In-silico screening of naturally derived phytochemicals against SARS-CoV Main protease.

Coronavirus disease 2019 (COVID-19) is a rapidly growing pandemic that requires urgent therapeutic intervention. Finding potential anti COVID-19 drugs aside from approved vaccines is progressively going on. The chemically diverse natural products represent valuable sources for drug leads. In this study, we aimed to find out safe and effective COVID-19 protease inhibitors from a library of natural products which share the main nucleus/skeleton of FDA-approved drugs that were employed in COVID-19 treatment guidelines or repurposed by previous studies. Our library was subjected to virtual screening against SARS-CoV Main protease (Mpro) using Molecular Operating Environment (MOE) software. Twenty-two out of those natural candidates showed higher binding scores compared to their analogues. We repurpose these natural products including alkaloids, glucosinolates, and phenolics as potential platforms for the development of anti-SARS-CoV-2 therapeutics. This study paves the way towards discovering a lead used in the treatment of COVID-19 from natural sources and introduces phytomedicines with dual therapeutic effects against COVID-19 besides their original pharmacological effects. We recommend further in vitro evaluation of their anti-COVID-19 activity and future clinical studies.

Introduction

SARS-CoV-2 is highly contagious virus, rapidly spreading and causing a global outbreak named “COVID-19” (Chen et al. 2020; Huang et al. 2020, Zhu et al., 2020 2019). Now, the coronavirus COVID-19 is affecting almost all countries around the world necessitating the rapid development of effective drugs especially with the emergence of new hot spots of COVID-19 infection within each continent in the world. Several vaccines are in the last phases of clinical trials before approval from FDA (Conte et al. 2020; Hodgson et al. 2020; Jain et al. 2020; Kaur and Gupta 2020; Lin et al. 2020; Sanchez-Felipe et al. 2020; Xia et al. 2020) and few vaccines have already been FDA approved and currently used (Knoll and Wonodi 2020; Ramasamy et al. 2020; Voysey et al. 2020; van Doremalen et al. 2020; Logunov et al. 2020; Polack et al. 2020; Folegatti et al. 2020).

Massive and fair distribution of COVID-19 vaccines represents a logistic problem challenged by the rapid emergence of new variants of COVID-19 that could escape the developed vaccines (Rubin 2021; Harvey et al. 2021), aside from increased concerns about the side effects of the approved vaccines which influence the global compliance (Wang et al. 2020, Russell and Greenwood 2020, Flanagan et al. 2020, Ledford 2021) and impart more challenges on developing potential COVID-19 antiviral drugs (Sharpe et al. 2020, Thanh et al., 2020). Pharmaceutical therapeutic strategies adapted from previous respiratory viral infections have been followed by some local guidelines, some of them have provided considerable therapeutic efficiency against COVID-19 (Rahman et al. 2020).

Virtual screening has been employed in drug discovery against emerging and fatal diseases including SARS CoV proteases (Sirois et al. 2004), hepatitis C virus RNA polymerase (Elhefnawi et al. 2012), dengue virus (Zhou et al. 2008), and ebola virus (Raj and Varadwaj 2016).

Several studies have recently employed in silico investigation of large sets of natural active constituents of plant origin using online databases in attempt to discover new hits against potential molecular targets (Shaldam et al. 2021, Al Naggar et al. 2021, Sayed et al. 2020, Owis et al. 2020, Divya et al. 2020, Rosales-Mendoza 2020, Tahir ul Qamar et al. 2020, Orhan and Senol Deniz 2020). ACE2 acts as a portal gate for the SARS-CoV-2, to enter the cell, by interacting with the viral glycoprotein spikes; this step represents a milestone for designing therapeutic interventions against COVID-19 (Behl et al. 2020). The known coronavirus 3-chymotrypsin-like protease (3CLpro), also known as Mpro, is the main protease, which is required for proteolytic maturation of the corona virus (Zhou et al. 2019; Berry et al. 2015). Targeting SARS-CoV Main protease (Mpro) enzyme will inhibit the viral maturation and enhance the host innate immune response against COVID-19 (Elmorsy et al. 2021, Shaldam et al. 2021, Jin et al. 2020, Tahir ul Qamar et al. 2020, Nguyen et al. 2012).

