In-silico analysis of the inhibition of the SARS-CoV-2 main protease by some active compounds from selected African plants.

OBJECTIVES: Over the years, Azadirachta indica, Mangifera indica, and Moringa oleifera have been shown to possess some antiviral characteristics. This study applies molecular docking techniques to assess inhibitory effects of some bioactive compounds from the plants mentioned above against the main protease (Mpro), a key protein involved in SARS-CoV-2 replication. Furthermore, adsorption, distribution, metabolism, excretion, and toxicity (ADMET) profiles for screened compounds were predicted in silico.
METHODS: The crystal structure of Mpro was retrieved from the Protein Data Bank, while the plant bioactive compounds were retrieved from Pubchem. Drug-likeness of the selected compounds and a control drug (hydroxychloroquine) were assessed, and the compounds that satisfied the drug-likeness rule were docked against Mpro. The docked complexes were analyzed using LigPlot and the protein-ligand profiler server. The top five compound hits were subjected to ADMET screening using the ADMETSar server.
RESULTS: A total of 17 out of 22 screened compounds passed Lipinski's assessment. Additionally, the most active compounds from the investigated plants exhibited relative inhibitory potentials against Mpro compared with hydroxychloroquine, which alludes to their possible involvement in inhibiting the SARS-CoV-2 main protease replication process.
CONCLUSIONS: In our study, most of the active phytocomponents of the investigated plants exhibited relative inhibitory potentials against Mpro of SARS-CoV-2 and preferred pharmacological features when compared with hydroxychloroquine. These findings indicate these compounds are potentially antiviral candidates against SARS-CoV-2.

Introduction

Since the beginning of the 21st century, novel Coronaviruses (CoVs) have caused major outbreaks of fatal pneumonia globally. Severe acute respiratory syndrome coronavirus (SARS-CoV) emerged and spread to five continents in 2003, while the Middle East respiratory syndrome coronavirus (MERS-CoV) broke out in the Arabian Peninsula in 2012. These outbreaks produced mortality rates of 10% and 35%, respectively.1, 2, 3, 4, 5 However, both SARS-CoV and MERS-CoV are zoonotic viruses, and their hosts are bats or civets and dromedary camels, respectively. To date, no specific therapeutic drug or vaccine has been approved for the treatment of human coronavirus; hence, human health is at significant risk due to the outbreaks of coronaviruses.

In December 2019, several pneumonia cases emerged due to outbreak of an unprecedented coronavirus later called Severe Acute Respiratory Syndrome Coronavirus 2 (SARS-CoV-2) in Wuhan, China.7, 8, 9, 10 The whole genome of this novel coronavirus was sequenced, and it revealed that it is 96.2% identical to coronavirus in bats and possess 79.5% and <50% sequence similarities with SARS-CoV and MERS-CoV, respectively,, The novel coronavirus was termed SARS-CoV-2 by the International Committee on Taxonomy of Viruses, and the associated pneumonia was designated as COVID-19 by World Health Organization (WHO) on February 11, 2020. The disease not only spread rapidly to most provinces of China but also to over 200 countries and territories across all continents and was announced as a global health emergency by the WHO on March 11, 2020.

The COVID-19 viral genome comprises approximately 30,000 nucleotides, and its replicase gene encodes two overlapping polyproteins, pp1a and pp1ab, essential for viral replication and transcription.,, These functional polypeptides are then released from the polyproteins by extensive proteolytic cleavage, with the aid of a 33.8-kDa main protease (Mpro), alternatively called 3C-like protease and papain-like proteases (PLPs). Biochemically, Mpro cleaves the polyprotein at 11 conserved sites to release non-structural proteins 4 to 16 (nsp4 - 16), starting with its self-cleavage from pp1a and pp1ab.,, Interestingly, recent studies have highlighted the advantage linked to the identification of Mpro from SARS-CoV-2 as a target attractive for antiviral drug design due to the absence of related homologs of Mpro in humans.,

Regrettably, there are no specific therapies approved by recognized international health agencies, such as the WHO and US Food and Drug Administration (FDA) for SARS-CoV-2.17, 18, 19 Presently, several agents, such as chloroquine, hydroxychloroquine, remdesivir, lopinavir/ritonavir, azithromycin, and tocilizumab are undergoing clinical trials under compassionate use protocols based on promising inhibitory effects exhibited against SARS-CoV-2 and related viral diseasesin vitro is now a milestone for drug repurposing against this viral malady. Despite their success in clinical reports, none have gained approval for use against SAR-CoV-2 because their curative outcomes remain very limited and their toxic side-effects cannot be ignored. Conversely, a recent study from Xu et al. and Ashfaq et al. revealed the relevance of molecular docking techniques to find newer and potential inhibitors of Mpro. As such, alternative novel evidence based on the molecular docking approach could become useful for the advancement of novel drugs for the management and treatment of SARS-CoV-2.

