Sage Journals: Discover world-class research

Abstract

In our quest to discover effective inhibitors against severe acute respiratory syndrome coronavirus 2 helicase, a diverse set of more than 300 naturally occurring antiviral metabolites was investigated. Employing advanced computational techniques, we initiated the selection process by analyzing and comparing the co-crystallized ligand (VXG) of the severe acute respiratory syndrome coronavirus 2 helicase protein (PDB ID: 5RMM) to identify compounds with structurally similar features and potential for comparable binding. Through structural similarity and pharmacophore research, 13 compounds that shared important characteristics with VXG were pinpointed. Subsequently, these candidates were subjected to molecular docking to identify seven compounds that demonstrated favorable energy profiles and accurate binding to the severe acute respiratory syndrome coronavirus 2 helicase. Among these, mycophenolic acid emerged as the most promising candidate. To ensure the safety and viability of the selected compounds, we conducted ADMET tests, which confirmed the favorable characteristics of mycophenolic acid, and the safety of atropine and plumbagin. Building on these results, we performed additional analyses on mycophenolic acid, including various molecular dynamics simulations. These investigations demonstrated that mycophenolic acid exhibited optimal binding to the severe acute respiratory syndrome coronavirus 2 helicase, maintaining flawless dynamics throughout the simulations. Furthermore, the Molecular Mechanics Poisson–Boltzmann Surface Area tests provided strong evidence that mycophenolic acid successfully formed a stable connection with the severe acute respiratory syndrome coronavirus 2 helicase, with a calculated free energy value of −294 kJ mol⁻¹. These encouraging findings provide a solid foundation for further research, including in vitro and in vivo studies, on the three identified compounds. The potential efficacy of these compounds as treatment options for coronavirus-19 warrants further exploration and may hold significant promise in the ongoing fight against the pandemic.

Keywords

coronavirus-19 docking molecular dynamics simulations Molecular Mechanics Poisson–Boltzmann Surface Area severe acute respiratory syndrome coronavirus 2 helicase structural similarity

Introduction

As of 23 July 2023, the World Health Organization (WHO) reported an alarming total of 768,237,788 confirmed cases of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection worldwide. Tragically, among these cases, 6,951,677 individuals have lost their lives.¹ These distressing figures highlight the urgent need for extensive efforts from scientists in the quest for a cure, underscoring the crucial role of research and collaborative efforts in combatting this global health crisis.

The use of the computer as a tool in drug discovery (computer-aided drug discovery (CADD)) has been an essential approach in pharmaceutical and medicinal research since the 1970s and before. Progressively, it drew more attention to being a trustworthy technique that can expect the potential of a compound to be biologically active, saving time, and money.² There are two main types of CADD techniques. The first is the structure-based that starts with the target protein and measures the ability of various ligands to bind with that protein through the calculation of binding energies.³ Contrastingly, the ligand-based method uses ligands (mostly a group of molecules) checking their possible activities against an identified ligand through pharmacophore, structural similarity, or quantitative structure–activity relationship (QSAR) models.^4,5 Regarding coronavirus-19 (COVID-19), the CADD approach was applied in several reports.^6–8

Throughout history, nature has been a generous source of remedies, offering humans a vast array of cures for various diseases and infections. From the earliest records of human existence, our ancestors relied on their surroundings to find plants, herbs, and other natural substances that could alleviate ailments and restore health.^9,10

In our team’s efforts to counteract the impact of COVID-19, a sophisticated multistage in silico CADD strategy was implemented. The objective was to discern natural inhibitors and propose the most efficacious one tailored to specific COVID-19 enzymes. Among a cohort of 310 antiviral natural metabolites, promising inhibitors for the main protease,^11,12 nsp10,¹³ and the papain-like protease¹⁴ have been pointed. Furthermore, among 3009 Food and Drug Administration (FDA)-approved drugs, the most potential inhibitors against the RNA-dependent RNA polymerase¹⁵ and the SARS-CoV-2 nsp16-nsp10 2′-o-methyltransferase¹⁶ have been indicated. Also, the exploration of 5956 traditional Chinese medicine to identify potential inhibitors for the SARS-CoV-2 helicase enzyme was conducted.¹⁷

Viral helicases are a sort of enzyme that is responsible for the unpack of the duplex oligonucleotides (RNA or DNA) into their constituent single strands besides other significant functions in the replication process.¹⁸ Consequently, the inhibition of that vital enzyme could be a crucial step in the fight against the target virus.

The aim of this study is to single out a natural compound to be an anti-COVID-19 candidate through the inhibition of SARS-CoV-2 helicase protein. The compounds were subjected to various in silico filtration processes. We herein report different advanced ligand and structure-based computational studies for a set of 310 reported antiviral natural compounds (Fig.S.1. and Table S.1. in the Supplementary data) to determine the most convenient SARS-CoV-2 helicase inhibitor. The used compounds were selected based on deep research on the scientific websites for natural metabolites that exhibited antiviral activities. The selected compounds belonged to diverse classes, such as alkaloids, peptides, flavonoids, anthraquinones, steroids, and carbohydrates.

The rationale of this work comprises seven successive computational steps to reach a promising candidate against SARS-CoV-2 helicase. First, a set of 310 naturally occurring compounds was examined for their structural similarity with the co-crystallized ligand of SARS-CoV-2 helicase. Then, the most similar 30 compounds were subjected to the pharmacophore study to select the most related compound to the co-crystallized ligand of SARS-CoV-2 helicase. Next, a set of 13 compounds that showed good fit values with the generated pharmacophore was docked against the crystal structure of SARS-CoV-2 helicase. Seven compounds that showed good fitting and binding mode at the active pocket of SARS-CoV-2 helicase were subjected to in silico physicochemical profiling assessment according to Lipinski’s and Veber’s rules. The tested compound that obeyed Lipinski’s and Veber’s rules (seven compounds) was subjected to ADMET study to reach the most convenient compounds for druggability and drug likeness. After that, the most convenient four compounds were evaluated for their expected toxicity potential. One compound showed an accepted range of toxicity and was subjected to an MD simulation experiment to assess its stability in the active pocket of SARS-CoV-2 helicase (Figure 1).