Kandeel and Al-Nazawi docked 487 FDA-approved drugs against Mpro and listed the top nineteen as effective candidates against COVID-19 (Kandeel and Al-Nazawi 2020). Relying on their study, clinical data from different guidelines of COVID-19 treatment in different countries, and feedback from published in vitro, in silico, or clinical studies that demonstrate effect or expected effect of repurposed drugs against COVID-19 (Kandeel and Al-Nazawi 2020; Oliveira et al. 2020; Wu et al. 2020), we determined the natural or plant derived substitute of these recently investigated repurposed drugs based on harboring a similar chemical nucleus or skeleton. Our similar chemical nucleus-based approach identified several phytochemical active agents of flavonoids, coumarins, alkaloids, terpenes, phenols, or glucosinolate origin. We decided to perform deep in silico investigation via molecular docking of these natural analogues against COVID-19 main protease to test possible activity of those naturally derived compounds compared to their chemical relatives which are already proved to have Mpro binding affinity by molecular docking. Our chemical structure-based approach identified 32 bioactive compounds of pure natural origin and reported to exhibit broad medicinal effects. These naturally existing compounds shared structural similarity (low to high) with their synthetic analogues and 14 of them displayed enhanced Mpro binding affinity greater than their analogues. Most of these chemical active constituents of plant or natural origin or their derivatives are well identified to possess wide range of pharmacological activities like anticancer, antioxidant, antihypertensive, hypoglycemic, antipyretic, antimalarial, bronchodilator, and antispasmodic which represent further advantage to treat the associated complications of COVID-19 (Nagu et al. 2021; Kashyap et al. 2021; Ghosh et al. 2021) or to provide dual therapeutic effect in patients with chronic diseases (hypertensive, diabetic, asthmatic, gouty arthritic, and hepatic) or cancer patients that have been infected with COVID-19.

Materials and methods

Molecular docking of the compounds in this study was carried out using Molecular Operating Environment 2009 (MOE) as previously described by Nagah et al. (Nagah et al. 2021). Briefly, the compounds were constructed in 3D structure, and their energies were minimized and saved to MDB file. COVID-19 main protease X-ray crystallographic structure coded as 6LU7 was downloaded together with its natural ligand (N-[(5-methylisoxazol-3-yl)carbonyl]alanyl-l-valyl-n ~ 1 ~ -((1r,2z)-4-(benzyloxy)-4-oxo-1-{[(3r)-2-oxopyrrolidin-3-yl]methyl}but-2-enyl)-l-leucinamide) from the Protein Data Bank. Hydrogens were added to the protein structure and missed connections and their types were corrected automatically. The receptor and its atoms potential were fixed. The active site of the enzyme was determined based on co-downloaded natural ligand and using site finder, dummy atoms were created to assign pocket of activity after removing natural ligand to free the pocket.

The constructed compounds’ database was docked against Mpro using the following parameters: Site of docking, dummy atoms; placement, triangle matcher; scoring, London dG with ten retains; refinement, forcefield. The resulted poses were investigated based on their energy, root-mean square deviation (rmsd), and formed interactions (bonds).

Results and discussion

Natural substitutes identification based on shared chemical nucleus

We specified highly ranked FDA-approved drugs recently repurposed against COVID-19 (Kandeel and Al-Nazawi 2020) analyzed by virtual screening and molecular docking against COVID-19 Mpro and we look through the natural library of bioactive compounds to find similar substitutes based on shared chemical nucleus or similar chemical skeleton. The identified compounds were ranked to possess high, moderate, or low degree of structural similitude to their compared partners (Table 1).

Table 1

Natural chemical analogues of FDA-approved drugs repurposed for treatment of COVID-19, their common uses, and source

Melatonin and artemisinin while approved and repurposed for COVID-19 in some studies, they belong to naturally or biologically produced compounds (Kandeel and Al-Nazawi 2020; Shneider et al. 2020; Zhang et al. 2020; Parlakpinar et al. 2020; El-Missiry et al. 2020; Cheong et al. 2020; Li et al. 2020; Gonzalez-Paz et al. 2020).

