Repurposing immune boosting and anti-viral efficacy of Parkia bioactive entities as multi-target directed therapeutic approach for SARS-CoV-2: exploration of lead drugs by drug likeness, molecular docking and molecular dynamics simulation methods

Abstract The COVID-19 pandemic has caused adverse health (severe respiratory, enteric and systemic infections) and environmental impacts that have threatened public health and the economy worldwide. Drug repurposing and small molecule multi-target directed herbal medicine therapeutic approaches are the most appropriate exploration strategies for SARS-CoV-2 drug discovery. This study identified potential multi-target-directed Parkia bioactive entities against SARS-CoV-2 receptors (S-protein, ACE2, TMPRSS2, RBD/ACE2, RdRp, MPro, and PLPro) using ADMET, drug-likeness, molecular docking (AutoDock, FireDock and HDOCK), molecular dynamics simulation and MM-PBSA tools. One thousand Parkia bioactive entities were screened out by virtual screening and forty-five bioactive phytomolecules were selected based on favorable binding affinity and acceptable pharmacokinetic and pharmacodynamics properties. The binding affinity values of Parkia phyto-ligands (AutoDock: −6.00–−10.40 kcal/mol; FireDock: −31.00–−62.02 kcal/mol; and HDOCK: −150.0–−294.93 kcal/mol) were observed to be higher than the reference antiviral drugs (AutoDock: −5.90–−9.10 kcal/mol; FireDock: −35.64–−59.35 kcal/mol; and HDOCK: −132.82–−211.87 kcal/mol), suggesting a potent modulatory action of Parkia bioactive entities against the SARS-CoV-2. Didymin, rutin, epigallocatechin gallate, epicatechin-3-0-gallate, hyperin, ursolic acid, lupeol, stigmasta-5,24(28)-diene-3-ol, ellagic acid, apigenin, stigmasterol, and campesterol strongly bound with the multiple targets of the SARS-CoV-2 receptors, inhibiting viral entry, attachment, binding, replication, transcription, maturation, packaging and spread. Furthermore, ACE2, TMPRSS2, and MPro receptors possess significant molecular dynamic properties, including stability, compactness, flexibility and total binding energy. Residues GLU-589, and LEU-95 of ACE2, GLN-350, HIS-186, and ASP-257 of TMPRSS2, and GLU-14, MET-49, and GLN-189 of MPro receptors contributed to the formation of hydrogen bonds and binding interactions, playing vital roles in inhibiting the activity of the receptors. Promising results were achieved by developing multi-targeted antiviral Parkia bioactive entities as lead and prospective candidates under a small molecule strategy against SARS-CoV-2 pathogenesis. The antiviral activity of Parkia bioactive entities needs to be further validated by pre-clinical and clinical trials.

The life cycle of SARS-CoV-2 consists of four phases: viral entry and attachment, genome translation, replication and transcription, and assembly and release.The S glycoprotein, composed of two major subunits that are proteolytically cleaved by TMPRSS2 (transmembrane protease serine 2, for S protein priming) into the receptor binding domain (S1 subunit: S1-CTD/RBD, which facilitates coronavirus access into the host cells, Bosch et al., 2003) and S2 subunit, which aids in membrane fusion of the virus with the host cell membrane for infection (Hoffmann et al., 2020).The viral fusion process with the host cells is activated once the S1 subunit binds with the ACE2 (angiotensin-converting enzyme 2) of the host cell entry receptor.The cleavage of the S protein is executed by furin in the viral cells, resulting in the S1 domain binding with the ACE2 to form a RBD/ACE2 complex (domain liable for viral binding by ACE2) with the help of TMPRSS2 of the host cell receptor (Hoffmann et al., 2020).During the infection cycle, the replicase gene is translated and encodes two polyproteins (pp1a and pp1ab) that are necessary for viral replication and transcription (Wu et al., 2020a,b).Both the pp1a (1-11 nsps) and pp1ab (1-16 nsps) are cleaved into individual non-structural proteins (nsps) (Ziebuhr et al., 2000).Papine-like protease (PLPro, processing of viral polyproteins, replicase complex and viral spread), main protease (MPro, maturation stage of virus replication) and RNA-dependent RNA polymerase (RdRp, viral replication and transcription) are encoded by nsp3, nsp5, and nsp12 of pp1ab through autocleavage, respectively (B� aez-Santos et al., 2015;Gao et al., 2020;Jin et al., 2020).Several SARS-CoV-2 variants and their genetic lineages have been evolving and spreading around the world since the onset of the COVID-19 pandemic.The Food and Drug Administration (FDA) approved drugs either have mild-to-moderate efficacy or have failed to halt the progress of the disease.Hence, no effective and specific treatments are available to date.
COVID-19 is a multi-factorial disease that involves complex treatment modalities with synchronized modulation of a network of interactions between target proteins of SARS-CoV-2.Hence, a multi-target-directed ligand (MTDL) strategy (drugs that simultaneously interact with multiple targets) has been employed as a potent therapeutic method for treating COVID-19 (Joshi et al., 2020;Trezza et al., 2022).Recent reports have recommended that phytochemicals comprising polyphenols, terpenoids, carotenoids, steroids, fatty acids, and other active components pose a ray of hope and provide defense against SARS-CoV-2 infection due to their promising antiviral properties and effectiveness in treating COVID-19 victims as alternative and complementary phytotherapy approaches (Anand et al., 2021;Shree et al., 2022).High throughput screening of the phyto-medicines through molecular docking and dynamics can empower the shortlisting of lead compounds to meet the critical need for repurposing phyto-medicines along with the treatment of COVID-19 (Joshi et al., 2020(Joshi et al., , 2021)).
The purpose of this study was to repurpose a custommade library of Parkia bioactive entities via high-throughput virtual screening and recognize novel multi-targeted drug-like molecules (MTDLs) against the SARS-CoV-2 target receptors (S-protein, ACE2, TMPRSS2, RBD/ACE2, RdRp, MPro, and PLPro) using ADMET, molecular docking, and molecular dynamics simulation studies.We screened a library of one thousand ligand molecules belonging to polyphenols, terpenoids, steroids, fatty acids, and other active components from Parkia species to assess their ADMET properties, drug-likeness, and to determine the binding affinity of the Parkia bioactive compounds using three docking strategies (AutoDock, FireDock, and HDOCK).Finally, after high-throughput virtual screening, fortyfive ligand molecules out of one thousand molecules of Parkia bioactives were screened based on their pharmacokinetic and pharmacodynamic profiles, as well as the binding affinity of the unique MTDLs of Parkia to inhibit the seven SARS-CoV-2 target receptors compared to FDA-approved antiviral medicines.The top hit ligand molecules were further subjected to molecular dynamics simulation and MM-PBSA analysis against ACE2, TMPRSS2, and MPro protein receptors for evaluation of their stability, flexibility, and compactness.Our results will contribute to beneficial outputs for the exploration and development of multi-targeted antiviral phytomolecules from Parkia as lead and prospective candidates under a small molecule strategy against SARS-CoV-2 pathogenesis.Consequently, cell line and animal model studies are warranted to assess their antiviral activity followed by clinical trials (Joshi et al., 2020(Joshi et al., , 2021)).
The chemical structures of the ligand molecules of the Parkia bioactive entities (45) and FDA-approved antiviral drugs (6) targeting seven SARS-CoV-2 receptors (Table 1) Table 1.Details of polyphenols, terpenes, steroids and fatty acids compounds from Parkia species (Fabaceae: Caesalpinioideae) used as a ligand against SARS-CoV-2 receptors and their physicochemical and biological activities.
Ligand name    Bailly and Cotelle (2005) Cinnamic acid Ruwizhi and Aderibigbe (2020) Antioxidant, Anti-inflammatory were retrieved from the PubChem database (https://pubchem.ncbi.nlm.nih.gov/) in 3D and 2D SDF formats (structure-data file).For the molecular docking study, the collected chemical structures of the ligands in the SDF file formats were loaded in the PyRx 0.8 software (Python Prescription Virtual Screening Tools, https://pyrx.sourceforge.io)using the Open Babel plug-in tool (Dallakyan & Olson, 2015).Energy minimization was carried out for the ligand molecules using the Universal Force Field (UFF) energy minimization parameter along with the conjugate gradient descent optimization algorithm for obtaining stable conformations of the ligands.Before the execution of molecular docking with the AutoDock Vina tool, the input ligand molecules were transformed into the PDBQT (Protein Data Bank, Partial Charge and Atom Type) file format (Trott & Olson, 2010).