Similarly, bioactive plant products play a key role in deterring disease and promoting treatment by enhancing antioxidant activity, inhibiting microbial growth, and modulating genetic pathways.26, 27, 28 Also, the global interest in the health-promoting benefits offered by phytoconstituents has increased recently compared with other synthetic drugs is due to their cheap, ease of access coupled with their few side effects. Numerous pharmacologically active medications with minimal or no side effects are derived from natural sources, such as medicinal plants and microbes. Furthermore, bioactive compounds from some plants possess antiviral activities., Therefore, natural products possessing antiviral activity can be used as templates in the search for potential bioactive compounds against SARS-CoV-2. In this regard, molecular docking can predict the underlying mechanism, emphasizing the receptor protein interactions with bioactive compounds.

In recent years, Azadirachta indica (neem), an evergreen tree belonging to the family Meliaceae, has been regarded as an essential medicinal and fast-growing perennial tree found in Africa, America, and India. It has gained global significance due to its wide-ranging therapeutic activities.,, It has considerable antinociceptive, emollient, antiviral, anti-inflammatory, antiseptic, antifungal, astringent, insecticidal, anti-helminthic, and antibacterial properties.,, Over 140 active constituents with established therapeutic usefulness have been identified in various parts of this plant. There are many constituents in A. indica, such as limonoids, nimbin, nimbidin, and nimbolide that can aid in disease management by modulating numerous pathways of genetic relevance and other activities. However, quercetin and β-sitosterol were the first polyphenolic flavonoids purified from fresh leaves of neem and were known to have antifungal and antibacterial activities.,,

Mangifera indica of the Anacardiaceae family is one of the main tropical trees originating from Asia, Brazil, China, India, Mexico, Nigeria, Pakistan, the Philippines, and Thailand, with India being the premier mango cultivating country. Despite the common use of mango fruit as food, various parts of mango trees have also been used in folkloric medicine in most South-East Asian and African countries., Barretto et al. characterized and quantified a wide-range of polyphenolic compounds in M. indica. Mangiferin, gallic acid, catechins, quercetin, kaempferol, protocatechuic acid, ellagic acids, propyl and methyl gallate, rhamnetin, and anthocyanins are the major polyphenolic compounds found in M. indica. According to Al Rawi et al., extracts from M. indica show antiviral activity against influenza virus H9N2 grown on cultured cells, with minimal cytotoxicity. Additionally, the bark, flowers, fruits, and leaves of M. indica have also been reported to possess antiviral properties.,,

Another African plant, Moringa oleifera of Moringaceae, originates from northern India and can be cultivated in tropical and subtropical regions.40, 41, 42 The diverse parts of the plant have traditional use in managing various ailments, such as arthritis, diabetes, fever, asthma, epilepsy, wounds, body pains and weakness, cough, blood pressure, skin infections, chronic anemia, cancer, viral infections, malaria, haemorrhage, tuberculosis, and other respiratory diseases., Earlier reports revealed its inhibitory activity against a few infectious disease-causing viruses: human immunodeficiency, herpes simplex, hepatitis B, and Epstein–Barr.,, M. oleifera possess a very high antioxidant activity., In addition, the leaves are very rich in vitamins, carotenoids, polyphenols, phenolic acids, flavonoids, alkaloids, glucosinolates, isothiocyanates, tannins, saponins, oxalates, and phytates.,,

Having considered the potential of these plant's bioactive compounds in the management and treatment of many diseases, with emphasis on their limitations, it is therefore essential to explore an evidenced-based molecular docking approach to assess the inhibitory potential of some of the active compounds from A. indica, M. indica and M. oleifera leaves against Mpro and also screen the top hit compounds for adsorption, distribution, metabolism, excretion, and toxicity (ADMET) in silico.

