key: cord-0813033-hcp7v7eu
authors: Muhammad, Shabbir; Hira Hassan, Sayyeda; Al-Sehemi, Abdullah G.; Abdullah Shakir, Hafiz; Khan, Muhammad; Irfan, Muhammad; Iqbal, Javed
title: Exploring the New Potential Antiviral Constituents of Moringa oliefera for SARS-COV-2 Pathogenesis: An in Silico Molecular Docking and Dynamic Studies
date: 2021-01-26
journal: Chem Phys Lett
DOI: 10.1016/j.cplett.2021.138379
sha: 7c0ad161738a2480292eafae1e9c171334f46427
doc_id: 813033
cord_uid: hcp7v7eu

The interactions of two crucial proteins of COVID-19 have been investigated with potential antiviral compounds from Moringa oliefera using quantum chemical, molecular docking and dynamic methods. The results of the present investigation show that ellagic acid and apigenin possess the highest binding affinities of -7.1 and -6.5 Kcal.mol-1against nsp9 and -6.9 and -7.1 Kcal.mol-1 against nsp10, respectively. The dynamic behavior of individual proteins and their respective best docked ligand-protein complexes are also studied at 30ns timescale. Both of these compounds also show the highest intestinal absorption and total clearance rate as compared to the other compounds under present investigation without any toxicity.

In December 2019, the virus SARS-CoV-2 emerged in Wuhan city of Hubei province of China and spread speedily throughout the world [1] . The world health organization (WHO) declared it as pandemic on March 12, 2020 [2] . The worldwide number of COVID-19 cases and deaths has reached to 95,482,598 and 2,039,653 respectively as reported by January 17, 2021 [3] . This novel virus belongs to the family coronaviridae and spherical in shape. It spreads primarily via saliva, droplets, or nasal discharges of an infected individual after coughing or sneezing [4] . These droplets can spread upto 1-2 m and can remain viable on surfaces for days [5] .

Clinical features of COVID-19 ranges from asymptomatic state to severe respiratory distress syndrome. The common clinical symptoms include fever, sore throat, cough, fatigue, headache and breathlessness [6] . The laboratory result of patients with severe COVID-19 presents increased neutrophil count, the most abundant type of white blood cells in human body [7] . They play a major role in line of defense as the first responders against acute infection. Neutrophil chemotaxis is stimulated by chemical signals such as interleukin-6/8 (IL-6/8) and recruit to the site of infection where they assemble and release more pro-inflammatory molecules which in response stimulate further inflammatory signals, progressing to cytokine storms that leads to failure of vital organs [8] . Furthermore, a large number of neutrophils were reported to found in alveolar cavities of serious patients suffering from COVID-19, which validates their roles in cytokine storms [9] . In this way, aberrant neutrophil activity and cytokine response contributes to SARS-CoV-2 pathogenesis in COVID-19. It is reported that two non-structural proteins 9/10(nsp9/nsp10) promote the IL-6/IL-8 mediated neutrophils chemotaxis and stimulate the host inflammatory response as observed in case of COVID-19 [10] . In a study, two anti-inflammatory drugs, pemirolast and eriodictyol were docked against nsp9 with the resulting binding energy of -6.5 kcal/mol and further suggest their role in inhibiting infection. It has been further reported that retonavir and remdesivir are potential drugs against nsp10, suggesting their ability to inhibit viral transcription [11] . Thus, both proteins can be proved as a potential therapeutic target in combating SARS-CoV-2 pathogenesis.

Today, computational drug discovery program has been gaining popularity and success because of its potential to discover robust molecule prior to its synthesis, whereas any traditional drug discovery approach is expensive and may take decades to accomplish [12, 13] . It is not affordable to wait till the discovery of any potential drug for COVID-19. Molecular docking is considered more useful in designing, evaluation and comparison of new drugs as it allows the examination of interaction of molecules in 3-dimensional (3D) space by considering the different forms of molecules and determines the factors that play role in vital pharmacological interaction [14] . Other approaches including in silico ADMET analysis, prediction of drug likeness and toxicity are also adopted to analyze potential drugs from several databases. These computational mediated approaches save the experimental cost and time in the process of drug discovery [15] . Moreover, frontier molecular orbitals (HOMO and LUMO) energies and molecular electrostatic potential (MEP) are widely used to determine their chemical reactivity of ligands. The current study is based on in silico analysis which was aimed to evaluate the interaction of selected components of Moringa oliefera against nsp9 and nsp10. M. oliefera is a highly valued plant because of its medicinal value. This plant has showed its efficacy against HIV, HBV, HSV, EBV, NDV, FMDV and AIDS [16] . The present study will provide several powerful insights for the researchers in quest for therapeutic intervention against COVID-19 pathogenesis.

