Exploring Phytochemicals of Traditional Medicinal Plants Exhibiting Inhibitory Activity Against Main Protease, Spike Glycoprotein, RNA-dependent RNA Polymerase and Non-Structural Proteins of SARS-CoV-2 Through Virtual Screening

Severe Acute Respiratory Syndrome Corona Virus 2 (SARS-CoV-2) being a causative agent for global pandemic disease nCOVID’19, has acquired much scientific attention for the development of effective vaccines and drugs. Several attempts have been made to explore repurposing existing drugs known for their anti-viral activities, and test the traditional herbal medicines known for their health benefiting and immune-boosting activity against SARS-CoV-2. In this study, efforts were made to examine the potential of 605 phytochemicals from 37 plant species (of which 14 plants were endemic to India) and 139 antiviral molecules (Pubchem and Drug bank) in inhibiting SARS-CoV-2 multiple protein targets through a virtual screening approach. Results of our experiments revealed that SARS-CoV-2 MPro shared significant disimilarities against SARS-CoV MPro and MERS-CoV MPro indicating the need for discovering novel drugs. This study has screened the phytochemical cyanin (Zingiber officinale) which may exhibit broad-spectrum inhibitory activity against main proteases of SARS-CoV-2, SARS-CoV and MERS-CoV with binding energies of (−) 8.3 kcal/mol (−) 8.2 kcal/mol and (−) 7.7 kcal/mol respectively. Amentoflavone, agathisflavone, catechin-7-o-gallate and chlorogenin were shown to exhibit multi-target inhibitory activity. Further, Mangifera indica, Anacardium occidentale, Vitex negundo, Solanum nigrum, Pedalium murex, Terminalia chebula, Azadirachta indica, Cissus quadrangularis, Clerodendrum serratum and Ocimum basilicumaree reported as potential sources of phytochemicals for combating nCOVID’19. More interestingly, this study has highlighted the anti-viral properties of the traditional herbal formulation “Kabasura kudineer” recommended by AYUSH, a unit of Government of India. Short listed phytochemicals could be used as leads for future drug design and development. Genomic analysis of identified herbal plants will help in unraveling molecular complexity of therapeutic and anti-viral properties which proffer lot of chance in the pharmaceutical field for researchers to scout new drugs in drug discovery.


INTRODUCTION
Novel Coronavirus disease (nCOVID'19) caused by SARS-CoV-2 virus has become a global threat and WHO has declared it as a pandemic. nCOVID'19 is the third life threatening virus in the SARS family of viruses after SARS-CoV occurred during 2002-03 and MERS-CoV during 2012 (Lee et al., 2003;Cheng et al., 2007;Zaki et al., 2012;De Groot et al., 2013). It is named as a novel Coronavirus as it shares significant dissimilarity against other members of the SARS family of viruses viz., SARS-CoV (30%) and MERS-CoV (60%) . Its unique genetic makeup has made it not responsive to available medical treatments and necessitated search for novel targets for vaccine development and drugs for effective prevention and treatment of nCOVID'19.
Exploding increase in the nCOVID'19 affected cases has brought this globe to a halt. Scientific community is trying to unravel genome complexity of nCOVID'19 for identifying novel targets for development of vaccines, screen available anti-viral drugs for effective management and short listing effective botanicals for therapeutic interventions. This has resulted in the accumulation enormous genomic information of nCOVID'19 in the public domain (https://www.ncbi.nlm.nih. gov/genbank/sars-cov-2-seqs/). Genomic analysis of nCOVID'19 revealed that it is approximately 30 kb in size (NCBI Accession # NC_045512) and further investigations identified three key genes viz., 1) coronavirus main protease (3CL pro )/papain-like protease (PL pro ); 2) RNA-dependent RNA polymerase (RdRp) and 3) spike glycoprotein (S protein) as potential targets for drug designing (Elfiky, 2020;Sampangi-Ramaiah et al., 2020).
Screening of existing antiviral drugs including interferon α (IFN-α), lopinavir/ritonavir, chloroquine phosphate, ribavirin, chloroquine, hydroxychloroquine and arbidol is in progress and many of these experiments require pre-clinical and clinical validation (Elfiky, 2020). Ineffectiveness of existing anti-viral drugs have made the doctors to resort using traditional medicines in nCOVID'19 treatments (Yang et al., 2014). Several attempts have been made to exploit the potential of several herbal products having potential to inhibit the main protease (Mpro)/chymotrypsin-like protease (3CLpro) using molecular modeling and docking studies (Colson et al., 2020;Dong et al., 2020;Touret and de Lamballerie, 2020). Elfiky (2020) made an attempt to screen 27 different ligands present in commonly used herbals of Indian cuisines against SARS-CoV-2 Main protease and identified 15 different ligands effective in binding the viral protease. Yang et al. (2014) made a systematic review of herbal drugs used in the effective treatment of SARS-CoV and MERS-CoV and emphasized the urgent need for evolving procedures involving complementary and alternative treatments in managing nCOVID'19. Studies conducted so far have made attempts by using limited number of ligands which may hinder discovery of effective viral inhibitor in the herbal gene pool. In this context, short listing potential herbal drugs effective against nCOVID'19 through in silico docking of globally available ligands and validating them through laboratory and clinical trials is one of the viable approaches in managing this pandemic. India is one of the richest biodiversity centers in the world and known for its vast repository of medicinal plants. Considering India's richest biodiversity of herbal medicinal plants and regular use of such medicinal plants in Indian health care system, the present study was undertaken to screen about 744 ligands including small molecules and phytochemicals from 37 different Indian medicinal plants against seven different protein targets of nCOVID'19 through molecular docking. Protein-Ligand interactions were analyzed carefully to shortlist potential small molecules and phytochemicals for drug development.

