Introduction

Non-steroidal anti-inflammatory drugs (NSAIDs) are widely used in the treatment of rheumatoid arthritis and inflammatory diseases. However, long-term use of NSAIDs has been associated with gastrointestinal bleeding and nephrotoxicity (Omar et al., 1996). Anti-inflammatory activity of non-steroidal anti-inflammatory drugs (NSAIDs) is mediated by inhibition of Cyclo-oxygenases, which results in decrease production of prostanoids, i.e. prostaglandins, prostacyclins, and thromboxane. This mechanism is believed to account for both therapeutic as well as adverse effects of NSAIDs. Two forms of COXs have been identified—COX-I and COX-II. While COX-I is expressed in most of body tissues, COX-II is present with low or undetectable levels in some tissues (Sandhu, 2003). The beneficial anti-inflammatory and analgesic activities are based on the inhibition of COX-II, but the gastro-intestinal toxicity and other side effects are due to the concurrent inhibition of COX-1 (Gupta, 1999). Agent which inhibits COX-II, while sparing COX-1 represents a new therapeutics goal for development of safer NSAIDs (Hawkey, 1999; Khannaa et al., 1997; Boehm and Smietana, 1996; Li et al., 1995; Manivannan et al., 2004). Pharmacological inhibition of COX can provide relief from pain and inflammation (en.wikipedia.org/wiki/cyclooxygenase). Pharmacological inhibition of PGs not only play a central role in inflammation, but also regulate other critical physiological responses, i.e. blood clotting, ovulation initiation of labour pain, nerve growth, and development (Singh et al., 2010) (Fig. 1).

Fig. 1
figure 1

Binding mode of compound I into COX-II

Full size image

Methods for predicting activity modes of organic molecules to protein receptors are widely used within drug discovery efforts. Ligand docking is typically achieved by generating a number of orientations (or poses) of a ligand within the active site, scoring of poses and identification of one or more poses that closely approximate the bioactive conformation. The bioactive conformations are determined by X-ray crystallography (Poulsen et al., 2008). Docking algorithms are also used for identifying putative binders from virtual chemical databases and for estimating the binding affinity of protein–ligand complexes (Prasanna et al., 2009).

Butenolides, also known as furanones and pyrrolones, are reported to have anti-inflammatory, analgesic, antiviral, and many other medicinal properties (Husain et al., 2005, Khan et al., 2002). Using this background information and our current interest in the ligand–protein binding, we attempted in-silico study of various hypothetical butenolides as COX-II inhibitors, by computational docking method. The compounds which were found to have high glide score and low energy were synthesized and evaluated for anti-inflammatory and analgesic activities. However, only docking studies are presented in this research paper (Figs. 2, 3).

Fig. 2
figure 2

Binding mode of compound II into COX-II

Full size image
Fig. 3
figure 3

Binding mode of compound V into COX-II

Full size image

Materials and methods

Docking is a computational method to determine possible binding modes of a ligand to the active site of a receptor. Docking studies has been performed with a set of hypothetical butenolide derivatives using Maestro 9.0 on COX-II (PDB ID 3HS5). The X-ray structures of which were accessed from the protein data bank (PDB). The basic structure of analogues is shown in Table 1. The COX-II (PDB ID 3HS5) X-ray structures were accessed from the protein data bank (PDB) (Fig. 4).

Table 1 Chemical structures of hypothetical compounds used for docking studies
Full size table
Fig. 4
figure 4

Binding mode of compound VI into COX-II

Full size image

Ligand preparation

The molecules were built using Maestro 9.0 and converted to 3D structure from the 2D using Lig Prep version 5.5(Maestro version 9). Lig Prep is a robust collection of tools designed to prepare high quality, all—atom 3D structures for large numbers of drug-like molecules, starting with the 2D structures in SD or Maestro format. The resulting structures were saved in Maestro format. The simplest use of Lig Prep produces a single, low energy, 3D structure with correct chiralities for each successfully proposed input structure. While performing this step, chiralities were determined from 3D structure and original states of ionization were retained. Tautomers were generated using Macromodel method discarding current conformers (Fig. 5).

Fig. 5
figure 5

Binding mode of compound VII into COX-II

Full size image

The conformational space was searched using the Monte Carlo (MCMM) method as implemented in Macromodel version 9.6 (Halgren, 1999). Each search was continued until the global energy minima were carried out using the truncated Newton Conjugate Gradient (TNCG) and the MMFFs force field as implemented in Macromodel. The conformational searches were done for aqueous solution using the generalized born/solvent accessible surface (GB/SA) continuum salvation (Hasel et al., 1988; Still et al., 1990) (Fig. 6).

Fig. 6
figure 6

Binding mode of compound IX into COX-II

Full size image

Protein preparation

The COX-II (PDB ID 3HS5) X-ray structures were accessed from the protein data bank (PDB). The protein structure with polar hydrogen was prepared using the protein preparation wizard in Maestro (Maestro version, 2010). In this step, bond orders were assigned, all hydrogen’s in the structure were added, and bonds to metals were deleted and adjust the formal charge on the metal and the neighboring atoms and deleting waters that were more than the 5 Å specific distance. With generated Het states options, prediction of ionization, and tautomeric states of the het group at pH 7 was achieved. In protein preparation, optimization of hydrogen bond network was carried out by reorienting hydroxyl group, water molecules, and amino acids. The final step in the preparation process was to refine the structure, with the help of restrained minimization. It was initiated in the imperfect minimization with the 0.3 Å RMSD for the minimization OPLS_2001 force field (Figs. 7, 8).