Figure 1.

Rationale of the discovery of a promising candidate against SARS-CoV-2 helicase.

Results and discussion

Molecular similarity

The molecular similarity is an in silico ligand-based type of calculation that is used to explore the similarity between two molecules following the scientific principle that the structurally similar molecules are predicted to have related biological properties.¹⁹ Molecular similarity targets the molecular properties of the explored compounds, such as the LogP and molecular weight. In addition, the molecular similarity mostly depends on the distances between atoms in space, structural and physicochemical descriptors.²⁰

Although molecular similarity can also be defined using molecular properties, such as molecular weight, and LogP, molecular similarity is most often defined in terms of a distance measure in a structural or physicochemical descriptor space. There many different distance measures available. In this work, we used Euclidean distance method that is available in Discovery Studio.

In this research, a molecular similarity study has been preceded for 310 naturally isolated compounds that showed antiviral activities before (Supplementary data) against the co-crystallized ligand of SARS-CoV-2 helicase protein (VXG) (PDB ID: 5RMM) using Discovery studio software. The used molecular properties include the number of rotatable bonds, number of cyclic rings, number of aromatic rings, number of hydrogen bond donors (HBD), number of hydrogen bond acceptors (HBA), partition coefficient (ALog p), molecular weight (M. Wt), and molecular fractional polar surface area (MFPSA). The examined compounds were subdivided into six sets (1–6). The first to fifth groups were consisted of 50 candidates, while the sixth consisted of 60 compounds (Figure 2).

Figure 2.

The similarity analysis between the tested molecules and (VXG). Green ball = reference molecule, red balls = similar compounds, blue balls = not similar compounds. (a) Compounds 1–50, (b) compounds 51–100, (c) compounds 101–150, (d) compounds 151–200, (e) compounds 201–250, and (f) compounds 251–310.

In total, 30 compounds (Figure 3 and Table 1) exhibited similar values of molecular properties with (VXG).

Figure 3.

The most structurally similar 30 compounds with (VXG).

Table 1.

The most structurally similar 30 compounds with (VXG) and their molecular properties.

Comp.	ALog p	M. Wt	HBA	HBD	Rotatable bonds	Rings	Aromatic rings	MFPSA	Minimum distance
6	0.827	264.383	2	1	1	4	0	0.113	0.702
8	0.61	264.363	2	0	0	4	0	0.143	0.674
13	2.331	248.236	3	0	0	4	2	0.222	0.682
14	0.67	147.131	2	1	0	2	1	0.339	0.556
15	1.579	251.301	4	1	3	2	2	0.292	0.502
52	0.16	290.377	3	2	5	3	1	0.172	0.736
53	0.737	304.361	4	2	5	4	1	0.217	0.922
56	2.189	230.216	3	0	1	3	2	0.216	0.942
58	2.17	246.215	4	0	2	3	2	0.242	0.935
60	2.175	230.216	3	0	1	3	2	0.216	0.940
112	2.161	314.333	5	2	4	2	1	0.249	0.419
123	1.515	174.153	3	1	0	2	1	0.331	0.400
124	1.962	188.179	3	1	0	2	1	0.292	0.400
125	2.509	242.27	3	0	0	3	1	0.176	0.483
127	2.466	244.243	4	1	1	3	2	0.235	0.450
172	1.597	244.243	4	0	3	2	0	0.208	0.465
178	1.436	258.269	4	2	2	3	1	0.274	0.477
188	1.287	355.273	3	2	0	4	2	0.282	0.686
189	1.287	355.273	3	2	0	4	2	0.282	0.686
200	2.331	248.236	3	0	0	4	2	0.222	0.706
217	2.846	267.111	1	1	0	3	2	0.142	0.819
222	2.154	242.699	3	0	4	1	0	0.164	0.663
234	2.24	332.434	3	2	1	4	1	0.197	0.777
235	2.24	332.434	3	2	1	4	1	0.197	0.777
236	2.526	332.434	3	2	1	4	1	0.198	0.802
288	1.683	319.329	6	1	6	2	1	0.28	0.363
289	0.371	292.33	4	2	3	2	0	0.323	0.421
290	1.941	238.283	4	1	3	1	0	0.259	0.416
298	2.716	343.443	4	1	5	3	2	0.252	0.495
305	2.639	206.281	2	1	5	1	1	0.156	0.473
VXG	−0.763	232.255	3	0	2	2	1	0.244	–

HBA: hydrogen bond acceptors; HBD: hydrogen bond donors.

Pharmacophore study

Generation of 3D-pharmacophore model

The main objective of the pharmacophore-based drug design is to generate a 3D pharmacophore model based on the known receptor–ligand pharmacophore generation approach.²¹ In this approach, various features would be expected from the study of the binding of the ligand with the target protein.^22,23 These features are essential for the activity. Consequently, any compound that contains these features is expected to be active.²⁴ The resulted pharmacophore model was generated based on receptor–ligand interactions. At first, a set of features from the binding ligand was identified. The feature types that considered during generation process are HBA, HBD, hydrophobic groups, positive and negative ionizable groups, and aromatic rings. The features that matched the interactions between the binding ligand and the protein were used to build the pharmacophore models. Excluded volumes were added based on the locations of atoms on the protein. Second, the pharmacophore models were ranked based on a measure of sensitivity and specificity and the top models are returned. The pharmacophore models were validated based on a genetic function approximation (GFA) model.²⁵ The resulted 3D pharmacophore model consisted of three features: two HBA and one hydrophobic center (Figure 4). The generated model was used to examine the tested compounds as possible SARS-CoV-2 helicase inhibitors.

Figure 4.

(a) The generated hypothetical 3D-pharmacophore geometry with four features; two HBA (red) and one hydrophobic (yellow). (b) Mapping of the co-crystallized ligand on the generated pharmacophore (Fit value = 36.08). (c) The calculated distances between the different features.