Phylogenetic Analysis of Main Protease of nCOVID'19
Protein sequence of SARS-CoV-2 encoding for main protease (PDB ID: 5R81) was used for PSI BLAST (NCBI) (Altschul et al., 1997) search to identify its homologs for understanding the evolutionary relationship with main proteases of other viruses. Multiple sequence alignment and phylogenetic analysis of SARS-CoV-2 main protease with other viral proteins was done using MAFFT server (Katoh et al., 2019). different target proteins of the nCOVID'19 genome ( Table 1). In addition, virtual screening was also performed against M Pro of SARS CoV (PDB ID: 2GZ9) and MERS CoV (PDB ID: 5C3N) for identifying inhibitors against main protease of all three viruses and inhibitors specific to nCOVID'19.

Ligand Library Preparation
Chemical structures of all the small molecules were retrieved from Dukes database , PubChem (Kim et al., 2019) and DrugBank (Wishart et al., 2018). From the DrugBank database, 99 chemical structures of drugs approved for the treatment of respiratory diseases and compounds exhibiting antiviral activity were collected. Forty chemical compounds with COVID 19 antiviral property retrieved from pubchem database were also included in the ligand dataset. Information regarding the origin, traditional use and protocol were obtained from literature, Indian medicinal plants database (http://www. medicinalplants.in/) and Indian Medicinal Plants, Phytochemistry and Therapeutics (https://cb.imsc.res.in/ imppat/home) database. Structures of 605 phytochemicals belonging to 37 different herbals and spices used in South Indian Traditional Medicine were also used for virtual screening (Supplementary Table S1). Among the 37 herbals, 14 herbals were found native to India ( Known active ingredients of eight herbal plants included in the Tamil traditional medicine "Kabasura Kudineer" (meaning water capable of boosting immunity) were also included in the screening. Overall, a total of 744 small molecules/ligands (Supplementary Table S1: a-c) were used for virtual screening against seven different protein targets. Interactions of phytochemicals were compared with drugs such as hydroxycholoroquine (Pubchem ID: 3652), chloroquine (Pubchem ID: 2719) and ivermectin (Pubchem ID: 6427057) retrieved from pubchem database.

