MOLECULAR MODELING AND DOCKING ANALYSIS OF SWINEFLU (H1N1) NEURAMINIDASE PROTEIN AGAINST THE PHYTOCHEMICALS OF ANDROGRAPHIS PANICULATA

H1N1

prediction because a structure of a protein is conserved compared to the amino acids, hence a query sequence structure can be predicted with high accuracy using homology modeling using a sequence of known structure (template) (Biasini et al., 2014).Energy minimization of the modeled protein was performed using Swiss PDB viewer (Dammalli et al., 2014;Laskowski et al., 1993).

Validation
The refined model reliability was evaluated through SAVES server: PROCHECK, verify 3D and PROVE.These validation methodologies show the quality of protein structure and gives scores based on the protein quality (Eisenberg et al.,1997;Colovos and Yeates, 1993;Dundas et al.,2006).

Active Site Identification
Active site determination was done for the modelled protein using CASTp to further work on its docking studies.Active site determination is important for preparing grid box before docking (Chao and Lin, 2010).

Preparation of protein and ligand structure
The final energy minimized protein was chosen for docking.Before docking studies, the protein was prepared by adding polar hydrogen atoms and kollman charges using Autodock 4.2.The active compounds from Andrographis paniculata were chosen for docking studies.A total of 16 test compounds: neoandrogrpholide, andographin, andrographolide, wogonin, stigmasterol, andropanoside, octadeconoic acid, beta sitosterol, 14-deoxyandrographolide, andrographiside, chlorogenic acid, cinnamic acid, andrograpanin, lupeol and apigenin were chosen.All these compounds are proved to have many medicinal benefits (Tanet al., 2016;Rigsby and Parker, 2016).The structure of these ligands was downloaded from PubChem Database in the SDF format was first converted to the PDB format using Pymol (Benet et al.., 2016;Hari, 2019).The ligands and their SDF structures were given in Table1.

Drug Scan
All the ligands were tested for Lipinski's rule of five parameters such as molecular weight, log P, and number of hydrogen bond donors and number of hydrogen bond acceptors using molinspiration server (Adejoro et al., 2016).Molinspiration and Lipinski Filters were applied for knowing the druggability of ligands by analyzing the number of hydrogen bond acceptor, the amount of hydrogen bond donors, Log P, and the molecular mass of the drugs (Morris et al., 2009).

Grid box
AutoDock 4.2software was used to generate grid box based on the active site of the protein.The grid-box was created that was large enough to cover the entire protein binding site and accommodate all ligands to move freely in it.The number of grid points in x, y, and z axes were set to 40×40×40.The X, Y and Z coordinates were also adjusted based on the active site of the protein (Vyas et al., 2008).

Ligand docking
AutoDock 4.2 was used for docking of ligands to the catalytic triad of protein.
Binding potential and possible conformations of ligand binding to the NA protein binding site can be assessed by doxking studies.To perform docking, initially the protein was fixed by adding polar hydrogen atoms and Kollman charges.Lamarckian Genetic Algorithm (LGA) method was used for ligand flexible protein-fixed docking studies.The best conformation was chosen with the maximum hydrogen bondings.Standard docking settings were used, and the 10 energetically most favorable binding poses were outputted (Rognan, 2011).

Modeling and validation of H1N1 NA protein
The structure of H1N1 NA protein was unavailable in Protein Data Bank and hence it was modeled using Swiss Model.The sequence of H1N1 NA protein was retrieved from NCBI with ID: ADJ40637.1.The structure building of protein was performed using Swiss Model.Swiss Model is a fully automated tool for protein structure prediction that works based on the principles of homology modeling.
Homology modeling method of protein structure prediction is based on a homologous template with known 3D structure.The template 5NWE was chosen based on the BLAST similarity search with the template having a identity score of 99.23%.The template 5NWE is the mutated crystal structure of NA, which a homotetramer with one amino acid residue (ASN 146) interacting with the ligand2-Acetamido-2-Deoxy-Beta-D-Glucopyranose (NAG).The final model was saved and viewed under Pymol molecular viewer (Figure 1).