Activity prediction

The test set of the selected 30 compounds and the co-crystallized ligand (VXG) were examined against the generated 3D-pharmacophore model. In this process, fit values (that quantitatively represent the existence of that essential features) between the tested compound and the generated 3D-pharmacophore were generated.^26,27

In total, 13 compounds (14, 52, 58, 123, 124, 125, 127, 172, 234, 235, 246, 288, and 290) were proved to have the main essential features of SARS-CoV-2 helicase inhibitor (Table 2). Interestingly, some compounds showed high fit values. Figure 5 showed the mapping of the preferred compounds against the generated 3D-pharmacophoric model.

Table 2.

Essential features, fit values, and relative fit of the tested compounds and the co-crystallized ligand (VXG) based on the generated pharmacophore model.

Ligand	No. of essential features	Pharmacophore fit score	Relative fit^b
Pharmacophoric queries	4	–	–
14	3	37.36	1.04
52	3	38.23	1.06
58	3	36.90	1.02
123	3	37.12	1.03
124	3	37.29	1.03
125	4	38.25	1.06
127	3	36.99	1.03
172	3	35.04	0.97
234	3	38.19	1.06
235	3	37.44	1.04
236	4	38.40	1.06
288	3	38.15	1.06
290	3	37.72	1.05
VXG	3	36.08	1.00

Hydrogen bond acceptor Hydrophobic center

Relative Fit = Fit value of a compound/Fit value of co-crystallized ligand (VXG).

Figure 5.

Mapping of the tested compounds on the generated pharmacophore. (a) Compound 14 (Fit value = 37.36), (b) compound 52 (Fit value = 38.23), (c) compound 58 (Fit value = 36.90), (d) compound 123 (Fit value = 37.12), (e) compound 124 (Fit value = 37.29), (f) compound 125 (Fit value = 38.25), (g) compound 127 (Fit value = 36.99), (h) compound 172 (Fit value = 35.04), (i) compound 234 (Fit value = 38.19), (j) compound 235 (Fit value = 37.44), (k) compound 236 (Fit value = 38.40), and (l) compound 288 (Fit value = 38.15).

The specified compounds are metabolites of natural sources with proved antiviral potentialities. In detail, isatin or tribulin (14) is an indole alkaloid that was reported from various Isatis species, such as Isatis tinctoria²⁸ and Couroupita guianensis.²⁹ Considerable reports were published about the antiviral activities of isatin and its derivatives^30,31 besides the in silico anti-COVID-19 effect.^32,33

Atropine (52), the famous tropane alkaloid of Solanaceae family,³⁴ inhibited new infectious herpes simplex virus (HSV)-1 production at a concentration of 200 µg mL⁻¹.³⁵ Also, atropine could stop the multiplication of polio, influenza, HSV, and vaccinia viruses.³⁶

The furanocoumarin, isopimpinellin (58), was isolated from different plants of family Apiaceae³⁷ and inhibited murine cytomegalovirus, Sindbis virus, and bacteriophage T4³⁸ in addition to its potentiality to inhibit SARS-CoV 3CL^pro dose-dependently.³⁹ Juglone (123), plumbagin (124), and β-lapachone (125) are naphthoquinone derivatives that are found in Juglandaceae family⁴⁰ and reported to have antiviral activities.^41–43 Isoeleutherol (127), the naphthalene of Lygodium microphyllum⁴⁴ exhibited in silico binding affinity of −6.9 against COVID-19 main protease.⁴⁵ Fulvoplumerien (172) is the iridoid of Plumeria rubr showing an inhibition potential against HIV through the inhibition of reverse transcriptase with an IC₅₀ value of 45 µg mL⁻¹.⁴⁶ Spongiadiol (234), epispongiadiol (235), and isospongiadiol (236) are tetracyclic furanoditerpenes that were isolated from a sponge of Spongia species^47,48 and inhibited HSV-1 (IC₅₀ values were 0.25, 12.5, and 2 µg mL⁻¹), respectively.^49,50

Mycophenolic acid (288) was discovered for the first time in 1893 from the fungus Penicillium glaucum.⁵¹ Excitingly, mycophenolic acid reduced the RNA level of SARS-CoV-2 by 100-fold with an EC₅₀ of 0.87 µM.⁵² The isolation of flutamide (290) from Delitschia confertaspora and its potential against influenza virus through the inhibition of flutranscription endonuclease was reported.⁵³

Docking studies

The molecular docking process was conducted for the most favored 13 compounds against SARS-CoV-2 helicase (PDB codes: 5RMM) using MOE 19.0901 software as described in previous reports.^54–59 In this test, a flexible docking procedure was applied for both the co-crystallized ligand and the tested molecules at the same pocket of the co-crystallized ligand with the same dimensions.

Validation of docking process

A redocking process of (VXG) in the active site of SARS-CoV-2 helicase has been conducted and showed a root-mean-square deviation (RMSD) value of 0.66 Å indicating a valid docking algorithm (Figure 6).

Figure 6.

Superimposition of the docked and original poses of the co-crystallized ligand in SARS-CoV-2 helicase active site.

Molecular docking of the Crystal ligand

The binding mode of (VXG) showed an energy binding of −6.60 kcal mol⁻¹ against SARS-CoV-2 helicase (Table 3). Inside out, the 4-phenylpyrrolidine moiety formed two pi–pi and pi–alkyl interactions with His554. Furthermore, the 3-carboxylate group interacted with Ser486, Asn516, and Asn177 by four hydrogen bonds with distances of 1.80, 2.04, 1.94, and 1.84 Å, respectively (Figure 7).

Table 3.

Docking score of 13 compounds against SARS-CoV-2 helicase (PDB ID: 5RMM).

