3β-Acetoxy-21α-H-hop-22(29)ene, a novel multitargeted phytocompound against SARS-CoV-2: in silico screening

Abstract The present pandemic disease COVID-19 demands an urgent need for more efficient antiviral drugs against SARS-CoV-2. Computational drug designing and discovery enable us to explore ethnomedicinal plants as a source of various lead molecules that can be used against present and future pathogens. Adiantum latifolium Lam., a common fern, is resistant to pathogens mainly due to the presence of various phytochemicals having antimicrobial properties. In our previous study, 3β-acetoxy-21α-H-hop-22(29)ene, a terpenoid has been characterized from the methanol extract of leaves of A. latifolium. The manuscript evaluates the antiviral potency of the compound against SARS-CoV-2 through molecular docking method. Proteins essential for SARS-CoV-2 multiplication in host cells are the target sites. The study revealed strong binding affinity of the compound for all the ten proteins selected, including seven nonstructural proteins, two structural proteins and one receptor protein, with a binding energy of −4.67 to −8.76 kcal/mol. MDS and MMPBSA analysis of the best ranked complex further confirmed the results. The multitargeted compound can be considered as a natural lead molecule in drug designing against COVID-19, but requires wet-lab experimentation and clinical trials. Communicated by Ramaswamy H. Sarma


Introduction
The present worldwide pandemic disease COVID-19 which emerged in December 2019 in Wuhan city of China is caused by a novel corona virus SARS-CoV-2. The virus is spherical in shape with proteins called spikes protruding from their surface. Spike (S), membrane (M), envelop (E), and the nucleocapsid (N) proteins are the most important structural proteins of SARS-CoV-2. Virus entry into the host cell is facilitated by S protein which binds to the host cell receptor angiotensin-converting enzyme 2 (ACE2), present in different types of tissues in the human body, abundant in mucosa. Sixteen nonstructural proteins (NSPs) with multiple enzymatic functions are also essential for virion assembly inside the host cells. Papain-like protease (PL pro , NSP3), chymotrypsin-like protease (3CL pro , NSP5) or main protease (M pro ), RNA-dependent RNA polymerase (RdRp, NSP12), ADP-ribosephosphatase (ADRP, NSP3) and 2 0 -O-methyltransferase (2'OMTase, NSP16) and endoribonuclease (NSP15) are the major enzymes. PL pro and M pro are involved in cleavage of translated polyproteins pp1a and pp1ab to form NSPs inside the host cells. ADRP activity plays a regulatory role in the replication process of the virus (Frick et al., 2020). 2'OMTase plays a key role in methylation of ribose 2 0 -O position of the first and second nucleotide in viral mRNA for evading host immune system (Frick et al., 2020;Putics et al., 2005;Sharma et al., 2020). The endonuclease is involved in RNA processing and the components of replicase complex (Kim et al., 2020). RNA binding protein (NSP9) is a conserved protein essential for replication and transcription of the viral genome. N protein is also a conserved protein in corona viruses as a component of replicase-transcriptase complexes.
Though repurposed drugs such as remdesivir, chloroquine, hydroxychloroquine, ritonavir and lopinavir or their combinations have been used for the treatment of COVID-19 pandemic, they do not play a satisfactory role in the treatment. Hydroxychloroquine and chloroquine (Cq) are reported to cause serious heart rhythm problems, kidney injuries and liver problems (FDA, 2020). Lopinavir and ritonavir cause abdominal pain, weakness, nausea, diarrhea, vomiting, headache, anemia and so on (Ogbru, 2020). The SARS-CoV-2 mutates faster and therefore previously effective antivirals might become ineffective (Jiang, 2020). So more potent and safer drugs have to be developed to combat the challenges raised by SARS-CoV-2.
Plants are a source of diverse bioactive secondary metabolites and primary source of pharmaceuticals. In modern medicine the isolated and characterized bioactive phytocompounds are used in brand new drug discovery by their chemical modifications (Veeresham, 2012;Wang et al., 2012). Favipiravir, chloroquine and hydroxychloroquine that have been used are also plant derived drugs (Joshi et al., 2021;Oliveira et al., 2009). Isolation and identification of such phytocompounds are the primary steps in the development of a new medication. Computational drug designing and discovery enable us to explore ethnomedicinal plants as a source of various lead molecules that can be used against the present pathogens in modern medicine. After molecular docking and cytotoxic studies many natural compounds were approved for clinical trials in Egypt based on their effect on viral main protease (Seadawy et al., 2021). Such in silico methods are economic, time saving and aid in limiting the use of animal models in pharmacological evaluation studies.