Autodock Vina (ADV) [17] was used for molecular docking studies to calculate the binding affinity and type of interactions between the ligands and targets (COVID-19 nsp9/10). Ligands and proteins were prepared for docking analysis by using Autodock 1.5.6 tool from MGL Tools (The Scripps Research Institute, Molecular Graphics Laboratory, 10550 North Torrey Pines Road, CA, 92037). The substrate or protein macromolecules were treated as rigid target while ligands were kept as flexible with different rotatable bonds. The quantum chemical studies for the molecular structural analysis of ligands were performed using GAUSSIAN 09 suit of program [18] while orbitals and electrostatic potentials are visualized by GaussView5 [19] . For quantum chemical calculations of ligand structures, their chemical structures were optimized (the lowest energy conformation) by density functional theory (DFT) method with B3LYP/6-311G* level of theory. The substrate or protein macromolecules were treated as rigid target while ligands were kept as flexible with different rotatable bonds. There are various docking protocols whose details are given below:

Crystal structures of nsp9 (PDB ID: 6W4B) and nsp16/10 complex (PDB ID: 6WVN) were retrieved in .pdb format from the Protein Data Bank (https://www.rcsb.org/). PDB is a worldwide archive to access the 3D structure of biological macromolecules [20] . 6WVN is a heterodimer containing nsp16 and nsp10 in a complex form [21] . In this study, we used a single chain of nsp10 for docking analysis. The MGL Tools were used for protein preparation. The water molecules and co-crystallized ligands were deleted from the macromolecule and polar hydrogens were added.

The initial 3D structures of selected ligands were retrieved in .sdf format from PubChem (https://pubchem.ncbi.nlm.nih.gov/). PubChem is a public accessible repository for chemical substances and related biological activities [22] . The optimized structures were then converted into .pdb format.

The PDB files of both ligands and proteins were converted in an extended PDB format, termed PDBQT for performing molecular docking analysis using ADV. Docking was performed with the 

ADMET (Absorption, distribution, metabolism, excretion and toxicity) profile exhibits vital importance in pharmaceutical industry and required to reduce undesired effects [23] . In this study, we used online server of pkCSM to predict drug likeness and ADMET profile. The pkCSM is an open database server which provides information regarding drug pharmacokinetics [24] . The SMILES of the selected ligands were taken from PubChem and submitted to pkCSM. The server database was run by selecting toxicity ADMET module.

Molecular docking was done to depict the mode of action of antiviral constituents of M. oliefera against nsp9 and nsp10. M. oleifera is known as "miracle tree" because it is highly nutritious and has the ability to improve the immune system with antiulcer, antipyretic, antiplasmodic, antioxidant, antibacterial properties [25] . It has been shown that M. oliefera contains flavonoids.

Flavonoids are proved to have antiviral activity, especially against syncytial virus [26, 27] . A total of six ligands were selected based on their reported antiviral activity. Their chemical structures can be seen in Figure S1 . Apigenin, quercetin, myricetin and chrysin are flavonoid constituents of M. oliefera with the reported ability to hinder viral growth. Ellagic acid and chlorogenic acid, being phenolic compounds, has been suggested as anti-HIV/HSV candidates [16, 28, 29] .

The pharmacological properties of ligands are also related to their electronic structures. It is also important to investigate the energy and distribution patterns of their orbitals especially the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO), which are usually call frontier molecular orbitals (FMOs). To shed light on the electronic structural properties of our selected ligands, we have optimized their structures using density functional theory (DFT) method at B3LYP/6-311G* level of theory. The energy gap between HOMO and LUMO orbitals is important factor for assessing thermal stability and reactivity of a molecule [30] .