SARS-CoV-2 receptors
SARS-CoV-2 receptor proteins were selected based on the life cycle of the coronavirus from entry to exit of the host cells, namely S protein (PDB ID: 6vsb, viral attachment to the host cell, Figure 2A), ACE2 (PDB ID: 1r42, cell entry receptor, Figure 2B), TMPRSS2 (PDB ID: 5CE1, vital host protease that assists entry of the SARS-CoV-2, Figure 2C), RBD/ACE2-complex (PDB ID: 6LZG, vital functional domain within the S1 subunit responsible for binding of SARS-CoV-2 by ACE2, Figure 2D), RdRp (PDB ID: 6M71, replication of the viral genome, Figure 2E), MPro (PDB ID: 6LU7, maturation stage of viral replication and transcription, Figure 2F), and PLPro (PDB ID: 4OW0, processing of viral polyproteins, generation of a functional replicase complex, and support of viral spread, Figure 2G), and were retrieved in PDB format from the RCSB PDB database (Research Collaboratory for Structural Bioinformatics, Protein Data Bank, http://www.rcsb.org.).
The seven selected target receptors of SARS-CoV-2 were further processed to remove preoccupied ligand molecules, water molecules, metal ions, and HET atoms from the receptors, and to add polar hydrogen atoms and partial charges to the receptors using the UCSF Chimera 1.15 (https://www.cgl.ucsf.edu/chimera/)tool (Pettersen et al., 2004).Additionally, truncated side chains in the receptor were remodeled and gasteiger charges were added to the target receptor proteins using the in-built dock prep program in the Chimera tool (Opo et al., 2021).The results of the target receptor protein preparation were saved in a PDB file for method validation and docking.

Pharmacokinetic and pharmacodynamics properties
Parkia and antiviral drugs ligand molecules were obtained from the Pubchem database (https://pubchem.ncbi.nlm.nih.gov/) in the isomeric SMILES format (Simplified Molecular Input Line Entry System).SwissADME (http://www.swissadme.ch) and Molinspiration Cheminformatics (https://www.molinspiration.com)tools were employed to predict the ADMET features (Table 2), lipophilicity, water solubility, drug likeness bioactivity score (DLBS) and pharmacokinetics-bioactivity (Table 3), and medicinal chemistry, drug likeness- targeting RBD/ACE2-complex receptor targeting PLPro receptor Narayanan et al. (2022) bioavailability score, and pharmacokinetics (Table 4) of the selected ligand molecules (Daina et al., 2017;Kolodziejczyk-Czepas et al., 2018).et al., 2014;Ertl et al., 2000;Lipinski, 2004;Lipinski et al., 2001;Veber et al., 2002).The drug likeness bioactivity score (DLBS) of the ligand molecule was predicted with the parameters of GPCR ligand, ion channel modulator, nuclear receptor ligand, kinase inhibitor, protease inhibitor, enzyme inhibitor and cytochrome P450 inhibitors and scored as the following: score >0 was considered an active drug, À 5.0 to À 0.0 as moderately active drugs and < À 5.0 as an inactive drug (Molinspiration cheminformatics server; Daina & Zoete, 2016).Pharmacokinetics bioactivity properties such as gastrointestinal (GI) absorption, blood brain barrier (BBB) permeability and P-glycoprotein substrate (P-gp) were computed using the SwissADME tool.The skin permeability (Log Kp) was predicted with the SwissADME tool and a more negative value of Log Kp signifies less permeation to the skin and is acceptable as a drug candidate (Potts & Guy, 1992).The zero alert of PAINS and Brenk filter (assists in categorizing the ligand molecule) indicates whether the ligand molecule has a response or not toward the biological assays (PAINS), and further ensuring the accepted toxic level, chemical reactivity, and metabolically unstable or enduring properties accountable for poor pharmacokinetics (Brenk), respectively (Baell & Holloway, 2010;Brenk et al., 2008).The synthetic accessibility score is a method to measure the ease of synthesis of drug-like compounds and it is computed as a score of 1 being very easy to synthesize, whereas a score of 10 being very difficult to synthesize (Ertl & Schuffenhauer, 2009).A bioavailability score value of > 0.10 indicates oral bioavailability based on Lipinski's rules to be considered a sufficiently absorbable molecule orally (Martin, 2005).