Figure 1
Structure of NA protein (ID: ADJ40637.1)obtained using SWISS MODEL and viewed using Pymol Viewer.
The obtained model was then energy minimized using Swiss PDB viewer and the final energy of the model was found to be -87624.516KJ/Mol (Table 2).The predicted model was then validated using Verify-3D and PROCHECK Ramachandran plot.Ramachandran plot provides information about the dihedral angles like φ (phi), ψ (psi) and ω in a protein.The ProCheck analysis revealed that 88.3% amino acid were in the most favored region and no amino acids were in the disallowed region which was shown in Ramachandran plot (Table 3 and Figure 2).Verify-3D shows information on the compatibility between the 3D atomic model and the protein aminoacid sequence and compares the results to valid structures.According to verify 3D, a score above zero is considered good model.In this study, verify 3D provided a good compatible score with about 97.81% of the residues having averaged 3D-1D score >= 0.2, which was within the native confirmation of the crystals (Figure 3) and the profile score above zero indicated the acceptable environment of the model.In a similar modeling of lipase protein, models with average of 80% residues with verify 3D score >= 0.2 was considered a good model (Sahoo et al., 2019).The overall quality of the predicted model was analyzed using PROVE which showed satisfactory results of highly reliable model with compatible Z-score values (Figure 4).The modeled structure was highly similar to the crystal structure of the mutated NA protein that was available.However, in the crystal structure only one amino acid has been found to interact with the ligand, but the modeled structure most of the amino acids in active site has been found to interact with the ligands.In our present investigation, we have modeled the structure of NA to perform docking analysis against the bioactive compounds of Andrographis paniculata.Similar study was performed using Andrographolide, a major constituent of the plant Andrographis paniculata to identify its binding mechanism through virtual screening and molecular docking approaches (Liu et al., 2007).The active site of the modelled protein was predicted using CASTp server and was shown in Table 4 and Figure 5.The total area and the volume of active site based on solvent-accessible surface was 187.500 and 152.523 respectively.The details of about 16 compounds of Andrographis paniculata were obtained from PubMed literatures and those compunds were tested for their drug likeness property using Lipinski's rule of five.Among them, 10 compounds satisfied Lipinski's rule and they were further subjected to docking studies against H1N1 NA protein (Table 5).6 lists the results of docking analysis of the ligands with NA protein.The andrograpanin compound has the highest binding energy with the active cavity of NA protein; the docked confirmation is given in Figure 6.

Symbol '-'indicates no violation
The docking analysis of the compounds with the modeled protein revealed that andrograpanin, a diterpene lactone isolated from Andrographis peniculata have the highest binding energy of -8.61 kcal/mol forming a hydrogen bond at Arg 152 (Figure 7a).Andrograpanin have already been proved for its antiinflammatory property in lipopolysaccharide induced macrophage cells.Neoandrographolide which is also a diterpene lactone possess a good binding energy of -7.39 kcal/Mol with the active site region of H1N1 NA receptor forming four hydrogen bonds at Arg 118, Arg 152, Glu 228 and Arg 368 (Figure 7b).The compound has good antiinflammatory property which has been studied by invivo methods (Sahoo et al., 2016).The terpenoid compound andrographolide and its derivative 14deoxyandrographolide have good binding energies of -4.38 kcal/Mol (Figure 7c) and -3.81 kcal/mol (Figure 7d) respectively.Both the compounds bind in the active site region of NA receptor forming hydrogen bonds at Arg 152 and Glu 288.Similar results were obtained when andrographolide was docked against Influenza A NA protein (Seong et al., 2018).Further, compounds such as apigenin, wogonin, eugenol, andropanoside and andrographin were also found to inhibit NA of H1N1 with binding energy in the range of -5.59 kcal/mol to -4.12 kcal/mol (Figure 7e-7i) and the results were compared with the standard drug Ostelmivir (Figure 7j).The major constituents of the plant andropanoside and andrographin form hydrogen bond at the residues Arg 152, Glu 277 and Arg 293 which forms the active site pocket of NA protein.Wogonin, a flavonoid from Scutellaria baicalensis has anti-viral activities influenza virus in human lung epithelial cells (Liu et al., 2008).Apigenin, an active flavonoid compound was proved for its highest antiviral activity against the infuenza virus (H3N2) (Marchese et al.,2017).Similary, eugenol showed the binding energy of -4.19 kcal/mol at the active site region of neuramindase.The compound eugenol has diverse biological activities such as antibacterial, antifungal, and antiviral properties (Carrasco et al.,2012;Benencia and Courreges, 2000).The standard compound Ostelmivir has a binding energy of -3.35 kcal/mol which was comparatively similar to our bioactive compounds used in this study.Thus, these compounds could effectively be used as leads for treating H1N1 infection.In vitro studies showed that many plants are potent inhibitor of NA of H1N1 like the Rubia yunnanensis root extract showed a maximum inhibition of 10%.The aqueous extract of whole plant of Achyranthes aspera showed an inhibition percentage of 43.67 again NA.
Carthamus tinctorius flowers showed a maximum inhibition of 51%.Root extract of Geranium strictipes showed 85% inhibition against NA.The whole plant of Balanophora involucrate showed around 64% inhibition while the whole plant of Euphorbia hirta showed around 61% inhibition.The roots of Paeonia delavayi showed a maximum inhibition of about 92%.The root tuber of Fagopyrum dibotrys showed 70% inhibition, while root tubers of Polygonum multiflorum showed 78% inhibition.The roots ofPolygonum aubertii showed an inhibition of around 86%.The rhizomes ofCurcuma longa showed an inhibition percentage of around 77.This shows that many plant compounds are potent inhibitors and can be used to treat H1N1 disease   In silico studies of plant compounds helps in understanding the molecular mechanism of inhibition of virulence targets of viruses.In this present investigation, the major compounds of the plant Andrographis paniculata which have been traditionally used for swine flu infection (H1N1) were studied for their mode of interaction with the virulence target of swine flu virus.NA, the antigenic determinants of H1N1 virus was subjected to docking studies against the compounds of the plant Andrographis paniculata.All the natural compounds showed good binding interactions with NA protein at the active site region.Of all the natural inhibitors, andrograpanin have shown to bind with good binding energy of -8.61 kcal/mol, which could be used for invitro and invivo studies to develop as drug for treating swine flu.