Ligand preparation
The structure of Ah was explained in our previous study (Pradeep Kumar et al., 2019b) (Figure 1). A voucher specimen (No.107) of A. latifolium has been deposited in the Herbarium of Government College for Women, Thiruvananthapuram. The Chemdraw MDL Molfile (MOL) format of Ah was converted to PDBQT file format with 3 D coordinates using Open Babel software (O'Boyle et al., 2011). Physicochemical characters and drug-likeness of the compound were predicted using SwissADME (Daina et al., 2017).

Active site prediction
Active sites of selected targets were predicted using Computed Atlas of Surface Topography of proteins (CASTp) web server (Tian et al., 2018).

Molecular docking
Molecular docking of the ligand into selected drug targets were carried out using Autodock 4.2 and the subsequent file generation, visualization and analysis of molecular structures by MGL tools (Morris et al., 2009).
The grid box parameters were selected for each protein in such a way as to cover the entire three-dimensional structure of the protein (Table S1) and saved in grid parameter file (GPF) format. Grid generation was done using Autogrid software. Then the proteins and the ligand were subjected to blind docking to find out all the possible protein-ligand interactions using Lamarckian Genetic Algorithm (LGA) in combination with grid-based energy evaluation. The optimization conditions were 50 genetic algorithm runs, population size 300 with a maximum number of evaluation 2.5 Â 10^6. These parameters were saved in Docking Parameter File (DPF) format. The output from docking analysis was saved in a Docking Log File (DLG). The conformations with maximum binding energy were selected and such target-ligand complexes were saved in PDBQT format. Selected repurposed drug was also docked in the same parameters for comparison.
Docking protocol validation was done using the same grid box parameters. The re-docked complex (N3 peptide inhibitor against SARS-CoV-2 M pro ) was superimposed on to the native co-crystallized N3-M pro using LS align (https:// zhanglab.dcmb.med.umich.edu) and the root mean square deviation for flexible superimposition was calculated.

Visualization of target-ligand complex and ligand interactions
The visualization of PDBQT format of target-ligand complexes was done using Chimera 1.15rc. Ligand interactions were further analyzed using BIOVIA Discovery Studio 2020 for hydrogen, hydrophobic bonds and van der Waals interactions between ligand and the target protein.

Molecular dynamics simulation studies and MMPBSA analysis
Molecular dynamics simulation of best ranked complex (ADRP(apo)-Ah) was performed using Gromacs 2018 employing CHARMM36 force field for 100 ns with 300 K temperature and 1 atm pressure maintained in the system. Then simulated complex's binding free energy was calculated with 100 ns MD trajectories using g_MMPBSA package. Binding free energy of protein-ligand complex is calculated at 20 ns interval of entire simulation.

Prediction of antiviral spectrum of ligand
Chemdraw MDL Molfile (MOL) format of the ligand was used to predict the antiviral spectrum employing PASS online software (http://way2drug.com/passonline/predict.php) (Khanal et al., 2021).

Physicochemical characters and drug likeness
The structure details of the ligand 3b-acetoxy-21a-H-hop-22(29)ene were already explained in our previous study (Pradeepkumar et al., 2019b). SwissADME analysis gave the molecular weight of the Ah as 468.75 g/mol with 34 heavy atoms, three rotatable bonds, two hydrogen bond acceptors and one hydrogen bond donor. Molar refractivity is 144.88 and TPSA 26.30 Å. Ah obeys the Lipinski rule of drug likeness.