Physico-chemical features and active site prediction of the SARS-CoV-2 receptors
The physicochemical features [number of amino acids, molecular weight, isoelectric point (pI), amino acid composition, negatively (R-) and positively (Rþ) charged amino acids, instability, aliphatic index, and grand average of hydropathicity (GRAVY)] of the SARS-CoV-2 target receptors (Table S2) were determined using the ProtParam Expasy tool (http:// web.expasy.org/protparam/)with the FASTA format (retrieved from RCSB PDB database) of the amino acid sequence of the target receptors as an input (Wilkins et al., 1999).
Figure 2. Preparedness of the SARS-CoV-2 receptor target proteins retrieved from PDB database and identification and characterization of the predicted active site (green colour) of the respective receptors by the fpocket web server for molecular docking studies.The selected SARS-CoV-2 receptors and their active pockets are spike protein (S protein-A, H), angiotensin-converting enzyme 2 (ACE2-B, I), transmembrane protease serine 2 precursor receptor (TMPRSS2-C, J), receptorbinding domain/angiotensin-converting enzyme 2 complex (RBD/ACE2 complex-D, K), RNA-dependent RNA polymerase receptor (RdRp-E, L), main protease receptor (MPro-F, M) and papain-like protease receptor (PLPro-G, N), respectively.geometry-based algorithm, which identifies possible active sites of the target receptor in accordance with sequential clustering steps to detect Voronoi tessellation and a-spheres, and characterizes and ranks the active pockets (Schmidtke et al., 2010).The best-ranked active pocket 0 was selected for each receptor (S protein: The active site amino acid profiles were categorized on the basis of their charged residues (His, Arg, Lys, Glu, and Asp), polar residues (Gln, Thr, Ser, Asn, Cys, Tyr, and Trp), and nonpolar residues (Gly, Phe, Leu, Met, Ala, Ile, Pro, and Val) (Petrova & Wu, 2006).

AutoDock Vina
AutoDock Vina 4.2 (which uses a Lamarckian genetic algorithm, LGA) was used for a molecular docking study in the PyRx virtual screening software (Eberhardt et al., 2021).The number of genetic algorithm (GA) runs and the number of energy evaluations (eval) were tuned to 250 and 25,000,000, respectively.Additional docking parameter options were maintained at their default settings.The grid map employing a grid box was arranged with the support of the AutoGrid program.The seven target SARS-CoV-2 receptors and the ligand molecules were loaded using the 'Vina wizard control' option.The active site residue of the respective target SARS-CoV-2 receptor molecules was selected, displayed, and labeled using the 'molecules' option in the PyRx virtual screening software for point-specific molecular docking.After labeling, a grid box appeared over the respective target SARS-CoV-2 receptor protein structure.(Fuhrmann et al., 2010).These clusters were systematically graded according to the lowest binding energy (the most stable conformer) and ranked in a hierarchical order.The docking results were viewed under the 'analyze results' tab and saved as a CSV format (Comma Separated Values).The receptor-ligand complexes were visualized using BIOVIA Discovery Studio Visualizer (https:// discover.3ds.com/discovery-studio-visualizer-download)(Biovia, 2015) and PyMol software (Schr€ odinger & DeLano, 2020).

PatchDock and FireDock
A two-tier methodology was employed to predict the SARS-CoV-2 target receptor-ligand complexes using the PatchDock algorithm (https://bioinfo3d.cs.tau.ac.il/PatchDock/), which is based on shape complementarity criteria (molecular shape representation, surface patch matching, filtering, and scoring), and the FireDock algorithm (Fast Interaction REfinement in molecular DOCKing; https://bioinfo3d.cs.tau.ac.il/FireDock/), which comprises the optimization of side-chain conformations, rigid-body orientation, refinement, and rescoring of rigid-body docking solutions.The energy-minimized PDB files of the seven SARS-CoV-2 receptor proteins and the selected ligand molecules of Parkia phytocompounds (45) and antiviral drugs (6) were uploaded in place of the 'receptor molecule' and 'ligand molecule' options displayed in the PatchDock online server, respectively.The clustering RMSD was set to 1.5 Å for protein-small molecule docking and the complex type was specified as 'protein-small ligand'.In the PatchDock algorithm, the first step is to perform a coarse refinement using RISCO (restricted interface side-chain optimization) with atomic radii (0.8 scale) to permit steric clashes.By the rigid-body optimization (RBO) process, scoring and ranking are performed for the refined docked complexes on the basis of an energy function and an output is generated.After submission of the receptor and ligand molecules, the docked complexes (100 top-hits) for each protein receptor along with their scores, areas, ACEs, transformation scores, and atomic contact energies (ACEs) were retrieved from the PatchDock server (Duhovny et al., 2002;Schneidman-Duhovny et al., 2005).Analysis of the presence of possible SARS-CoV-2 receptor-ligand interactions was performed using a Discovery Studio visualizer.
The option 'refine best solutions with FireDock' from the PatchDock web server was employed to rerun the FireDock using the ten top-hit docked complexes (solutions) for full refinement.In this reiterated run, each obtained docked complex from the PatchDock was consequently refined by full interface side-chain optimization (FISCO, atomic radii: 0.85 scale).The refinement process using FISCO comprises (1) assortment of the movable residues, (2) rotamer energy calculation for side-chain flexibility, (3) construction of the side-chain optimization problem deciphered by integer linear programming for rearranging the side-chains, and ( 4) refinement of the relative position of the docking partners by Monte Carlo minimization, acquiring the binding score function.The refined docked complexes were ranked based on the binding score including ACE, van der Waals interactions (attractive and repulsive), partial electrostatics, and supplementary approximations of the binding-free energy (Andrusier et al., 2007;Kingsford et al., 2005;Mashiach et al., 2008).