Comp.	RMSD value (Å)	Docking score (kcal mol⁻¹)	Interactions
Comp.	RMSD value (Å)	Docking score (kcal mol⁻¹)	Hydrogen bond	Pi interactions
VXG	1.16	−6.60	4	2
14	1.42	−6.30	4	3
52	1.70	−7.30	3	-
58	1.26	−7.01	3	2
123	1.98	−6.70	4	1
124	1.99	−6.80	4	3
125	1.89	−7.80	1	3
127	1.62	−7.50	1	3
172	1.32	−6.70	2	2
234	1.62	−8.70	4	4
235	2.25	−5.77	3	3
236	1.90	−8.65	1	6
288	1.63	−7.80	5	2
290	1.61	−6.79	4	1

Figure 7.

VXG docked in SARS-CoV-2 helicase, hydrogen bonds (green), and the pi interactions (purple lines) with mapping surface showing VXG occupying the protein’s active pocket.

Molecular docking of the selected compounds

Compound 14 exhibited an energy binding of −6.30 kcal mol⁻¹ against SARS-CoV-2 helicase. More closely, the indoline ring created three pi–pi stacked interactions with Tyr515 and His554. In addition, the two C=O moieties interacted with His554, Asn516, and Asn177 by three hydrogen bonds with distances of 2.92, 3.36, and 2.45 Å, respectively. Finally, the amino group formed one hydrogen bond with Ser485 with a distance of 2.20 Å (Figure 8).

Figure 8.

Compound 14 docked in SARS-CoV-2 helicase, hydrogen bonds (green), and the pi interactions (purple lines) with mapping surface showing 14 occupying the protein’s active pocket.

The binding mode of candidate 52 exhibited an energy binding of −7.30 kcal mol⁻¹ against SARS-CoV-2 helicase. The nitrogen atom of 8-azabicyclo[3.2.1]octan moiety created one hydrogen bond with Pro514 with a distance of 2.30 Å besides an ionic interaction with Asp534. Moreover, the hydroxyl group of 3-hydroxy-2-phenylpropanoyl moiety formed two hydrogen bonds with Asn516 and Asn177 with distances of 2.68 and 2.20 Å, respectively (Figure 9).

Figure 9.

Compound 52 docked in SARS-CoV-2 helicase, hydrogen bonds (green), and the pi interactions (purple lines) with mapping surface showing 52 occupying the protein’s active pocket.

The binding mode of compound 58 exhibited an energy binding of −7.01 kcal mol⁻¹ against SARS-CoV-2 helicase. In detail, the 7H-furo[3,2-g]chromen ring formed two pi–pi interactions with His554 and one hydrogen bond with Asn516 with a distance of 2.36 Å. Also, the carbonyl group of 7H-furo[3,2-g]chromen-7-one interacted with Asn516 and Ser486 by two hydrogen bonds with distances of 2.11 and 1.80 Å, respectively (Figure 10).

Figure 10.

Compound 58 docked in SARS-CoV-2 helicase, hydrogen bonds (green), and the pi interactions (purple lines) with mapping surface showing 58 occupying the protein’s active pocket.

Compound 124 exhibited occupied the active site of SARS-CoV-2 helicase with an energy binding of −6.80 kcal mol⁻¹. In further detail, the 2-methylnaphthalene moiety formed three pi–pi and pi–alkyl interactions with Pro175 and His554. Besides, the two C=O groups interacted with Asn516 and Ser486 by two hydrogen bonds with distances of 3.01 and 2.62 Å, respectively. Finally, the hydroxyl group interacted with Ser485 and Ser486 by two hydrogen bonds with distances of 2.02 and 2.03 Å, respectively (Figure 11).

Figure 11.

Compound 124 docked in SARS-CoV-2 helicase, hydrogen bonds (green), and the pi interactions (purple lines) with mapping surface showing 124 occupying the protein’s active pocket.

Regarding compound 172, it exhibited an energy binding of −6.70 kcal mol⁻¹ against SARS-CoV-2 helicase. Particularly, the but-2-en-1-ylidene moiety formed two pi–alkyl interactions with Leu412 and Pro408. In addition, the carbonyl group in 4-carboxylate moiety interacted with Asn516 and Ser486 by two hydrogen bonds with distances of 3.48 and 2.49 Å, respectively (Figure 12).

Figure 12.

Compound 172 docked in SARS-CoV-2 helicase, hydrogen bonds (green), and the pi interactions (purple lines) with mapping surface showing 172 occupying the protein’s active pocket.

With regard to the binding mode of compound 234, it showed binding energy of −8.70 kcal mol⁻¹ against SARS-CoV-2 helicase. In detail, the decahydrophenanthro[1,2-c]furan moiety formed four pi–alkyl interactions with His554, Tyr180, and Pro175. Also, the hydroxyl group of the 7-hydroxy decahydrophenanthronyl moiety interacted with Asn177 and Ser485 by two hydrogen bonds with distances of 2.36 and 2.64 Å, respectively. Furthermore, the carbonyl group of the decahydrophenanthro[1,2-c]furan-8(4H)-one moiety interacted with Asn516 and Asn177 by two hydrogen bonds with distances of 2.34 and 2.33 Å, respectively (Figure 13).

Figure 13.

Compound 234 docked in SARS-CoV-2 helicase, hydrogen bonds (green), and the pi interactions (purple lines) with mapping surface showing 234 occupying the protein’s active pocket.

Concerning compound 288, it docked into the active site of SARS-CoV-2 helicase exhibiting binding energy of −7.80 kcal mol⁻¹. In depth, the methyl group of the 7-methyl-1,3-dihydroisobenzofuran-5-yl ring created a pi–alkyl interaction with the Pro175 amino acid. As well, the hydroxyl and carbonyl groups of the 4-hydroxy-3-oxo-1,3-dihydroisobenzofuran-5-yl ring interacted with Ser486 and Asn516 by three hydrogen bonds with distances of 1.84, 3.45 and 2.38 Å, respectively. Likewise, the carboxylate moiety interacted with Asn179 and Thr532 by two hydrogen bonds with distances of 1.83 and 2.31 Å, respectively (Figure 14).

Figure 14.