Virtual Screening
Virtual screening was performed using Python Prescription Virtual Screening tool (PyRx 0.8) containing AutoDock Vina module (Dallakyan and Olson, 2015). Protein structure was prepared using SWISS PDB Viewer by adding hydrogen atoms and energy minimization. Prepared protein structure was fed into the PyRx tool along with the structure of 744 ligands. Both the ligands and protein molecules were converted to pdbqt file using the AutoDock module of PyRx tool. Inhibitors are expected to bind in the active site/binding site of the protein to inhibit the function of the protein target. In the present study, binding sites were predicted using CASTP server (Tian et al., 2018) and the predicted sites were used for setting grid (XYZ dimensions: 25*25*25) in the AutoDock Vina for virtual screening experiment with the exhaustiveness value of 8. Predicted binding sites were also verified with protein-ligand binding sites of experimental structures for accuracy. Furthermore, phylogenetic analysis of SARS-CoV-2 M Pro was carried out using PSI-BLAST (NCBI) (Altschul et al., 1997) and MAFFT server (Katoh et al., 2019). Top 10 ligand hits against each of the seven protein targets were taken for further analysis. 2D and 3D interactions between the protein-ligand were analyzed using Schrodinger Maestro visualizer and BIOVIA discovery studio visualizer 2020 (Accelrys Inc. San Diego, CA, United States) software. ADME properties of top 10 ligands screened against individual protein targets were predicted using SWISSADME server (http://www.swissadme.ch) (Supplementary Table S2).

Phylogenetic Analysis on Coronavirus Main Proteases
Main protease (M pro , also called 3CL pro ) is considered as one of the important molecular targets for designing novel drugs against coronaviruses (Anand et al., 2003). With a view to design drugs/ inhibitors specifically targeting main protease of nCoVID'19, insilico analysis was performed using main protease sequences of SARS-CoV-2, SARS-CoV and MERS-CoV. Multiple sequence   Fruit, seed, seed essential oil Antihypertensive and antiplatelets, antioxidant, antitumor, antiasthmatics, antipyretic, analgesic, antiinflammatory, anti-diarrheal, antispasmodic, anxiolytic, antidepressants, hepato-protective, immuno-modulatory, antibacterial, antifungal, anti-thyroids, antiapoptotic, anti-metastatic, antimutagenic, antispermatogenic, anticolon toxin, insecticidal and larvicidal activities Damanhouri and Ahmad, (2014) alignment identified 12 significant differences between main proteases of SARS-CoV and SARS-CoV-2 ( Figure 1). Out of the 12 differences, S45 to A45 was found to reside within in the binding site of SARS-CoV-2 main protease. This may play a crucial role in determining differential binding affinity of the two proteases. Phylogenetic analysis of SARS-CoV2M Pro with other CoV M Pro sequences sharing >50 percentage similarity revealed its significant genetic relatedness with SARS CoV (96.08% similarity) and bat CoV (76.84% similarity) ( Figure 2). Also it shared significant similarity with ORF1ab of Rousettus bat coronavirus. Main protease of nCoVID'19 shared only 50.65% similarity against main proteases of MERS-CoV. Above results clearly indicated the need for a highly specific novel drug specifically inhibiting main proteases of SARS-CoV-2.  Virtual screening of 744 ligands belonging to small molecules and active compounds from 37 medicinal herbs against seven major protein targets of nCOVID'19 predicted probable SARS-CoV-2 inhibitors. Information regarding the binding site residues predicted using CASTp server is provided in Table  3.
Binding affinity of drugs such as hydroxychloroquine, chloroquine and ivermectin are provided in Table 4 which was used as reference for comparison of phytochemicals. Top 10 hits reported with higher binding affinity for each target protein is considered for downstream analysis ( Table 5)
The phytochemical cyanin found in Zingiber officinale was found to have higher binding affinity with the main proteases of all three coronaviruses ( Figures 3A-C) and placed in the top 10 hits list. It showed a binding affinity value of −8.3 kcal/mol, −8.2 kcal/mol and −7.7 kcal/mol against SARS-CoV-2(PDB ID: 5R81), SARS CoV (PDB ID: 2GZ9) and MERS CoV (PDB ID: 5C3N) main proteases respectively. Hydrogen bond interactions for SARS-CoV-2 (THR 26, SER 46 and GLU 166), SARS CoV (THR 26, LEU 141 and CYS 145) and MERS CoV (THR 26, CYS 149 and GLU 169) were observed in the binding sites respectively.