Molecular docking study
Molecular docking study could reveal the binding interactions and affinity of ligands with target proteins. The binding energy of Ah with target proteins varies from À4.67 to À8.76 kcal/mol (Tables 1 and 2). It shows spontaneous and strong binding without consuming energy. More negative binding energy represents better stability of the ligand-target complex. Comparative studies with standard drug substantiate the results. Strong and weak hydrogen bonds between the target and ligand are formed during molecular interactions. Hydrogen bond length of Ah with target proteins varies from 1.94 to 3.76 Å whereas that of Cq 1.84 to 3.72 Å. Decreased bond length represents higher bond order. The inhibition constant (Ki) of Ah with all target proteins (except PL pro ) ranges from 0.38 to 58.51 lM. Inhibitors with Ki values less than 100 lM can be considered as potent inhibitors  (Zheng & Polli, 2010). Inhibition constant (Ki) is an indication of the potency of the ligand to inhibit the functioning of its target protein ie., the smaller the Ki value the higher the binding affinity. This means that smaller amount of ligand is sufficient to inhibit the target activity.

Binding with M pro
Ah showed strong binding with M pro with binding energy of À6.34 kcal/mol and Ki of 22.71 lM at 298 K. The compound interacts with 16 amino acids, all fall within the active site of the enzyme including the substrate-binding His41 Cys145 catalytic dyad (Table 1, Figure 2), indicating the potency of compound as an inhibitor. Other interacted residues Gly143 Cys 145 His 163 His 164 and Glu 166 are also reported to be important for antiviral drug design (Dai et al., 2020;Zhang et al., 2016). M pro is one of the best characterized target for drug development as it is essential for the cleavage of polyproteins and release of other non-structural proteins (Magro, 2020). This support the present finding 3b-acetoxy-21a-Hhop-22(29)ene as a candidate against SARS-CoV-2. Drugs like lopinavir, ritonavir, remdesivir, indinavir are also M pro targeted to treat COVID-19 ( Cui et al., 2020). The corresponding binding energy of chloroquine is À7.24 kcal/mol (Table 1, Figure 3).

Binding with ADRP
Ah has more binding affinity with ADRP apo form than Cq. Binding energy of apo and complex forms of ADRP is À8.76 kcal/mol (Ki 0.38 lM) and À6.27 kcal/mol (Ki 25.31 lM) respectively (Table 1, Figures 2 and S2). The apo form interacts with 18 amino acids (16 of active site) of the target through 22 interactions whereas complex form with 14 amino acids (14 of active site) through 17 interactions. Trifluperidol is a known novel inhibitor of ADRP (Pandey, 2020). The high binding affinity of Ah with ADRP would disrupt virus multiplication.  Figure S4). NSP9 binds to RNA which is involved in viral RNA synthesis. It is a single-stranded protein unique to corona virus and its dimer structure is essential for viral viability (Egloff et al., 2004). Interaction of Ah with NSP9 may induce confirmatory changes disrupting its activity.

Binding with S protein
Ah showed high affinity to S protein than Cq with a binding energy of À5.77 kcal/mol and Ki of 58.51 mM. The compound makes 18 interactions with 16 amino acids of active site of the protein (Table 2, Figure S1). S protein is the first line structural protein target for SARS-CoV-2 control as its S 1 subunit binds to the host cell receptor ACE2 (Wrapp et al., 2020). Nafamostat is one such drug blocking spike-mediated membrane fusion of the virus with host cell (Wondmkun & Mohammed, 2020).

Binding with ACE2 receptor
Ah interacted with 16 amino acid residues (19 interactions) of active site of the enzyme with a binding energy of À5.98 kcal/mol while that of Cq is -4.65 kcal/mol with 9 amino acid interactions (Table 2, Figure S5). ACE2 receptor has significant biochemical roles in regulating blood pressure, wound healing and inflammation, through the reninangiotensin-aldosterone system pathway (Zhang et al., 2017). S-ACE2 interaction is critical in determining the infectivity of the virus and this makes it an important therapeutic target for the control of virus infection (Hoffmann et al., 2020). The drug Griffithsin isolated from the red algae Griffithsia is one such drug used to prevent and treat HIV, HCV, HSV, SARS-CoV, and SARS-CoV-2 infections (Decker et al., 2020). Ah has more binding affinity with ACE receptor than Cq showing the potency of the compound to prevent S-ACE2 interaction.