Compound 288 docked in SARS-CoV-2 helicase, hydrogen bonds (green), and the pi interactions (purple lines) with mapping surface showing 288 occupying the protein’s active pocket.

Pharmacokinetic profiling study

Seven compounds (14, 52, 58, 124, 172, 234, and 288) that showed good fitting and binding mode at the active pocket of SARS-CoV-2 helicase were subjected to in silico physicochemical profiling assessment according to Lipinski’s and Veber’s rules.

Lipinski’s rule describes the oral absorption of a drug if the molecule fulfills at least three of the four principles (HBD ⩽ 5; HBA ⩽ 10; molecular weight < 500; LogP < 5). In addition, Veber’s rule deals with the bioavailability of a compound if the molecule fulfills its two principles: rotatable bonds < 10 and polar surface area ⩽ 140 Å. Compounds with 10 or fewer have a high probability of good oral bioavailability.^60,61 As shown in Table 4, all tested molecules and reference ligand VXG obeyed Lipinski’s and Veber’s rules.

Table 4.

Pharmacokinetic profiling of the tested molecules according to Lipinski’s and Veber’s rules.

Compound	Lipinski’s rule of five					Veber’s rule
Compound	LogP	Mole. Wt.	HBD	HBA	Violationof Lipinski’s rule	Number of rotatable bonds	TPSA
14	0.67	147.131	1	3	0	0	46.17
52	1.721	289.369	1	4	0	5	49.77
58	2.17	246.215	0	5	0	2	57.9
124	1.962	188.179	1	3	0	0	54.37
172	1.597	244.243	0	4	0	3	52.6
234	3.157	320.337	2	6	0	6	93.06
288	2.24	332.434	2	4	0	1	70.67
VXG	0.711	233.263	1	4	0	2	57.61

HBA: hydrogen bond acceptors; HBD: hydrogen bond donors.

Computational ADMET studies

Six ADMET parameters were assessed using Discovery studio software using amenamevir (reference drug) as described in the literature.^62–66 The results were summarized in Table 5 and Figure 15. The results revealed that blood penetration levels of all compounds are ranging from medium to low, indicating less possibility for CNS side effects. Additionally, all compounds had good or optimal solubility levels. Furthermore, all the tested molecules showed good absorption levels and appeared to be non-inhibitors for cytochrome P450. Concerning hepatotoxicity, compounds 52, 124, 234, and 288 were predicted to be non-hepatotoxic, unlike compounds 14, 58, and 172. Finally, all compounds except compound 288 were predicted to bind plasma protein by less than 90%. These results indicated that compounds 52, 124, 234, and 288 have good pharmacokinetic properties and accordingly were preferred for further investigations.

Table 5.

Predicted ADMET for the elected metabolites and the reference compound.

Comp.	BBB level^a	Solubility level^b	CYP2D6 prediction^d	Hepatotoxicity^e	PPB prediction^f
14	3	4	F	T	F
52	3	4	F	F	F
58	2	3	F	T	F
124	2	3	F	F	F
172	2	3	F	T	F
234	3	3	F	F	F
288	3	3	F	F	T
Amenamevir	4	3	F	T	T

Blood-brain barrier (BBB) penetration level, 4 = very low and 0 = very high.

4 is optimal while 1 is very low.

0 is good, and 3 is very poor.

CYP2D6, the inhibition of cytochrome P2D6, T = inhibitor, F = non inhibitor.

Hepatotoxicity, T = toxic, F = non-toxic.

Plasma protein-binding (PBB), binding with plasma protein, F: < 90%, T: > 90%

Figure 15.

The expected ADMET study.

Computational toxicity studies

Virtual toxicity studies against different toxicity models have been carried out using Discovery studio software as described before⁶⁷ using amenamevir as a reference molecule. The results were summarized in Table 6 and detailed in the Supplementary data.

Table 6.

Toxicity properties of compounds.

Comp.	FDA rodent carcinogenicity	TD₅₀^a	MTD^b	LD₅₀^b	LOAEL^b	Ocular irritancy^c	Skin irritancy^c
52	Non-carcinogen	208.906	0.118	0.427	0.010	N	I
124		846.689	0.254	0.765	0.175	I	N
234		6.115	0.074	0.755	0.031	I	I
288		350.730	0.096	0.586	0.042	N	I
Amenamevir		4.690	0.026	3.077	0.034	I	N

TD: tolerated dose; MTD: maximum tolerated dose

Unit: mg kg⁻¹ day⁻¹.

Unit: g kg⁻¹.

I = irritant, N = non-irritant.

Compound 234 was eliminated because it was predicted to have carcinogenic potency. Furthermore, it showed low carcinogenic potency (TD₅₀) and low rat maximum tolerated dose (MTD) values of 6.115 mg kg⁻¹ day⁻¹ and 0.074 g kg⁻¹, respectively.

However, compounds 52, 124, and 288 showed high TD₅₀ values of 208.906, 846.689, and 350.730 mg kg⁻¹ day⁻¹, respectively. Also, it exhibited high MTD values of 0.118, 0.254, and 0.096 g kg⁻¹, respectively. All compounds showed rat oral LD₅₀ values ranging from 0.427 to 0.765 g kg⁻¹ body weight which were less than the reference drug (LD₅₀ = 3.077 g kg⁻¹). As regards the rat chronic lowest-observed adverse-effect level (LOAEL) model, compounds 124, 234, and 288 showed high values compared to the reference molecule. Finally, compounds 52 and 288 were predicted to be non-irritant against the ocular irritancy model, while compound 124 was anticipated to be non-irritant against skin irritancy. Based on the aforementioned results, compounds 52, 124, and 288 are expected to have a low toxicity profile and were selected out for further studies.