Materials and Methods

Protein target selection, preparation, and active site prediction

The three-dimensional (3D) X-ray crystal structure of Mpro found in SARS-CoV-2 (PDB ID: 6lu7) was obtained from the Protein Data Bank (PDB) (www.pdb.org/pdb). Using UCSF-Chimera© (version 1.13) software tools (http://www.cgl.ucsf.edu/chimera), the protein target was prepared for docking by removing the co-crystallized ligand and additional water molecules to make it a nascent receptor, after which hydrogen charges were added to the protein. The DogSite module of the protein-plus web server (http://proteinsplus.zbh.uni-hamburg.de), which is based on the druggability score of identified pockets, was applied to predict the binding site of Mpro.

Ligand preparations

Three-dimensional (3D) structures of the selected molecules were obtained in a simple data file from the PubChem server (https://www.ncbi.nlm.nih.gov/pccompound) and were converted to mole files using MarvinSketch© (ver. 15.11.30). The molecules were optimized for docking using the Merck molecular force field (MMFF94) in Avogadro (ver. 1.10). A total of twenty-two compounds from A. indica, M. indica, and M. oleifera was selected for the study, while hydroxychloroquine was adopted as a control drug.

Drug likeness screening

The selected compounds were screened for drug-likeness as described by Lipinski et al.. The compounds were analysed using DruLiTo software to predict their logP, molecular weight, and hydrogen bond donor and acceptor values. Lipinski's rule of five was applied to screen for the probable molecules that can serve as oral drugs.

Validation of the in-silico docking protocol

The docking protocol was validated to substantiate the accuracy and reliability of the docking results, according to Gregory et al. The goal was to accurately reproduce the binding pore and the molecular interactions of the co-crystallised ligand of the experimentally-crystallized protein structure. Accordingly, the native ligand of the x-ray protein was separated from the protein and then prepared for docking in UCSF-Chimera. The ligand was then docked back into Mpro's active site using Auto Dock Vina in PyRx. The docked complex was superimposed with the X-ray resolved crystal Mpro bearing the co-crystallized ligand to generate the root mean square deviation (RMSD) value in PyMOL. Notably, the RMSD value ranging from (0–2) Å is appropriate for docking and is indicative that the protocol could be subsequently used for other small molecule inhibitors.

In silico molecular docking

The docking was executed using a flexible docking protocol as described by Trott and Olson with slight modifications. In brief, Python Prescription 0.8, a software suite of Auto Dock Vina, was utilized for the molecular docking analysis of the selected ligands with Mpro. The protein data bank, partial charge, and atom type (PDBQT) files of the protein were generated (using their previously created PDB files as inputs). The enzyme's target site was set with the help of a grid box with dimensions 27.601 Å × 32.911 Å x 32.182 Å, and the centres were adjusted based on the active site of the enzyme. Once the molecular docking was completed, 10 configurations for each protein-ligand complex were generated for all the ligands. Text files of the scoring results were also created for manual comparative analysis. All through this in silico investigation, an exhaustiveness of 10 was used for docking to achieve results with high accuracy and reliability. The protein-ligand complexes as well the molecular interaction were all visualized in 3D and 2D using PyMOL© Molecular Graphics (version 1.3, 2010, Shrodinger LLC), LigPlot+© Roman Lakoskwi, European Bioinformatics Institute, Hinxton, Cambridge, United Kingdom (version 2.1.), and the Protein-ligand interaction profiler's (PLIP) webserver (https://projects.biotec.tu-dresden.de/plip-web/plip) to check for additional interactions between the ligands and the Mpro.

Prediction of ADMET properties

The ADMET properties of the compounds with the highest hits from the molecular docking studies were executed using the ADMETSar webserver (http://lmmd.ecust.edu.cn/)., This is a web platform in which the hit compounds were accessed for their aqueous solubility, pharmacodynamics, and pharmacokinetic properties using the various models.