Effect of Suggested FDA Drugs on SARS-CoV-2 Protein Targets
Hydroxycholoroquine, chloroquine and ivermectin drugs were selected as positive controls to compare the binding interaction of phytochemicals for the assessment of anti-viral activity (Caly et al., 2020;Principi and Esposito, 2020). Hydroxychloroquine was reported to show promising inhibitory activity against nCOVID'19 spike protein (Gautret et al., 2020;Liu et al., 2020). Our results revealed that hydroxycholorquine and chloroquine showed less binding affinity against all the seven targets of nCOVID'19  Table  S4). Analysis of hydrogen bond interaction with ivermectin, showed three hydrogen bonds for RdRp target (ARG 403, TYR 453 and TYR 489) and two hydrogen bonds for spike protein (ASN 497 and SER 814), each one hydrogen bond for NSP3 (TYR 161) and NSP9 (ARG 40) respectively. It is also important to note that the ivermectin was reported as best hits for the above mentioned targets compared to hydroxychloroquine and chloroquine against SARS-CoV-2 targets. In the following section, binding of ivermectin was compared with amentoflavone and agasthisflavone phytochemicals as they were predicted to inhibit multi-target SARS-CoV-2 proteins based on the docking score.
Frontiers in Pharmacology | www.frontiersin.org July 2021 | Volume 12 | Article 667704 21 ( Figure 6A). 1,8-Dichloro-9,10-diphenylanthracene-9,10-diol from Carica papaya was found to exhibit significant binding affinity against spike glycoprotein (−8.2 kcal/mol). GLY 496 residue was found to be involved in the formation of hydrogen bond with the 1, 8-Dichloro-9, 10diphenylanthracene-9,10-diol. Earlier, leaf extracts of Carica papaya was reported to have significant effect in combating dengue virus infection (Rajapakse et al., 2019) and its exact FIGURE 8 | Network diagram showing the interaction of plant phytochemcials with SARS-CoV-2 protein targets. Degree of connectivity represents number of SARS-CoV-2 that may be inhibited by each plant. In this network, NSP 15 stands first in the order where many plants connected (degree of connectivity) which indicates that metabolites from the connected plants has shown highest binding affinity (top ten screened compounds). Likewise, Vitex negundo was predicted to inhibit highest number of SARS-CoV-2 targets in the virtual screening. Plants name with * indicates their presence in Kabasura kudineer.
Residues are colored based on their physiochemical properties.
Another small molecule, friedelin from Vitex negundo and Acorus calamus was also found to exhibit significant binding affinity of −9.6 kcal/mol against NSP9 (PDB ID: 6W4B) ( Table 5). Even though, many hydrophobic interactions were observed, no hydrogen bond interaction was found in the binding site of NSP9. It is very interesting to observe that five out of the top ten inhibitors are from a single plant source Solanum nigrum. As evidenced from other studies, Solanum nigrum is one of the traditionally known medicinal plants known for its use in treatment ofseizure, pain, ulcer, inflammation, diarrhea, eye infections, jaundice and oxidative stresses (Jain et al., 2011;Javed et al., 2011;Wang et al., 2013;Zaidi et al., 2014).

Molecules Exhibiting Inhibitory Activity Against Multiple Protein Targets of nCOVID'19
Phytochemicals showing strong interactions against multiple targets of viruses are expected to confer durable protection to the patients. This will be more beneficial in situations where the virus is developing mutations in one of the targets. Small molecules namely, amentoflavone, agathisflavone, catechin-ogallate and chlorogenin exhibited significant binding affinity toward multiple targets of nCOVID'19.
Amentoflavone showed docking score of RdRp (−9.3 kcal/ mol) ( Figure 4C), NSP9 (−8.3 kcal/mol) ( Figure 5F), NSP3 (−7.4 kcal/mol) ( Figure 5C), NSP10-NSP16 (−8.5 kcal/mol) and spike glycoprotein (−8.2 kcal/mol). In all the docked complexes, the target -ligand binding affinity was greater than (−8.0 kcal/mol) except NSP3. 3D and 2D Protein-ligand interactions exhibited by the small molecules amentoflavone and agathisflavone are shown in Figure 4, Figure 5 and Supplementary Table S3. Molecular interactions such as hydrogen bond, vander waals interaction, pi interaction were observed to be higher for amentoflavone incomparison with the drug ivermectin. Amentoflavone is a naturally occurring biflavonoid reported to be found in more than 120 plants (Yu et al., 2017). Many of these plants have been used in traditional medicine for several thousand years in different parts of the world. Several studies have reported that amentoflavone possess anti-inflammatory, anti-oxidative, anti-diabetic, anti-tumor, anti-viral and anti-fungal activities (Yu et al., 2017). Evidences have been reported for amentoflavone exhibiting anti-senescence activity in the cardiovascular and central nervous system (Park et al., 2011). Further, Amentoflavone isolated from Torreya nucifera was demonstrated to possess inhibitory activity against SARS-CoV3CL Pro (Ryu et al., 2010).

