MOLECULAR DOCKING AND MOLECULAR DYNAMICS STUDIES OF ACALYPHA INDICA L. PHYTOCHEMICAL CONSTITUENTS WITH CASPASE-3

Objective: This study investigated the structure-based molecular interactions between phytochemical constituents of Acalypha indica L. and caspase-3. Methods: Thirty-three phytochemical constituents of A. indica were screened against caspase-3. The X-ray crystal structure of human caspase-3 was retrieved from https://www.rcsb.org/structure/. The molecular interactions of the phytochemicals were studied using the AutoDock 4.2 software and followed by molecular dynamics (MD) simulations using the Amber18 software. Results: From this study, 25 screened phytochemicals were found to have a better binding mode than the native ligand. Moreover, the binding stability of the top four hits evaluated by MD indicated that the hydrogen bonds in MD were quite different from the molecular docking results due to the massive receptor and ligand movement in the MD simulations. However, with the exception of stigmasterol, all ligands were able to stabilize the protein. Conclusion: This study suggested that γ -sitosterol acetate , β - sitosterol acetate, and γ -sitosterol might be able to induce caspase-3, thereby activating apoptosis. These high-affinity compounds can bind to caspase-3 more efficiently and were able to stabilize the protein. Therefore, they have the potential to be used as lead compounds in the treatment of cancer.

The computational method has been regarded as a simple and costeffective way of discovering compounds with high biological activity for drug development. It gives a thorough assessment of the relevant characteristics and interactions of compounds, minimizing mistakes and undesired findings during in vitro studies [21]. One of the targets of anticancer drugs is apoptosis (programmed cell death) [22]. Caspases play an important role in the regulation of apoptosis [22]. While caspase-3 is an essential component in some apoptosis pathways [23]. This study aimed to investigated the structure-based molecular interactions between the phytochemical constituents of A. indica and caspase-3.

Hardware and software
The hardware used in this study was a computer unit equipped with an Intel® Core™ i7-7200U CPU @2.50 GHz (4 CPUs. 2.7 GHz), 8 GB of RAM, and a dual-boot OS: Windows 10 Pro 64-bit operating system, ×64-based processor, and Linux Ubuntu 16.04 LTS 64-bit.
Visualization of docking poses and interacting were carried out by using the Discovery Studio Visualizer [24]. Preparation of macromolecule and ligands, as well as docking studies, were carried out using AutoDock 4.2 [25][26][27] and exploring ligand stability in protein crystal structures was carried out using Amber18 [28].

Macromolecule preparation
The crystal structure of caspase-3 was taken from the Protein Data Bank with PDB ID 1NME [29,30]. The small molecules and water molecules in the macromolecule were removed, and polar hydrogens and Kollman charges were added.

Molecular docking studies
Preparation of caspase-3 macromolecular coordinates and ligands, including polar hydrogen atoms, was carried out using AutoDock Tools 1.5.6. Pre-calculation of atomic affinities was performed using AutoGrid. A grid box comprised of 40 × 40 × 40 points with a spacing of 0.375 Å and was focused on the dimensions of 101.360 × 87.332 × 31.445 points. The Lamarckian Genetic Algorithm (LGA) and specifications were 100 runs, elitism of 1, the mutation rate of 0.02, the population size of 100, a crossover rate of 0.08, and 2.500.000 energy evaluations. The docking results were visualized using the Discovery Studio Visualizer 3.5 package [24].

MD studies
Ligands with top hits scores were selected for further MD studies. MD simulations were performed with the Amber18 software. The Amber ff14SB force field was applied to caspase-3 macromolecule, and charges were added to the macromolecules using the software's database. The general AMBER force field (GAFF2) was applied to the structures with the top hit score and NL, and the AM1-BCC method was taken to determine its partial charge in the GAFF2 force field [31]. The ligand topology and atomic charge files were generated by applying the Antechamber suite. The root mean squared deviation (RMSD), root mean squared fluctuation (RMSF), and the interaction energy of all protein-ligand complexes was used to assess the stability of the system after 100 ns of MD simulation.

RESULTS
In this study, 33 phytochemical constituents of A. indica, collected from the reference [1], were docked onto the active site of the caspase-3 macromolecular target and their binding affinities were evaluated (table 1). The RMSD value of the overlay of the docked pose and the X-ray crystal structure of the NL was 1.34 Å. The binding energy and inhibitory constant of the NL were-4.28 kcal/mol and 724.43 µM, respectively. The binding mode of NL showed hydrogen bond interactions with the amino acid residues of Arg64, Gly122, Cys163, Gln161, Ser205, and Arg207 ( fig. 1A).  . 1B-C). Both of these compounds formed no pi sigma-interaction. Compounds 3 and 4 formed no hydrogen bonds and one pi sigma-interaction with Phe250 and Phe252 residues, respectively ( fig. 1D-E).
The MD simulation was carried out to determine the stability of the protein-ligand complex. The binding stability studies on the top four ligands 1, 2, 3, and 4 with the highest affinity for caspase-3 showed the RMSD values of backbone atoms on the ligands ( fig. 2). Residual fluctuation behavior studies showed that, in general, the RMSF value of each residue in the ligand complex was less than 2 Å. However, it can also be observed that there was the instability of Asp146 residue in the caspase-3 complex with all compounds shown with RMSF values of 8.28 ( fig. 3)