According to Koopmans' theorem [31] , the HOMO energy presents the ability to donate an electron and the LUMO energy is related to accept an electron. Both of these properties (HOMO and LUMO) have significant roles in thermal stability and reactivity of a compound [32] . The values of HOMO and LUMO energies and their energy gap (ΔE HL ) calculated in aqueous phase can be seen in Figure 1 . Overall, there are no drastic changes found among the ΔE HL values of selected ligands where the maximum difference of 0.29 eV was found between the ΔE HL values of Quercetin and Chrysin. The increasing order of the ΔE HL is Quercetin ˂ Myricetin < Chlorogenic acid < Ellagic acid < Apigenin < Chrysin. It can be seen that Ellagic acid, Apigenin and Chrysin possess higher energy gaps and better thermal stabilities in aqueous phases.

The presence of partial charges on ligand and protein molecules plays a crucial role in leading the docking of ligand with protein. Quantum chemically, the molecular electrostatic potential (MEP) diagram can be helpful to get insights into the 3-dimentional structural and topological features of ligands. Molecular electrostatic potential on the total density surface of a compound indicates whether the effect of nuclei or electrons are dominant over a specific point of molecular geometry [39] . It is important to pen down here that the MEPs are used in a semi quantitative way to study the 3-dimentional structural features. For instance, blue region represents the most positive electrostatic potential, red color shows the areas with the most negative electrostatic potential and green areas with nearly zero potentials. To be more specific, the electrostatic potential decreases in the order of blue ˃ green ˃ yellow ˃ orange ˃ red color coding. In our selected ligands, the most of the negative regions (red) are around carbonyl groups owing to their electron abundance environment, which are also attractive sites for electrophilic attack. On the other hand, more positive regions (blue) are mainly towards the protonic H-atom of hydroxyl groups (OH) which will potentially act as H-bond donors in protein-ligand intermolecular interactions [40] .

The basics of designing rational drug is to utilize the knowledge of the protein-ligand binding mechanisms and structural data to explore the potential of uncovering new drug candidates. So, an extensive understanding of the molecular interaction/recognition nature is also of great value, which will provide insights about designing, developing and discovering drugs [41] . Molecular docking is an extensively used computational tool for predicting binding modes in protein-ligand interaction [42] . In present study, the molecular docking was done by adopting a grid based technique as implemented AutoDock Vina. All six ligands were docked explicitly against nps9 ( Figure S2a ) and nsp10 ( Figure S2b ) of COVID-19. Table 1 revealed that binding affinity of our ligands with COVID-19 nsp9 ranged from -6.1 to -7.1 kcal/mol. The binding affinity of antiviral ligands for nsp10 of COVID-19 was only slightly different from that of nsp9. The binding affinities for nsp10 of COVID-19ranged from -6.8 to -7.1 kcal/mol. Among all compounds, ellagic acid showed the highest binding affinity to nsp9 ( Figure S3 ) with an inhibition constant (k i ) of 5.98 µM. The k i can be used for accurate appreciation of inhibitory potential of the drug. The K i value reflects the amount of drug required to reach the 50% inhibition [43] . Whereas, apigenin exhibited the highest affinity in case of nsp10 ( Figure S4 ) with same k i value as in prior case. These results showed that all selected ligands exhibit good binding affinity with our target molecules. Moreover we selected Remdesivir as a standard drug [44] and Niaziminin, well known for its anticancer properties [45] , docked against the both targeted proteins. The results were compared with our selected antiviral ligands (Table S2) , which further suggest their affinity with the targeted proteins.

To analyze the detailed interaction of best conformation, free version of Biova Discovery Studio 

After completing molecular docking studies by using rigid crystal structure of nsp9 and nsp10, MD simulation was done to analyze the dynamic behavior of targeted proteins and lead compounds. We performed MD simulations for 30 ns (3×10 6 femtosecond) time scale for the best docked protein-ligand complexes comprising of ellagic acid-nsp9 and apigenin-nsp10 complexes.

