MOLECULAR DOCKING OF CURCUMIN ANALOGUES AS SERCA INHIBITORY AGENTS

Сurcumin derivatives were virtually screened for inhibitory activity towards SERCA by computational docking. A detailed characterization of the inhibitor binding site at the molecular level and determination of the amino acids involved in interactions with curcumin and its derivatives have been provided. Crucial enzyme/inhibitor interactions were identified by analyzing the docking-predicted binding poses of active compounds. The loss of hydrophilic group by curcumin leads to an increase in binding energy. Some of curcumin derivatives showed better docking energies than curcumin indicating that they could be potent enzyme inhibitors. Additional binding energy was provided by extensive hydrophobic interaction between the hydrophobic parts of the ligands and the nonpolar residues at the binding site. Curcumin derivatives satisfy Lipinski’s Rule of five which testifies to their druglikeness (absorption, distribution, metabolism and excretion) and possible pharmacological activity. PubChem Compound database. Curcumin analogs were simulated by a variation of substituents on phenyl rings followed by optimiztion of the structures at the B3LYP/6-31+G* level using Gaussian 09. The Graphical User Interface program Auto-Dock Tools 1.5.4 was used to simulate the protein and ligands. The grid boxes size was set at 90, 86 and 126 Å (x, y, and z) to include all the amino acid residues that present in rigid macromolecules, and at 18, 26 and 30 Å to include all the amino acid residues that present in binding site. The spacing between grid points was 1.0 Å. Rigid as well as flexible docking was used. The AutoDock Vina 4.0 was used to run and analyze the docking simulations. During the docking process, a maximum of 9 conformers was considered. Autodock results were analyzed to study interactions and binding energy of the docked structure. Hydrophobic and hydrogen bonds and binding distance between atoms were measured for the best conformers. The results were visualized by PyMol.

Introduction. The sarco/endoplasmic reticulum calcium ATPase (SERCA) is an ion transport protein present in the membranes of intracellular calcium storages [1][2][3]. These storages act as sources for rapid calcium release through calcium-specific channels and ryanodine receptors that constitute an integral part of various signaling pathways. The rapid release of calcium ions triggers a variety of physiologically important functions, such as muscle contraction. Blockage of SERCA leads to malfunction of the calcium homeostasis in living cells and apoptosis. In contrast to chemotherapeutics SERCA inhibitors provoke apoptosis at all stages of the cell cycle. Due to its high potency and specificity, thapsigargin (TG) is the most commonly used SERCA inhibitor [4]. As a natural product of considerable structural complexity, TG is not easily synthesized, being therefore rather expensive agent. Due to significant value as research tools and medicinal potential as an emerging new class of anticancer agents, the development and design of novel SERCA inhibitors has become an area of considerable interest in the recent discovery of a sizeable repertoire of inhibitors with good potency. Examples include the fungal metabolite cyclopiazonic acid (CPA), di-2,5-tertbutylhydroquinone (BHQ), terpenolides, clotrimazole, tetrabromobisphenol, and curcumin [5][6][7][8][9][10]. In addition, inhibitors of SERCA are invaluable tools for the study of the enzyme's physiological functions.
The SERCA is an integral membrane protein that consists of three cytoplasmic domains and ten transmembrane helices (M1-M10) (Fig. 1). A phosphorylation domain contains the invariants ASP 351 residue. A nucleotide binding domain contains the nucleotide binding pocket. An actuator domain communicates the conformational rearrangements that regulate the binding and release calcium [11]. Two inhibitor binding pockets were revealed: the first one is for TG and the second one for BHQ and CPA [4; 7; 9]. The binding site of TG is the cavity surrounded by M3, M5 and M7. TG effectively prevents the conformational changes leading to Ca 2+ binding by immobilization of M3 and of ASN 768 (one of the Ca 2+ ligand residues at site) [4]. BHQ and CPA occupy a pocket surrounded by M1-M4. These compounds inhibit SERCA by blocking the calcium access channel and immobilizing a subset of transmembrane helices. A binding site for curcumin is not known yet, therefore a mechanism of SERCA inhibition by this compound is not fully understood [2; 10].