Results

In this in silico study, 22 active compounds from the study plants and the control drug (Figure 1) were assessed for their inhibitory properties against Mpro (Figure 2) of SARS-CoV-2, the key protein implicated in the pathogenesis of COVID-19. Furthermore, the active site of Mpro was predicted using the DogSite scorer on the protein-plus web server, which is a fully automatic algorithm for predicting pockets and their druggability. The higher the drug score of a predicted pocket, the more suitable it is for adoption as the binding site of a target protein. At the end of the prediction, 37 amino acid residues were identified in the pocket with the highest drug score and simple score. The amino acid residues are; Gln19, Thr25, Thr26, Leu27, Asn28, Arg40, His41, Cys44, Asp48, Met49, Pro52, Tyr54, Cys117, Tyr118, Asn119, Gly120, Phe140, Leu141, Asn142, Gly143, Ser144, Cys145, His163, His164, Met165, Glu166, Leu167, Pro168, His172, Phe181, Val186, Asp187, Arg188, Gln189, Thr190, Ala191 and Gln192. The platform predicted 24 out of the established 25 amino acids from previous studies.,

Figure 1

The structures of active compounds from Azadirachta indica, Mangifera indica, and Moringa oleifera, and the control drug (hydroxychloroquine).

Figure 2

The structure of the ribbon pattern (a) and surface presentation, (b) of the main protease from SARS-CoV-2.

A total of 17 out of the 22 active compounds initially selected for this study passed Lipinski's rule of five. In contrast, the other five (azadirachtin, isoquercitrin, quercitrin, rutin, and salanin) that violated at least two of the rules were eliminated from further investigations (Table 1). Before docking, the protocol for docking was validated as elucidated in the methodology. This process was carried out to provide assurances of the deployed protocols and docking tools to accurately provide correct binding signatures with quality interactions between the target protein and the bioactive compounds investigated in this study. This experiment was considered successful since the docked complex precisely reproduced the original pose as the native ligand (N-leucinamide) (Figure 3) with an RMSD value of 1.049 Å. The results of molecular docking between the assessed active compounds and Mpro is presented in Table 2. The docking outputs revealed that catechin, chlorogenic acid, ellagic acid, epicatechin, gedunin, kaempferol, lupeol, mangiferin, myricetin, nimbandiol, nimbinene, nimbolide, and quercetin had better binding positions with Mpro than with hydroxychloroquine, as adjudged by their relatively lower binding energy values (also see Figure 4). The control drug, hydroxychloroquine, has a binding energy of −6.4 kcal/mol while forming hydrogen bonds with Ser144, Cys145, Gln189 and Thr190 in the active site of Mpro, and it established a hydrophobic interaction with Phe140, Leu141, Asn142, Gly143, Ser144, Met165, Glu166, Arg188, and Gln189 in the active site of Mpro. Based on the docking outputs, mangiferin had the best binding position, with a binding energy of −8.4 kcal/mol. It interacts with the target enzyme to yield hydrogen bonds with nine amino acid residues in the active site of Mpro: His41, Leu141, Asn142, Gly143, Ser144, Cys145, Arg188, Thr190, and Gln192. Furthermore, it establishes hydrophobic interactions with His41, Met49, Leu141, Asn142, Gly143, Ser144, Cys145, Met165, Glu166, Arg188, Gln189, Thr190, and Gln192 in the Mpro active site. Conversely, gallic acid had the lowest binding affinity compared with other screened compounds and our control drug, as predicted from its relatively higher binding energy (−5.5 kcal/mol) after it was docked against Mpro. Gallic acid formed hydrogen bond interactions with Leu141, Gly143, Ser144, Cys145, His163 and Glu166 in the active site of Mpro and hydrophobically interacted with Mpro by binding with Phe140, Leu141, Asn142, Gly143, Ser144, His163, Met165, and Glu166 amino acids in the active cavity of the target protein. The relatively higher binding energy exhibited by gallic acid is equivalent to the lowest binding affinity compared with other compounds and the control drug. Thus, the available data from the pool of investigated active compounds suggest that mangiferin and gallic acid may possess the most and least inhibitory effect against SARS-CoV-2 Mpro, respectively.

Table 1

Lipinski's drug-likeness screening of the control drug and the active compounds from Azadirachta indica, Mangifera indica, and Moringa oleifera leaves.