HDOCK
HDOCK (http://hdock.phys.hust.edu.cn/), a hybrid algorithm of template-based modeling and ab initio free docking (Yan et al., 2017a), was employed to execute molecular docking between the energy-minimized SARS-CoV-2 receptors (7) and the ligand molecules (Parkia phytocompounds: 45 and antiviral drugs: 6).The workflow of HDOCK consists of five steps: (1) uploading of input receptor and ligand molecules in the PDB file format, (2) selection of the best receptor (HHSearch) and ligand (FASTA) templates by sequence similarity search against the PDB sequence database, (3) structure homology modeling using MODELLER for receptor and ligand, (4) fast Fourier transform (FFT) based HDOCK lite global docking, and (5) visualization of docking models and template-based model as an output (Huang & Zou, 2008, 2014;Yan et al., 2017b;Yan & Huang, 2020).Of the ten top-ranked poses predicted by HDOCK based on the energy scores, the top-1 receptor-ligand complex was selected on the basis of evaluation metrics including the docking score, ligand RMSD (Å), and receptor interface residue (polar) and compared.

Molecular dynamics simulation (MD simulation)
Three SARS-CoV-2 receptors (ACE2, TMPRSS2, and MPro) were employed in simulation studies based on the mechanisms of SARS-CoV-2 transmission and pathogenesis.The top hit (pose1) receptor-ligand complex was chosen from each of the three molecular docking strategies followed, based on the lowest binding energy, including AutoDock (ACE2_didymin, TMPRSS2_rutin, and MPro_didymin), FireDock (ACE2_ lupeol, TMPRSS2_ursolic acid, and MPro_campesterol), and HDOCK (ACE2_epigallocatechin gallate, TMPRSS2_didymin and MPro_rutin) for MD simulation studies.The GROMACS (GROningen MAchine for Chemical Simulations, v5.1.5)software was used for MD Simulation (Abraham et al., 2015;Berendsen et al., 1995).Its workflow includes: (i) preparation of the protein receptor (pdb2gmx) and ligand topology (Chimera: addition of H atoms, Pettersen et al., 2004) using the Antechamber [parametrizing molecules using GAFF (general amber force field, AMBER 99SB-ILDN, Maier et al., 2015)], and acpype (a Python interface to Antechamber writing GROMACS topologies, Sousa da Silva & Vranken, 2012) to generate force field and parameter files for SARS-CoV-2 receptor protein and ligand, respectively (Hornak et al., 2006;Wang et al., 2006); (ii) definition of a grid box (cubic, 10 Å) and solvation (a three-point water model, TIP3P); (iii) addition of ions (Naþ/Cl‾ ions were added as counter ions to neutralize the system); (iv) energy minimization (1000 steps and 10 kJ/mol/nm tolerance) was performed using the steepest descent minimization algorithm to eliminate the steric clashes (atomic position) and structural errors in the bond length and bond angle; (v) equilibration (Zhang et al., 1997;Zhang et al., 2013) comprises two stages, namely phase-I (apply restraints to the ligand) where restrained constant number of atoms, volume, and temperature (NVT) ensemble equilibration was executed for 500 ps at 300 K with Berendsen thermostat coupling, and in the phase-II (treatment of temperature coupling groups), the constant number of atoms, pressure, and temperature (NPT) ensemble equilibration was then carried out for 1000 ps (1 ns); (vi) generation of MD includes the Particle Mesh Ewald (PME) approximation (to calculate long-range electrostatic interactions by setting a cutoff radius of 10 Å, Darden et al., 1993), LINear Constraint Solver (LINCS) algorithm (constraint of bond lengths), and unrestrained production [at 100 ns, 300 K temperature (using Langevin coupling, Washio et al., 2018) and 1 bar pressure (using Nose � ı-Hoover Langevin Piston algorithm, Tu et al., 1995)] of MD simulations for data collection.Each simulation was run for 100 ns with a 4-fs time step, and each trajectory was retained after 50 ps.Periodic boundary conditions (PBC) correction was made to the receptor-ligand complex trajectory and the system was equipped to its start position on the basis of the receptor's backbone before further analysis was executed.
VMD (Humphrey et al., 1996) and PyMol (Schr€ odinger & DeLano, 2020) visualization tools were employed to picturize the simulation trajectories (DeLano, 2002).Finally, simulation snaps were employed to compute the root-mean-square deviation (RMSD, gmx rms tool), root-mean-square fluctuation (RMSF, gmx rmsf tool), the radius of gyration (Rg, gmx gyrate tool), the number of hydrogen bonds between the ligand and the receptor, and the solvent-accessible surface area (SASA, gmx sasa tool) with the GROMACS utilities.Poses with an RMSD and RMSF of �2 Å are considered stable, compact, and flexible.Graphical illustrations were prepared using Microsoft Excel spreadsheet software (https://www.microsoft.com/en-in/microsoft-365/).

Binding-free energy calculation (MM-PBSA)
The binding-free energy of the SARS-CoV-2 receptor-ligand complexes (acquired from AutoDock, FireDock, and HDOCK) was computed using Molecular Mechanics Poisson-Boltzmann Surface Area (MM-PBSA) protocols executed by the g_mmpbsa suite to compute molecular mechanics potential energy (electrostatic and Van der Waals interactions) and solvation-free energy (polar and non-polar solvation energies).The free energy calculation was computed as: GX (receptor-ligand complex) ¼ EMM (MM potential energy in vacuum) þ GP (polar solvation energy) þ GNP (non-polar solvation energy).MM-PBSA was calculated using a short trajectory of the last 80 ns from the total MD trajectory (100 ns), and trajectory profiles were kept back for every 50 ps.A SASA model was employed to estimate the non-polar solvation energy.The contribution of the individual residue binding energy impact on the total binding energy was analyzed using the Python script 'MmPbSaDecomp.py'(g_mmpbsa suite) (Kumari et al., 2014).