Fig. 1. Structure of SERCA
Our study was aimed at providing detailed characterization of the inhibitor binding site at the molecular level and identification of amino acids involved in interactions with curcumin. We also undertook virtual screening of forty-three curcumin derivatives by computational docking for inhibitory activity towards SERCA. Crucial enzyme/inhibitor interactions were identified by analyzing the docking-predicted binding poses of active compounds.
Results and Discussion. Docking SERCA with curcumin using a grid space involving all amino acid residues of the protein enabled us to find a binding site for curcumin. It is a cavity surrounded by M3, M5 and M7 helices. TG and curcumin binding sites overlap (Fig. 2). It is known AutoDock predicts poses of TG correctly. The RMSD between the docking-predicted and the experimentally observed heavy atom positions of TG is 2.2 Å [12]. A small grid box involving all residues of predicted binding pocket was used for further docking of curcumin and its derivatives. Docking for curcumin led to a cluster with binding energy -8.7 kcal/mol. It has one hydrogen bond and nine hydrophobic interactions (Fig. 3). The hydrogen bond was formed between hydroxyl of curcumin and carbonyl of hydrophobic amino acid ALA 306 through a distance of 2.0 Å (O-H…O). The hydrophobic bonds were formed between phenyl ring of curcumin and methyl group of VAL 769, methoxy group of curcumin and methyl group of LEU 260, carbon atoms of heptanoic chain and phenyl ring of PHE 834 and PHE 256. which was severely displaced from the position predicted by rigid-docking, and thus presumed unrealistic. In addition, the decrease in binding energy was observed. These findings are not surprising. It was reported earlier that flexible docking simulations using AutoDock for the TG can lead to erroneous ligand and receptor conformations and are in need of further improvement to be of practical use [12]. Given the lack of improvement in docking performance and increased computation time, simulating a flexible SERCA binding site does presently not offer any real advantage for curcumin.
Docking results for some curcumin derivatives show binding energy of -8.8 --9.0 kcal/mol (Fig. 4). Additional binding energy was provided by extensive hydrophobic interactions between hydrophobic parts of the inhibitors and nonpolar residues in the binding site (Table 1). Further the molecular descriptor analysis was performed to identify druglikeness of the curcumin derivatives c1-c4 by Lipinski's Rule of five, which describes molecular properties important for a drug pharmacokinetics in the human body. According to the obtained results (Table 2) the derivatives c1-c4 obey Lipinski's Rule.

Fig. 4. Docked molecules with the lowest binding energies
Curcumin was found to have good binding to SERCA enzyme. The curcumin binding site is a hydrophobic pocket surrounded by M3 and M5. It was revealed that curcumin interacts with SERCA mainly through hydrophobic interactions and stacking with PHE 256. These interactions cause the distortion of the phenyl rings. Molecule is pressed against the M5 and M3 helices, thereby probably preventing the movements that are necessary for Ca 2+ binding. Docking runs with a conformationally flexible binding site produced no significant improvement of the results. The loss of hydrophilic group by curcumin leads to increased binding energy. Some of curcumin derivatives showed better docking energies than curcumin indicating that they could be potent enzyme inhibitors. Curcumin derivatives satisfy Lipinski's Rule of five which testifies to their druglikeness (absorption, distribution, metabolism and excretion) and possible pharmacological activity. Therefore, the compounds c1-c4 may be potentially effective anticancer agents.
Computational Methodology. The three dimensional structure of SERCA protein [PDB: 2C88] was obtained from Protein Data Bank. The structure was determined using X-ray Diffraction. Curcumin, the principal curcuminoid of the popular Indian spice turmeric is used as the ligand. Compound ID: 969516 was retrieved from NCBI PubChem Compound database. Curcumin analogs were simulated by a variation of substituents on phenyl rings followed by optimiztion of the structures at the B3LYP/6-31+G* level using Gaussian 09. The Graphical User Interface program Auto-Dock Tools 1.5.4 was used to simulate the protein and ligands. The grid boxes size was set at 90, 86 and 126 Å (x, y, and z) to include all the amino acid residues that present in rigid macromolecules, and at 18, 26 and 30 Å to include all the amino acid residues that present in binding site. The spacing between grid points was 1.0 Å. Rigid as well as flexible docking was used. The AutoDock Vina 4.0 was used to run and analyze the docking simulations. During the docking process, a maximum of 9 conformers was considered. Autodock results were analyzed to study interactions and binding energy of the docked structure. Hydrophobic and hydrogen bonds and binding distance between atoms were measured for the best conformers. The results were visualized by PyMol.