S/N	Plant	Phytochemicals	Chemical formulae	Molecular weight	LogP	Number of HB acceptor	Number of HB donor	No of violation
1	Azadirachta indica	Azadirachtin	C35H44O16	720.26	−0.683	16	3	2
		Gedunin	C28H34O7	482.23	3.557	7	0	1
		Nimbandiol	C26H32O7	456.21	−0.311	7	2	0
		Nimbin	C30H36O9	540.24	1.276	9	0	1
		Nimbinene	C28H34O7	482.23	0.928	7	0	0
		Nimbolide	C27H30O7	466.2	1.035	7	0	0
		Nimbolinin	C35H46O10	626.31	3.137	10	1	1
		Salanin	C34H44O9	596.3	3.089	9	0	1
		Quercetin	C15H10O7	302.04	1.834	7	5	0
2	Mangifera indica	Catechin	C15H14O6	290.08	0.852	6	5	0
		Ellagic acid	C14H6O8	302.01	1.366	8	4	0
		Epicatechin	C15H14O6	290.08	0.852	6	5	0
		Gallic acid	C7H6O5	170.02	0.964	5	4	0
		Kaempferol	C15H10O6	286.05	1.486	6	4	0
		Lupeol	C30H50O	426.39	11.901	1	1	1
		Mangiferin	C19H18O11	422.08	−0.631	11	8	2
		Quercetin	C15H10O7	302.04	1.834	7	5	0
		Rutin	C27H30O16	610.15	−0.735	16	10	3
3	Moringa oleifera	Caffeic acid	C9H8O4	180.04	0.888	4	3	0
		Catechin	C15H14O6	290.08	0.852	6	5	0
		Chlorogenic acid	C16H18O9	354.1	−0.7	9	6	1
		Ellagic acid	C14H6O8	302.01	1.366	8	4	0
		Epicatechin	C15H14O6	290.08	0.852	6	5	0
		Gallic acid	C7H6O5	170.02	0.964	5	4	0
		Isoquercitrin	C21H20O12	464.1	0.099	12	8	2
		Kaempferol	C15H10O6	286.05	1.486	6	4	0
		Myrecitin	C15H10O8	318.04	2.182	8	6	1
		Quercetin	C15H10O7	302.04	1.834	7	5	0
		Quercitrin	C21H20O11	448.1	0.802	11	7	2
		Rutin	C27H30O16	610.15	−0.735	16	10	3
4	Control Ligand	Hydroxychloroquine	C18H26ClN3O	335.18	1.548	4	2	0

Figure 3

Molecular docking protocol validation. This crucial process can enhance the accuracy and reliability of an in-silico docking experiment. A comparison of the binding modes for the re-docked ligand (blue) vs. the co-crystallized ligand (red), shown as a stick representation. a) The molecular docking protocol accurately regenerated the binding configuration of a crystallographically determined protein-ligand complex with an RMSD value of 1.049 Å using PyMOL. Amino acid residue interactions with b) the re-docked and c) the co-crystallized ligand executed in LigPlot+.

Table 2

The binding energies and molecular interaction profiles of the control drug and the active compounds from Azadirachta indica, Mangifera indica, and Moringa oleifera leaves against the main protease of SARS-CoV-2.