Principal component analysis (PCA)
PCA moderates the complexity of the data accessed on SARS-CoV-2 receptor-ligand complexes to signify the biologically pertinent protein domains shifting from the inappropriate confined movements of the atoms.The principal components convert the correlated variables into uncorrelated variables by which to extrapolate the variance.In PCA, a covariance matrix was created by detaching the translational and rotational motions of the protein domain and atoms from the MD system using gmx covar from GROMACS.A covariance matrix with a positive value indicates correlated motion while a negative value denotes anti-correlated motion between the two Ca atoms.The procedure of diagonalization of the covariance matrix (gmx covar) was followed to compute the eigenvectors and eigenvalues.The eigenvectors that correlate with the largest eigenvalues (PCA) exemplify the most relevant collective motions.The original trajectory was filtered and the most significant eigenvectors 1 and 2 were estimated using the gmx anaeig from the GROMACS utilities (Chitongo et al., 2020).The overlap between the computed PCs and coordinates of the trajectory was calculated using the gmx anaeig tool, and 2D projections were generated using PAST (version 1.86b) software (Hammer et al., 2001).

ADME, pharmacokinetics, drug likeness, and medicinal chemistry of the Parkia ligand molecules
The physicochemical and pharmacological properties of the virtually screened ligand molecules (phytocompounds of Parkia spp., 45 out of 1000 ligands) were evaluated to identify a potent drug candidate as an inhibitor of SARS-CoV-2, and their efficacy was compared with that of antiviral drugs (Tables 2-4) for drug design using molecular docking and MD simulation strategies.

Prediction of active pockets and active site amino acid profiles of the SARS-CoV-2 receptors
The shape and size of the active pocket of the SARS-CoV-2 protein receptors (Figure 2H-N) determine the interactions of the ligand molecule with the receptor.The pocket score of the active sites is computed based on the shape metrics (number of vertices, alpha-sphere radius, number of alphaspheres, B-factor, hydrophobicity score, polarity score, real volume score, charge score, local hydrophobic density score, number of apolar alpha sphere and proportion of apolar alpha sphere residues) to determine the possible interactions with the ligand molecules.The number of vertices, mean alpha-sphere radius and mean number of alpha-spheres (the size of the cavity) of the SARS-CoV-2 receptors fluctuated within the range of 85-329, 3.52-4.01,and 0.43-0.46,respectively (Table S3, supplementary material).The mean Bfactor of the binding pocket (atoms that are contacted by at least one a-sphere) determines the flexibility of the receptors and ranged between 0.16-0.42.The real volume, charge, polarity, hydrophobicity and local hydrophobicity scores of the active pocket of target receptors were found to be within the range of 3.95-4.75,1-6, 14-29, 21.26-35.07,and 17.27-50.33that measure the relative volume and charge of all amino acids in contact with one a-sphere of the pocket as well as residue-based polar and non-polar character of the active pocket as a whole or local parts that are relatively hydrophobic, respectively (Table S3, supplementary material).
and their HDOCK scores were higher than the reference drugs (Figures S8-S10).

MD simulation and MM-PBSA energy calculations
Due to the high binding affinity of the ligand molecule, as per the AutoDock, FireDock, and HDOCK docking scores, and the mechanisms of SARS-CoV-2 transmission and pathogenesis, ACE2, TMPRSS2, and MPro receptors and their interacted ligands, including AutoDock: ACE2_didymin, TMPRSS2_rutin, and MPro_didymin; FireDock: ACE2_lupeol, TMPRSS2_ursolic acid, and MPro_campesterol; and HDOCK: ACE2_epigallocatechin gallate, TMPRSS2_didymin, and MPro_rutin complexes, were chosen for the MD simulation.RMSD (stability), RMSF (flexibility, compactness, and stability), Rg (stability, compactness, and folding of the structure), and non-bonded interactions (stability, H-bonds, and -EvdW, Eelec, Epolar, and Enonpolar) were computed in a time-dependent manner (100 ns).Concurrently, the respective SARS-CoV-2 receptors with their empirically validated anti-viral inhibitors (ACE2_ mozenavir, TMPRSS2_nafamostat, and MPro_nelfinavir) were also compared to authenticate the simulation results.The executed simulations were validated by the quality check factors in relation to temperature, pressure, and potential/kinetic energy, and were found to be stable throughout the simulation period of 100 ns (Figure S11, supplementary material).

Secondary structure evolution analysis
The stability of the ACE2 receptor_didymin/lupeol/epigallocatechin gallate (Figure 10), TMPRSS2_rutin/ursolic acid/didymin (Figure 11), and MPro_didymin/campesterol/rutin (Figure 12) complexes were assessed by the secondary structure evolution analysis at intervals of 20 ns during the 100 ns MD simulation.The secondary structure elements, namely, a-helices and b-sheets of the ACE2/TMPRSS2/MPro receptor, were conserved continuously during the MD simulation process, indicating the stability and consistency of the receptors after binding to the phyto-ligands (Figures 10-12).A slight shifting of the N-and C-terminal coils was detected, but no significant changes were observed structurally, while the phyto-ligands were persistently bound to the active site of the receptor without any structural alterations, inferring the stability of the phyto-ligands.