Docking study

Structurally analogous bioactive compounds were investigated by molecular docking against the first resolved COVID-19 crystal structure (main protease), at the same time molecular docking of FDA analogous alternatives was conducted in the same way against the same target. Results of interaction energies with Mpro pocket of COVID-19 virus and receptor amino acids involved in these interactions were extracted from molecular simulation analysis and various bioactive compounds were compared against their structural FDA-approved analogues according to binding stability of the formed complex denoted by interaction energies with Mpro (Supplemental 1, Table 2). Most natural substitutes that exhibit high resemblance to FDA chemical analogues show motivating binding scores that may attribute ligand substrate interaction to functional groups or chemical skeleton shared between the FDA and the natural analogues. Considerable number (14) of the investigated phytomedicines exhibited higher affinity and stable binding to Mpro enzyme 1.34 to 2.51 folds more than their FDA alternatives (Table 2, Supplemental 2).

Table 2

Binding scores and amino acids interactions of tested natural analogues and their FDA-approved analogues against COVID-19 main protease Mpro

Name of synthetic drug	Score	Amino acid interactions	Name of natural drug	Score	Amino acid interactions	FC*
Chromocarb	− 11.6991	Glu166, Arg188, Thr190	Khellin	− 20.1254	Glu166	1.7203
			Visnagin	− 16.8687	Glu166	1.4419
			Quercetin	− 16.3019	His41, Leu141	1.3934
			Silybin	− 22.6614	Thr24, Phe140	1.9370
Aminophylline	− 12.4880	Gly143	Caffeine	− 13.6623	His41, Gly143	1.0940
			Theophylline	− 12.4880	Gly143	1
			Theobromine	− 10.1135	Gly143	0.8099
Methazolamide Zonisamide	− 9.0162 − 13.9920	Thr45, Ser46 Glu166	3-Hydroxypropyl-glucosinolate	-19.2458	His163, His164	2.1346 1.3755
			3-Methylsulfinylpropyl-glucosinolate	− 22.6547	Phe140, His164	2.5127 1.6191
			3-Benzoyloxypropyl-glucosinolate	− 20.4349	Thr26, Leu141, Ser144	2.2665 1.4605
			3-Indolylmethyl-glucosinolate	− 21.7720	Thr26, Leu141, Asn142, Ser144, His163	2.4148 1.5560
Methoxamine Octapamine	− 13.3139 − 12.0179	Glu166 Thr190	Nor-Ephedrine	− 8.5017	Glu166	0.6386 0.7074
			Ephedrine	− 5.8499	His41	0.4394 0.4868
			Vanillin	− 11.8509	Gly143	0.8901 0.9861
			o-Coumaric acid	− 13.8211	Glu166	1.0381 1.1500
			p-Coumaric acid	− 12.1056	Leu141, Gly143	0.9092 1.0073
			Caffeic acid	− 14.0479	Thr190	1.0551 1.1689
Nicotinamide	− 7.4521	His41, His164	Nicotinic acid	− 6.9997	His41, Gly143	0.9393
Aminosalicylate Sodium	− 11.0481	Glu166	Salicin	− 15.0911	Phe140, Leu141	1.3659
Bemegride	− 11.0712	His163	Piperine	− 16.4044	Thr45, Ser46, Gly143	1.4817
Pyrazinamide	− 8.1357	His41, His164	Colletotrichumine A	− 7.8977	His41	0.9707
Melatonin	− 17.8233	Gly143	Vincristine	− 18.5459	Asn142, Gly143	1.0405
			Vinblastine	− 23.9685	His164, Gln189	1.3448
			Reserpine	− 26.0568	Gly143	1.4620
			Ergotamine	− 19.9235	His41, Gly143	1.1178
			Ergometrine	− 16.1153	Phe140, Glu166	0.9042
Propylthiouracil	− 14.3637	Gly143	Uracil	− 11.9928	His164	0.8349
Artemisinin	− 11.5254	His163, Glu166	Colchicine	− 18.4888	Gly143, Gln189	1.6042
			Demecolcine	− 5.8776	Glu166	0.5100
			Ascaridole	− 12.1891	Glu166	1.0576
			Cineole	− 10.4338	Gly143	0.9053
Triflusal	− 13.9774	His41, Gly143	Paclitaxel	− 25.0555	Glu166, Gln189	1.7926

* Fold change of natural compound compared to its FDA-approved analogue(s), compounds in bold show fold change more than 1.3 compared to their synthetic analogues.