Ligands	Plant Source	Binding Energy kcal/mol	No. of H-Bond	Residues involved in Hydrogen bond	Residues involved in hydrophobic interaction	Residues involved in π-stacking	Residues involved in Salt Bridge
Caffeic acid	M. oleifera	−5.6	3	Leu141,Ser144 and His163	Phe140,Leu141,Asn142,Gly143,Ser144,Cys145,His163,Met165,Glu166 and Gln189	–	–
Catechin	M. indica and M. oleifera	−7.2	4	Glu166,Asp187,Thr190 and Gln192	His41,His164,Met165,Pro168,Asp187,Arg188,Gln189,Thr190 and Gln192	His41	–
Chlorogenic acid	M. oleifera	−7.2	5	Cys145,His163,Arg188,Thr190 and Gln192	Leu141,Asn142,Gly143,His163,His164,Met165,Glu166,His172,Arg188,Gln189 and Thr190	–	His41, His163 and His172
Ellagic acid	M. indica and M. oleifera	−7.3	3	His41,Arg188 and Thr190	His41,Cys145,His164,Met165,Glu166,Asp187 and Arg188	–	–
Epicatechin	M. indica and M. oleifera	−7.0	3	Ser144,His163 and Gln189	His41,Met49,Phe140,Leu141,His163,His164,Met165,Glu166,Asp187,Arg188 and Gln189	–	–
Gallic acid	M. indica and M. oleifera	−5.5	6	Leu141,Gly143,Ser144,Cys145,His163 and Glu166	Phe140,Leu141,Asn142,Gly143,Ser144,His163,Met165 and Glu166	–	–
Gedunin	A. indica	−7.3	1	Asn142	His41,Asn142,Gly143,Cys145,His164,Met165,Glu166,Pro168,Gln189 and Thr190	–	–
Hydroxychloroquine	Control ligand	−6.4	4	Ser144,Cys145,Gln189 and Thr190	Phe140,Leu141,Asn142,Gly143,Ser144,Met165,Glu166,Arg188 and Gln189	–	–
Kaempferol	M. indica and M. oleifera	−7.8	3	Leu141,Ser144 and Gln189	Met49,Leu141,Cys145,His163,Met164,His165,Glu166,Asp187,Arg188 and Gln189	–	–
Lupeol	M. indica	−7.6	–	–	His41,Asn142,Gly143,Cys145,His164,Met165,Glu166,Leu167,Pro168,Gln189 and Thr190	–	–
Mangiferin	M. indica	−8.4	9	His41,Leu141,Asn142,Gly143,Ser144,Cys145,Arg188,Thr190 and Gln192	His41,Met49,Leu141,Asn142Gly143,Ser144,Cys145,Met165,Glu166,Arg188,Gln189,Thr190 and Gln192	–	–
Myrecitin	M. oleifera	−7.7	–	–	Thr26,Leu27,His41,Met49,Tyr54,Gly143,Cys145,His164,Met165,Asp187,Arg188 and Gln189	–	–
Nimbandiol	A. indica	−7.1	2	Thr26 and Gly143	Thr25,Leu27,His41,Met49,Phe140,Leu141,Asn142,Gly143,Cys145,Gln166 and Gln189	–	–
Nimbin	A. indica	−6.2	2	Arg40 and Arg188	Arg40,Tyr54,Glu55,Met82,Asn84,Cys85,Gly179,Asn180,Phe181,Val186,Asp187 and Arg188	–	Arg41 and Arg188
Nimbinene	A. indica	−6.5	2	Asn142 and Gly143	Thr25,Thr26,Leu27,Met49,Phe140,Leu141,Asn142,Gly143,Cys145 and Gln189	–	–
Nimbolide	A. indica	−7.6	3	Glu55,Phe181 and Arg188	Arg40,Asn53,Tyr54,Glu55,Met82,Cys85,Asn180,Phe181,Val186 and Arg188	–	Arg40 and Arg188
Nimbolinin	A. indica	−6.2	1	Arg40	Arg40,Tyr54,Glu55,Met82,Asn84,Cys85,Asn180,Phe181,Phe185,Val186 and Arg188	–	Arg40 and Arg188
Quercetin	A. indica, M. indica and M. oleifera	−7.5	4	Leu141,Ser144,His163 and Gln189	His41,Met49,Phe140,Leu141,His163,Met165,Glu166,Asp187,Arg188 and Gln189	–	–

Figure 4

The binding configuration of ligands showing their poses and interactions in the binding site of the main protease of SARS-CoV-2. (a) Hydroxychloroquine, (b) mangiferin, (c) kaempferol, (d) lupeol, (e) nimbolide, and (f) quercetin. The interaction analysis shows hydrogen bonds (dashed green lines) and hydrophobic interactions (curved red lines) as ligands (purple) interact with the amino acid residues in the active site of Mpro.