Fig. 7
figure 7

Binding mode of compound XII into COX-II

Full size image
Fig. 8
figure 8

Binding mode of compound XIII into COX-II

Full size image

Docking

Glide searches for favorable interactions between one or more ligand molecules and a receptor molecule, usually a protein. The shape and properties of the receptor were represented on a grid by several different sets of field that provide progressively more accurate scoring of the ligand poses. For receptors that adopted more than one conformation on binding, grids were prepared for each conformation, to ensure that possible activities were not missed. Ligand molecule was picked so it could be excluded from the grid generation with Vander waals radius scaling 1.00 and partial charge cutoff of 0.25. The compounds were docked using Glide with standard settings in standard precision (SP) mode. Grids were generated using Glide version 5.0 following the standard procedure recommended by Schrodinger. The docked poses discussed in it were not necessarily the highest scoring, but were selected as the highest scoring pose with a reasonable conformation and binding mode as judged by the modeler (Fig. 9).

Fig. 9
figure 9

Structure of compound XII

Full size image

Calculation of the conformational energy

The docked conformation was minimized with Macro Model using default restraining force constant of 100 kJ/mol. This allowed the docked conformations to relax (adjust) to the MMFFs force field. Without the relaxation the energy calculated by MMFFs would be meaningless. This relaxation did not change the conformation as RMS between the docked and relaxed structure are <0.1 Å.

Results and discussion

Most important features of docking are the logical interaction of the ligand with the putative-binding site of the enzyme. The molecules built using Maestro 9.0 were converted to 3D structure using Lig Prep version 5.5 (Maestro version 9). Using Lig Prep, a single, low energy, 3D structure with correct chiralities for each successfully proposed input structure was designed. Tautomers were generated using Macromodel method, discarding current conformers, and conformational space. Each search was continued until the global energy minima were carried out. The protein structure with polar hydrogen was prepared and optimization of hydrogen bond network was carried out. The favorable interactions between one or more ligand molecules and a receptor molecule was carried out by preparing grids where ever required, to ensure that possible activities were not missed. The process of evaluating a particular pose was achieved by counting the no. of favorable interactions that might be hydrogen, hydrophobic or electrostatic bonding. Out of fourteen compounds included in this study, two compounds (I and XII) were found to have best results which showed hydrogen bond interaction with amino acids of COX-II protein. The compounds (XIV, IX, VII, VI, V, and III) did not show any hydrogen bond interaction with pocket but good glide score between −9.14 and −8.03 and reasonable conformational energy. The compounds III and XIV did not showed hydrogen bonding but glide score and conformational energy comparable to I and XII, it may be due to some other type of bond interaction, other than hydrogen. From these studies it was observed that binding pocket within the COX-II structure are formed mainly by three amino acid residue HOH902, THR212, and THR206, as shown by most potent compounds I and XII. The results of docking studies by Schrodinger provided very useful information as given in Table 2.

Table 2 Ligand–receptor interaction data of butenolides using Schrodinger software
Full size table

The docking studies revealed useful information with respect to interaction of butenolide derivatives with COX-II receptor. Oxygen of the methoxyl group of compound I showed one hydrogen bond interaction to hydroxy group of amino acid residue HOH902 and same group also formed another hydrogen bond with residue THR212. They are highlighted in yellow colour. H-bond with the HOH902 had a distance of 2.248 and with THR212 had a distance of 1.935. The compound I had Glide score of −6.55.

The nitrogen atom of the quinoline ring of compound XII, formed a strong hydrogen bond with THR206, and showed the distance of 2.221. The compound XII had highest glide score (−9.3) and lowest energy, i.e. −95.66 kJ/mol.

Glide score of standard drug Diclofenac was found to be 10.33, whereas the glide scores of selected compounds III, V, VI, VII, IX, XII, and XIV were found as −9.14, −8.66, −8.03, −8.12, −8.47, 9.39, and 9.03, respectively. These glide scores data indicated that interaction of selected compounds was comparable to standard drug Diclofenac. Moreover, from the literature we found that Glide score of Celecoxib (Bhandari et al., 2009) was found to be −11.37, which was also comparable to hypothetical compounds. The results of glide score were further validated by calculation of conformational energy of Diclofenac (−33.57 kJ/mol) and Celecoxib (−43.7 kJ/mol), which was higher than the hypothetical compounds used in the study. The compounds that had highest glide score and lowest conformational energy, can be used for further drug designing and also provides a way to synthesize new potent compounds in laboratory. The most potent compound of our study was XII, which was having highest glide score and lowest conformational energy.

From the graph shown in Fig. 10, we found that there was a good correlation between glide score and conformational energy. As the conformational energy decreases, glide score increases in the same passion. The highest is glide score, lowest is conformational energy and compound is more stable in binding pocket.

Fig. 10
figure 10

Plot of glide score of the docked poses against the conformational energy

Full size image

Conclusion

Docking studies have helped us to know about the binding modes of the furanones to elicit their COX-II inhibitory activity. These investigations were found to be very useful during the synthesis of selected compounds, which were more potent and selective COX-II inhibitors. Moreover it was also proved from above discussion that geometry of receptor plays important role in defining drug action.