Binding simulation of investigated compounds

By analyzing the substrate ligand binding mode of top ranked drugs, hydrogen bonding and hydrophobic interactions were the leading drive for binding. Generally, compounds that interact with several amino acids and side chains of the docked polypeptide like 3-indolylmethyl-glucosinolate are more able to form hydrogen bonds between free hydroxyl groups in the backbone and the interacting amino acids of the binding pocket or between hydroxyl/amino groups in amino acids and polar atoms from the chemical compound backbone. In addition to hydrogen bonding, hydrophobic interactions like noncovalent interactions between π systems and arene arene stacking interactions between aromatic rings in both ligands and substrates like the alkaloids ergotamine and ephedrine and the flavonoid quercetin, or aromatic rings and heterocycles like imidazole in the backbone of the alkaloid caffeine or pyrazine in the alkaloid colletotrichumine A or pyrimidine in case of nicotinic acid.

The analysis of stable protein–ligand architecture and the involvement of a specific amino acid in the ligand binding represent valuable aspects to identify and discriminate the hot-spots. From the binding simulation of tested compounds, about 15 amino acids dispersed in the Mpro binding pocket are involved in the ligand binding, most of them associate to tested compounds by hydrogen bond formation, His41 is observed to be crucial for arene arene stacking interaction, and the residues Glu166, Gly143, His41, His163, His164, Ser144, and Thr26 are the common key residues that are observed from most protein–ligand interactions.

High score alkaloids

Alkaloids that share analogous chemical nucleus like indole alkaloids reserpine from rauwolfia, vincristine and vinblastine of vinca, ergotamine and ergometrine alkaloids from ergot, terpenoid alkaloid paclitaxel of taxol, piperine from pepper and colchicine from colchicum (Evans et al. 2009) were among the top score candidates. Hydrogen bonding is mostly involved in alkaloid binding simulation to the Mpro pocket, and amino acid residues Gly143, Glu166, Gln189, and His164 of the docked polypeptide are mostly involved in the interaction with the investigated alkaloids (Figs. 1 and 2, Supplemental 2). In this study, two promising compounds reserpine and paclitaxel while exhibiting low similitude to the comparable synthetic compound unexpectedly possess high binding scores to Mpro enzyme (-26.0568 and -25.0555, respectively). Reserpine was able to dock via interaction with Gly143 of Mpro while paclitaxel exhibited two hydrogen bonds between its hydroxyl groups and Glu166 and Gln189 amino acids of Mpro (Supplemental 2). Vinblastine showed affinity to Mpro via interaction with His164 and Gln189 at a binding score of − 23.9685 (Fig. 1, Table 2). Colchicine represents another example from this group that formed hydrogen bonds with Gly143 and Gln189 via its methoxy and amido groups (Fig. 2, Table 2). Cautions should be paid for the use of such alkaloids specially the anticancer ones to avoid their side effects; however, cost benefit ratio will play a role in their use depending on the severity of the case or in case of COVID-19 cancer patients.

Fig. 1

Amino acid interactions of COVID-19 main protease Mpro with vinblastine alkaloid in two-dimensional configuration (up) and three-dimensional configuration (down) using MOE

Fig. 2

Amino acid interactions of COVID-19 main protease Mpro with colchicine alkaloid in two-dimensional configuration (up) and three-dimensional configuration (down) using MOE

Glucosinolates

Glucosinolates glycosides were recognized in our screening as natural structural substitutes to the carbonic anhydrase inhibitor methazolamide indicated in the treatment of increased intraocular pressure and the antiepileptic zonisamide and both drugs are repurposed against COVID-19 (Kandeel and Al-Nazawi 2020). Glucosinolates show promising binding scores and form stable complex through interaction with a cluster of amino acid residues Phe140, His164, Asn142, His163, Thr26, Leu141, and Ser144 which are assigned behind ligand substrate binding in the binding simulation (Fig. 3, supplemental 2). 3-Methylsulfinylpropyl-glucosinolate comes on the top of this group with a binding score of − 22.6547, it interacted with Phe140 and His164. On the other hand, 3-indolylmethyl-glucosinolate was the most interesting compound in this study as it stabilize itself in Mpro pocket via six different bonds with five amino acids namely Thr26, Leu141, Asn142, Ser144, and His163. It is obvious from these interactions that sulfate and glucose groups of glucosinolates are essential for activity (Fig. 3, Table 2, Supplemental 2).