MD simulations studies

Molecular dynamics (MD) simulations can provide plentiful information regarding the dynamical and structural changes that happen on a protein after binding to a ligand. Also, it affords generous energetic information about the protein–ligand complex.⁶⁸ The obtained results could be very useful to understand the interactions between the ligand and the target protein on an atomic resolution.⁶⁹

The conformational changes and variations, and the atomic dynamic movements of backbone atoms in the mycophenolic acid–SARS-CoV-2 helicase complex were computed by RMSD to verify the stability of the mycophenolic acid–SARS-CoV-2 helicase on both apo and ligand bonded states. It was noticed that mycophenolic acid, SARS-CoV-2 helicase, and the mycophenolic acid–SARS-CoV-2 helicase complex exhibited very low RMSD values without major fluctuations demonstrating the great stability (Figure 16(a)). The flexibility of SARS-CoV-2 helicase, during the 100 ns of the experiment, was computed in the term of RMSF to investigate closely the region of proteins that have been fluctuated during the MD simulation. The root-mean-square fluctuation (RMSF) denoted that the binding of mycophenolic acid makes the SARS-CoV-2 helicase slightly flexible in 150–250 residue areas (Figure 16(b)). To study the compactness of mycophenolic acid–SARS-CoV-2 helicase complex, the radius of gyration (R_g) was investigated. Figure 16(c) designated the decrease of the R_g values of the mycophenolic acid–SARS-CoV-2 helicase complex than the starting period. Such results marking the high degree of compactness of the mycophenolic acid–SARS-CoV-2 helicase complex. Aiming at the analysis the conformational changes occurred during the mycophenolic acid–SARS-CoV-2 helicase interaction, the solvent accessible surface area (SASA) was measured over 100 ns. Interestingly, the SARS-CoV-2 helicase displayed a decrease in the surface area giving relatively low SASA value at the end of experiment than the starting period (Figure 16(d)). Finally, hydrogen bonding in the mycophenolic acid–SARS-CoV-2 helicase was computed. Advantageously, the results verified the high number of conformations of SARS-CoV-2 helicase formed up to eight hydrogen bonds (Figure 16(e)). Such interesting results authenticate the stability in with the mycophenolic acid–SARS-CoV-2 helicase complex.

Figure 16.

MD simulations results; (a) RMSD values of mycophenolic acid, SARS-CoV-2 helicase complex and mycophenolic acid–SARS-CoV-2 helicase complex. (b) RMSF, (c) R_g, (d) SASA of SARS-CoV-2 helicase, and (e) hydrogen bonding mycophenolic acid–SARS-CoV-2 helicase complex in the MD run.

Molecular Mechanics Poisson–Boltzmann Surface Area results

The binding-free energy of mycophenolic acid–SARS-CoV-2 helicase complex was computed over the last 20 ns of the MD run with a 100-ps interval using Molecular Mechanics Poisson–Boltzmann Surface Area (MM/PBSA) method. Mycophenolic acid displayed a low binding-free energy value of −294 kJ mol⁻¹ with SARS-CoV-2 helicase (Figure 17(a)). Furthermore, the total binding-free energy of the mycophenolic acid–SARS-CoV-2 helicase complex was disintegrated into per amino acid residue share energy. The output of that experiment gave us further details about the primary (essential) residues that share favorably in the binding of mycophenolic acid to the SARS-CoV-2 helicase. ARG-178, LYS-202, LYS-414, and ARG-560 residues of SARS-CoV-2 helicase were found to be the higher contributors with free energies that more than −20 kJ mol⁻¹ (Figure 17(b)).

Figure 17.

MM/PBSA results of mycophenolic acid–SARS-CoV-2 helicase complex; (a) free binding energy over the 100 ns, and (b) crucial amino acids that were contributed in the mycophenolic acid–SARS-CoV-2 helicase complex.

In 1969, mycophenolic acid exhibited significant in vitro antiviral activities against DNA and RNA viruses using agar diffusion plaque reduction and tube culture methods. The inhibition power was noticed against vaccinia virus, adenovirus, herpes, and measles.⁷⁰ Mycophenolic acid protected the virus-infected cells with 100% at a concentration of 10 μg mL⁻¹.⁷¹ In another experiment, it was active against orthopoxviruses through the inhibition of the viral cellular inosine monophosphate dehydrogenase.⁷² As well, it inhibited the viral RNA replication of the dengue virus.⁷³ In addition, mycophenolic acid inhibited in vitro and in vivo hepatitis C virus (HCV) infection and augmented interferon-stimulated gene expression.⁷⁴ Besides, the antiviral activity of 288 against rotavirus,⁷⁵ hepatitis E⁷⁶ and B⁷⁷ viruses, HIV (in vivo and in vitro),^78,79 the Japanese encephalitis virus,⁸⁰ viral hemorrhagic septicemia, and the infectious pancreatic necrosis viruses.⁸¹ Interestingly, mycophenolic acid and 6-Mercaptopurine/6-thioguanine inhibited MERS-CoV papain-like protease in a synergistic way.⁸² Surprisingly, mycophenolic acid decreased the viral RNA level of SARS-CoV-2 by 100-fold with an EC₅₀ of 0.87 µM.⁵²

Method

Molecular similarity detection

Discovery studio 4.0 software was used for 310 natural metabolites against VXG, the co-crystallized ligand of SARS-CoV-2 helicase protein (PDB ID: 5RMM) to select the most similar 30 metabolites (see method part in Supplementary data).

Pharmacophore studies

Discovery studio 4.0 software was used for the most similar 30 metabolites as (see method part in Supplementary data).

Docking studies

Docking studies were done for the selected 13 metabolites and VXG against SARS-CoV-2 helicase protein (PDB ID: 5RMM) using Discovery studio and MOE softwares⁸³ (see method part in Supplementary data).

ADMET analysis

Discovery studio 4.0⁸⁴ was used for the selected seven metabolites (see method part in Supplementary data).

Toxicity studies

Discovery studio 4.0⁸⁵ was used for the selected four metabolites (see method part in Supplementary data).

Molecular dynamics simulation

The mycophenolic acid–SARS-CoV-2 helicase system was prepared using the web-based CHARMM-GUI^86–88 interface using CHARMM36 force field⁸⁹ and NAMD 2.13⁹⁰ package. The TIP3P explicit solvation model was used (see supporting data).