DISCUSSION
The binding affinity between the 33 ligands and the crystal structure of caspase-3 was predicted by the scoring function. RMSD parameters were used to verify the position of the binding site for the NL, which was performed by overlaying the docked pose and the X-ray crystal structure of the NL. The RMSD value of 1.34 Å indicates that the docking process can reproduce the correct binding pose of the NL ligand to its receptor. The binding mode of NL with caspase-3 macromolecules is quite similar to that which has been reported by Erlanson, et al., 2003 [29].
Molecular docking results showed that compound 4, as well as compound 1, is an epimer of β-sitosterol and compound 2, respectively. Compounds 1 and 2, which have the same substituent but different conformation, have different free energy of binding and number of hydrogen bonds (2:1) and hydrophobic (7:9). Compound 4, which has no substituent, has a lower free energy of binding and a different number of hydrogen bonds (0) and hydrophobic (6) compared to compounds 1 and 2. This indicates that the acetate substituent plays an important role in the formation of hydrogen bonds in the three sitosterol compounds. Compound 4 has potential anti-cancer activity through growth inhibition, cell cycle arrest, and apoptosis in breast and lung cancer cells [32], and is cytotoxic against colon and liver cancer cells [33]. It is necessary to conduct MD simulations for compounds 1, 2, 3, and 4 to investigate the stability of the protein-ligand interactions.
MD simulations with the molecular mechanics generalized Born surface area (MM/GBSA) methods were used to investigate the interaction dynamics between ligands 1, 2, 3, 4 and their targets. The goal of these simulations was to investigate the impact of ligand binding on the residues of protein targets, particularly at the binding areas. Strong binders tend to reduce the mobility of the atoms they bind to and, as a result, stabilize the protein's binding area. This was investigated by determining the RMSD of the protein binding site atoms throughout the course of a 100 ns simulation. Furthermore, the RMSF was used to calculate the total fluctuation of each atom in the binding site relative to the reference structure.
The complex of caspase-3 with NL showed stability at 10-100 ns, although there was a small movement with an RMSD value of less than 2 Å. The complex of caspase-3 with ligands 1 and 4 achieved stability from the start of 0 ns with an average RMSD value of 1.104 Å. Meanwhile, the complex of caspase-3 with ligand 2 starts to reach stability around 5-100 ns, but at 5 ns it is suspected that the ligand has moved from its initial position as indicated by an RMSD value of more than 2.5 Å. The stability of this complex begins to be achieved after 5 ns at a new position.
Moreover, the complex of caspase-3 with ligand 3 had an average RMSD value of 2.233 Å. This ligand did not achieve stability and was likely to escape from the receptor at 25 ns, as indicated by its RMSD value of more than 2 Å. After the test time was added, the possibility of re-release occurred at around 55 ns and no stable interaction occurred. All ligands 1 st Bandung International Teleconference on Pharmacy (BITP), 2021 | 214 were able to stabilize the receptor as seen from the average RMSD of the receptor of 2.317 Å. The receptor that binds to NL have an average RMSD of 2.657 Å due to the very complex structure of the receptor with a very large number of atoms, so that the slightest movement of atoms from the receptor will cause the RMSD value to change greatly.
RMSF was used to explore the stability of the residue at the binding site during the 100 ns simulation. The RMSF of all residues around the ligands was counted in the last 40 ns pass by using the Discovery Studio. The complex as a result of the simulation is calculated through visualization and transformation in the mean structure ( fig.  3). The RMSF of each residue in the complex was less than 2 Å, which indicates that the binding site was quite stable during the simulation [34]. The instability of Asp146 residues in the caspase-3 complexes with all compounds were possible because the position of the Asp146 residue is not at the side chain binding site but at the end of the protein structure.
The binding free energies (ΔGbind) of the complexes of caspase-3 with NL or the compounds 1, 2, 3, and 4 consist of the van der Waals interaction energy (ΔEvdw), the electrostatic interaction energy (ΔEele), the polar solvation free energy (ΔGp), and the non-polar solvation free energy (ΔGnp). It can be concluded that van der Waals provides the main driving force for the complexes of caspase-3 with compounds 1, 2, and NL. However, the polar and non-polar solvation energies are unfavorable for the binding free energies. The binding stability of the top four hits evaluated by MD simulations indicated that the hydrogen bonds in MD were quite different from the molecular docking results due to the massive receptor and ligand movement in the MD simulations.

CONCLUSION
This study suggested that γ-sitosterol acetate, β-sitosterol acetate, and γ-sitosterol might be able to induce caspase-3 thereby activate apoptosis. These high-affinity compounds can bind to caspase-3 more efficiently and were able to stabilize the protein. Therefore, they have the potential to be used as lead compounds in the treatment of cancer.

ACKNOWLEDGEMENT
The present study is part of a dissertation research that was financially supported by the Indonesia Endowment Fund for Education through the Indonesian Education Scholarship for Doctoral Program, 2016-2020 for EF.