All molecular dynamic calculations were performed using CHARMM force-field [46] in NAMD [47] while trajectory, RMSD and other analysis were done by VMD program [48] as developed at NIH Center for Macromolecular Modeling and Bioinformatics, Theoretical and Computational Biophysics group. Further details about molecular dynamic calculations and its adopted input parameters can be seen in supporting information of the article. The values of RMSDs for the C-α atoms of apo-forms (only proteins without docked ligands) of nsp9 and nsp10 proteins and their respective complexes were evaluated to analyze thermodynamic conformational stability during 30 ns time period of MD simulations. The RMSD graphs of ellagic acid-nsp9 complex and its apoform nsp9 were shown in Figure 9 (a) while for the apigenin-nsp10 complex and its apo-form nsp10 were shown in Figure 10 (a) . A careful analysis of RMSD plots shows a different trend in the RMSD plots for both proteins and their respective ligand complexes. For example, the RMSD of apo-form of nsp9 indicates that after 1 ns, it is equilibrated from 1 to 30 ns and fluctuates mainly from 1.6 to 2.7Å. Unlike its apo-form protein, the RMSD of ellagic acid-nsp9 complex indicates a different behavior where it rose upto 3.3Å in the first 10 ns and then it is equilibrated from 10 to 30 ns and fluctuates mainly from 3.6 to 4.7Å (see Figure 9 (a)). The RMSD analysis of ellagic acid-nsp9 complex indicates a profound effect of ellagic acid docking, which after 10 ns shows relatively stable ligand-protein complex. Similarly, the RMSD graphs of apigenin-nsp10 complex and its apo-form nsp10 were shown in Figure 10 Figure 9 (c) and 10(c) while their respective single protein structures were also shown in Figure 9 (b) and 10(b). A comparative analysis of structural changes among Figure 9 (b) and 9(c) illustrates that nsp9 show relatively more prominent changes where more fluctuations or randomness is seen towards the right side of nsp9 where ellagic acid is docked. While on the other hands, the structural changes among Figure 10 (b) and 10(c) are less obvious which might be due to less intense interactions/effects of apigenin over nsp10. Somewhat similar findings were also observed in their RMSDs which have shown more fluctuations in ellagic acid-nsp9 complex as compared with apigenin-nsp10 complex. Thus, it might be reasonable to say that the effect of ellagic acid is more obvious on nsp9 as compared to apigenin for nsp10 as studied by molecular docking and molecular dynamic studies here.

A candidate molecule is of no use and does not matter how good it fits with the receptor, if its absorption is poor or if it is excreted too slowly from the body. So, it is always necessary to predict the drug likeness, for which Lipinski's rule of 5 was used. According to this rule, any ligand is considered as a drug-like if it meets certain requirements including molecular weight < 500 Dalton, number of H-bond acceptor < 10, number of H-bond donors < 5 and lipophilicity represented as log P < 5 [49] . All our selected ligands obeyed this rule with just one violation in case of chlorogenic acid and myricetin (H-bond donors > 5) (Table S1) , which strongly suggests their suitability as a drug.

In drug designing process, it is very important to predict human pharmacokinetic properties which help in identification and progression of candidate molecules that can be successful in the clinic [50] .

Absorbance in the small intestine is one of the vital processes for determining the drug bioavailability after oral administration [51] . It is worth mentioning that all selected constituents of M. oliefera have intestinal absorbance more than 30%, especially of apigenin and chrysin, which describes their ability to absorb easily. Moreover, all these constituents showed water solubility (log S) greater than -5, which reflect their solubility in water at 25˚C (Table 2 ) [24] . Our results also showed that none of our selected drug is P-glycoprotein I/II inhibitor which is a very essential aspect in pharmacokinetics study. Any drug that can inhibit these transporters can lead to interfere with the pharmacokinetics of other drugs [52] . So, it can be suggested that these drugs can be safe and effective adjuvant agents with any pharmaceutical drug against Sars-CoV-2 pathogenesis.