Discussion
Several SARS-CoV-2 variants and their genetic lineages have been evolving and spreading around the world since the start of the COVID-19 pandemic.Nine vaccines (Comirnaty, Covishield, Janssen, Moderna, Sinopharm, CoronaVac, Covaxin, Covovax, and Nuvaxovid) and several antiviral drugs (remdesivir, favipiravir, ribavirin, lopinavir, ritonavir, darunavir, arbidol, chloroquine, hydroxychloroquine, tocilizumab and interferons) authorized by the WHO and FDA are available, but no effective and specific treatments are available to date.Both vaccines and the medications for COVID-19 can cause adverse potential outcomes and further in-depth elucidation is warranted in terms of the side effects, safety, and efficacy.Therefore, to manage the current situation, it is necessary to identify an alternative strategy, i.e. drug repurposing (exploration of existing and approved compounds with known preclinical, pharmacokinetic, and pharmacodynamic profiles for new therapeutic practices) for ascertaining a therapeutic agent to combat COVID-19 (Singh et al., 2020).Under these circumstances, the repurposing of natural phytocompounds as therapeutic targets for SARS-CoV-2 should be explored, because of their proven nutraceutical properties such as antiviral, antioxidant, antiinflammatory, and immune-boosting.Seventy species of Parkia have been reported all around the world, of which P. biglobosa, P. speciosa, P. javanica, P. bicolor, P. biglandulosa, P. filicoidea, and P. clappertoniana are the common species and have been consumed (flower, leaf, stem, pod, seed, root, bark, seed oil) as a functional food in the form of vegetables, salads, and curry (Nayak et al., 2022).Plant and seed extracts, and pure compounds of Parkia comprise polyphenols, terpenoids, fatty acids, phytosteroids, and many volatile compounds, which possess antioxidant, anti-inflammatory, antimicrobial, anticancer, antihypertensive, antiulcer, antidiabetic, antimalarial, hepatoprotective, and antidiarrheal activities (Saleh et al., 2021).The Figure 9. Principal component analysis (PCA) scatter plots exhibit the cluster conformations and display the stability and finest binding affinity of the phytoligands toward the SARS-CoV-2 receptors by using scrutinizing the eigenvalue and cumulative variance.The eigenvalues of the didymin (A), lupeol (B), and epigallocatechin gallate (C) against the ACE2 receptor, followed by rutin (D), ursolic acid (E), and didymin (F) toward the TMPRSS2 receptor, and finally by didymin (G), campesterol (H), and rutin (I) upon the MPro receptor, signifying the distinct conformation and most stable compounds.The cluster conformation in the PCA plot labelled with blue, red and green denotes the chosen docked complexes derived from the AutoDock Vina, FireDock and HDOCK tools, respectively.ACE2: Angiotensin-converting enzyme 2; TMPRSS2: Transmembrane protease serine 2 precursor; MPro: Main protease.
phytoligands chosen from the Parkia species are polyphenolic, terpenoid, steroid, and fatty acid compounds (1000 phytocompounds) that possess therapeutic properties as an immune system modulator, viral entry and replication inhibitors (Adem et al., 2021;Jasso-Miranda et al., 2019;Lalani and Poh 2020;Vazquez-Calvo et al., 2017;Z� ıgolo et al., 2021).Hence, an attempt has been made to repurpose the Parkia phyto-ligand molecules as a multi-target-directed ligand approach to alleviate the pathological consequences caused by the SARS-CoV-2.Consequently, the use of polyphenolic, fatty acid, and steroid compounds from the Parkia may contribute to alternative prophylactic and therapeutic provisions along with the therapy for SARS-CoV-2.Phytocompounds show encouraging therapeutic effectiveness against several receptor proteins of SARS-CoV-2 from entry to exit (Parihar et al., 2022).The pathogenesis of SARS-CoV-2 involves viral entry and attachment to the host (S-protein, ACE2, TMPRSS2, and RBD/ACE2 complex receptors), viral replication and transcription (RdRp receptor), and viral maturation and spreading (MPro and PLPro receptors), resulting in suppression of the innate immune response, augmentation of oxidative stress, and acute inflammation (Mrityunjaya et al., 2020).Hence, a thorough understanding of the pathophysiological mechanisms of COVID-19 and the receptor structure-function interactions are vital criteria for finding drug targets.Earlier findings revealed that several bioactive molecules from plants served as prospective drug candidates to inhibit the structural and non-structural proteins of SARS-CoV-2 infectivity (Adem et al., 2021;Mrityunjaya et al., 2020;Parihar et al., 2022;Z� ıgolo et al., 2021).

ADMET and drug-likeness
The basics of ADMET are based on oral and intestinal absorption, bioavailability, tissue-specific distribution, metabolism, excretion, and toxicity of the compound.Prediction of ADMET features of phyto-ligands is necessary in the initial stage of drug discovery development, which reduces cost, time, labor, and collectively enhances the success rate during pre-clinical and clinical studies (Egbuna et al., 2021;Guan et al., 2018;Jia et al., 2020).LOR 5 is a criterion of five (molecular mass, hydrogen bond donor, hydrogen bond acceptor, and log P) that helps in screening of phytocompounds to determine if they possess drug-like characteristics or not (Shree et al., 2022).The water solubility of bioactive molecules plays a crucial role in the execution of physiological functions (Lee et al., 2017).Phyto-ligands with good water solubility may ensure the possibility of high biological availability.Parkia phyto-ligands fulfilled the criteria of ADMET properties, including molecular weight (50-500), octanol-water partition coefficient/lipophilicity (0.7-5.0), topological polar surface area (20-130), the number of H-bond acceptors (0-10), the number of H-bond donors (0-5), the number of rotatable bonds (0-5), the number of heavy atoms (15-50), molecular volume , and molar refractivity (40-130).The Parkia phyto-ligands exhibited drug-likeness properties, including high gastrointestinal and oral absorption, water solubility (<6), bioavailability (0.11-0.85) and drug likeness bioactivity (À 5.0-> 0), low permeation to the skin (negative value), low synthetic accessibility score (1.00-6.52),no blood brain barrier permeability, no PAINS and brenk alerts, modulation of GPCR and nuclear receptor ligands and ion channels, inhibition of kinases and proteases, and CYP1A2, CYP2C19, CYP2C9, CYP2D6, and CYP3A4 inhibitors.The Parkia phyto-ligands druglikeness properties were comparable to those of the anti-viral drugs.All the selected phytoligands of Parkia were found to be suitable for being prospective drug candidates based on their biological activity and pharmacologically distinguishing characteristics as reported earlier in the virtual screening of phytochemicals by targeting multiple proteins of SARS-CoV-2 (Azeem et al., 2022;Joshi et al., 2022;Khasamwala et al., 2022).