MM/PBSA studies

The g_mmpbsa package of GROMACS⁹¹ was used to calculate the MM/PBSA of mycophenolic acid–SARS-CoV-2 helicase system (see supporting data).

Conclusion

This research used a set of 310 reported antiviral natural compounds as an examination set to single out the most convenient SARS-CoV-2 helicase inhibitor. The chemical structural similarity test against the co-crystallized ligand assigned the most similar 30 compounds. Then, the study of the pharmacophoric features selected the most suitable 13 compounds. Next, the molecular docking against the SARS-CoV-2 helicase protein (PDB ID: 5RMM) filtered seven candidates. Furthermore, the ADMET and toxicity studies elected three metabolites. The Density functional theory (DFT) study suggested mycophenolic acid (288) as the most convenient SARS-CoV-2 helicase inhibitor. The MD simulation studies over 100 ns (RMSD, RMSF, Rg, SASA, and H-bonding) verified the right binding of mycophenolic acid to SARS-CoV-2 helicase. Further in vitro and in vivo studies should be conducted on mycophenolic acid as a potential treatment for SARS-CoV-2 through the inhibition of helicase protein.

Supplemental Material

sj-pdf-1-chl-10.1177_17475198231221253 – Supplemental material for Computer-aided drug discovery of natural antiviral metabolites as potential SARS-CoV-2 helicase inhibitors

Supplemental material, sj-pdf-1-chl-10.1177_17475198231221253 for Computer-aided drug discovery of natural antiviral metabolites as potential SARS-CoV-2 helicase inhibitors by Eslam B Elkaeed, Ibrahim H Eissa, Abdulrahman M Saleh, Bshra A Alsfouk and Ahmed M Metwaly in Journal of Chemical Research

Footnotes

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This research was funded by Princess Nourah bint Abdulrahman University Researchers Supporting Project number (PNURSP2024R142), Princess Nourah bint Abdulrahman University, Riyadh, Saudi Arabia. The authors extend their appreciation to the Research Center at AlMaarefa University for funding this work.

ORCID iD

Ahmed M Metwaly

Supplemental material

Supplemental material for this paper is available online.

References

WHO. WHO Coronavirus (COVID-19) Dashboard, https://covid19.who.int/ (2023, accessed 23 July 2023).

Macalino

SJY

Gosu

Hong

, et al. Archiv Pharma Res 2015; 38: 1686–1701.

Zhang

Drug Design Discov 2011; 716: 23–38.

Sliwoski

Kothiwale

Meiler

, et al. Pharmacol Rev 2014; 66: 334–395.

El-Zahabi

Sakr

El-Adl

, et al. Bioorgan Chem 2020; 104: 104218.

Jalmakhanbetova

Suleimen

Oyama

, et al. J Chem 2021; 2021

Alesawy

Abdallah

Taghour

, et al. Molecules 2021; 26: 2806.

El-Demerdash

Metwaly

Hassan

, et al. Biomolecules 2021; 11: 460.

Metwaly

Ghoneim

Eissa

, et al. Saudi J Biol Sci 2021; 28: 5823–5832.

10.

Han

Yang

Metwaly

, et al. J Ethnopharmacol 2019; 239: 111942.

11.

Elkaeed

Youssef

Eissa

, et al. Int J Mol Sci 2022; 23: 6912.

12.

Elkaeed

Eissa

Elkady

, et al. Int J Mol Sci 2022; 23: 8407.

13.

Metwaly

Ghoneim

Eissa

, et al. Saudi J Biol Sci 2021; 28: 5823–5832.

14.

Elkaeed

Metwaly

Alesawy

, et al. Life 2022; 12: 1407.

15.

Elkaeed

Elkady

Belal

, et al. Processes 2022; 10: 530.

16.

Eissa

Alesawy

Saleh

, et al. Molecules 2022; 27: 2287.

17.

Metwaly

Elwan

El-Attar

A-AMM

, et al. J Chem 2022; 2022: 1–23.

18.

Habtemariam

Nabavi

Banach

, et al. Archiv Med Res 2020; 51: 733–735.

19.

Bender

Glen

RC.

Organic Biomol Chem 2004; 2: 3204–3218.

20.

Eckert

Bajorath

Drug Discover Today 2007; 12: 225–233.

21.

Schaller

Šribar

Noonan

, et al. Comput Mol Sci 2020; 10: e1468.

22.

Kim

K-H

Kim

Seong

B-L.

Expert Opinion Drug Disc 2010; 5: 205–222.

23.

Seidel

Wieder

Garon

, et al. Mol Inform 2020; 39: 2000059.

24.

Liu

Yin

Yao

, et al. Front Cell Infect Microbiol 2020; 10: 118.

25.

Zhang

Xiang

M-L

Liang

J-Y

, et al. Mol Divers 2013; 17: 767–772.

26.

Jade

Pandey

Kumar

, et al. J Biomol Struct Dynam 2020; 2020: 1–17.

27.

Temml

Kaserer

Kutil

, et al. Future Med Chem 2014; 6: 1869–1881.

28.

Marcelo

Gontier

Dauwe

Phytochemistry 2019; 163: 89–98.

29.

Begum

Motobayashi

Hasan

, et al. Agronomy 2020; 10: 1388.

30.

Jarrahpour

Khalili

De Clercq

, et al. Molecules 2007; 12: 1720–1730.

31.

Varma

Nobles

WL.

J Pharmaceut Sci 1975; 64: 881–882.

32.

Liu

Sun

, et al. Europ J Med Chem 2020; 206: 112702.

33.

Boozari

Hosseinzadeh

Phytotherapy Res 2021; 35: 864–876.

34.

Dimitrov

Metcheva

Boyadzhiev

Separat Purif Tech 2005; 46: 41–45.

35.

Alarcón

González

Carrasco

Antimicrobial Agent chemo 1984; 26: 702–706.

36.

Yamazaki

Tagaya

J General Virol 1980; 50: 429–431.