Binding with N protein
Ah strongly binds with N protein with binding energy À6.59 kcal/mol, Ki 14.69 mM, 18 interactions with 18 amino acids. All amino acids belong to the predicted active site of the protein. The binding energy of Cq is À6.26 kcal/mol (Table 2, Figure S7). N protein plays a vital role in assembling virus genetic material into CoV particles (Ma et al., 2010). Drugs like simeprevir and grazoprevir are reported to have the same mode of action (Prajapat et al., 2020).

Binding with endoribonuclease (NSP15)
Ah has binding affinity (-5.81 kcal/mol) with NSP15 with Ki of 55.10 lM. There are 19 interactions with 14 amino acid residues,8 interacted amino acid lies within the active site of the enzyme. Cq does not show any interactions with amino acids of active site (Table 2, Figure S8). Pyridone compounds were reported as endoribonuclease inhibitors .

Binding with PL pro
Ah-PL pro complex is with binding energy À4.67 kcal/mol, Ki 376.69 lM, 13 interactions with 11 amino acids (2 of active site). Binding energy of chloroquine is found to be À4.07 kcal/mol (Table 2, Figure S9). The binding of Ah with PL pro and M pro may have a synergistic effect against SARS-CoV-2 as they are involved in translation processes. On Autodock validation, N3 interacted with THR25, GLY143, SER144, HIS172, MET165, GLU166, PRO168, GLN189, the active site residues of M pro , with RMSD of 1.96 Å. RMSD value from the known conformation up to 2 Å is considered to have good validation (Hevener et al., 2009).

Molecular dynamics and MMPBSA analysis
The RMSD (root mean square deviation) of ADRP (apo)-Ah complex fluctuates with 0.36 nm at 25 ns, 0.44 nm at 50 ns and peak value of 1.12 at 80 ns. The average value of RMSD is 0.52 ± 0.25 nm ( Figure S10). Lower the RMSD, greater the stability of the protein (Shivanika et al., 2022). This supports the time-dependent stability of the protein in presence of ligand in the active site. The fluctuations observed in RMSD at various time periods are attributed to the conformational changes in the protein molecule (Sharma et al., 2020).
The amino acid residues of ADRP (apo) displaying high fluctuations in RMSF (root mean square fluctuation) values for ADRP (apo)-Ah complex ranges from 31 to 39, 71 to 90, 154 to 177, 218 to 243 and 278 to 293, highest fluctuation at the residue 236 ( Figure S10). The average value is 0.49 ± 0.17 nm. RMSF fluctuations indicate increased potential of the protein to interact with its ligand (Sharma et al., 2020;Shivanika et al., 2022).

Prediction of antiviral spectrum of ligand
Ah was predicted as an antiviral agent against Hepatitis B, Herpes, Influenza and Rhinoviruses (Table  3). Pharmacological activity (Pa) > pharmacological inactivity (Pi) (confidence > 0) are considered as possible for the compound.
The study proved 3b-acetoxy-21a-H-hop-22(29)ene as a multitargeted phytocompound against SARS-CoV-2. High binding affinity of the compound to the target proteins and low Ki value are important features. Multitargeted drugs are more effective and safer than single targeted drugs (Zhang et al., 2017). Novel compounds targeted towards multiple proteins of the pathogens can overcome drug resistance and show light upon new mode of action of available drugs (Watanabe & Kawaoka, 2015). So, we argue that the phytocompound 3b-acetoxy-21a-H-hop-22(29)ene can be considered as a natural lead molecule in drug designing against SARS-CoV-2. The antiviral potency of the compound can be augmented by its chemical modifications. Clinical trials under in vitro as well as in vivo conditions, pharmacological evaluation in vivo models are required for its therapeutic application. Pa-pharmacological activity, Pi-Pharmacological inactivity.