MD simulation studies
Several phytocompounds from plants strongly interact with several structural and nonstructural proteins of SARS-CoV-2, causing conformational changes.The structural and chemical array of plant molecules serve as precursors for scaffoldings with strong binding affinity with the active site of SARS-CoV- plot.A highly stable cluster occupies less phase space, whereas an unstable cluster has more space (Dey et al., 2021).The 2D projections of ACE2/TMPRSS2/MPro complex systems occupied less phase space, had more well-defined clusters, more restricted motions of the backbone Ca atoms, and established a more stable complex compared with the antiviral drugs.The secondary structure evolution analysis of the receptor-ligand complex revealed that the a-helices and b-sheets of the ACE2/TMPRSS2/MPro receptors were conserved continuously during the MD simulation process, indicating the stability and consistency of receptors after binding to the phyto-ligands.
Even though no FDA-approved drugs are currently available for the treatment and prevention of COVID-19, several drugs are being repurposed and are under clinical trials, such as remdesivir, favipiravir, and darunavir (antivirals); anti-HCV; sofosbuvir, IDX-184, and ribavirin (nucleotide inhibitors); and imatinib (kinase inhibitor).Provisional approval has been given to dexamethasone and remdesivir for COVID-19 treatment, but undesirable results have been observed after their use, consequently highlighting the immediate need to discover novel molecules that are therapeutically effective and safe for humans.COVID-19 disease is interconnected to diet, and sustaining a healthy diet regimen during a pandemic increases the survival rate and maintains the body healthy, while also helping to avoid oxidative stress and inflammatory conditions.Therefore, this study was performed to demonstrate the efficacy of polyphenols, terpenoids, steroids, fatty acids, and other active components present in the Parkia species as immune boosters.These docking and simulation studies contribute evidence of the binding affinity of polyphenols, terpenoids, steroids, fatty acids, and other active components existing in Parkia species with the SARS-CoV-2 target protein receptors (S-protein, ACE2, TMPRSS2, RBD/ACE2, RdRp, MPro, and PLPro).All the virtually screened phytocompounds were more potent inhibitors of SARS-CoV-2 proteins and found to be lead compounds to boost the immune system.The phytomolecules present in Parkia are a main repository for exploiting therapeutic properties using ADME and drug properties in relation to pharmacodynamics and pharmacokinetics.The oxidative impairment is safeguarded by quenching and free radical scavenging and, as a result, upsurges the antioxidant reserves and prevents oxidative damage.Parkia phytomolecules also trigger the interruption of procarcinogens growth by using drug biotransformation systems, hence being potential antitumor agents (Kamisah et al., 2013;Ralte et al., 2022).The phytocompounds of Parkia species are used extensively in traditional and alternative system medicine for treating inflammatory, respiratory, circulatory, and digestive system disorders (Odounharo et al., 2022;Ralte et al., 2022;Saleh et al., 2021).Consumption of Parkia in the daily diet can moderate the severity of the symptoms of COVID-19.Phytocompounds of Parkia may act as a natural immune booster to strengthen the immune system, and people around the world are using Parkia (leaves, stems, pods, roots, bark, seeds, and edible parts) in various cuisines.These foods are responsible for boosting the immune system due to their strong anti-inflammatory and antioxidant properties, which intensify the immune response and provide protection as a second line of defense by keeping the required antioxidant reserves.The future scope of the study includes formulating antioxidant and anti-inflammatory nutraceuticals or food supplements.These formulated supplements can be used in suitable dosages and consumed regularly to boost the immune system and provide the body with antioxidants and anti-inflammatory agents that promote the production of antibodies in the blood and encourage remedial measures for COVID-19.The future work also includes the use of different ethnobotanical compounds from Parkia species to find the scope as lead candidates whose bioavailability is high and to provide a second line of defense in avoiding health issues related to COVID-19 and extending lifespan.Therefore, this in silico study exhibited the possibility of therapeutic effectiveness of Parkia species in combating COVID-19 and managing the pandemic.

Conclusions
This work delivers scientific insights for the application of Parkia bioactive ligands as promising drug candidates to expand the SARS-CoV-2 remedial treatment.Altogether, bioactive molecules of Parkia (didymin, rutin, epigallocatechin gallate, epicatechin-3-0-gallate, hyperin, ursolic acid, lupeol, stigmasta-5,24(28)-diene-3-ol, ellagic acid, apigenin, stigmasterol, campesterol) followed the pharmacokinetic and pharmacodynamic parameters and exhibited a safe and permissible range of ADMET and drug-like profiles.These multi-targetdirected ligands of Parkia demonstrated strong binding affinity with SARS-CoV-2 receptors (S-protein, ACE2, TMPRSS2, RBD/ACE2, RdRp, MPro, and PLPro).The compactness, flexibility, and stability of the ACE2, TMPRSS2, and MPro receptors formed favorable conformations with strong binding interactions with the respective Parkia phyto-ligands, validated by MD simulation and MM-PBSA studies.An in silico investigation using the phytoligands of Parkia spp.showed that polyphenols, terpenoids, steroids, fatty acids, and other active components present in Parkia having the capability to act on multiple targets and are more potent in blocking the seven SARS-CoV-2 protein receptors (S-protein, ACE2, TMPRSS2, RBD/ACE2, RdRp, MPro, and PLPro).Therefore, regular consumption of Parkia (leaf, stem, pods, root, bark, seed and edible parts) as part of a diet during the pandemic may augment the immune system by providing anti-inflammatory and antioxidant properties to combat oxidative stress and inflammation, and also act as a second line of defence by replenishing the system with the required antioxidants and anti-inflammatory compounds to boost mitochondrial and cellular functions.Furthermore, this practice can be an effective preventive measure against COVID-19.Nevertheless, in vitro and in vivo experiments will be necessary to corroborate these findings.Henceforth, the nutrients present in the different plant parts of Parkia can be used as immuneboosting food supplements along with the COVID-19 treatment.