The steady state volume of distribution (VDss) shows the dose of drug needed to distribute uniformly as in blood and plasma. It is considered low if log VDss is less than -0.5 and high if it is more than 0.45. Higher VDss value presents greater drug distribution in plasma rather than tissue and lower VDss shows the poor ability of drug to transverse or diffuse across the cell membrane [53] .Our results show that ellagic acid and chrysin are ideal candidates in respect to distribute uniformly in blood plasma (see Table 2 ).

Another factor that can affect the pharmacokinetics interactions is drug metabolizing enzymes.

Cytochrome P450 is an important detoxification enzyme in human body, mainly in liver. Any alteration in activity of these enzymes can lead to disturbance in pharmacokinetics of drugs that are metabolized by these enzymes [54] . There are two most important isoforms of cytochrome P450: cytochrome P2D6 and cytochrome P3A4 [55] . Table 2 shows that those compounds don't inhibit or affect these enzymes. So, it can be interpreted that all these compounds are safe as adjuvant agents with other drugs.

Total clearance quantifies the removal of drug from blood or plasma. Clearance (drug elimination process) results from both kidney and liver [56] . Total clearance Log (CL tot ) predicts combination of renal clearance and hepatic clearance. It is linked to bioavailability and vital to achieve steadystate concentrations by determining dosing rates [24] . A steady-state level is reached when the drug is administered at appropriate concentration and bioavailable. The higher the CL tot value, the faster the excretion process of the compound [55] . Table 2 shows the total clearance values from which their rate of excretion can be predicted. Apigenin and ellagic acid has the highest CL tot value as compared to the others. Moreover, it can be seen that all of the compounds are not substrate of renal OCT2. OCT2 (organic cation trasnsporter 2) is a protein transporter having vital role in renal drug clearance. Substrates of OCT2 can interact with OCT2 inhibitors and may lead to adverse reactions [24] .

Toxicity of the compound can be predicted by AMES toxicity which predicts if any compound has mutagenic potential or not [24] . 

Thus, a combined quantum chemical, molecular docking and dynamic study was performed for nsp9 and nsp10 of COVID-19 with antiviral constituents of M. Oliefera to assess a new potential therapeutic drug candidate. Unlike previous studies, we have firstly studied the structural chemistry of ligands in the form of their optimized geometries, HOMO-LUMO orbital energies, reactivity parameters, the 3-D orbital distribution patterns and MEPs to get insights into the structure-reactivity of ligands. All the potential ligands were docked with nsp9 and nsp10 nonstructural proteins of COVID-19. These results revealed that all selected ligands form the stable complexes with the targeted proteins and showed the highest binding affinity values of apigenin (-7.1 kcal/mol) for nsp10 and ellagic acid (-7.1 kcal/mol) for nsp9, which are in line to our previous quantum chemical results. The dynamic behavior of individual proteins and their respective best docked ligand-protein complexes are also studied at 30ns timescale, which showed different effects of ligands on docking with these proteins as assessed by their respective RMSDs.

Based on quantum chemical results, apigenin (4.17 eV) and ellagic acid (4.32 eV) showed lower HOMO-LUMO energy gaps and predicted for better chemical reactivity. Additionally, ADMET analysis by pkCSM revealed that both of these compounds also have the highest absorption and clearance rate in comparison with other ligands. The present investigation highlighted the significant potential of studied ligands for antiviral activity against SARS-COV-2. These results suggest further in vitro and in vivo studies for establishing a strong experimental evidence of activity of antiviral constituents of M. oliefera against SARS-COV-2.

The authors declare no conflict of interest Table 2 : ADMET analysis of selected ligands using pkCSM online database server 

Final report confirms remdesivir benefits for COVID-19

Visual molecular dynamics

Antiviral constituents of Moringa oliefera were docked against non structural proteins

The structural chemistry of ligands was investigated by studying of their optimized geometries, HOMO-LUMO orbital energies, the 3-D orbital distribution patterns and molecular electrostatic potentials

Pharmacokinetic properties were evaluated by ADMET analysis Credit Author Statement

Al-Sehemi have performed calculations, H. A. Shakir and M. Khan played important role to study docking interactions. While M. Irfan and J. Iqbal helped a lot in writing results and discussion. Besides this, S. Muhammad and S. H. Hassan declare the following financial interests/personal relationships

The