37.

Peroutka

Schulzová

Botek

, et al. J Sci Food Agricult 2007; 87: 2152–2163.

38.

Hudson

Towers

Photochem Photobiol 1988; 48: 289–296.

39.

Park

J-Y

J-A

Kim

, et al. J Enzyme Inhibit Med Chem 2016; 31: 23–30.

40.

Hirakawa

Ogiue

Motoyoshiya

, et al. Phytochemistry 1986; 25: 1494–1495.

41.

Vardhini

SR.

J Recep Signal Trans 2014; 34: 456–457.

42.

Susithra

Chamundeswari

Srikanth

, et al. J Ravishankar Univ 2016; 29.

43.

Karthikeyan

Med Aromatic Plant 2016; 5: 239–246.

44.

Kuncoro

Farabi

Rijai

, et al. Makara J Sci 2018; 22: 4.

45.

Prasetio

Kepel

Bodhi

, et al. eBiomedik 2021; 9: 101–106.

46.

Tan

Pezzuto

Kinghorn

, et al. J Nat Prod 1991; 54: 143–154.

47.

Yong

Garson

Bernhardt

PV.

Acta Crystall Sec C: Crystal Struct Commun 2009; 65: o167–o170.

48.

Minale

Marine Natur Product Chem Biol Perspect 2012; 1: 175–240.

49.

Che

CT.

Drug Develop Res 1991; 23: 201–218.

50.

Donia

Hamann

MT.

Lancet Infect Dis 2003; 3: 338–348.

51.

Bentley

Chem Rev 2000; 100: 3801–3826.

52.

Kato

Matsuyama

Kawase

, et al. Microbiol Immunol 2020; 64: 635–639.

53.

Hensens

Goetz

Liesch

, et al. Tetrahedron letters 1995; 36: 2005–2008.

54.

El-Adl

Sakr

Yousef

, et al. Bioorgan Chem 2021; 114: 105105.

55.

Eissa

Ibrahim

Metwaly

, et al. Bioorgan Chem 2021; 107: 104532.

56.

El-Adl

El-Helby

A-GA

Ayyad

, et al. Bioorgan Med Chem 2021; 29: 115872.

57.

Eissa

El-Helby

A-GA

Mahdy

, et al. Bioorgan Chem 2020; 105: 104380.

58.

Nasser

Eissa

Oun

, et al. Organic Biomol Chem 2020; 18: 7608–7634.

59.

Abbass

Khalil

Mohamed

, et al. Bioorgan Chem 2020; 104: 104255.

60.

Veber

Johnson

Cheng

H-Y

, et al. J Med Chem 2002; 45: 2615–2623.

61.

Lipinski

Lombardo

Dominy

, et al. Adv Drug Deliver Rev 2012; 64: 4–17.

62.

Parmar

Soni

Guduru

, et al. Bioorgan Chem 2021; 115: 105206.

63.

Yousef

Sakr

Eissa

, et al. New J Chem 2021; 45: 16949–16964.

64.

Alesawy

Al-Karmalawy

Elkaeed

, et al. Archiv Der Pharmazie 2021; 354: 2000237.

65.

El-Metwally

Abou-El-Regal

Eissa

, et al. Bioorgan Chem 2021; 112: 104947.

66.

Eissa

Dahab

Ibrahim

, et al. Bioorgan Chem 2021; 112: 104965.

67.

Suleimen

Metwaly

Mostafa

, et al. J Chem 2021; 2021: 1–10.

68.

Sousa

Fernandes

Ramos

MJ.

Proteins: Struct Funct Bioinform 2006; 65: 15–26.

69.

Hollingsworth

Dror

RO.

Neuron 2018; 99: 1129–1143.

70.

Planterose

J General Virol 1969; 4: 629–630.

71.

Cline

Nelson

Gerzon

, et al. Appl Microbiol 1969; 18: 14–20.

72.

Smee

Bray

Huggins

JW.

Antiviral Chem Chemother 2001; 12: 327–335.

73.

Diamond

Zachariah

Harris

Virology 2002; 304: 211–221.

74.

Pan

de Ruiter

Metselaar

, et al. Hepatology 2012; 55: 1673–1683.

75.

Yin

Wang

Dang

, et al. Antiviral Res 2016; 133: 41–49.

76.

Wang

Zhou

Debing

, et al. Gastroenterology 2014; 146: 1775–1783.

77.

Gong

De Meyer

Clarysse

, et al. J Viral Hepatit 1999; 6: 229–236.

78.

Chapuis

Rizzardi

D’agostino

, et al. Nature Med 2000; 6: 762–768.

79.

Kaur

Klichko

Margolis

AIDS Res Human Retrovirus 2005; 21: 116–124.

80.

Sebastian

Madhusudana

Ravi

, et al. Chemotherapy 2011; 57: 56–61.

81.

Marroquí

Estepa

Perez

Antiviral Res 2008; 80: 332–338.

82.

Cheng

K-W

Cheng

S-C

Chen

W-Y

, et al. Antiviral Res 2015; 115: 9–16.

83.

Baez-Santos

Mesecar

2021, https://www.rcsb.org/structure/4OW0.

84.

Metwaly

Elkaeed

Alsfouk

, et al. Plants 2022; 11: 1886.

85.

Suleimen

Jose

Mamytbekova

, et al. Plants 2022; 11: 2072

86.

Kim

Iyer

, et al. J Comput Chem 2008; 29: 1859–1865.

87.

Brooks

III Mackerell

Jr. , et al. J Comput Chem 2009; 30: 1545–1614.

88.

Lee

Cheng

Swails

, et al. J Chem Theory Comput 2016; 12: 405–413.

89.

Best

Zhu

Shim

, et al. J Chem Theory Comput 2012; 8: 3257–3273.

90.

Phillips

Braun

Wang

, et al. J Comput Chem 2005; 26: 1781–1802.

91.

Suleimen

Jose

Suleimen

, et al. Molecules 2022; 27: 1216.

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

3.19 MB