Table 4 .
absorption; BBB: Blood brain barrier permeability; P-gp: P-glycoprotein substrate.b Default range: Log S: <6; DLBS: score >0 was considered to be active drug, À 5.0 to À 0.0 as moderately active drug and < À 5.0 as inactive drug.Pharmacokinetics and medicinal chemistry properties of the phytocompounds from Parkia species (Fabaceae: Caesalpinioideae) using SwissADME online software tool.P450 1A2 inhibitor; CYP2C19: Cytochrome P450 2C19 inhibitor; CYP2C9: Cytochrome P450 2C9 inhibitor; CYP2D6: Cytochrome.P450 2D6 inhibitor; CYP3A4: Cytochrome P450 3A4 inhibitor; Log Kp: Skin Permeation (cm/s).a Log Kp: the more negative value indicates the less skin permeation and the acceptable range of Log Kp for drug candidate; PAINS and Brenk: zero alert means the ligand molecule having no interference toward the biological assays (PAINS), and ensuring the accepted toxic level, chemical reactivity, and metabolically unstable or to endure properties accountable for poor pharmacokinetics(Brenk); Synthetic accessibility score: score 1 means very easy to synthesize and score 10 means very difficult to synthesize; Bioavailability score: > 0.10, indicates oral bioavailability based on Lipinski's rules to be considered a sufficiently absorbable molecule orally.

Figure 7 .
Figure 7. Evaluation of stability and compactness of the three docked SARS-CoV-2 receptor_ligand complexes (ACE2, TMPRSS2 and MPro) derived from AutoDock Vina, FireDock and HDOCK tools by molecular dynamics (MD) simulation for 100 ns calculating the deviation (nm) about function of time and residue index.The root mean square deviation (RMSD: A, E, I), the root mean square fluctuation (RMSF: B, F, J), the radius of gyration (Rg: C, G, K) and the number of H-bond (D, H, L) analyses of the docked SARS-CoV-2 receptor-ligand complexes (A: ACE2_didymin, ACE2_lupeol, and ACE2_epigallocatechin gallate; E: TMPRSS2_rutin, TMPRSS2_ ursolic acid, and TMPRSS2_didymin; I: MPro_didymin, MPro_campesterol and MPro_rutin) with respect to 100 ns simulation time, signifying restricted mobility of Ca carbon atoms, flexibility of amino acid residues, the degree of rigidity, compactness and stability of the receptor_ligand complexes.Blue, red and green colours denote the docked receptor_ligand complexes ascertained from the AutoDock Vina, FireDock and HDOCK tools, respectively.ACE2: Angiotensin-converting enzyme 2; TMPRSS2: Transmembrane protease serine 2 precursor; MPro: Main protease.

Table 2 .
Description about physicochemical properties of the ligand molecules from Parkia species (Fabaceae: Caesalpinioideae) using Molinspiration and SwissADME online software tools.

Table 3 .
Lipophilicity, water solubility, pharmacokinetics and bioactivity prediction characteristics of the phytocompounds from Parkia species (Fabaceae: Caesalpinioideae) using Molinspiration and SwissADME online software tool.
the grid box center coordinates accorded to the Ca atom of the glutamic acid 536 residue, chain A: x (1.4516), y (7.0199), and z (À 20.0728) with dimensions of the grid box of 30 � 30 � 30 and point spacing of 0.0404 Å; (5) RdRp SARS-CoV-2 receptor: the grid box center coordinates concurred to the OH atom of the threonine 319 residue, chainA: x (134.4783),y(138.8945), and z  (122.1931)withdimensions of the grid box of 30 � 30 � 30 and point spacing of 0.0216 Å; (6) MPro target receptor: a grid box with dimensions of 30 � 30 � 30 placed at the grid box center coordinates (x, y, z) of À 12.9723, 24.5383, and 54.1290 linked to the Ca atom of the glutamine 14 residue of chain A was employed to incorporate the entire active site of MPro, and the grid spacing value was adjusted to 0.0387 Å; and (7) PLPro target receptor: the grid box center coordinates concurred to the OH atom of the lysine 196 residue, chain A: x (À 20.3550), y (29.0141), and z (À 32.7026) with dimensions of the grid box of 30 � 30 � 30 and point spacing of 0.0785 Å.The AutoDock program was set and run after the grid mapping generation.The AutoDock LGA employed a tight clustering analysis by using a root mean square deviation (RMSD) tolerance of 1.0 Å to produce 200 conformers for each ligand The grid box was adjusted and resized to encompass the entire active site of the target receptor molecule.The details of the grid box center coordinates, dimensions, and point spacing of the SARS-CoV-2 receptors are specified as follows: (1) S protein target receptor: the grid box center coordinates conformed to the Ca atom of the glutamic acid 613 residue, chain B: x (251.0468),y(216.9108), and z (219.8380) with dimensions of the grid box of 30 � 30 � 30 and point spacing of 0.0401 Å;(2) ACE2 protein receptor: the grid box center coordinates coincided with the oxygen atom of the glycine 537 residue for chain A: x (36.9301), y (70.110), and z (53.6511) with dimensions of the grid box of 30 � 30 � 30 and point spacing of 0.0367 Å; (3) TMPRSS2 protein receptor: a grid box with dimensions of 30 � 30 � 30 placed at the grid box center coordinates (x, y, z) of 11.2766, À 7.4317, and 27.05003 compared to the H atom of the serine 376 residue of chain A was employed to incorporate the entire active site of TMPRSS2, and the grid spacing value was adjusted to 0.0367 Å; (4) RBD/ACE2 receptor:

Table 5 .
Binding affinity between the ligands (phytochemical compounds from Parkia species) and the SARS-CoV-2 protein receptors using the Autodock vina tool.

Table 6 .
Global free energy affinity between the ligands from Parkia species and the SARS-CoV-2 protein receptors using Firedock online server.

Table 7 .
Docking score of the SARS-CoV-2 protein receptor_ligand docked complexes using the HDOCK online server.