Fig. 3

Amino acid interactions of COVID-19 main protease Mpro with 3-indolylmethyl-glucosinolate in two-dimensional configuration (up) and three-dimensional configuration (down) using MOE

Flavonoids, coumarins and phenolic phytocompounds

Significant number of flavonoids and phenolic natural products came out in our screening, interestingly most of them were proposed in other studies to possess Mpro affinity and previously documented to have antiviral effect. Quercetin, caffeic acid, and o-coumaric acid were recently repurposed against COVID-19 in several studies (Colunga Biancatelli et al. 2020; Bachevski et al. 2020; Mani et al. 2020; Bhowmik et al. 2020; Elfiky 2020, Sayed et al. 2020). Flavonoids and phenolics were identified as analogues to the vasoprotective chromocarb and the phenylethanolamine octapamine used for treatment of hypotensive regulatory and circulatory disorders. Aside from these previously assigned compounds, new members like the major flavonolignan of silymarin (silybin) and khellin coumarin from Ammi visnaga show promising binding scores for COVID-19 Mpro enzyme (− 22.6614 and − 20.1254, respectively), silybin exhibited hydrogen bonds interactions through its hydroxyl groups and Thr24 and Phe140, while khellin interacted with Glu166 via one of its methoxy groups. Another interesting component from this class is salicin that exceeded the binding score of its synthetic analogue (aminosalicylate sodium) by about 1.36, it showed interactions with Phe140 and Leu141 amino acids of Mpro pocket (Fig. 4, Table 2, Supplemental 2). Like alkaloids and glucosinolates, repurposed flavonoids, coumarins, and phenolics stabilize their interaction through hydrogen bonding with amino acids including Thr26, Thr24, Phe140, Leu141, and Glu166 within the Mpro pocket (Fig. 4, supplemental 2).

Fig. 4

Amino acid interactions of COVID-19 main protease Mpro with silybin, in two-dimensional configuration (up) and three-dimensional configuration (down) using MOE

Relation to oseltamivir (Tamiflu®)

Antiviral therapy of COVID-19 patients with lopinavir/ritonavir showed considerable efficiency according to local guidelines (Negrut et al. 2021). Oseltamivir (an antiviral agent) is approved in COVID-19 treatment protocol in many countries including Japan, China, and Egypt. Molecular docking of this compound against Mpro exhibited binding with His164 and Gln189 at a score of − 17.299. Based on in silico analysis of the docked natural products relative to oseltamivir, we found that ten compounds including khellin, silybin, 3-hydroxypropyl-glucosinolate, 3-methylsulfinylpropyl-glucosinolate, 3-benzoyloxypropyl-glucosinolate, 3-indolylmethyl-glucosinolate, vinblastine, reserpine, colchicine, and paclitaxel showed better binding scores. These results reflect the importance of natural products as drug leads for treatment of COVID-19. We highly recommend carrying out in vivo study for these compounds to confirm their importance as new and effective treatment for this COVID-19 infection and determine their effective doses.

Limitations

Our study was based on previous in silico studies of FDA-approved drugs that were virtually screened against COVID-19 Mpro; we only considered natural substitutes of these drugs that possess considerable structural similarity. A more comprehensive approach has to be considered to cover more natural candidates and phytochemicals and target other COVID-19 functional proteins in the future screening approaches.

Conclusion

We screened a library of phytochemicals based on similar analogous chemical structure or shared functional groups to FDA-approved drugs recently employed or repurposed for COVID-19 treatment. Interestingly, our screening identified some candidates of natural active agents that have been recently repurposed by virtual screening as COVID-19 Mpro inhibitors like quercetin, colchicine, piperine, and caffeic acid. Looking at compounds with high interaction energy in our screening, we identified novel bioactive compounds including antihypertensive alkaloid reserpine, potential anticancer like paclitaxel, vincristine and vinblastine alkaloids, well-known hepatoprotective silybin and other flavonoids and phenolic compounds, and more important cluster of glucosinolate glycosides naturally occurring in many pungent plants such as mustard, cabbage, broccoli, rocket, and horseradish.

Future prospects

While we propose these newly screened phytomedicines based on their binding scores as drug leads for design of potential COVID-19 Mpro inhibitors, we recommend further in vitro examination and clinical evaluation of their virtual anti COVID-19 activity.

Below is the link to the electronic supplementary material.

Supplementary file1 (PDF 425 KB)

Supplementary file2 (PDF 6088 KB)