ADMET properties were predicted using the pharmacokinetic parameters of the ADMETSar web server (http://lmmd.ecust.edu.cn/) for six candidate compounds with hydroxychloroquine (control) after a successful docking study (Table 3). All the hit compounds possessed low absorption in the intestine via Caco-2 permeability except hydroxychloroquine, which might be due to the molecular size, whereas all possessed higher human intestinal absorption. Only the control, lupeol, and nimbolide exhibited blood–brain barrier penetrating potential. They all exhibited high plasma protein binding rates that might affect their movement to the target site to exert pharmacological actions. Most of them (5) reside in mitochondria, and the remainder (2) are localized in lysosomes. For drug metabolism, none of the compounds would be metabolized by CYP2D6 except hydroxychloroquine, and none of them would likely inhibit CYP2D6. Kaempferol, myricetin, and quercetin display inhibition of CYP1A2, which might indicate liver metabolism inhibition, while kaempferol showed inhibition of CYP2C19 and 2C9. All compounds would serve as a substrate to CYP3A4 except myricetin. Kaempferol, myricetin, nimbolide, and quercetin would likely inhibit CYP3A4. All compounds were found to be toxic in the Salmonella typhimurium reverse mutation assay (AMES), a preliminary drug screening assay to analyse whether a drug can cause mutations in Salmonella typhimurium., None were predicted to be carcinogenic. All were predicted to cause hepatotoxicity in humans except hydroxychloroquine and lupeol. Hydroxychloroquine and nimbolide are p-glycoprotein substrates, indicating their ability to release phosphate from adenosine triphosphate (ATP). All compounds show good aqueous solubility, which might be due to the number of hydroxyl groups in them. Only hydroxychloroquine and nimbolide could block human ether-a-go-go (herG), a potassium ion channel.

Table 3

Absorption, distribution, metabolism, elimination, and toxicity predictions for kaempferol, mangiferin, myrecitin, nimbolide, and quercetin.

Models	Hydroxychloroquine	Kaempferol	Lupeol	Mangiferin	Myrecitin	Nimbolide	Quercetin
Ames mutagenesis	+	+	–	+	+	–	+
Acute Oral Toxicity (c)	III	II	III	IV	II	III	II
Blood Brain Barrier	+	–	+	–	–	+	–
Biodegradation	–	–	–	–	–	–	–
Caco-2 Permeability	+	–	–	–	–	–	–
Carcinogenicity	–	–	–	–	–	–	–
CYP1A2 inhibition	–	+	–	–	+	–	+
CYP2C19 inhibition	–	+	–	–	–	–	–
CYP2C9 inhibition	–	+	–	–	–	–	–
CYP2C9 substrate	–	–	–	–	–	–	–
CYP2D6 inhibition	–	–	–	–	–	–	–
CYP2D6 substrate	+	–	–	–	–	–	–
CYP3A4 inhibition	–	+	–	–	+	+	+
CYP3A4 substrate	+	+	+	+	–	+	+
CYP inhibitory promiscuity	–	+	–	–	+	+	+
Hepatotoxicity	–	+	–	+	+	+	+
Human either-a-go-go inhibition	+	–	–	–	–	+	–
Human Intestinal Absorption	+	+	+	+	+	+	+
Human oral bioavailability	+	–	–	–	–	–	–
Acute Oral Toxicity (Kg/mol)	2.6650	1.7388	3.8523	2.9792	2.3758	3.7769	2.5588
P-glycoprotein inhibitior	–	–	–	–	–	+	–
P-glycoprotein substrate	+	–	–	–	–	+	–
Plasma protein binding (100%)	0.7560	1.0613	1.0168	0.9697	1.1616	0.7761	1.1748
Subcellular localization	Lysosomes	Mitochondria	Lysosomes	Mitochondria	Mitochondria	Mitochondria	Mitochondria
UGT catalyzed	+	+	+	+	+	–	+
Water solubility LogS	−3.5660	−3.1423	−4.4139	−2.3978	−2.9994	−4.4552	−2.9994

Discussion

Lipinski's rule of five is a rule of thumb that evaluates a compound's drug-likeness to determine if the compound possesses specific biological or pharmacological activity conferring oral bioavailability in humans. According to Usha et al. and Singh and Konwar; any compound that exceeds molecular weight (Mw) > 500 Da, calculated log P > 5, hydrogen-bond donors >5 and hydrogen-bond acceptors >10 is unlikely to be further pursued as a potential oral drug because it would likely be deficient in properties essential for the absorption, distribution, metabolism, and excretion.

