinity amongst all the screened compounds

Molecular Dynamics (MD) simulation studies provide the validation of the docking experiments supporting multi-targeting ability of Diosmin with non-structural proteins of SARS-CoV-2

Poonam is thankful to Department of Science and Technology, Government of India for financial support (DST/TDT/AGRO-54/2019). BR acknowledges Science and EngineeringResearch Board for financial support under CRG scheme (CRG/2020/005800). SK and CU are highly grateful to CSIR for senior research fellowship, PPS acknowledges DBT, Govt. of India for junior research fellowship.