Figure 4 -
Figure 4-Results of PROVE showing Z-score values Docking analysis of H1N1 NA protein with the bioactive compounds of Andrographis paniculata.Natural compounds play an important role in reducing the virulence of an infection by minimizing the side effects of commercial drugs and many researches focuses on understanding the mechanism of interaction of these natural plant compounds with the drug targets of infectious diseases using in silico approaches(Meenambiga et al., 2018;Gupta et al., 2013, Seniya et al.,2014).Very few in silico studies have been performed with NA against natural inhibitors to recommend natural ligands as drugs targeting NA of Swine flu.In our present

Figure 5 A)
Figure 5 A) Active site residues of NA protein predicted using CASTp server.B) Letters highlighted in blue indicate the amino acids in the active site pocket of NA protein.

Figure 6 -
Figure 6-Docked confirmation of the compound Andrograpanin with NA Table6lists the results of docking analysis of the ligands with NA protein.The andrograpanin compound has the highest binding energy with the active cavity of NA protein; the docked confirmation is given in Figure6.

TRP179Figure 7 -
Figure 7-3D interactions of modeled NA protein with A) Andrograpanin, B) Neoandrographolide, C)Andrographolide, D) 14, deoxyandrographolide, E) Apigenin, F) wogonin, G) eugenol, H) Andropanoside, I) Andrographin, J) Ostelmivir CONCLUSION Medicinal plants have been traditionally used for treating viral infections due to their anti-viral properties.In silico studies of plant compounds helps in understanding the molecular mechanism of inhibition of virulence targets of viruses.In this present investigation, the major compounds of the plant Andrographis paniculata which have been traditionally used for swine flu infection (H1N1) were studied for their mode of interaction with the virulence target of swine flu virus.NA, the antigenic determinants of H1N1 virus was subjected to docking studies against the compounds of the plant Andrographis paniculata.All the natural compounds showed good binding interactions with NA protein at the active site region.Of all the natural inhibitors, andrograpanin have shown to bind with good binding energy of -8.61 kcal/mol, which could be used for invitro and invivo studies to develop as drug for treating swine flu.

Table 1
Details of phytochemicals of Andrographis paniculata

Table 3 -
Results Figure 2 Ramachandran plot of predicted NA model (the red, dark yellow, and light-yellow regions represent the most favored, allowed, and generously allowed regions)

Table 4
Active site residues of modelled protein of H1N1 NA predicted using CASTp server.

Table 6 -
Molecular docking analysis of phytochemicals of Andrographis paniculata against H1N1 NA protein