According to Liu and Wang, dozens of proteins are coded by the coronavirus, some of which are involved in viral replication and entry into cells. Mpro is an essential enzyme for coronaviral replication, making it an ideal target for drug design and development since humans do not possess a homologue. This enzyme's role is to cleave polyproteins 1a and 1 ab at 11 conserved sites to generate non-structural proteins 4 to 16 (nsp4-16), starting by self-cleavage (nsp5).,, The Mpro adopted for this study was retrieved in its 3-D crystal structure (6LU7), co-crystallized with an N-leucinamide inhibitor with a resolution 2.16 Å, an amino acid sequence length of 306, and a homodimeric protein., The catalytic site of Mpro consists of 25 amino acid residues after extensive literature mining vis: Thr25, Thr26, His41, Cys44, Met49, Tyr54, Phe140, Leu141, Asn142, Gly143, Ser144, Cys145, His163, His164, Met165, Glu166, Pro168, His172, Phe185, Val186, Asp187, Arg188, Gln189, Thr190, and Gln192. Also, the active site exhibits a CysHis (Cys145 and His41) catalytic dyad., In vitro and in vivo studies on mangiferin, a C-glucosyl xanthone, have revealed various pharmacological activities that include antiviral properties.,,, The binding affinities between the active compounds (ligands) and Mpro in this study were stabilized by non-covalent interactions, such as hydrogen bonds, hydrophobic bonds, pie–stack interactions, and salt bridge formation (Figure 3 and Table 3). Similarly, Mohapatra et al. and Salentin et al. reported that the binding affinity would be higher (binding energy will be lower) if ligand's ability to form hydrophobic interactions with hydrophobic amino acid residues in the binding site is higher. Stojanovic and Zari reported the significance of hydrophobic interactions in many systems with strong intermolecular forces. This evidence could be responsible for the appreciable binding affinity of the active compounds docked against Mpro in this study. Interestingly, the influence of a hydrogen bond in stabilizing the molecular interaction between the ligands and the protein cannot be downplayed because of its critical roles in enzyme catalysis, protein-substrate, and protein-inhibitor complexes as well as structural stability of various biological molecules., To this end, analysis of molecular interactions exhibited by the active compounds assessed in this study as judged by their binding energies, hydrogen bonding, and hydrophobic interactions, with the surrounding amino acids in the active pocket of Mpro, revealed the positive impact of their ligand–protein complex status, invariably enhancing the inhibition of Mpro from SARS-CoV-2.

The purpose of performing ADMET screening of any chemical compound within the human body is to determine the pharmacological and pharmacodynamic properties of a candidate drug compound within the biological system. Mangiferin was predicted to have high plasma protein binding and not be a p-glycoprotein substrate. Thus, mangiferin might be transported to its target site in an effective dose. The aqueous solubility predicted for mangiferin might be because of the numerous hydroxyl groups found in its structure. The liver is the leading site of metabolism for various drugs and xenobiotic agents, making it highly vulnerable to harmful effects; hence, hepatotoxic effects resulting from drugs or xenobiotic metabolic processes may result in severe organ (liver) collapse and ultimately death. However, our in-silico study results are insufficient to ascertain the efficacies of the studied plant compounds in the management of COVID-19. This is the major limitation associated with this study. Hence, there is a great need for further in vitro and in vivo studies and clinical trials to validate the study plants' inhibitory potentials against SARS-CoV-2 primary protease replication.

Conclusions

In conclusion, most of the active phytocomponents of the study plants exhibited relative inhibitory potentials against Mpro of SARS-CoV-2 and preferred pharmacological features when compared with hydroxychloroquine, making them potential antiviral candidates in the fight against COVID. However, there is a need for further studies to ascertain the efficacies of these active compound(s), as the inhibitory properties displayed in this study may present an appropriate first step in the development of novel drugs for the management or treatment of coronavirus.

Recommendations

Since the pandemic remains a major health crisis and no cure has been found, an urgent need for further studies to establish the efficacies of these active compound(s) as their inhibitory properties displayed may present an appropriate leading step in the repositioning, repurposing, and development of novel drugs for the management or treatment of SARS-CoV-2.

Source of funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Conflict of interest

The authors have no conflict of interest to declare.

Ethical approval

There are no ethical or financial issues, conflicts of interest, or animal experiments related to this research.

Authors' contributions

HIU and AA conceived and designed the study, conducted research, and collected and organized the data. SSJ and PTS analyzed and interpreted the data. TOJ and JBD wrote the initial and final drafts of the article. HIU and SSJ reviewed the final version of the manuscript submitted for publication. All authors provided logistical support and approved the final draft, and are accountable for the manuscript's content and similarity index.