DISCUSSION
Among the 605 phytochemicals originating from 37 plant species, 33 (6% approx.) phytochemicals from 22 plants ( Figure 8) were found to be the best hits with higher binding affinities against all the seven targets ( Table 5). Among those 22 plants, four plants were found to be the ingredients of a traditional siddha herbal formulation namely "Kabasura kudineer" recommended by AYUSH Board of Government of India for boosting immunity. Vitex negundo was reported to possess 32 different phytochemicals that were included in this study. Out of the 33, five different compounds namely, luteolin 7-O-beta-D-glucoside, luteolin 7-O-(6'-malonylglucoside), agnuside, luteolin-7-o-betad-glucopyranoside and friedelin were found to exhibit significant binding affinity against five different protein targets of SARS-CoV-2 namely, spike glycoprotein, SARS-CoV-2 main protease, NSP3, NSP9, NSP15 in the SARS-CoV-2.
Pedalium murex was reported to have Diosgenin (NSP9), Lupeol acetate (NSP16-NSP10), Urosolic acid (NSP15) and Rubusic acid (SARS-CoV-2 M pro , NSP3) which might be probable SARS-CoV-2inhibitors. It is noteworthy that maximum of five targets were predicted to be inhibited by the compounds from Pedalium murex. In spite of its role as anti-ulcerogenic, nephroprotective, hypolipidemic, aphrodisiac, anti-oxidant, anti-microbial and insecticidal activities, Pedalium murex has been traditionally used in treating ailments like cough and cold as a regular practice (Patel et al., 2011a).
Medicinal plants namely, Azadirachta indica, Terminalia chebula, Cissus quadrangularis, Clerodendrum serratum and Ocimum basilicum reported more than two phytochemicals that might possibly inhibit SARS-CoV-2 targets. These herbal plants might be the potential targets for future research toward developing herbal formulations against SARS-CoV-2. Intensive genomics and proteomics research may lead to identification of novel drugs against this pandemic disease.

CONCLUSION
Generation of improved knowledge and understanding biochemical and molecular basis of herbals used in traditional Ayurveda and siddha medicine will accelerate development of effective drugs in controlling emerging diseases. In this study, comparative analysis of main proteases of MERS-CoV (PDB ID: 5C3N), SARS-CoV (PDB ID: 2GCZ9) and SARS-CoV-2 (PDB ID: 5R81) revealed significant differences between the three homologs which were confirmed by differential binding affinity exhibited by 744 phytochemicals/small molecules against the three main proteases. Cyanin from Zingiber officinale was screened as the best phytochemical with highest binding energy against main proteases of all the three viruses. Popular fruit tree, mango (Mangifera indica) and, cashew nut (Anacardium occidentale) rich in amentoflavone and agathisflavone were showing possible inhibitory activity against multiple targets of SARS-CoV-2. Vitex negundo, Solanum nigrum, Pedalium Murex, Terminalia chebula, Azadirachta indica, Cissus quadrangularis, Clerodendrum serratum and Ocimum basilicum were also found to contain phytochemicals that may have possible inhibitory activity against SARS-CoV-2 proteins. More interestingly, this study has picked up Carica Papaya that may possess inhibitory activity against spike glycoprotein and M Pro of SARS-CoV-2 which was known for its protective role against dengue virus in humans (Rajapakse et al., 2019). Overall, this study has shortlisted potential phytochemicals that may have inhibitory activity against SARS-CoV-2 which could be taken for further testing, formulation and discovery of novel drugs.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.

AUTHOR CONTRIBUTIONS
SN, JM, CR, KA, and BN performed experimental, data analysis, and drafting of original manuscript. KN, GR, RM, MS, and KN conceptualized the experiment, refined the manuscript, funded the project, and helped in revising the manuscript. All authors read and approved the final version of the manuscript.